anemll-quantized-qwen3.8-27b-for-CoreAI / weights_digest_0.2.json
anemll's picture
Release 0.2 export: three-bit MLP, token-mixer rotations, distilled weights
e1cd2ca verified
Raw History Blame Contribute Delete
255 kB
{
"0/input_layernorm.weight": "float32(5120,):31885d9b9484c5fa90df3a97d3c6c61f4e83966cba35c4d16208e31b97dfd632",
"0/linear_attn.A_log": "float32(48,):402e0e023407e5c3d239aec23f8f1f7abfe84a922d8adfd913573dcb8ebc262f",
"0/linear_attn.conv1d.weight": "float32(10240, 1, 4):5dc2cb4ce90d89a3f04c213cc4305ca3cf32627e43fee441d23087603b1fbf15",
"0/linear_attn.dt_bias": "float32(48,):0b37fd757a73b32927fc9c38716c9cb5af84356bda0967e8551fb672373fcffe",
"0/linear_attn.in_proj_a.weight": "float32(48, 5120):b14bb3d7092ab649bcfffee745653a5642eea916781a828c1547dd42b2a0ee04",
"0/linear_attn.in_proj_b.weight": "float32(48, 5120):4aea9bbeeeb82b8df0941961c17efeb36d43611f7c7596365acbc3701ccabc87",
"0/linear_attn.in_proj_qkv.weight/idx": "uint8(5120, 5120):43f4124c48ad3599429f26c98bdf1f7d34d522e0b2e3421f498f2dd736d64434",
"0/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):5e54694d146f6a9184d5542f2a8729f6dde7b59491d83d94a8507aabb3928624",
"0/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):54876441e3774990813622fe6b10b45b3538fff4e63c9e30ba5716c82ab95dfa",
"0/linear_attn.in_proj_qkv.weight/lut": "float16(16, 2):696d6f90e2c35edafdb549d7adee5a2f26f9bff836a20d0c73a4ec7ac61fcac8",
"0/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):daadba0b8b582f14b1b2441d2b18cc1dcc30c7d26c6a24da1c2495367289c8d4",
"0/linear_attn.in_proj_z.weight/idx": "uint8(3072, 5120):b4d23070ddc8146bb059f8f8f6fb481f593f7dedca513e7d458c6f41bb8ea78d",
"0/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):efd2e24dd5d90734efc98a12d3155e2d7bcdb6e7caff886f85a6c51280bafc85",
"0/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):beea5a6f3296ae65bf5f3934c0fab6cf460d0bdb75a5f72092afddfd4763297b",
"0/linear_attn.in_proj_z.weight/lut": "float16(16, 2):05a34e55d251ad0519e8ff2b54b9f6b20032373f2476a7572f3fb233d576a987",
"0/linear_attn.in_proj_z.weight/scale": "float16(6144,):5cf3d342902f8c252f21f99e3d81717a4666a4f525916268f271d3ee9f3c32e7",
"0/linear_attn.norm.weight": "float32(128,):9b99b9e43d21529320cfb320ad2ffc630b0d0d373369011261b623fe9de66891",
"0/linear_attn.out_proj.weight/idx": "uint8(2560, 6144):49a6ce64d95209f12efc2665808eb42c07a5a75c0020abe9a0dc60acd4cb4b32",
"0/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):436a35cd2754aee45248f6db2966c9f86940a153d1b318442fe2c01763e82ded",
"0/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):4f6dd258ec20c4de55d51dfd6f6eff22b36dd920f4894b90fc3e1c2ae9900617",
"0/linear_attn.out_proj.weight/lut": "float16(16, 2):cb7991b55dd75428380a68044220e67fcd2db1592ad57b338ade498df5965be3",
"0/linear_attn.out_proj.weight/scale": "float16(5120,):e63530e838a770c13b92ef8653f0226dd32f59ff41c022745e1137ee4ffe63f3",
"0/mix.rotation": "int64(2,):ae19bd9756b2b7504b73412971fbeabfe8301f53a0ebb448a0ddfaee5cc75fbd",
"0/mlp.down_proj.weight/idx": "uint8(2560, 17408):d4a20c1b8029efb777c867c7c520caac35a222f63155cca58b5b7a3e17bf0bc6",
"0/mlp.down_proj.weight/lut": "float16(64, 2):cdd2da69955cc2202518d5a3e11a332e80a981385d443465da5d3e1ce2401170",
"0/mlp.down_proj.weight/scale": "float16(5120,):bf75c6ba5218ea60829f1f40d327b4b68d63c680e7b564b8ed82bfaed8b494b1",
"0/mlp.gate_proj.weight/idx": "uint8(8704, 5120):881207a7ced46f2ee2a6d2b617447f9b9b9c87429b86430e0ee2232c3acb4e06",
"0/mlp.gate_proj.weight/lut": "float16(64, 2):88340bdd38a202df4bf2d27543716bcd7759c3a06a7b6828e27d63f745f0b13b",
"0/mlp.gate_proj.weight/scale": "float16(17408,):06dc95fe69de654fba565d9958c3f7a9c40b05bc5a9509885a452c0dd2046639",
"0/mlp.rotation": "int64(2,):65a3abf28eed7c1f8ffd6e9c52076d9d2828ea37e468a2ba09d5353ea032ada6",
"0/mlp.up_proj.weight/idx": "uint8(8704, 5120):5966923626602b91461009414e0f243350c0c5d0b29e82c2881e126eebba3bdb",
"0/mlp.up_proj.weight/lut": "float16(64, 2):07daa57f4d0bcf3723ab16519df95755df9e45e6a27a34bc47e616639f72c1ad",
"0/mlp.up_proj.weight/scale": "float16(17408,):894d55a5c0ac440e4af5bc176e4b2ee2a5f158932bb3348c632ca10c766b592b",
"0/post_attention_layernorm.weight": "float32(5120,):d63767c9d77692a71027ddef848a1835658ec0811bb7779dc67bff104be7deb4",
"1/input_layernorm.weight": "float32(5120,):a9ae688e1fd4218b96dfd89fe6c1a650a0c21219700d7045ff3454565d4cfdf5",
"1/linear_attn.A_log": "float32(48,):db4f2e809561a21e9659d19daa1b55c8f110ebf21257d298deb972fe9d0dc337",
"1/linear_attn.conv1d.weight": "float32(10240, 1, 4):82c3cea22310adbbc8d89b652e7ac4bdf70a6671ae63cdbca479c9f1cd6fa0e1",
"1/linear_attn.dt_bias": "float32(48,):a0a09442fb9549b94e6d6f700adc40ce0c709edf2e72546d4c8a771673490f93",
"1/linear_attn.in_proj_a.weight": "float32(48, 5120):971648e546f26b1977cc466340bdf93588963d6b3fd9d3bfc4c07a0a75eac792",
"1/linear_attn.in_proj_b.weight": "float32(48, 5120):65d8947f50187522e5b89c09d7b03e4243f821308361a85ba2fb31164e94342f",
"1/linear_attn.in_proj_qkv.weight/idx": "uint8(5120, 5120):1361cd6685c68d07d4ed6654291e5a83dc9f503c9c79bdffb4937a6e1edd7a68",
"1/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):11864dab0f14dec693d50dd8befb091d67e3965785b6cbfbff1017968d2909a4",
"1/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):e11118ad2c3523fa99f52f825c868622c478cafd9ee99ac0062039b858375c67",
"1/linear_attn.in_proj_qkv.weight/lut": "float16(16, 2):2c0a3e06b9586cc663a87efebf2a8baddf69e4a8ab8526f8e3e03a058a990800",
"1/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):a4b14eaec1eec5b91e8c3b685d4c54b3868bf03bfba2f052ac6c404f0e19411b",
"1/linear_attn.in_proj_z.weight/idx": "uint8(3072, 5120):568a12c615fdba7c7f767a30abb96b0479d09e76a3697ab0b5c991873dff8da0",
"1/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):73a8649f18328eaec3908de4d83dc2b28a76861b66a4ee6679c24156caff3d24",
"1/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):2c95a9ac250b7c286a8ca78a87424dddef8460c1012608a340afba5a16bb7003",
"1/linear_attn.in_proj_z.weight/lut": "float16(16, 2):d79e89e9f9565b8f6fd328de51a478308311154d2d49e64959e075c990763308",
"1/linear_attn.in_proj_z.weight/scale": "float16(6144,):be8a984407e93f94ee2ab1f7e3dccf79ba8eaf172b00a1f4894eb0a018169ed4",
"1/linear_attn.norm.weight": "float32(128,):8d4d738c2eb28c05f7e5e214253c2d0823bc88882cb7c43f95cab008d354f45b",
"1/linear_attn.out_proj.weight/idx": "uint8(2560, 6144):e58ad4eeea9de561601c52b642f2f34597915f7f620f8e6314b3d492e6138dbd",
"1/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):c9da55ea1eb9dba93e3789cd15e6844f93eafc02cc573e0125364e3af6be3f88",
"1/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):38e196e331f6df5414903c27624f85b0ab4c39e4bce3098f42281f8bb40da71d",
"1/linear_attn.out_proj.weight/lut": "float16(16, 2):d06ba5a335a0f45b9fc8a7c0673dab164325279d43c4cce6591bb61c6e722174",
"1/linear_attn.out_proj.weight/scale": "float16(5120,):736e82678d404073a3f47756b30b3fae20e92c6b9e7639177501a923531f23dd",
"1/mix.rotation": "int64(2,):acf3be3838b2c4dc97ebd1d66dd11184c840da0c9eb34358141da034f9bae36d",
"1/mlp.down_proj.weight/idx": "uint8(2560, 17408):0249bc4e106994ce366316cbd7995b4ec2692138203ffe87f03a2930493021d5",
"1/mlp.down_proj.weight/lut": "float16(64, 2):4028a6e598cd7f31d7299d213aaeb363422424f9b97f55e1d9aa8fc4345ce532",
"1/mlp.down_proj.weight/scale": "float16(5120,):5be4932ab9baca89db7ba0540a61fd8d32673134dfb5fbbc1bbeca67dc04e652",
"1/mlp.gate_proj.weight/idx": "uint8(8704, 5120):2a2133c4543582b137694228bc71d71528be5877aac4f8a28bcbc711fd86549b",
"1/mlp.gate_proj.weight/lut": "float16(64, 2):16061c10311997bded95f599c65f4a755c9646cce1241ac6d8da5c854152a167",
"1/mlp.gate_proj.weight/scale": "float16(17408,):9eb20287e2763a5662b892f883a3c4563271c811762af126725ed679ae951b34",
"1/mlp.rotation": "int64(2,):2d488b329d9ff8d53cf3376856e7031affd354997d54b4fe67fd39d67a05b8e4",
"1/mlp.up_proj.weight/idx": "uint8(8704, 5120):d256d4cd23d94b86ff6b2bf2c0d4ec6bd2848aa91cdad364be6a5a779247edf4",
"1/mlp.up_proj.weight/lut": "float16(64, 2):a16c5d536c7d39c9fdcafa2139be84c227d6329082f4768f73c0636cc2de79e9",
"1/mlp.up_proj.weight/scale": "float16(17408,):5e89ef2344a1a7891f5bd5230913085742c01743aaedc9dde87e5febd40e1337",
"1/post_attention_layernorm.weight": "float32(5120,):830e16fe00fe6eb7445c7c3c459ed4c93605872d0bd5b81150209f367b4b2c12",
"10/input_layernorm.weight": "float32(5120,):13bada352cb447bdca62998bf2f92bbcb9ee843ec4c936d0116edd6d679cb24a",
"10/linear_attn.A_log": "float32(48,):5720713d6e7f6130525acd42ec433cbeeec1310f30b54688e3a138cbf397dce4",
"10/linear_attn.conv1d.weight": "float32(10240, 1, 4):e0385d35038e1de093d3eb7a01709b18c172606e83a58e17c8e95eff796130af",
"10/linear_attn.dt_bias": "float32(48,):bbd853cfc32d0a0d79d17eb1be751af137369f08fca13a1c474b11a11c8f7dd3",
"10/linear_attn.in_proj_a.weight": "float32(48, 5120):602c56b77d3e8e7753d6647ec3cbc6c522ae21f77836f05ec4279d729de48358",
"10/linear_attn.in_proj_b.weight": "float32(48, 5120):21d1555106ffa7305c25f5bd34b533e96842df3fa4c5e7cf5a85eaf85a4a920c",
"10/linear_attn.in_proj_qkv.weight/idx": "uint8(5120, 5120):15b487e1d1830ecb3255c9963c1afb5f3e230d2df007bd4dc650255d73f0fa69",
"10/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):01b84c4a566658c22eba310acc689999e9a016c262c1e5e00b66f3cc548ae090",
"10/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):79d0b7958934eec4ddaa8ebd72cf268ba5d583e79b9ef76811351b25ff31a9bb",
"10/linear_attn.in_proj_qkv.weight/lut": "float16(16, 2):a661faf0eb986bab4c5cd5bffa3546aaecd397e562bb98bbf90b09ea0c3e74e6",
"10/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):4ff820694a883785a21a351bcbe55e9203909b06951e4c3ceebf9c9445932199",
"10/linear_attn.in_proj_z.weight/idx": "uint8(3072, 5120):e37e2fb284d049fe2a17e75200a664e643b58836ea53e0abf7dedca891db7d27",
"10/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):16c5600d364dec5b6ace58ff034ace711757cbc170b88245adcd7b0cf67ed34a",
"10/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):58e6043fa1ca250a3b512a8206f73557fbd00813ed2932d21aa5fcb60d365f1a",
"10/linear_attn.in_proj_z.weight/lut": "float16(16, 2):48739cbc16e163330134e6c257680d390db63b8ae7403543ce7b7c0b085ee10f",
"10/linear_attn.in_proj_z.weight/scale": "float16(6144,):164fb9dc9b4708f2e722d7c8352cc2e63fce99623154b79177e318355f0cdf59",
"10/linear_attn.norm.weight": "float32(128,):9d57b6c1bfa6e132eeea48fe90eacf15720959a4b12853f8dd527da5b519f9ae",
"10/linear_attn.out_proj.weight/idx": "uint8(2560, 6144):b22f0c72c4c6347464f1d8210444904d3d885f4d29d60ff3485253ae3b01c54e",
"10/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):d572ee0af17c83a7198f390b1c81be8c0d2e75b398476721f88df524f2ed7588",
"10/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):64620702c8a714fddfa08680eccf36ab7337911546c95d599e17febfa0ba1589",
"10/linear_attn.out_proj.weight/lut": "float16(16, 2):baaa1f7bd066704adae3ee2d5dd276fce0ad84b6d935f053577c365ab19658b4",
"10/linear_attn.out_proj.weight/scale": "float16(5120,):c540b310dc774a1e40b853a47f2a82108e93ced7841e14e4644a3ee4e28dd7ac",
"10/mix.rotation": "int64(2,):f13ae67811f2100a023783c1c42414bdda89151f925cd18fee8803d3ef2ad0b8",
"10/mlp.down_proj.weight/idx": "uint8(2560, 17408):f509dee9e968f2e50ee35a587ce8c8c9bcbc8a2b26bd65e1fefa5492e6bbd1be",
"10/mlp.down_proj.weight/lut": "float16(64, 2):94a0d5b993b1f6674fae1d3d49097c83b5d34b88f4275e6dc641a0ba798fcb4b",
"10/mlp.down_proj.weight/scale": "float16(5120,):b58f7f35ada2a4189754d48cd46c923aba286fa946c8581e6d2d44709227115c",
"10/mlp.gate_proj.weight/idx": "uint8(8704, 5120):4dfd2478df2cab457a74ea22ab72e75eed49393260d1d41c8b9f0b76c6a74b82",
"10/mlp.gate_proj.weight/lut": "float16(64, 2):9ca48326190f9d2ec7f61cccf02f9a8805a120c7929a2cc0cb3606a7d277c40f",
"10/mlp.gate_proj.weight/scale": "float16(17408,):8ba8369aad33e4433d0476d2993705ed05afb3aaabb1eae8d7395a12b09c505a",
"10/mlp.rotation": "int64(2,):b05da690820c3b344149d1ae9923687ddb460941501e7d7aac87a4f8258d800d",
"10/mlp.up_proj.weight/idx": "uint8(8704, 5120):7081837de97a6d121108ec6a1c040c295e827ab75e0b362b25090edf10093fdc",
"10/mlp.up_proj.weight/lut": "float16(64, 2):197eba7c4ced85798f511ca8497c8e46606728a01fc09031781b54fa3693f5fb",
"10/mlp.up_proj.weight/scale": "float16(17408,):6ec4117789a8242d99c6742de38f5e30e5df98e9962c8c31bc0d0c0206dd4858",
"10/post_attention_layernorm.weight": "float32(5120,):465f144840dad37e963bdd527f1f7d278db7311a0b0a398caee87404518b2ff2",
"11/input_layernorm.weight": "float32(5120,):86f243f26be99fe9fa771bf124c39fb4e2265554c72477f41eca63784519dd5f",
"11/mix.rotation": "int64(2,):57e467c1546e3f17bf07dc30ef64f376c0ee01276c8ea38c91e645d21916f43e",
"11/mlp.down_proj.weight/idx": "uint8(2560, 17408):bced91a925ded2d40c7ef1b30dc85957326b4db10e8b9e327864a61e004f6eed",
"11/mlp.down_proj.weight/lut": "float16(64, 2):44137d6fca749a93b121932d80872c52d46555f2fdfd879d830452444778bb1e",
"11/mlp.down_proj.weight/scale": "float16(5120,):1a3f989e826f7b117c894fe4516ef36c0fb2c2dbc154a9b7188f61720aefb818",
"11/mlp.gate_proj.weight/idx": "uint8(8704, 5120):84212231ad2e3c16b750a5e9a1dedd96855e7400a3ba0dd8897566a0324c9314",
"11/mlp.gate_proj.weight/lut": "float16(64, 2):d778b064697811e8568a51a664f64e5bc56e26663019774fe67216d723cea44a",
"11/mlp.gate_proj.weight/scale": "float16(17408,):3868fd3d26a9ff6fb8ba17cb8d27891e34b6b7a86085a58b1b07c7a0b7769cb3",
"11/mlp.rotation": "int64(2,):eba6af984de90a2a0f3f3cf75fcedfaf651fda3d9262b38f3b2e445baa92cdf9",
"11/mlp.up_proj.weight/idx": "uint8(8704, 5120):e536f234c110d2428da5d66bd4bb9e99c5db030fe0ce3367da3ade6d7704301c",
"11/mlp.up_proj.weight/lut": "float16(64, 2):a7996a830f7436a72aca5e3a07944bb8cdf75dbb04d6c91a691135117455d391",
"11/mlp.up_proj.weight/scale": "float16(17408,):7639c5970e423ab944f9cccf33b33950d3351692417fbc2fef3df58c3a9144dc",
"11/post_attention_layernorm.weight": "float32(5120,):55ac33ce1a50770c7adf7dc87e77d6b38c6875d0faafc3e416ad98fd4b84800d",
"11/self_attn.k_norm.weight": "float32(256,):9d8a946bbfab73eef0fba5207a7137336e1f59c55af41c5a3decc58b266aab47",
"11/self_attn.k_proj.weight/int8": "int8(1024, 5120):5b9529722ba7bba01a673049dd19d781161b4d09a5606044f5145a86320ae93d",
"11/self_attn.k_proj.weight/lr_a": "float16(1024, 64):524697889a3356d9e9355bf24e0cb6b7022dc6d0de2b46adefd9dad76ed71542",
"11/self_attn.k_proj.weight/lr_b": "float16(64, 5120):e9ab76695e3a90329a0378065f6ceace06d128053a857e67ae72344bace662b0",
"11/self_attn.k_proj.weight/scale": "float16(1024,):502140c4a010babc8ea0700dc11a1ce522f610f108f5d1a55f60680030e5142a",
"11/self_attn.o_proj.weight/idx": "uint8(2560, 6144):3698b079743687639ee40544547995b59d80e7449033d4e83b51ac62cf916aa0",
"11/self_attn.o_proj.weight/lr_a": "float16(5120, 64):72cb62a004f2c653e9c7678814204428cd3ddbb7fa070883ae9dd581c6fd9c7d",
"11/self_attn.o_proj.weight/lr_b": "float16(64, 6144):9b25ba97d573ef652602260d661c008f575da720e01c771f9a3fd16b290a29f3",
"11/self_attn.o_proj.weight/lut": "float16(16, 2):1d306c20c5f42d43f31103aed901723b1448e8204c0eb1b70a85a583c21367df",
"11/self_attn.o_proj.weight/scale": "float16(5120,):f3e6e4dc958c6b1722c3b22a2a5db399740bf0b9799501e5b29fcf17b6e1423f",
"11/self_attn.q_norm.weight": "float32(256,):d89ae98d0032e32b0a0cc53da75d9b1bfc363910feb298431deda997fd7927b1",
"11/self_attn.q_proj.weight/idx": "uint8(6144, 5120):e51e8b39f3840127e7f133bb7907845c4e5d94c0ad21ef990fbe3c473f05f35b",
"11/self_attn.q_proj.weight/lr_a": "float16(12288, 64):62b1cc9afa3d4bd95e5a64fd9251343487a8f66315acefbd7358766ea1fcc4db",
"11/self_attn.q_proj.weight/lr_b": "float16(64, 5120):4c9d2d7220b3c924191a8e7335b4459616cbeda58e7ccc3c94668c54cb736261",
"11/self_attn.q_proj.weight/lut": "float16(16, 2):bbd1409d08fd2d664e0da3530094a4b778bf19d0e2b9a80affde2eca4d91f171",
"11/self_attn.q_proj.weight/scale": "float16(12288,):ee930bde50effeb6252bf54c460714ff4a56cb8136c670ad75e6181347982351",
"11/self_attn.v_proj.weight/int8": "int8(1024, 5120):1e3bf3f47d36bc57bc3e0ffee2379c0a886d54aa44543a0b781059aa3c8bdefd",
"11/self_attn.v_proj.weight/lr_a": "float16(1024, 64):88ee3846eede028993f504c23cd58229bc31bee825945423d762f109ac88a035",
"11/self_attn.v_proj.weight/lr_b": "float16(64, 5120):ce160707419599aaa387223b30ffedc55cff4a5ececf6eac05b42a25e097d8d8",
"11/self_attn.v_proj.weight/scale": "float16(1024,):d1ef9847fbed211301f418eeb2161b1d940cbfd6aefadc2520bc28aeaa25d7b0",
"12/input_layernorm.weight": "float32(5120,):0de33c13daef2719bcdabccc359d44da59159451a105becda4a29b3f99bda9b2",
"12/linear_attn.A_log": "float32(48,):64ad545f745f59a86435908d0eca52c418994c0489e7a3e9770322d368f62ab7",
"12/linear_attn.conv1d.weight": "float32(10240, 1, 4):4df347715d76ffe9815e2540e252eb201fdc041896574209348fcc2f458fd977",
"12/linear_attn.dt_bias": "float32(48,):19bc00370501d952f442bb250afb1ed5929975585bacae3b467e7717acd26bd6",
"12/linear_attn.in_proj_a.weight": "float32(48, 5120):140dc4bae40055f65cfaf53dec8a7f4734a2a350ac0eb52e8f6a0ad5b88bfec2",
"12/linear_attn.in_proj_b.weight": "float32(48, 5120):be9ba59655545013162a258208da7a20a4c1ad3b82ee4f435ec15b779ebf3761",
"12/linear_attn.in_proj_qkv.weight/idx": "uint8(5120, 5120):e7f159e16c4209ce764dfe2901d1546c4d7f3f9ea2e288841ffb7bb20e95aa9f",
"12/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):48814842f55961fafe24a623ac683ec042e364f04d61cfa92cf6470c92116902",
"12/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):fbbe6c90d5f79462bea2f930d41720c04528a7d3ef80a1a7dd5f8611229e6625",
"12/linear_attn.in_proj_qkv.weight/lut": "float16(16, 2):4593346b596f1c59644896fc8290ed1b21c7d48055bc43da04f414c6adbc04c5",
"12/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):581197e577d3c19065021ad49df872cbde393a9eaf420dbb9505a138b870546b",
"12/linear_attn.in_proj_z.weight/idx": "uint8(3072, 5120):ec9932731cc8a4b4ca8a4e770f8f4a5eb866b15883be8e7dc4c15c76a4aec0f7",
"12/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):d2a4d8ccc0d24f2d18698f4bb4f72db4aa036a10a653a9666c1bc179597c460c",
"12/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):da7a9d7a372a78729690efb55c3f08b65b09f0d14906c7ff6a50ea1178db4e6a",
"12/linear_attn.in_proj_z.weight/lut": "float16(16, 2):02f6383156da5868723cc97c0711a1065312c12094630997ebfaaa39956cd830",
"12/linear_attn.in_proj_z.weight/scale": "float16(6144,):1df1306681c9f8f5a8287f8ca2ad0831306ba52a24eb8d6262b711ba12c6d656",
"12/linear_attn.norm.weight": "float32(128,):dd850c3453f04a106a21f5de845207920a366103cc0e4afcbba72c23b7621bbd",
"12/linear_attn.out_proj.weight/idx": "uint8(2560, 6144):1892b878efd8a157d4b313da7c2c3ec7df62b7b659e19f8e59b1222bce5c6c7b",
"12/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):f1437dd7f3a4eb254f3e9bc334829b1dcfd89f87b44b94fb32e1eb766360392f",
"12/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):7bb9764a1a1963a1f16d7e80a17b4432d346ba8957f1448f98bd6310558abceb",
"12/linear_attn.out_proj.weight/lut": "float16(16, 2):9839959e41af5c2563641e835ab0ed99e52d64dc8614fdc92be79875a3cc3c81",
"12/linear_attn.out_proj.weight/scale": "float16(5120,):adc69bcb2222d048ed8789e5a55f52dd5372841b423ad3aba372f7fd64205119",
"12/mix.rotation": "int64(2,):54baf4aa5cc8d16c5b4f04ea8e96427ea89e4a9aff40d2df2c8310a985c3c503",
"12/mlp.down_proj.weight/idx": "uint8(2560, 17408):b4db56e999d693cda6eb2652a6a4589930124604116d2caaa5d67ea0bfdc80c2",
"12/mlp.down_proj.weight/lut": "float16(64, 2):961274acdbe9e6f0765b7179ea0699551611f66a2793ccdd323cc7e112cb899c",
"12/mlp.down_proj.weight/scale": "float16(5120,):7b0d9b459713b8085197c3ca5386b84eaddec90c752bd6c38aaed3aa47e3ce8e",
"12/mlp.gate_proj.weight/idx": "uint8(8704, 5120):65f748a72bf11d72443e91fc8ea8713e6773fd1ccb52dd2d970f217aef88491d",
"12/mlp.gate_proj.weight/lut": "float16(64, 2):c0e313411562ee3a0666670ac53fe4f5e703490000b461a14d557ab239507cc5",
"12/mlp.gate_proj.weight/scale": "float16(17408,):e83ece1962c1515ebcf0fe8531ece70f9943fa9d1a03399f7f1cd8ac24631039",
"12/mlp.rotation": "int64(2,):bf81b6ce13bd296405ca14ba735195f61d548ab117c59719734582d95ad56c3d",
"12/mlp.up_proj.weight/idx": "uint8(8704, 5120):a59193a17c66407c33a2dbacb8d7f3d6c104384d2bb085cca672ca871e84cd47",
"12/mlp.up_proj.weight/lut": "float16(64, 2):29a2ce8c65b357fe0da74b86f5a2e3ff22d9ab0995a10a77003afa839ae6695a",
"12/mlp.up_proj.weight/scale": "float16(17408,):40b42d24cd71e705b99a3e73c92f9c097ffa60c1a0aa54252e5d481e5bbf7064",
"12/post_attention_layernorm.weight": "float32(5120,):09802ff3142021002755f111c6f44ee229d7ea652de6c25e257e96947a8ba683",
"13/input_layernorm.weight": "float32(5120,):85abd34c430755107f0656b573cfaadac78b2cca6fac58929af3a7b54a1ac17e",
"13/linear_attn.A_log": "float32(48,):4ca9faf53a17162c6bd0947fb15811629b68e7674f966f68e958d16e2a554c1f",
"13/linear_attn.conv1d.weight": "float32(10240, 1, 4):6a79cd69f4b3e986c499989e60df95f67cecf27ae99a9845bc1d1cdb0a63bcc8",
"13/linear_attn.dt_bias": "float32(48,):4f0a4e32b4b26da40b9bc9ef36f9a054f5efd9a8163b6b2c6c558a13f09d41ed",
"13/linear_attn.in_proj_a.weight": "float32(48, 5120):6a241987c61045ae19a5f27e85fc71ab01539b9c217d9ca23c1719c009943d15",
"13/linear_attn.in_proj_b.weight": "float32(48, 5120):2b2179134076bf22881c0a0cbfb37f0d5fd0180f3e3d3e91fe26f252813d3518",
"13/linear_attn.in_proj_qkv.weight/idx": "uint8(5120, 5120):8ea270b995970468054f9f6b58be56685dc6d8310c71fd3b449b471d8a2aced1",
"13/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):2d9b69e8ad321b5b308f855e63e0f68d0e23cb0da25e9ef45b48d37755857ed9",
"13/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):15461ae0eb67acbeb323acd97557ec5b9356d9eb2213ef07d641064e80c66efb",
"13/linear_attn.in_proj_qkv.weight/lut": "float16(16, 2):e029598408c13e6f2328a73aff299764c434ff022fb14df7cb3823c8a19d8acd",
"13/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):bfc8f2eb4b4dee81405b95e134c75d3cd728ff0b3f54c6d2d4c871951c1a4c65",
"13/linear_attn.in_proj_z.weight/idx": "uint8(3072, 5120):60fddaf7d70d13dcf8d4276755b4009cc0e9b5f20e7dc8d5234feb93c9069d5b",
"13/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):3752d69dea446e80f3620e40f6cbb973e84330aabe64c1c1ae6192c17fc70d43",
"13/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):abe85dcc96c49d41a96bf32ba23b180fe861e69e4a0e476ba1abde4098abdc83",
"13/linear_attn.in_proj_z.weight/lut": "float16(16, 2):4ce7b21239fb7f4c107e3054f37c06b8ab5d777c3a6218ca8cde70e4fe6a243e",
"13/linear_attn.in_proj_z.weight/scale": "float16(6144,):bad3f9137cd2c27758bd1be2b8eec8dfc892e32203b3798af276d569e9699d70",
"13/linear_attn.norm.weight": "float32(128,):07d8d99ff711e9eb4209ca138f9cd87abafc8a0cbe5eb16bcbd93110493f6425",
"13/linear_attn.out_proj.weight/idx": "uint8(2560, 6144):09107150a17a333f5949dbd1577e2ab87ed399ecb35890224fbd88237173ccb3",
"13/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):0d15b87b9e2c03e0f465e116224422aee02f68ce79499985301c589deeb5f850",
"13/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):40b003f82ebabb6858d4cc4345f469fe59c8333bdddc9f5fa78770635ff63ee0",
"13/linear_attn.out_proj.weight/lut": "float16(16, 2):eacebc80cade5499862052b0ee1c3d85dd948b639933280b548a98964352bc88",
"13/linear_attn.out_proj.weight/scale": "float16(5120,):0b2a467ffc8332f949198640ac5682c2d2a04b4f60892fc4f7f33b4ed2c97335",
"13/mix.rotation": "int64(2,):e8309c073623abc8d58ff78bfe8c44d7dd6676c0338ae30fe9638aa21cbe5325",
"13/mlp.down_proj.weight/idx": "uint8(2560, 17408):4d0dc5ecb6b5f1e3b67b5a5545f463c8e61cf6fedadac4e255eb2cbb7ba9f8ec",
"13/mlp.down_proj.weight/lut": "float16(64, 2):a70d59736284e0e690afe8bd878f9ea1d2c4bebf613bcda9e4d8487dd21d132d",
"13/mlp.down_proj.weight/scale": "float16(5120,):311cc75d6269f643974d7382918f19c8906a1323bdac8902738c55bdc7a5335c",
"13/mlp.gate_proj.weight/idx": "uint8(8704, 5120):621fb0b4271e70cfaf6b17dcfc4460ec8f76f03763828b20bc855b0ba5b540dd",
"13/mlp.gate_proj.weight/lut": "float16(64, 2):6287bc5094d1ea0d6889e0118aaadbae03a6a411b4d9b60a85f3a0a8c20505e6",
"13/mlp.gate_proj.weight/scale": "float16(17408,):d26a9e07c5fe5072f818ebaf17b737d3232150a51c9870ca47a7385331df8c44",
"13/mlp.rotation": "int64(2,):7849024379e99da3ed864c1918bf229ddb33e86e3732ff8a3f273a586aaff283",
"13/mlp.up_proj.weight/idx": "uint8(8704, 5120):25b6f61455e18746ad18d02f0d943b278dd619a0bb19726a92864ad0170e7f0d",
"13/mlp.up_proj.weight/lut": "float16(64, 2):f585abf4105ca51cc320873669650efbfa73ed8be13db21d3f90d0a212b568be",
"13/mlp.up_proj.weight/scale": "float16(17408,):4d2f6558597d73616973233c1ed7aa3f632bc379e7d9158b4e7e09c9d2d005b7",
"13/post_attention_layernorm.weight": "float32(5120,):b4d82e231733e56780113c287e1bf8b80e6d4adac7f58ce1e1d33d58a8617562",
"14/input_layernorm.weight": "float32(5120,):78adb7cc39cda9cee38c00c7f00459605f9e9f8bb29025e28328e5ffb4d5dc21",
"14/linear_attn.A_log": "float32(48,):7c2c7a59acf8f616fa31e08636a3280a451f4ff5f1a1f512c75689b3e7665784",
"14/linear_attn.conv1d.weight": "float32(10240, 1, 4):92dfa53cd01c31c19e6fe1131ca5e94ef91484c91e696d3f3d4bfbc161a98b95",
"14/linear_attn.dt_bias": "float32(48,):d970ed462f0866711c446c7b096a07a80247aaf0aa2bf6d877d022af02642cbd",
"14/linear_attn.in_proj_a.weight": "float32(48, 5120):a520ee493887aaab1e6d3f8aaa833a8d722ece937810f8b7e3a7cfa5666328c3",
"14/linear_attn.in_proj_b.weight": "float32(48, 5120):da53e368075ff61d8fa374a46440523f1efdb032e13827527125e02f7c4b3a8c",
"14/linear_attn.in_proj_qkv.weight/idx": "uint8(5120, 5120):908d97fe8a791c850c0205c1940338a8a92e3b3142a28b892eb0dc7211936d78",
"14/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):5c386d669292f90b666d920ac967c639dbe68e63f31dc6a1e1a7869d750700d5",
"14/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):33c4c5e96748bd8db4a717cf491aaeb1a83a36b5bbbc5ab1065289c0a0cee392",
"14/linear_attn.in_proj_qkv.weight/lut": "float16(16, 2):274a86369e777ba89a499f0d29c63b31b1664b63f6243bcc40c64cbc5995c1f9",
"14/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):5be9733e8b9644df50ddcf37554e8f6cbe037120e7c79dde38e75e08afa1cd61",
"14/linear_attn.in_proj_z.weight/idx": "uint8(3072, 5120):fc1677e4bd15ce6b51d6480024fa48f771bfbd29d1de3a12a7c3c5ba70afb5c4",
"14/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):0fa18af86a856db40936858b1e1370bd55176177ab3927834fbb86accc4eef90",
"14/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):e4fc26f61c70dbe3539347a7a050a3ebad84231277f182ccbf70ecb9840783cf",
"14/linear_attn.in_proj_z.weight/lut": "float16(16, 2):681ccad3acd0292e75e99da204991f513ddd2cd9af2e00f2742d032532f8f89a",
"14/linear_attn.in_proj_z.weight/scale": "float16(6144,):1cb4bbca6227c6910b1ebf2354825a193ec350d3f2d22b680add0baf1cfcfa5f",
"14/linear_attn.norm.weight": "float32(128,):24d4329c64a46c9b73f599cceb46e49f8bd7c9f733340a0a5e8fdd5a7f1aa8c4",
"14/linear_attn.out_proj.weight/idx": "uint8(2560, 6144):263285dee4424f1bf06b69bda5d16765d0298b27f6f072a4df14da8b9b83bcec",
"14/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):840598c60c5866cbb1fe1bdb7d17790a97fbddd64307561d340735cf00e85e0a",
"14/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):3afc1adb441926ad498a87355e2920b027f72cc5678c854fd1d36ca1985a1f0b",
"14/linear_attn.out_proj.weight/lut": "float16(16, 2):9ce29250d5b3fb9191fe83972d655ec0e88fc3ad3c3fc13568910acfb1a5e474",
"14/linear_attn.out_proj.weight/scale": "float16(5120,):cc807287e32dd312a5a49a81a294d02534bf7736c0ce976fea1d84dc54777702",
"14/mix.rotation": "int64(2,):4ad5721e7f5c410006076a4859fc51b114a6d6c9e6bf26eff2f4040955b6dbab",
"14/mlp.down_proj.weight/idx": "uint8(2560, 17408):005c383739ed0ba94688760fbbd2444585387cc6756ef8ad48c47a4fbf82fb7f",
"14/mlp.down_proj.weight/lut": "float16(64, 2):c08bc30bcc14f40660f0d07dc2510740cb3b86c5f25ccc2d88ec9e9c2e86b028",
"14/mlp.down_proj.weight/scale": "float16(5120,):c6981b919df30109773acdc61f0c55f17a7882e66781199c0c66241bd7a162f3",
"14/mlp.gate_proj.weight/idx": "uint8(8704, 5120):eefcae212d36208179920c21de731dff01ffe4b044f12623f83295844f6b219e",
"14/mlp.gate_proj.weight/lut": "float16(64, 2):756cce489562c92250390143c4de178beade90e8d75f3eda29e3f2017275874e",
"14/mlp.gate_proj.weight/scale": "float16(17408,):7e1ad0960bf60a7855f1c8a185137438c2a881ce76ffc96fcaeadc27aed70cc0",
"14/mlp.rotation": "int64(2,):d56cfab05509e08ef923b345e54f8a4b03eaf77253f8f80245446c896ad8496f",
"14/mlp.up_proj.weight/idx": "uint8(8704, 5120):c642a76bc8e54e75b0d0227ebbe9d94504390f55c8fc891cd839a32efd26c453",
"14/mlp.up_proj.weight/lut": "float16(64, 2):af53931cf39b044dbc4049403a122d2146aefa5379a7e1a900dad145c83f8999",
"14/mlp.up_proj.weight/scale": "float16(17408,):287010e7de04798da99abd301dce378482bf67d5fb994da44debe31761383ec6",
"14/post_attention_layernorm.weight": "float32(5120,):07c16cafc4a8b33a3cb0fedfa172f4f86d1ecc6cf95d1f3e16139d5d34cf962e",
"15/input_layernorm.weight": "float32(5120,):899941678133dcc807b347600a76cd55698b070dfd6e32d5125b5948e593a494",
"15/mix.rotation": "int64(2,):1ea019ae0f069f6e396c45bff54a1d344e07f4540a26534a83c391699f22f3bd",
"15/mlp.down_proj.weight/idx": "uint8(2560, 17408):6965dfa08d10a1d5e64c2608a120575764edfbfbef8dc85ae03479010498ee03",
"15/mlp.down_proj.weight/lut": "float16(64, 2):b0b69ec9c800f3e6c69a0ec00fcabd705c749308553acba211ed8c5b69372de6",
"15/mlp.down_proj.weight/scale": "float16(5120,):d500e90ede920816027bf8f821bc4e286ea3ec47cca4f3e1b85204964ce2b91e",
"15/mlp.gate_proj.weight/idx": "uint8(8704, 5120):8481ca4ffdbb47fd2b244cf2deb0c67bc218b14dec73c89b19452d81c08d90ca",
"15/mlp.gate_proj.weight/lut": "float16(64, 2):3eda9809040ba538d5cd61c027154246b52ef0247a54beee92f9a987087541de",
"15/mlp.gate_proj.weight/scale": "float16(17408,):d4f8ffd2951904e9fc3a7606a9ff658a08de284bd8b7417e6335c93e87b57757",
"15/mlp.rotation": "int64(2,):8267c52a206c8e7350108da96adc7e2143e8a0aada645ee2361d306e7c65e612",
"15/mlp.up_proj.weight/idx": "uint8(8704, 5120):387bc1c6faeea05903924acd7f87336e9c8485c8e1e1042218877a18fb50ed27",
"15/mlp.up_proj.weight/lut": "float16(64, 2):1867877cce0357c8932c1b182dc88204ebcab4802eada24edf11e4f92b63e057",
"15/mlp.up_proj.weight/scale": "float16(17408,):39ade1f4f8bd81f2ffa510465462847b894c857d2aef0a5256aa850ec86b1ff3",
"15/post_attention_layernorm.weight": "float32(5120,):91219e60c934c8119049ee7f23ccadabd72f9ae0695819edc41006e03c6a0425",
"15/self_attn.k_norm.weight": "float32(256,):cb8133510e0260f2d7f9502b24b41f6cf14891541d2a79e91539199e28de07ad",
"15/self_attn.k_proj.weight/int8": "int8(1024, 5120):fdba0f397fae503f6eeac1edd03e87a184ef18aaabf4673d6383de456a999228",
"15/self_attn.k_proj.weight/lr_a": "float16(1024, 64):40bb68ae0be9417904a59c7caae8088bc90432bc0c14067242d77b19fe3bf94b",
"15/self_attn.k_proj.weight/lr_b": "float16(64, 5120):02e41f027cc374d78ccec0c66106d51df36166e8b5355d78d0f70586cc9a701b",
"15/self_attn.k_proj.weight/scale": "float16(1024,):ad0b08c047cb8e77e9c1424173df6aeed4d458a913edb44a8550d390b30f8f97",
"15/self_attn.o_proj.weight/idx": "uint8(2560, 6144):1b431574efddb7eb6533368632d0bec85877807c5288f701cfffe62c30054cb5",
"15/self_attn.o_proj.weight/lr_a": "float16(5120, 64):e8a158f578745caaabc574b65e035bf273d462f3e9309c2dc4cfab024ca0302e",
"15/self_attn.o_proj.weight/lr_b": "float16(64, 6144):7638747dd39ca9cfb8925b0e1ac1445ef04651379164bbfb4faad8ce3642651a",
"15/self_attn.o_proj.weight/lut": "float16(16, 2):04e6ac135bb80e18114c98e4b5b57c918af65c84a2c05a4e6226c64ae413ef65",
"15/self_attn.o_proj.weight/scale": "float16(5120,):0ca342da626198c1d2638b632e0d64cc1479dc64e8277d3cf9d130a1e2d29482",
"15/self_attn.q_norm.weight": "float32(256,):67453b73333d4a55129512ef46a799b788dc634b40bad54340d39aa3a9fa16ee",
"15/self_attn.q_proj.weight/idx": "uint8(6144, 5120):92f86f36af8abfe06bac6214ffc276237c29f850030163a29642bc10e7c7177e",
"15/self_attn.q_proj.weight/lr_a": "float16(12288, 64):05bc67f7c7088b2ec500e9cf36bed3b93a88da2fe71b18343123774b85ca78cc",
"15/self_attn.q_proj.weight/lr_b": "float16(64, 5120):7f7b061421ffff47bcef6ba152c086e6aa904994a5b42ef3b758d348a2a1b74f",
"15/self_attn.q_proj.weight/lut": "float16(16, 2):a09894df900b4ba2fa9145bd5417490fc5c2ce9a62ffa6830c3cf46f2c786706",
"15/self_attn.q_proj.weight/scale": "float16(12288,):808cadd30d8d141bb32546a84be5f8ffdda0f3cf404a3b6f9f356f2c990eed3a",
"15/self_attn.v_proj.weight/int8": "int8(1024, 5120):31f36ff6a5ff43477d8915b8c5f4f34314a8537dcf17b2cee1f41fff1072ca7a",
"15/self_attn.v_proj.weight/lr_a": "float16(1024, 64):a6577ad74dfa40775e2fa9210164736edf770648967c52307b1496cff40a6da3",
"15/self_attn.v_proj.weight/lr_b": "float16(64, 5120):8cde372a565bbaaaa05b9452c8a05f00d4e75e32f8b52121fe50d7191987d3d3",
"15/self_attn.v_proj.weight/scale": "float16(1024,):98e90ae1420c78ccf47e47f5540232e753bc5f0bdb6e24737c2f576a19f4fe7d",
"16/input_layernorm.weight": "float32(5120,):477191cceb0c84dace60ec37289ab53c5b745fe65afa55a44ef56c4d8ca9eb8a",
"16/linear_attn.A_log": "float32(48,):d840c424b5b8ec8fbfc86d6abd0d8b01a635fbe2ed5ffdaf3a5b88ba51cb1576",
"16/linear_attn.conv1d.weight": "float32(10240, 1, 4):841f86510c7da77e6739509fec1ec602fd8c7ad20bdb8b13a334d1d80896e470",
"16/linear_attn.dt_bias": "float32(48,):6f078a8ab9335ff4df16d405a853de74091303538b8772ff7e24d1b1ca043174",
"16/linear_attn.in_proj_a.weight": "float32(48, 5120):2c94f90fab146ee74004440e5477f89ad2275adff5f9de38db47a172133029dc",
"16/linear_attn.in_proj_b.weight": "float32(48, 5120):8ea7bc8394fcc1bfb68f26a9fd99216f82385b3fd9d93b581e86f43343ae1976",
"16/linear_attn.in_proj_qkv.weight/idx": "uint8(5120, 5120):c58f13af37db0f4c495b773c12106db5f8f4ec448d6750ed91adc1968d0c8da4",
"16/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):2e43fbddcc9dd6283b4fe2cb369c332c5a7b209b1663f63ef6b92b1791c028c6",
"16/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):e3ece4abec87593f8baed997e57f08bffc411a08673eb505c387bf7fb31797e9",
"16/linear_attn.in_proj_qkv.weight/lut": "float16(16, 2):5728ba1daa46d65f5baa3aca173b392700e3096279b32dd3f8094fb4255f62fc",
"16/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):453acb66a082436720585119e39958bb67a74ab6e06d770bb4f6bf8b0368efed",
"16/linear_attn.in_proj_z.weight/idx": "uint8(3072, 5120):031dddd6fc5f4947ea87fc066a510374b18aee5bd83465cb5f20f7715764f415",
"16/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):3c435c2d7c4d1ae4035fddd9054905f1f23ca83a12751bae3b94a56b460a05fe",
"16/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):7358f642f7e7ee36b2851413284333498829e7afd1e5abfb75e4a1cfd5c81f03",
"16/linear_attn.in_proj_z.weight/lut": "float16(16, 2):11a8e85df36533fcff3bf2a172b431ba6067af627e8746498f23c26a13ca3509",
"16/linear_attn.in_proj_z.weight/scale": "float16(6144,):25517b7e547de72ee65539271eb0e4457c6fa4386951a011fea96429f582fa78",
"16/linear_attn.norm.weight": "float32(128,):387ad397bef86b0a0fa520c6a4ddf97d5f110b1e8c1236a651abf4694528b3b9",
"16/linear_attn.out_proj.weight/idx": "uint8(2560, 6144):e2e884f3a8bae7824cb2e5cb97b7d00ebed25457e0b52ee7b07269b2312bd7ef",
"16/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):aa16e5912fc1f99729142ad0024d371b859fcaebec34959ba5e9c9160b094599",
"16/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):a5d7fd12964ec6f13446a6b16a0211097deee04c80b1fdee0965a599e8292ebe",
"16/linear_attn.out_proj.weight/lut": "float16(16, 2):83cf8f9121f16f8182fa186773a089b7405474ce9bca747af77baa056a7eac92",
"16/linear_attn.out_proj.weight/scale": "float16(5120,):9429217599b72a43c7dc94d521029319d3efb0b3de1b3aaabb1bb0f05daa6825",
"16/mix.rotation": "int64(2,):7cdbf4a9bdc87e340580949ee491d61f036c387d9719b10a00239c61e3f7c47c",
"16/mlp.down_proj.weight/idx": "uint8(2560, 17408):17367e09c41e26a1fdc88f8d6d12d7e94755ed2240e9b8e8e3a5562753907903",
"16/mlp.down_proj.weight/lut": "float16(64, 2):6dcef44471ef4151e695195486dc158b70c19780bcb1df2cc9f8136933a7326e",
"16/mlp.down_proj.weight/scale": "float16(5120,):3b5fa5f428060b2f31aff4ad069579427f97a7c160448719d5ca6a29e6c42e55",
"16/mlp.gate_proj.weight/idx": "uint8(8704, 5120):94b6cb4b317dc7d4fdc537bbe342ffb5804d7d977caae203ec0e3d3ef2e57447",
"16/mlp.gate_proj.weight/lut": "float16(64, 2):70b2890c38ef30e26f6e0bff33fecc85ac8d143103dc6d0e3da9ac1946770326",
"16/mlp.gate_proj.weight/scale": "float16(17408,):76278d11ecd1652c59726a51c12c08c9f38403dc7573cd0e290e2ea77cf01638",
"16/mlp.rotation": "int64(2,):69169c8712990edcf99c08df0e710d37613fc2f7834b75bb72ac857279c3266a",
"16/mlp.up_proj.weight/idx": "uint8(8704, 5120):f1baf93d9343cb35e5d1924f61ea52f30c1a33830f4060def60ec4e4c892ea8c",
"16/mlp.up_proj.weight/lut": "float16(64, 2):6370793b1d2cd3413deb9d22fe7de959eea09d145fd731f6727d97c2d2d5aadc",
"16/mlp.up_proj.weight/scale": "float16(17408,):11509bbeb3505f11e7f7927035b2fed138f1b9422f3847181418117a038273d4",
"16/post_attention_layernorm.weight": "float32(5120,):11c1d3e30ef0a6481a4766724e57459ca70fb44b8a3ffa9a01321d74eca8c33b",
"17/input_layernorm.weight": "float32(5120,):1545a67d0dbd54e32cdf5c8cf187cd2c262925f1844f65bd3224bb16948c5d1e",
"17/linear_attn.A_log": "float32(48,):75c4541d982cb98cf967fb7e7270bb1e22228ef9f39f3bfb8a77c0858051bff2",
"17/linear_attn.conv1d.weight": "float32(10240, 1, 4):7fcc1321a322bab451caa57fe0759ef8a39832a86b2bebd4cab407ccd4d933c3",
"17/linear_attn.dt_bias": "float32(48,):8873049acc40f8d2fdd791c7f9a84fad3f4c00083ad11ab7941ee6ade0ea1d31",
"17/linear_attn.in_proj_a.weight": "float32(48, 5120):d32517562ed9223ce619c6d8b22046e88d42076cb7f133d57f96c75a3f3b300a",
"17/linear_attn.in_proj_b.weight": "float32(48, 5120):0403cbb7d5867c0777bb6dd580f9b1807c7415d47d0024eecaaa9575a7c20305",
"17/linear_attn.in_proj_qkv.weight/idx": "uint8(5120, 5120):76eb73165025c7905049476273de9af664bee133cc983c6e54afb1e1a804a078",
"17/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):fd5b155372cf37aa429e5066f05dab46b608bebb966856147559af9d4226a93b",
"17/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):4e96e7cbde189df0285056b81e0db69d3c7f6decf9867fcd65a831dade93f429",
"17/linear_attn.in_proj_qkv.weight/lut": "float16(16, 2):21dce8f401b3cbae3dcddb5b5f54705ccc6f0795fdc9342f3c1fd4e1331ea49f",
"17/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):0633df68a011af83493869e5d215cfa83647b08b411496570e170d362b6c46a6",
"17/linear_attn.in_proj_z.weight/idx": "uint8(3072, 5120):5cd74ae2b8d7cec4619e9d7e56731a09695fd2bea4b0bbd79f4bff8f4458d6d2",
"17/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):979b98a79034cf42dc031168a69430cb56868c33ccb073b781f273e6efb2c104",
"17/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):7d0c635536303e1e8ec4d2a040c23effe19dd5ee43d381ee6ffc976d14b3d617",
"17/linear_attn.in_proj_z.weight/lut": "float16(16, 2):b55741644b71d749a9f3c30a2f9869f446240ba75cb7cb9ab764bd6fede0e37e",
"17/linear_attn.in_proj_z.weight/scale": "float16(6144,):2e8b2f0afe8979d3ec64ed0a080d9748fd7bf9b4daff6b45fbf3831a35fb1f2b",
"17/linear_attn.norm.weight": "float32(128,):7fdb4bc7b06b6ced929a0e0995b583f133fe60080467984defdc722ca633eb20",
"17/linear_attn.out_proj.weight/idx": "uint8(2560, 6144):f5ffdd495c10443809414191c38dad2ff28797c4aef1af4c05e91405b93ab2dc",
"17/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):6d7d17ae3bf46cca2ea624a70e10bf158e02ee39f99298246e290d9136652530",
"17/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):36a0689094c118be2cb808fa8c4135c265a8e0595da39a6cde68dce5bf69ec5c",
"17/linear_attn.out_proj.weight/lut": "float16(16, 2):6c34d0b2d6805d4971146aea1848a8d9e0b42d41da9f5463022a885334f434fe",
"17/linear_attn.out_proj.weight/scale": "float16(5120,):917b7a16b28d640b4ae6c0b88240469f8efa9c0087abbcf2d5e4f299326219f4",
"17/mix.rotation": "int64(2,):891d9f31a5db202b56d1c7513ed6c3d2337fa75a751598f2ae51ba76dceb18c5",
"17/mlp.down_proj.weight/idx": "uint8(2560, 17408):1d76a2284daa729b690c1035ec7f70c90466f6bbde912bb397db753fe42391b3",
"17/mlp.down_proj.weight/lut": "float16(64, 2):57ac3bf2700365ce6b5afa4a99df36946d3dbb721fe588d0b99db65797742950",
"17/mlp.down_proj.weight/scale": "float16(5120,):0e5af909635e7f0ae08fb72fac95a1f3c423542ebde25bb323beeed804e676b3",
"17/mlp.gate_proj.weight/idx": "uint8(8704, 5120):ce1d0a527edbe8c7046d66485a8488aedc7cbddcadec87bd6e9d668abc1ce245",
"17/mlp.gate_proj.weight/lut": "float16(64, 2):0f40643b9416e1ee05a24350c1f9e6b3e7053a7fc70601aad78403f492e8fa44",
"17/mlp.gate_proj.weight/scale": "float16(17408,):87d8fd51dd4eb5fd22324bbacca905575388050406009fc9c3c3a5d9d137b7d8",
"17/mlp.rotation": "int64(2,):227a884225f9dfc8ea9b96f045cd88371af72d11ae6796c2637a3420a56ff2a3",
"17/mlp.up_proj.weight/idx": "uint8(8704, 5120):747b823c9b48ff9f6d538983c3d46af76474161a8a1850c688f17d8c14549be4",
"17/mlp.up_proj.weight/lut": "float16(64, 2):66385b23ad98f79a051e3144ba7f26ba04d6b65e3ce9034ad8c64fb5237c3cce",
"17/mlp.up_proj.weight/scale": "float16(17408,):9924ab4ca3bf17b2b5842326c35794ccdb3609ff6d6ffc58c04ccb3581d5dcce",
"17/post_attention_layernorm.weight": "float32(5120,):b7c2d691774e32ffdcccb6aec2f62af925995da493dcaf188a0c5637ce2f4900",
"18/input_layernorm.weight": "float32(5120,):ec7c4c659d54b87afa7979e205d45725dfba198790ad9ef9889048fae7de27ad",
"18/linear_attn.A_log": "float32(48,):7b433ca625d07cbe0306ec8de65edf70d522db1c70b070818ed54ea934bd613c",
"18/linear_attn.conv1d.weight": "float32(10240, 1, 4):503fe9b8e0ecc3b8a62af6b58f8c15c22855f6726c8f959a9987d7efd647714f",
"18/linear_attn.dt_bias": "float32(48,):990e1a0ccf0023a76ac5f7b5b834f133a2ff4e62a09aee2f1ba77b2247d756ae",
"18/linear_attn.in_proj_a.weight": "float32(48, 5120):9111bf56b67cfd0753ea1b0fae49a2145774312666cae6d434d37deecc6c7cbf",
"18/linear_attn.in_proj_b.weight": "float32(48, 5120):1fd255351e3670cfe2ba24a19769965a2f39198da59d0027054317f920f93012",
"18/linear_attn.in_proj_qkv.weight/idx": "uint8(5120, 5120):7399a9caceada782ae50e8d1dbf644d61e961323cbb718ad4fe4963d082f0582",
"18/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):f8325380982a5f08331ee58ec12bd6bab324359ab5fabf8f527a553798fd8d50",
"18/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):02e2b916d8505ab72cc6f74d3e4315a0ad25ae720bebb721c38eae34087e2226",
"18/linear_attn.in_proj_qkv.weight/lut": "float16(16, 2):ef3e1974c5fb6b87e2768f470822c6873e0afa9c6f285869b7ff0c6dfdbd94a0",
"18/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):8d6ceb59f19e29e9d407331a0f2f4498a5283124b8b99b92cc36f502b8d53bce",
"18/linear_attn.in_proj_z.weight/idx": "uint8(3072, 5120):83c88a53ee24e876bf22dc49de0b34d91d127841bfe86883b75ea25a0ed85b87",
"18/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):026c947177e8e3d929b8019f27ebcf71d8cdeb2469d34617fc30a3d4b23e3b71",
"18/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):ec5105da05ff48c3a4df08a945dded29890a1b5e999b22b6f4abfbb5582f100b",
"18/linear_attn.in_proj_z.weight/lut": "float16(16, 2):30a298de2de167712217f751ce59db183eb07e194893c79f8082c90cb9911a05",
"18/linear_attn.in_proj_z.weight/scale": "float16(6144,):6d8e631d671d8960df4bb3331ac000aa52d2dff8b72929d1009e64083d44cf05",
"18/linear_attn.norm.weight": "float32(128,):9de1581342dcd07527419b2bb67feb654368e8ccd9381e9552317f6dcd7c249c",
"18/linear_attn.out_proj.weight/idx": "uint8(2560, 6144):df1f8ff74d88c456ca41cc152b5df30efe7600b9445fa65bb1980c5adaac8810",
"18/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):25b4a148f242a0c0ba70c6e663744af290ac53653ea32b9539116082ccc28cde",
"18/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):1a1b7063d7fb266b9dbe4c28df305632045dca2431c56be5b64ab2da61a9387c",
"18/linear_attn.out_proj.weight/lut": "float16(16, 2):80bc4422f3b939fcf67d97bf0073e43f64b0ac59d2fe3fd85af5403c7499b3f2",
"18/linear_attn.out_proj.weight/scale": "float16(5120,):e4b1c31400a0c3aafc069126f364341cc4bcacf32fe2dcc3d179a414d94119b7",
"18/mix.rotation": "int64(2,):71911fa383bcd7a4bb38b9a7c5413a1a96e0eba965282e20731fc417f27243aa",
"18/mlp.down_proj.weight/idx": "uint8(2560, 17408):ae03edae7deca79f6197a7af6e8d3b338a02a80fca51de8253cb6d7b0ac9da4e",
"18/mlp.down_proj.weight/lut": "float16(64, 2):298dd1e2c6e04a28db3d7bee78bf71629eb1066641a6449e90656a1fd5ddb3bb",
"18/mlp.down_proj.weight/scale": "float16(5120,):4a311704ff06dcdaacecb6df808d86a463ee05b99ba73a57b009f16d58f06a54",
"18/mlp.gate_proj.weight/idx": "uint8(8704, 5120):c6ee858765bce02c6dddc5fdd859a4b8bd896e0d4f4b8843a96c8545a9965a31",
"18/mlp.gate_proj.weight/lut": "float16(64, 2):33cc822116d3dc4d9dc586715ebff94de6d402784dca53fa25c04bfdbad0a38f",
"18/mlp.gate_proj.weight/scale": "float16(17408,):ed0be16ebd073afdd4f78e8caa3cd1387742c7c49b11e7ced45cf51511a0a5fc",
"18/mlp.rotation": "int64(2,):29a5e5fbe9e69e9d8d42d03a712b0b863081de1de1b0b78f77cd7e2b0535bcb7",
"18/mlp.up_proj.weight/idx": "uint8(8704, 5120):013159b718c92174dc7b9d25b07c30f139aff84394f04584ea64c15d94a8e7e9",
"18/mlp.up_proj.weight/lut": "float16(64, 2):e3d84041eef70314b2f7aa35c31e3e5a64fb00d6499f385abab1f425b244e93b",
"18/mlp.up_proj.weight/scale": "float16(17408,):970d1a9ca51bbb816d3cca8bc627d3824aa8cb4789970c4ff7aec49fff2427a7",
"18/post_attention_layernorm.weight": "float32(5120,):b87b7a5f9ab6e571b561a1b0a64d4c60cd653628ba90f9d7986284249c746a91",
"19/input_layernorm.weight": "float32(5120,):4aa053720de3054a7640d3e1bb2be69ebde861c51d4b73e8d2127059238f4ecb",
"19/mix.rotation": "int64(2,):45bb98ea9bb13ec60a36e1e508e101eec3b3497a242410dcb85aaea83a150050",
"19/mlp.down_proj.weight/idx": "uint8(2560, 17408):7ca009e238e058939129927bef0722d0301f6f8a1e1d1dcf9a509dd71c9556b4",
"19/mlp.down_proj.weight/lut": "float16(64, 2):0c873564650f0346b6631b4dc8195587e7d77e57b5fcec09660eea24a4377acf",
"19/mlp.down_proj.weight/scale": "float16(5120,):deaf57160b0abe5f2c658a9b013ffb86bc5a1908d43d40058212fafce85c95b5",
"19/mlp.gate_proj.weight/idx": "uint8(8704, 5120):be6e1e241f4e86fc78b2fba4560b018fc2f349dcf0f1376e8276f5024013c3fe",
"19/mlp.gate_proj.weight/lut": "float16(64, 2):d9570f21dc14691b360939e92eeb515c4b04770d1765e7d702aa54da807dc69d",
"19/mlp.gate_proj.weight/scale": "float16(17408,):3919bc9189f8e7ecf0741a8a331d84875d22fde069de5305da55f6fd7fde0b3f",
"19/mlp.rotation": "int64(2,):c74b11128c61527d3bd7f70c16ba9f719f33eb8d1c9a5fe385d7aaaf94437d3b",
"19/mlp.up_proj.weight/idx": "uint8(8704, 5120):9800e081ef8c947db44682e17546c96badba898746fce4e5418757e4730cf3b4",
"19/mlp.up_proj.weight/lut": "float16(64, 2):0d0d795a1c604bab1ebd908689373419dc4c3cbc6817df87c4da7b88a6eafaf1",
"19/mlp.up_proj.weight/scale": "float16(17408,):194008e251ce45fde6462bbc525600c1d602f0eb43c109f590ac5029be5a7258",
"19/post_attention_layernorm.weight": "float32(5120,):cdd235f02cf8168aeec79f7215b73733bc2328509a8f444e9b80a58fdc0a21fb",
"19/self_attn.k_norm.weight": "float32(256,):e2862a1bace433b3bf5952998be293d2272fe4bdfe911a6d1b3e5026f2e1bbc3",
"19/self_attn.k_proj.weight/int8": "int8(1024, 5120):eac843d9c98a17712405765d6a4ad3666fb3a9b9215eb757c6729fd3311a6073",
"19/self_attn.k_proj.weight/lr_a": "float16(1024, 64):d6d684cdb25ee85102b58a74167660e510ab88ba632b4663c4c5010f3705e49f",
"19/self_attn.k_proj.weight/lr_b": "float16(64, 5120):8319b0e1fb62298c19334b4aa54f9c9129d0bd7d8ba66afc346f1d9582093c8e",
"19/self_attn.k_proj.weight/scale": "float16(1024,):635d710346f8e8f908db09b9984ff0a814d76eb7bac2759367481725e75a1d95",
"19/self_attn.o_proj.weight/idx": "uint8(2560, 6144):bb01747b25120e598da59832289a869bed940acc5efa4b10ba72553cf6a794dc",
"19/self_attn.o_proj.weight/lr_a": "float16(5120, 64):c648a21d62f1f5155bedba565aed0e46a649010c4fc7dba62c5d906bee6d758c",
"19/self_attn.o_proj.weight/lr_b": "float16(64, 6144):59d8df09895e1fad09d0401e0584cb249b9bfc4619f41f453d8578dd0adc87b8",
"19/self_attn.o_proj.weight/lut": "float16(16, 2):0e7e806eab32ec07a1d43dd00084a28f450706d64a944cc412161623239ab963",
"19/self_attn.o_proj.weight/scale": "float16(5120,):ec09f0a0c6794487c88215354c3085efdff17a7e4c48a23353da37537416625d",
"19/self_attn.q_norm.weight": "float32(256,):c9f4f09bf48aaef386fec1a9ddd2d0ff7b9056e3b66aa9b35a457b44ab520f2b",
"19/self_attn.q_proj.weight/idx": "uint8(6144, 5120):8bf78d446ac87c51fcd3edd7e64054299e08fd894b867ada9fe2fe813cfe81f1",
"19/self_attn.q_proj.weight/lr_a": "float16(12288, 64):2a2a1a6b8ab931b55e209f34cd89ba81f02856efc657ad11dd2e9ae54540d48f",
"19/self_attn.q_proj.weight/lr_b": "float16(64, 5120):af3e51ac4a162e9769deafd61b0c3a02785ca7149ef7f3a70cec075c52b72376",
"19/self_attn.q_proj.weight/lut": "float16(16, 2):6ce00fa66aab74464cde513935e2b1e96c4a3a33faa6e4a1e59ee2aa26d17294",
"19/self_attn.q_proj.weight/scale": "float16(12288,):bf2326cdeacc6616d2652ef786d6c599ff08d8dc907b7e76aa3e13534a54d0c7",
"19/self_attn.v_proj.weight/int8": "int8(1024, 5120):58f0d96aaaab068302831dca7e3be8e7b927018dd36d15401b6495cb19b74e3e",
"19/self_attn.v_proj.weight/lr_a": "float16(1024, 64):1e526c9b04f649fd589dea9050d2175e7ee0dc5cf3e160a10c16b30d070f1e91",
"19/self_attn.v_proj.weight/lr_b": "float16(64, 5120):d7de187957c5ceeab50a139d9be1ffffda52752b1a4f4dfde88cf0ec92c0b207",
"19/self_attn.v_proj.weight/scale": "float16(1024,):808b05e70e41fbf5b5e0a86eb9a05f0c850f369e8987b116858f106e98bd55e2",
"2/input_layernorm.weight": "float32(5120,):20cd7a8eff968c98436475eeb728efac3aae682a5612dbc8a5591c2a3b9d737f",
"2/linear_attn.A_log": "float32(48,):7982f9fdf59fb7f2b4111b646063fc484d2009dc5866505984537e9c8886e372",
"2/linear_attn.conv1d.weight": "float32(10240, 1, 4):bc6eefc788dab0795ff168af8ec28eedebfd0eb8c04cc9550f0535adf6c7011e",
"2/linear_attn.dt_bias": "float32(48,):67fc93cd13f11e8dfb1188c6862c65be6ddc63ef5903ec0ffbc7ac39f48dc0f7",
"2/linear_attn.in_proj_a.weight": "float32(48, 5120):3737dc37a8edf6a798ba9d89d3cedacdbf146bad3ff3a932226b4bd2aa77eeed",
"2/linear_attn.in_proj_b.weight": "float32(48, 5120):18cece2f58e5cbefeb002018cc453b7eb4b461efc2b73b31dd2e579fb9530e76",
"2/linear_attn.in_proj_qkv.weight/idx": "uint8(5120, 5120):7ac27a02f7d90c7b216b0adc872fd76eea1ae23d82becf57062c007284fd1bd2",
"2/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):c824e58cb50799be0027dc05668c9e1f23ac06431587fde76e40ec3dd21a4f4a",
"2/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):8d2d4826a97521d8f42297bfb1cf611f38f2f82ccc2c309e5850b6e7d7ad490e",
"2/linear_attn.in_proj_qkv.weight/lut": "float16(16, 2):460623ccbd0d719fdefdfe98d87a252d500365e4cdcf53b460362baebeaada2f",
"2/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):ed6d4b3bbc62e874d01c056ebd5e9a886aa391144000111f0a43f4f4041b9c71",
"2/linear_attn.in_proj_z.weight/idx": "uint8(3072, 5120):bfda0d1b72c4da7b97026b4a0654afa2ea5d49a9c7c8b49d1865a9572ab8b450",
"2/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):1e4b79a40e41658725740fe8900d447940369c05106d44f3b9376d1ea1e2f04e",
"2/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):5e61368aee86a35c60353a8d29b63cb07b6fd1d946ae357d99aca641c26e1389",
"2/linear_attn.in_proj_z.weight/lut": "float16(16, 2):cbebaee182dc0b0898d9ce72764ee5b07ecf03d2513f8c08285cf8ed03b0ac3c",
"2/linear_attn.in_proj_z.weight/scale": "float16(6144,):9c45004e0f168fcbed67cf46d9c2c5893f5479909b01b1649ee8b551e3df821f",
"2/linear_attn.norm.weight": "float32(128,):3ca0f3c06fb888627198498c689988e51a4711e829b8fa6cd025eda105ec6e90",
"2/linear_attn.out_proj.weight/idx": "uint8(2560, 6144):5f833d83b9c0a95974f89f995008899b7fb3ce0652e6c0a645060aeb11086839",
"2/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):6392eb2ca86fdebe4dc5816aefe9bf8e4c7a3323637074d6194e48bce41a8799",
"2/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):17af94bef87ddcbb6fabf1db11dca1f8979829bf23e95f5ed12636146df5c7d8",
"2/linear_attn.out_proj.weight/lut": "float16(16, 2):fb14f811861b243f8438fc9fafc4df70535006cacfe75f89d68d7f84187bb1a7",
"2/linear_attn.out_proj.weight/scale": "float16(5120,):48804f8ac8be05fd6f0175c0bcedbed17c2bf9fe1e3df8ee7975cd48c27f6972",
"2/mix.rotation": "int64(2,):acceeea95f53e147d243e963de5f029111ec4b7bda2ebb92fcb010e496b91293",
"2/mlp.down_proj.weight/idx": "uint8(2560, 17408):e0eaffd715d25321925ef1ea2e8f8c1f749ddb585e060c4db0ab5e1141df7c3a",
"2/mlp.down_proj.weight/lut": "float16(64, 2):ae3497c5f97ee5641e6394396f0655119ff5e9ca84bc1633cd94f510145b5e08",
"2/mlp.down_proj.weight/scale": "float16(5120,):d3d3fe5e853cc3f2a366d79ea1e4aa3046c77c073c97307e354bf7c816fd81b7",
"2/mlp.gate_proj.weight/idx": "uint8(8704, 5120):000700212571eeca0cb28efc1f27239e0198b4f2d6d554adb728aeed244065a0",
"2/mlp.gate_proj.weight/lut": "float16(64, 2):65883beca84c2c897f544a445d411702ef4c07f62f47962fa1a2bde4abd7ada2",
"2/mlp.gate_proj.weight/scale": "float16(17408,):732ae7b0ec36ed7b2dfc7e3539296ce3a470572937172b869af20bf18d789201",
"2/mlp.rotation": "int64(2,):e50f77a92461e7d8a552c9cc985e9f7d56ffe8abd0dbb98f82d61b655c8bab9a",
"2/mlp.up_proj.weight/idx": "uint8(8704, 5120):4b49f03c25557db815c65697512c976b18ae7abb3ae68bb00debbd64ba4c8c14",
"2/mlp.up_proj.weight/lut": "float16(64, 2):451794bd91be7dc6d9f216c13c26fdc7214d9b63b239e555acc45cb342b30b27",
"2/mlp.up_proj.weight/scale": "float16(17408,):9b1a6e047e37109ed6946bbb8f9c58f5da40b21b75c049182ba89bceb6f2a134",
"2/post_attention_layernorm.weight": "float32(5120,):0ae8a701edf67499225f6a0a108f5039bd9115e9a2794f9e3fb3522c3070fa05",
"20/input_layernorm.weight": "float32(5120,):7417d7d44bcc845c0b35609f2f1f3811b92b623becdf6d605648a5a25c72fa1f",
"20/linear_attn.A_log": "float32(48,):0e862dfa772de794a2d465dec537edd04ac051ab041c250367c5f631a07ebfd5",
"20/linear_attn.conv1d.weight": "float32(10240, 1, 4):63591d8c854837339329bebde8311184f78d90120cd30873c50b561b5dc6e930",
"20/linear_attn.dt_bias": "float32(48,):5f5f16a07351457d82d9d8f603f1947d78a626e9d5d540ba08b3ef9622944bef",
"20/linear_attn.in_proj_a.weight": "float32(48, 5120):9ec16c537d3ba5287c578850291dcdaa864a1890756d47fe16675214ba4d9642",
"20/linear_attn.in_proj_b.weight": "float32(48, 5120):f2803869c833d77a14146d961ff610c91f7d365d31559ffdea0d403b8224ec84",
"20/linear_attn.in_proj_qkv.weight/idx": "uint8(5120, 5120):edaebe5716c5d90949a56f88d4cd57541cdd0f03e959cec7f1f2fb8d7e8ea2ec",
"20/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):3b6e99d069854ca3b378ca7d6c384b8a4a51251969bd1cd5854f9bd86d209a63",
"20/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):50e26b7f6bb99965ad84b409ecbb58f0d5435aeda801641a2a332372277a6183",
"20/linear_attn.in_proj_qkv.weight/lut": "float16(16, 2):c891436f248dbcf7a6665da74eee06652649bb860b0f339c765bbfb67f62fa52",
"20/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):7c5d3b74186b5e1eae95d593735192aab6d94368b69412dc2747921bdec07af8",
"20/linear_attn.in_proj_z.weight/idx": "uint8(3072, 5120):efdc6a419b865e0a2e0cea56e327bf5bb7a212dcf438f5cdcc2f640bb4fd74ae",
"20/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):98bd351518e070101f4edb1f8fc5ccd5634affbeaf204e4c3b6368537d275fc1",
"20/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):ee178a0c58a07830bda5835c77a94c7e83cf8f249f14482dd38d744c241678a1",
"20/linear_attn.in_proj_z.weight/lut": "float16(16, 2):50d3cfd8db621ba1cc76dc7d318a3da549dbf00044ce7dea9cc91eae6041e46b",
"20/linear_attn.in_proj_z.weight/scale": "float16(6144,):48ab6533d4174cbb19e99b18d5433d772e3e41a5cf5d67d2c68c541a1bb88090",
"20/linear_attn.norm.weight": "float32(128,):a7868b74be35bfc81a24f923d84101afce53929347773e5ee3fc5ca5c9ed7465",
"20/linear_attn.out_proj.weight/idx": "uint8(2560, 6144):a64fa7408ec236d4dce7659578881f8232659c8be30ccbbddf0477ee8c979bb0",
"20/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):1e3737385e95f1f6860cdf5538c448f0746a804df6ee4dc500a41f564a177afd",
"20/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):56ec9489d811bfb17e3e9955c28949da919929db3f45e2aef2fc7b54d8bc00f9",
"20/linear_attn.out_proj.weight/lut": "float16(16, 2):1a64968045e6e1d7eee6e69709b0d03a2d4d5efec2532ceda2af6af9157586d6",
"20/linear_attn.out_proj.weight/scale": "float16(5120,):a179d11e75741569abf74b4e25b8b515d28a496a98f799c620abbb45dfb70154",
"20/mix.rotation": "int64(2,):dd4e4c7ab7be41b58d3a160fd99af96b73513b548d007cca344fda72d2e6ee69",
"20/mlp.down_proj.weight/idx": "uint8(2560, 17408):5a2491990addb43b1df66013adb25864ca30e8224a5d6b575bf0caecb19ae1b5",
"20/mlp.down_proj.weight/lut": "float16(64, 2):5f5e68f952ed42e81eeb8a8d835861f17aa80bd6c8d55a55795959a704594d5c",
"20/mlp.down_proj.weight/scale": "float16(5120,):ad4ce8ddafd8b86b3b25d53df8ef077900490c820b09f7cc079e9e625c170c9b",
"20/mlp.gate_proj.weight/idx": "uint8(8704, 5120):41f7091f33377a18b0c1936ffbe4b529b1eb28c0965e7faa7a058fa9227d71d3",
"20/mlp.gate_proj.weight/lut": "float16(64, 2):909e920b5655cd3d5557d2eb6dd38dd29691740b972ac41f04bca0c71314f8ec",
"20/mlp.gate_proj.weight/scale": "float16(17408,):467004545039a076630e8cdda0e7b16e4a1168026f58aa39d0d685bd1e771511",
"20/mlp.rotation": "int64(2,):cd98d53929bcb0df98a82ac4c6f7eb2c9f060d4f210e163139f1ef3e721eb388",
"20/mlp.up_proj.weight/idx": "uint8(8704, 5120):63e54f1c71d0aa8929f1001b50f1cd1c003256bcef18ea4d85f8b9151aa4b9af",
"20/mlp.up_proj.weight/lut": "float16(64, 2):477f24039ef5a126e9596715208d4a63dd1a8b0ab8e45e8886e2b3b72e3bf938",
"20/mlp.up_proj.weight/scale": "float16(17408,):cc2511b35908def31921a5fb3184c01ddeb0ce3120075f2d53b7d62fc37a0e1f",
"20/post_attention_layernorm.weight": "float32(5120,):e4bb1ff1674b0b300ef97852b75f58965ac3914bc5aff0f248b78d568990d49e",
"21/input_layernorm.weight": "float32(5120,):c92524f9f70cfded8594f499a19292884302be3f9b4068c79d311a463a502c0a",
"21/linear_attn.A_log": "float32(48,):79caf7b490a9fe160a26630a9027078045bb30a4042291814be51352c955aaf8",
"21/linear_attn.conv1d.weight": "float32(10240, 1, 4):886fa3f4169c764a32be687f6395a8b2e3739beecd5cd6a26ae6ace5cc43019a",
"21/linear_attn.dt_bias": "float32(48,):f8c0428f5eb842eac003c7dbd2dcd6b3841eb01f335c0f949236ee2168c73a81",
"21/linear_attn.in_proj_a.weight": "float32(48, 5120):3c8e8dde2b898ea7a8a0fb6b537c637a0f52b16a1d7d602bbdf866bdad634118",
"21/linear_attn.in_proj_b.weight": "float32(48, 5120):b73980038f554b6db34100fbf36b9e8e9433d50bb090d350858d36c1be7da6c3",
"21/linear_attn.in_proj_qkv.weight/idx": "uint8(5120, 5120):3292c1a1cf300aa706bc406e322f04b8ac719cdfaf66ff162b00e5384bbd14e9",
"21/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):42fdd13ff72084c0254e88c4227903a0bfc5ce1738e8150f1a1339732aab3f8d",
"21/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):564829e063600d51ffbba5c80088751d9dafaa8a49351fc1814446839228ae18",
"21/linear_attn.in_proj_qkv.weight/lut": "float16(16, 2):803186baa6efd48a4bd8ac1e673edea872a0717c674cf18fad92761c219a82fe",
"21/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):edd609a5ec3ce56a4c0dc2ee10f9250e61c46d213b4d8be8ade0f264c7eca470",
"21/linear_attn.in_proj_z.weight/idx": "uint8(3072, 5120):3dd936f280d8d7d0005d1f6687397203ec8bd401059ad61e9a29dbd1aec71c0c",
"21/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):71de5470d4f740bc9e16da68584be7bee829bc3b24dd80207b1aff3950a85ce9",
"21/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):ae9d594f7d7e930bd081aa02ff5bfc6f44a31b68be346e3df4f24198dc68e4c1",
"21/linear_attn.in_proj_z.weight/lut": "float16(16, 2):7434b9c1da9c722ffc5cfb0dcfb3e076f7801b7f1ce7ed1d7efd812d01b43a8a",
"21/linear_attn.in_proj_z.weight/scale": "float16(6144,):bea08e043273c74d810d65710f5091e55bb16167e2ac0d87cce28a62d957bc06",
"21/linear_attn.norm.weight": "float32(128,):cdae9f45af0c3442a5e3ac9375369c80530f44173f876694e210b9b328864d27",
"21/linear_attn.out_proj.weight/idx": "uint8(2560, 6144):edcb5214e5263c5ef633bb56f835461a7b501f842a481a3063ffb866800c5c54",
"21/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):779fe907e91a96f2a4f42b3a075494dc9a739369cab518915a0cbe7ea9393e37",
"21/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):09b46cea08c55b75bda5a23661de9d72ede935425d4e0642a314c018952b4667",
"21/linear_attn.out_proj.weight/lut": "float16(16, 2):32f44cc5a453b012a45a1c2df3399aeb7008a49a16869d956f02c5034c16f5cc",
"21/linear_attn.out_proj.weight/scale": "float16(5120,):62f6e1466d6b54729ffd42360c96ca33ce32d0122836392733da1e903ee23bdd",
"21/mix.rotation": "int64(2,):c8a1f08f12ae79533ff3cb8567530324d160c08d5a73f23536f0cdf6f2cbd5f9",
"21/mlp.down_proj.weight/idx": "uint8(2560, 17408):a56fda91b6bf92f2e3c8b9792229f8d3dc54bb747fa4cbd939bdf8ecc38057b7",
"21/mlp.down_proj.weight/lut": "float16(64, 2):c823fba387438bcd8f733cece5c7dde6e6fb34e20db5c029b0a501a4ba596983",
"21/mlp.down_proj.weight/scale": "float16(5120,):7bd0f88a3abd20e1a5f3a23accf099e1a648de340aa9056eb2a7ea78d4818f1f",
"21/mlp.gate_proj.weight/idx": "uint8(8704, 5120):e4ad528db05872b46e9998b5f4246775ca38c4ac7d3200ee2ada79e65660a155",
"21/mlp.gate_proj.weight/lut": "float16(64, 2):6bb9cee44fc6784449cb6e88c839167cd5ddb7da05306c1890945a1922060b6a",
"21/mlp.gate_proj.weight/scale": "float16(17408,):b460a078290d94a2621dde1ccb31ed6648ab7dd5584124c69ef4b564e544354b",
"21/mlp.rotation": "int64(2,):4df8c17d79df2dd8c0e05dcf1fa5646de706dd385d24b24aec5275383dcb47e9",
"21/mlp.up_proj.weight/idx": "uint8(8704, 5120):2e0f7129c9d827ef9d10dd7c183e95b6a1e03e82884cfe093f398d6b78b75a4b",
"21/mlp.up_proj.weight/lut": "float16(64, 2):c66a64464a8259e0ed0fb977083c6154d2d380e7afcbe32d0811b7df59aee033",
"21/mlp.up_proj.weight/scale": "float16(17408,):73969d0184f2a916f9cc6560e294a103308c7200757945a7c87954541d6dab9c",
"21/post_attention_layernorm.weight": "float32(5120,):383bc8d98ad78e44db856efa110d0b046dc5e00e31f2805d5f06d50bec95dd5a",
"22/input_layernorm.weight": "float32(5120,):0032009af190b3f14db1c9c9f495ba0ebf6411df7c7751fc71b71988824395fe",
"22/linear_attn.A_log": "float32(48,):42624e419008152cfe3b9d5c7bcf8998d3cb6d348c028508305905f4134a0b7b",
"22/linear_attn.conv1d.weight": "float32(10240, 1, 4):e11b1fc31e4739367fa9329ab843649e88936296b2b509a8749e8e97ad7073cf",
"22/linear_attn.dt_bias": "float32(48,):3f52cac33d7d22fc847f0c5d570fa6220fca451eb5c3dd38ea7ca0a266cdeb99",
"22/linear_attn.in_proj_a.weight": "float32(48, 5120):ecdb7a0ef2a28d11ee22b691130fc45bda22db30f0a15c92c17339449791f46d",
"22/linear_attn.in_proj_b.weight": "float32(48, 5120):43cce9177eed85b39f1d79f911bf0b263a96b33f57a7af1f7e183ebcaf7d5607",
"22/linear_attn.in_proj_qkv.weight/idx": "uint8(5120, 5120):05910af0b2e6c0f5ebf1988127cc455fcd33468fedb200f978a36f680a3b1f67",
"22/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):8ac25a2b3599dc4657218b29d7fe886e0405d3870ef898338b5654e67d977200",
"22/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):50b05f60b7e15dc8a5c691c37746edf86069733ab7993930f938f27cf95b64f4",
"22/linear_attn.in_proj_qkv.weight/lut": "float16(16, 2):bb4c15c948bac920a16cb3fea2f51c69568bebb6740c1b08391876a19d97b0bf",
"22/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):3b7e11a48d360432459ddc1e200b1ca4780886c8753e8bd47b1528c41c371499",
"22/linear_attn.in_proj_z.weight/idx": "uint8(3072, 5120):bebe5f6b03fa002c1db948e6a66c3e67a7ef06b85e2d82747cb9fc3b84768510",
"22/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):6b8cdb212f7123508adec5f9942e46cf02eac509a0080189f5ac5f3114c76066",
"22/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):e0a6175d44ce9ead817f0e8c53e13497f8ad3a18d8507e7d02147be115a96915",
"22/linear_attn.in_proj_z.weight/lut": "float16(16, 2):20881bba66b3144a81616a246a514528f82cbb380137c1998f80a8ff5ac958f2",
"22/linear_attn.in_proj_z.weight/scale": "float16(6144,):ccd2aeea228bb8a7c820b70c05f71167e040ccab97f8ba65b618e1e1830e6380",
"22/linear_attn.norm.weight": "float32(128,):9f301b92defac70db25a6e2de2a26476769924f70c663e16d390b9032d33153e",
"22/linear_attn.out_proj.weight/idx": "uint8(2560, 6144):5b1732c7ecbd0553d53f9c795c1d92aa7e0651ee14327563cb81851e6ee59240",
"22/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):2125d4c84d30eaa09e51f1bdaf0aeb461fa135056f1b4aada16247a0c564de04",
"22/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):7cdeda68c669fca3d9d262262efe0c77d32d6df9bf45b20d8ac633079b40ebc9",
"22/linear_attn.out_proj.weight/lut": "float16(16, 2):202e386f6ecdf5eee4f5fbb4660c139d4a6b32f74c45557940b5d3deb573083a",
"22/linear_attn.out_proj.weight/scale": "float16(5120,):58d06ee2810ec7f1d3cc008cef77c47ec10588c9ac5d20a51af47d99ad5b64a8",
"22/mix.rotation": "int64(2,):afe039e0ee481280ea55b72df355b50c1ddcd8db7bdb6f8a0682a0a05f788a53",
"22/mlp.down_proj.weight/idx": "uint8(2560, 17408):9222f7dc49401564bae8ae2cf78b83fa645e72808179ad21bce4b6ed9f445e48",
"22/mlp.down_proj.weight/lut": "float16(64, 2):7714fd199cc0db0918796c03a5d8d37960053935a937cbe9427c0f2f59a4b023",
"22/mlp.down_proj.weight/scale": "float16(5120,):53e8595917265bf820a9aa1149dbcbbccaa8127bb7620a2a180f5c3261bd32f9",
"22/mlp.gate_proj.weight/idx": "uint8(8704, 5120):77bd8d78bec7362e567fa98a98020f288a757745a015210443c0f38a6dd119ec",
"22/mlp.gate_proj.weight/lut": "float16(64, 2):f3318ca6867aaa630a8dc023c73df24c9cc445eb0158a789f83a949806bb4ee6",
"22/mlp.gate_proj.weight/scale": "float16(17408,):cece9c72403e3131280422d972b0caaaae372438a6ceee360e108d370ef6d1ad",
"22/mlp.rotation": "int64(2,):add6c8586d8b44f3c2dcdc74448543de7993ef8ea0b9a9367c2f97381e62665a",
"22/mlp.up_proj.weight/idx": "uint8(8704, 5120):11ae762d5ad243a485a07e52e8a3c1ac0b4c7242eef7f824e38e8d9c153318b9",
"22/mlp.up_proj.weight/lut": "float16(64, 2):27e8502edb928455463fac2cfcdea78ba886ca11472dd33b4d633479cb5f934b",
"22/mlp.up_proj.weight/scale": "float16(17408,):088267c28cf7256fb96c98cfeb9a7363beea6046de419fc613708cae32d112a1",
"22/post_attention_layernorm.weight": "float32(5120,):b0e9dd4257a3ca5f2b0e404775771cacaeedc447fe0441bffe88059a91616065",
"23/input_layernorm.weight": "float32(5120,):f2cfa13c53061776d2edb0e8af687d103a9bee2a90ce75000f5046488ac3d151",
"23/mix.rotation": "int64(2,):3648273dc5999552d73da9113117f2b928468feb152030faac45c5d816e7200a",
"23/mlp.down_proj.weight/idx": "uint8(2560, 17408):6e8dafd8983cb9a608776739fd3c7625c615838cb45130881ccbe2d7fc783638",
"23/mlp.down_proj.weight/lut": "float16(64, 2):3abc5fc1cb8df07d71e101591ea7fc3fa7faa413501b0534579f1af2843d57ce",
"23/mlp.down_proj.weight/scale": "float16(5120,):fe69b5d4fbe6f7410c1b55c11dc549244694b1915d32d65ad9263bad9259056d",
"23/mlp.gate_proj.weight/idx": "uint8(8704, 5120):53677ade60c3105ad0e8e4e0a82039468345eba9844a81c4325a59901029bc56",
"23/mlp.gate_proj.weight/lut": "float16(64, 2):d050e35a8b61e5bb77e840ab487a11ffe4a37d37099387bdc381d53c3cb9d0d7",
"23/mlp.gate_proj.weight/scale": "float16(17408,):d54f427e41b6a82e1bcd35b3258a505f7ba9800bc769e8b4c1daed24e5523bb9",
"23/mlp.rotation": "int64(2,):35ee850143a2b5f66f65fb7c2d4242d7eb09b9ca47a535252294602c440259bf",
"23/mlp.up_proj.weight/idx": "uint8(8704, 5120):ebc35b109b059b589938f8eec4596ed34040f792a838f8ee38ce8b7f4c7f1f94",
"23/mlp.up_proj.weight/lut": "float16(64, 2):c47a38f5fc06befe43f85a23ef99964f3347ffa9e2ee73c3a93c2baec4d1df9b",
"23/mlp.up_proj.weight/scale": "float16(17408,):7176822824a61f50edad7a28841a0070d07123f77e590981940106176bf961bf",
"23/post_attention_layernorm.weight": "float32(5120,):9981c553581ba01d35e10b7de0206e67adc34e4608765dbb8006ad811a22ee34",
"23/self_attn.k_norm.weight": "float32(256,):36e75eb7659726b84895fcce967cff636f41b18a59bcb23832d02380e7d22e16",
"23/self_attn.k_proj.weight/int8": "int8(1024, 5120):25f395f734a0d9b91078616d9c327cdca2b72fd1046caa2ac31635f18e699479",
"23/self_attn.k_proj.weight/lr_a": "float16(1024, 64):5e40f9fef547483be0793f772db1585d53d0376b7b44a572b807065aa6db3769",
"23/self_attn.k_proj.weight/lr_b": "float16(64, 5120):dce57ee358f3de5bedcd3143697c4eb467ce0ca1a63419e5424729c03dec56a2",
"23/self_attn.k_proj.weight/scale": "float16(1024,):c41be9c8fef7511cba7c6f04264c5faf2cd0df21c9158b8886cfea7be5f8053b",
"23/self_attn.o_proj.weight/idx": "uint8(2560, 6144):2446e48e46cccf3e941e1088e2accf3c0451dae50b2016427ca2cb24df434395",
"23/self_attn.o_proj.weight/lr_a": "float16(5120, 64):ce49fe9168d67b93b9886e5ecd4c9f5bb60a7ffa52634b9ee2a54374b8823153",
"23/self_attn.o_proj.weight/lr_b": "float16(64, 6144):29014f95b3e219b17de415dcd88bd943deb33183e7222ad4d988ba92ed00b045",
"23/self_attn.o_proj.weight/lut": "float16(16, 2):1f27414343f8b949d5250e73558462f0759e0b52065f0b032675a3808ba79794",
"23/self_attn.o_proj.weight/scale": "float16(5120,):72c681f93d65c4e1d6d08ac2a9c7b4b9fc97002041a704257b8fbafe86399a97",
"23/self_attn.q_norm.weight": "float32(256,):edea8163e469654adf43dfe83652e00c42b0c1035cf0437e78638690e607441e",
"23/self_attn.q_proj.weight/idx": "uint8(6144, 5120):c84a2c711028bc49761951b266deabcc1e1438c4fb9967a095a6a23d4ede28b6",
"23/self_attn.q_proj.weight/lr_a": "float16(12288, 64):c758835f9737faeb544feaba5f95c551eccd75148c9ff16c39d5bd1267601958",
"23/self_attn.q_proj.weight/lr_b": "float16(64, 5120):483f154cee0ef85e65e3bcf425d44c28c479dc3c0b3fcf309e9d3a00ce46566d",
"23/self_attn.q_proj.weight/lut": "float16(16, 2):866968c4a7c332b2fa2485f7ca5c9f593fed23b5a85f93537c0cf9458d715bf0",
"23/self_attn.q_proj.weight/scale": "float16(12288,):13b911e03899f7cc4b998c87b4190f71699095bbcaa65751ee6985a9e06efdfa",
"23/self_attn.v_proj.weight/int8": "int8(1024, 5120):048a8c7b69d1840f764436e1ec6b8bafc83749ca66f8d0fe1984ffd717f1debf",
"23/self_attn.v_proj.weight/lr_a": "float16(1024, 64):33d916e5bcd86e078f371b1d6563ca2326fa5197d7f37a44f1fc5de12adcd1a8",
"23/self_attn.v_proj.weight/lr_b": "float16(64, 5120):e5faf895c8b8f9dd1e970de6642e9bf548c77ef4963d18dda045cb9d589ef252",
"23/self_attn.v_proj.weight/scale": "float16(1024,):7315d138d07381558c2ffad088e0446eedd6cce1d8fa2481aca51eb1e3b80bcd",
"24/input_layernorm.weight": "float32(5120,):c0ef96dc1aff53e3ce71eac311af73309186b17a0e8f78fdd5358bbd5ab49d0e",
"24/linear_attn.A_log": "float32(48,):96c23fe1c63c0b5804eb2886453d897348e73de33dcab7a4c922fb14c73519fe",
"24/linear_attn.conv1d.weight": "float32(10240, 1, 4):03b018a5f0b141129398c34526e1a677a1f55b5e4146b1c23e33a0a90e2f0746",
"24/linear_attn.dt_bias": "float32(48,):a230618956726045592e5b202d35321d15f5033c298274eeed4d4f5f44220c88",
"24/linear_attn.in_proj_a.weight": "float32(48, 5120):3c88ca7cb44149fd044ab43bc0f1d5a9e8e798bfe6975a4fb2a72db79210a64f",
"24/linear_attn.in_proj_b.weight": "float32(48, 5120):d8213fe25ee9e6377170b1c3376b4d926d68ec064ac391ad872a676d909f4fb8",
"24/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):7a61eabb9308ce9017ab9fb8fa2290631467b89b8af9187b7cfcda420fedb625",
"24/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):c98aae6ea42fe20261a03305252b529531a9336d06be1f746deed5522cdd940a",
"24/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):62d7868de0949516e6e6c3c9b8424bb98ec706cef5e199c42ff0d141338a87bf",
"24/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):e80ddf210660b67344a684a69f7a8c37ef01555701062a410f8c43ae7c700af8",
"24/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):02eb773b6ea4ff9c200580f77d871414fc0d33606a4754d2fc1267afe8da7969",
"24/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):85900c9efbe479d0a7ab62f324dbff5dc1ebfbc3760303e15b9964f60163328a",
"24/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):1ee42b719448c10bf164bbd48e05a34d21781540e192cead41dd6ece1fb7cea0",
"24/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):104023996977a719f144c873b6f57ea1e5166207262658e7a528b0cd020f0fbb",
"24/linear_attn.in_proj_z.weight/lut": "float16(16, 1):879d0e5e18cef1ad68170b435febe6e81b2d408940c7983f9a314590deb738d6",
"24/linear_attn.in_proj_z.weight/scale": "float16(6144,):db48883f848d39735c813e4f7f4835d1a6933c105faf720d619a682dbf4b638e",
"24/linear_attn.norm.weight": "float32(128,):dc27662be8d06acd9123b0fddd4556ef23c933415a14cd296fbf9264d4c285eb",
"24/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):f8b505d17ca0d8914834af40932d68f5fb21f34e3b6673115840ad269f13470c",
"24/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):dfb9628fda68d4c9ccdab1d1edfb44656ea0a04236dc790e0997cbdc9374a747",
"24/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):0ed21093bd387b85aba6ff44aedfccf086fc8843a9d30b9782ae78eab1a07cc9",
"24/linear_attn.out_proj.weight/lut": "float16(16, 1):28dc216579d85e5c66e22657b900b7d607b8794df0208eba3296f45727612032",
"24/linear_attn.out_proj.weight/scale": "float16(5120,):bf445485858cba6466646da436ff27e6cff6de223b57b30ccaeafbb103a4cf26",
"24/mix.rotation": "int64(2,):c8b4f817f9b53889d71d5c12e68309a053be4ca903ba4a15c470546fbb8851e0",
"24/mlp.down_proj.weight/idx": "uint8(2560, 17408):665bb7f0798adb19b59c1c874e41eca8eff39d28ac8bf4cbc8a3713d1823cb5b",
"24/mlp.down_proj.weight/lut": "float16(64, 2):dfcc29116c1b6c99f76d6da91f3dfc9d7c5b6769d334e02e71bb1c7cc795cccb",
"24/mlp.down_proj.weight/scale": "float16(5120,):4cefddb872f3031117f7f2c38a3f34e1df5c15f40ed4e441f8e5f85534a783b0",
"24/mlp.gate_proj.weight/idx": "uint8(8704, 5120):abbf4604494cdbad9d0a35616a4467d3f8bceaf9337ceae5f99e619a7f0e2f04",
"24/mlp.gate_proj.weight/lut": "float16(64, 2):1c98b20f4133f7bacdff515a1f79d0e9cf5a4229cee6a71a08521518343e29ee",
"24/mlp.gate_proj.weight/scale": "float16(17408,):d4a943f59a6503bc1e8a0b1383466d0b13981068d8c8bb10bc2718c6d7691e03",
"24/mlp.rotation": "int64(2,):876f69e1992aeb973e4f46edec05aa3c356b987109c9f57032aaac749144870e",
"24/mlp.up_proj.weight/idx": "uint8(8704, 5120):e3cb1d5ee31249a0e3fcefb6e137bcd773bdc531e06ac46de410b8e3e8e79234",
"24/mlp.up_proj.weight/lut": "float16(64, 2):e5c3a375284673b6bebf3b128178e385a2876230c87f0768547db705f0a264c5",
"24/mlp.up_proj.weight/scale": "float16(17408,):1bffbe62a448ba0b6d1e5404704668085051ddae2cfd4fa34fbd9e2a7b74188c",
"24/post_attention_layernorm.weight": "float32(5120,):7097203e539b6f0b9c2b1fad64993789c836620b210553740368254cc5c25637",
"25/input_layernorm.weight": "float32(5120,):5dfee4591d86b53c3cd4cfc681b1e1cce17ca7a5ffe1deadf73d4dbbb04b5aed",
"25/linear_attn.A_log": "float32(48,):d6b1c2f95cd1208a43d859567c06182e195fc4c6e0388b024436574f74da2aac",
"25/linear_attn.conv1d.weight": "float32(10240, 1, 4):5388d311c6bdb31b64e62c5897f8b613694037586898a9804fe97856163c32a8",
"25/linear_attn.dt_bias": "float32(48,):4d052f9d431c82bd5c37731829dd776cd5edecd4468f4127d167ab3f841da03a",
"25/linear_attn.in_proj_a.weight": "float32(48, 5120):25ea10fdad2e98a60663017a723a05ded536c5a02cecebe8b0e6f2bb301235ad",
"25/linear_attn.in_proj_b.weight": "float32(48, 5120):c645d63af9aec0d4e91add5c55e69b5d7e82e7344765134409f28cfb797195af",
"25/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):9ac706d9f39e6de3a48b217e3caffef1f1f118e8ef73b59a2c9a834677c4664c",
"25/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):244a072df6aa68eaff874f212878edb745284aa123610c73165254743bc10da3",
"25/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):f71a832d0e0e30b1cce612ac171b614fec63f8934c161f589e81135cd5a66b2e",
"25/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):03e9e807f8f4e65a430aa627760908728914e690aac89613fb759157f219d997",
"25/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):68d193ee68ba776919ae513e784d6074b606c454dc7c5c302af9ac91d9b74e1b",
"25/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):1029abb89dbf11fbc24111144af5b41e8deba79fc5e6891ed62f8994a45cf3fb",
"25/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):5f472e61532b8bedebe290c0e28c4d33995adb26d7cddfebca654fd25c03af42",
"25/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):d07ae79440453edbbda83dfba8b56778085613a7308461fb9f87a9205325e5a8",
"25/linear_attn.in_proj_z.weight/lut": "float16(16, 1):646f1b16cd0f8f9bc02109b71cf4aa8740f846b3f0d495568269b90836420aa6",
"25/linear_attn.in_proj_z.weight/scale": "float16(6144,):7049faa5ec11f3fc52978c84d53b4165c7b671860d2b583b0cac8355cc51cc63",
"25/linear_attn.norm.weight": "float32(128,):d38e2759896014317f9b3c9e7d6064a02edd33d5de1a7ccc878717aeebca0e96",
"25/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):b3a66641d24addfbecc290c7f38dbc81fd5ef96721092a862f092f5fd6901a3a",
"25/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):6ee748efc74d7a934cfdc774b6bc4b2c2ae951dd5b959dc3aed281f55e8bfaab",
"25/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):6dec064487abf28ddba897a961e15a824c0be8faf308ae2205b7c4d51d8ccb04",
"25/linear_attn.out_proj.weight/lut": "float16(16, 1):85f22662c6f1af66e4e8a7c96d039d048999798848eae3b6de7df95a06cd27fa",
"25/linear_attn.out_proj.weight/scale": "float16(5120,):92053dc3f6034cd20d80983e9a6430d28e6103a8c8de0226461dce1befbc2f49",
"25/mix.rotation": "int64(2,):f7cd09f7693e676d72782768deae1d25e173d30317809a22c55fc39cad3887fd",
"25/mlp.down_proj.weight/idx": "uint8(2560, 17408):b8e0600a794ed49b15f601cd35530c322c8589905879adbbb4073d27e99cabb4",
"25/mlp.down_proj.weight/lut": "float16(64, 2):2a8164e9a7510132b4070c06e9536dee031bfa03767d18e4264abb183742af2d",
"25/mlp.down_proj.weight/scale": "float16(5120,):e8857f72a333ab0dab2dd3ce8aa0ff96c4e8295b1bb8d14b55b210f3fc4cd445",
"25/mlp.gate_proj.weight/idx": "uint8(8704, 5120):0633a3a25a6ba9826e459caf19ffb9f7b0f5adc1b9793d4a82a96a98655eeb8e",
"25/mlp.gate_proj.weight/lut": "float16(64, 2):5c567d33155fabc0310a2a6218fb3eed32ac0c577cbabb909a549ad0f0479d88",
"25/mlp.gate_proj.weight/scale": "float16(17408,):2249996549900ab747afd9bc3db14a046d787603649ca812cceefe8c2b9ed8ca",
"25/mlp.rotation": "int64(2,):159356dbcd96cfe22bb5198caf785d971b83e431295a66f1924280408dfcb768",
"25/mlp.up_proj.weight/idx": "uint8(8704, 5120):e51539c3a5ed50c599bca5d5a2bfb0aa46f2c6bf50c4738827c65fe717a05f12",
"25/mlp.up_proj.weight/lut": "float16(64, 2):eeaf9a5a3915c016412d7300d5363a5b40c4606781d5ed6054ceba05c06916bf",
"25/mlp.up_proj.weight/scale": "float16(17408,):1ce5d51a8a58de6b54d9eed0db9a060e392c3e156bd9c8977160008dc15fb28e",
"25/post_attention_layernorm.weight": "float32(5120,):619de4c8ff96ccc1d83fb78210df0ecbbc73eee6abcfa4ea621a26484f9fce4e",
"26/input_layernorm.weight": "float32(5120,):da1fb63633389ed5b2db6bea6071c3f1d700e4d22f7574f2d3edd68ca515f83f",
"26/linear_attn.A_log": "float32(48,):fdd6767c1fa319cdb99440bc9e12ccb45738dc675e64cab34168ad007bb1af2e",
"26/linear_attn.conv1d.weight": "float32(10240, 1, 4):5fe544635bcbab3a474241bfb2ba3882020f27f8e9c0f8bfb2ec6a4c9efd2c57",
"26/linear_attn.dt_bias": "float32(48,):d059f6c86b8597adff18a323cdde4aa4598c6d5fa2233ce0dc95ae7fa16a295a",
"26/linear_attn.in_proj_a.weight": "float32(48, 5120):16e68aa71b29cbac1659ec1055df70062acf4a70679749ff4313d0fb4ce99dbf",
"26/linear_attn.in_proj_b.weight": "float32(48, 5120):14ea359d2a951f7f0d6b1a910cfa581bc45ea5033e8e3f99cb8029be8c39c4d5",
"26/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):bed528f889c88ee7ec04c93ee2aa65185877caf3a518c258861ce83ff038a3f0",
"26/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):d3e3239916ecc92e74ce0e825b73139287b60aea1235c668357b7f5832615493",
"26/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):7b5f446e19a134e5decb54812334be72dccf23171bc37a40c5201fe5ca11fe8a",
"26/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):1528b289a5e29aeb6e0defca5426cfe3ea022836624b1378659b38fa966862d8",
"26/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):5e304f7bd12c901b7394b930bc92b8f97a07859a67013de44988116cbe31f715",
"26/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):8e681ec1356a454e7400a28288533489b54554209b90ad31cb62e1d23432e225",
"26/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):f759e3e4e8b8f54545d1b521602977ef78e9d9cd56cc74ff6ac861fd69c40d4b",
"26/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):e5868d75eab36bb3fe17bdebc3e476c223820f1621ceab5ce6d2ade97c64522a",
"26/linear_attn.in_proj_z.weight/lut": "float16(16, 1):14a02402c0b7cf0987d6c7a445eef86e19579b72474545325e0f2926903cb249",
"26/linear_attn.in_proj_z.weight/scale": "float16(6144,):4be95be644b585b58e52596b2457a7211d70a82883f8052d80f155f13b1a991c",
"26/linear_attn.norm.weight": "float32(128,):8c28172ada1aff8138e5ed0e06da968e65c7f84037f3760d2b24297a449ea34b",
"26/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):1993a10c8eeb747227bb53e01b529cee9042e94cb882985cfc451e2a4e347cf8",
"26/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):b756e5413c4cd85df07db911ff96f3fd842c8d591cd2ddc2a8e6f2cae298b26d",
"26/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):5964c7d1dee9c21dd80502323de54898933f32266e4ed69b6e17aa6d57791c56",
"26/linear_attn.out_proj.weight/lut": "float16(16, 1):fe462362f8ad5533c821233280f7b295ecc1ed5a30bdced22f385b760df7a795",
"26/linear_attn.out_proj.weight/scale": "float16(5120,):86075116c540710b553302c36cb1c85e68ec84d2e44c78fbfababee14fae2a49",
"26/mix.rotation": "int64(2,):8aef732d243f8809f6367d28a64e904e4c8590bb078028dc3735c98f0f09d828",
"26/mlp.down_proj.weight/idx": "uint8(2560, 17408):617452f413ef6e2c64d9896bf4ee09de51cf405cd466e5dccc666816b47b91cc",
"26/mlp.down_proj.weight/lut": "float16(64, 2):c2ac4b8c20d930802eecb7a91b94d4453562baa27a2fe2c8c48e7437fe9c13b6",
"26/mlp.down_proj.weight/scale": "float16(5120,):0c368282c13a0afc726e6637ae6c0e9c77254fe23f40f69bfee45f05eeecaeaa",
"26/mlp.gate_proj.weight/idx": "uint8(8704, 5120):f6e999a3280e792b5415ee1ae6bd068e4f8e4c7025a32e6c3bf8745c481324a3",
"26/mlp.gate_proj.weight/lut": "float16(64, 2):c116852b5054f773bd3ff2f56858bf7c236030e135a18283317c9fa8ff9d16f2",
"26/mlp.gate_proj.weight/scale": "float16(17408,):53e90fafef707118c1d94635dd48364f411a9c8ba701e575069254b8b882eebd",
"26/mlp.rotation": "int64(2,):3c4a2bc69a80a41df4c8cc876846f583631366b0a791ab1902bc878fd961000f",
"26/mlp.up_proj.weight/idx": "uint8(8704, 5120):cea789f291ee2a2a984bc432d5a1c6be872785a6655ae95e55bf1777d41f2dfe",
"26/mlp.up_proj.weight/lut": "float16(64, 2):24a7a4380bb274c3e00b050ee5acccfed29d78261dd3b4a0b532c804798f1302",
"26/mlp.up_proj.weight/scale": "float16(17408,):22d873d1a457264efffcb92571ed241041babadca97d4027b22fed71637eaf55",
"26/post_attention_layernorm.weight": "float32(5120,):6b8ba6a28bb361218c577e9115812c83e25c9b642a76ae7f39a919ae56b05c96",
"27/input_layernorm.weight": "float32(5120,):53d3c041bb6ce5a607e801a5ae59fb062eedaa425f4210c70af1620a7e2915df",
"27/mix.rotation": "int64(2,):85862776d81535ee48c736d2f775b0e0c4b50b0e2f600a8a35ca7c32f8357881",
"27/mlp.down_proj.weight/idx": "uint8(2560, 17408):c78e977f6adba526555ddf01f0197353c9ba4b28886960de49181d8812a2b53b",
"27/mlp.down_proj.weight/lut": "float16(64, 2):1a28231603853df9a2072b1b0b35b4afdf97b35411fd0abf90802fb7fd4af98c",
"27/mlp.down_proj.weight/scale": "float16(5120,):e87830040478c715ebf60288805d8aac4c92d6434678711186351e0a3b44506e",
"27/mlp.gate_proj.weight/idx": "uint8(8704, 5120):2412743fb77944ca5e675ff9c6bafdbcaaf3f8a5dd8778eccdeeaf8665001074",
"27/mlp.gate_proj.weight/lut": "float16(64, 2):354d16a371cbe6ae75f4ed5cb9b0fda31d7e6bae193c0423bcd881fde679fe5b",
"27/mlp.gate_proj.weight/scale": "float16(17408,):c4b82ae2988e574f1a94f51eb838ed67f4b9e84d3030fd73d2172b8e1e0dc173",
"27/mlp.rotation": "int64(2,):44d65ad82ab6ab0a81f2c504a31b992b0ab70868426234a29805bebf986483b8",
"27/mlp.up_proj.weight/idx": "uint8(8704, 5120):95f2d466bb393319dfa1fed6cb2e15522bebc83c3f51a5bf25c20100010dc9bd",
"27/mlp.up_proj.weight/lut": "float16(64, 2):6b696ca3545a8ca9b887897562b843636f56c3929028328a77f5c696fe9c9984",
"27/mlp.up_proj.weight/scale": "float16(17408,):cc1efaef72c2a8c6756fd727dc5971b7d72706472d653f530c56cc914c586c8f",
"27/post_attention_layernorm.weight": "float32(5120,):106766d551ab28874fafd11d607e82caa4a34798d6520c2bf8f0bfc76a11939b",
"27/self_attn.k_norm.weight": "float32(256,):c474444a67aa1fec2c07caa8e574e2edf64062b0f71b0495d4372a96520fa55d",
"27/self_attn.k_proj.weight/int8": "int8(1024, 5120):838a6a4e3df12a4ec42e573b6bad14c3f10345574b786ffd0651a7799b4647ec",
"27/self_attn.k_proj.weight/lr_a": "float16(1024, 64):6c0a05132fc2d8d0ecf568beed28638b0abbf4a6c53b6faccfd38e99b6d8514e",
"27/self_attn.k_proj.weight/lr_b": "float16(64, 5120):e32de5bcda16a6e2ab377011eea543e43a8d5ce8f35ddede217dad2a5797eebd",
"27/self_attn.k_proj.weight/scale": "float16(1024,):2f0ce536ae2e53b41e5a54356b54f43c6bb425fbdce9faf727b34a39795ad69f",
"27/self_attn.o_proj.weight/idx": "uint8(5120, 6144):0ddf50cae13aa16e6515f6b319b31f4c75a40d7933fe179bf5ff43954f110d47",
"27/self_attn.o_proj.weight/lr_a": "float16(5120, 64):eef35813d8573264b742c315c048c32622d8bf7a0632e483565c1f42b3ce9dfe",
"27/self_attn.o_proj.weight/lr_b": "float16(64, 6144):c8f8f134129cadc7eab50506e13fc25f92ea32ed0349ea96e0becd2ace8f4a25",
"27/self_attn.o_proj.weight/lut": "float16(16, 1):a7fbd8225c224409023f290795d2d6c0d4b61d83c74311a49ff436a9f8fbb1a4",
"27/self_attn.o_proj.weight/scale": "float16(5120,):876017f024c0f1b76b3c1fa30414d948d7dab472543d4af909f18b45af47c871",
"27/self_attn.q_norm.weight": "float32(256,):78bbce129bb59cc49b7b3e6c2c7b85b48dc22593585ff3a1ce59c6fd96a611be",
"27/self_attn.q_proj.weight/idx": "uint8(12288, 5120):8013ccc3142f7ca3b2d05ce8ff8e05c2b38e94d11e0ae1d786154a5069540f7b",
"27/self_attn.q_proj.weight/lr_a": "float16(12288, 64):ccbedf40444e0ae0df07bdcb25ed49c5afb1d7c957a5d3deb4ddd788d343c6e1",
"27/self_attn.q_proj.weight/lr_b": "float16(64, 5120):731615485c692bbda7f79948d60bc0ce4eac63285f3dfb7f3c85fc2c3fc62646",
"27/self_attn.q_proj.weight/lut": "float16(16, 1):a4a10442c5768c0287d4520845ca8f3acd3bacf0b8a97a16e845583f849d40c8",
"27/self_attn.q_proj.weight/scale": "float16(12288,):e4ce9b9e8fd986ecb8d1f1854d71b018f78f5613ddf927eb48be1d041cede9a3",
"27/self_attn.v_proj.weight/int8": "int8(1024, 5120):6d0860c2b8c4d86713704ac5f3719e42519a096d3c2d6e52dfe6537e657e7167",
"27/self_attn.v_proj.weight/lr_a": "float16(1024, 64):d09774c55059ab98bc6eeae195ff22d59a5101e8c1df312917dac474ece9a911",
"27/self_attn.v_proj.weight/lr_b": "float16(64, 5120):c07b09c36e9ffd108a852651a7f79b10250d7bbfe69ba8700207b1781d848a93",
"27/self_attn.v_proj.weight/scale": "float16(1024,):87c40cd0721706d5d6d24dfdf841e99b56a3d37bd5e58e55538962f945c04479",
"28/input_layernorm.weight": "float32(5120,):e0240db5c854be85f15d624f081b294a5db37e5c80be9758d7aef09da73d40a0",
"28/linear_attn.A_log": "float32(48,):57d3d1be8d704c6b4c85a0366bb3e6cbbfd0d8068e70bae0a54c7ab34185b7f5",
"28/linear_attn.conv1d.weight": "float32(10240, 1, 4):506939ac0b06333426b0c1a69c8fae59d00112a811d76d8eb9db4aaa8b8e3a36",
"28/linear_attn.dt_bias": "float32(48,):6f1bf05647318fcc10f1720ad0bfb7ae7fe444a972fafbce64d9033657c2f4f9",
"28/linear_attn.in_proj_a.weight": "float32(48, 5120):6342e08349cdc9cd9d254eed15fd6ed25f5501a90716c58def1eb4366f140449",
"28/linear_attn.in_proj_b.weight": "float32(48, 5120):4742358233536e5dde3c9c5a5e6d7cef5d11003de3c13b0f46f25de98240014d",
"28/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):a5a496f6dbac169258f7f32c5e114b545a1d625959893d346275045ed8d3d34d",
"28/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):95fe6b2bf74ba4215b7d013fc40c8ecccf6f6706d316d2d506f856ecffcbac35",
"28/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):76bd7ad4a4986f1aaa1142809372eea695aecf5d04220208bf95f6ca006a717f",
"28/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):486d133511c224a9add05418db85a8ba133cae4155caa87129e78567cb52b444",
"28/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):6786aaeb693faf2cd3e5443e2b3cf92e2203658cb4c9b9b099c63b3c5e4ffb11",
"28/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):aa40d827a1ee84c7f5ff581b370db315f9e00da3890ecc35792ee45349cbea78",
"28/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):9cd65319dd8b2cf6f897c64919e2e9feee1bb9c420761881aab2ebd98a4c93a7",
"28/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):f4b95c552a656ef45cfcb39714037c1fc23359f45dae2c1d64eb4f5318348326",
"28/linear_attn.in_proj_z.weight/lut": "float16(16, 1):4e4a7967633564d91afa77c27391b79d59cbb079c73fbcf0214eb0b57a948e8d",
"28/linear_attn.in_proj_z.weight/scale": "float16(6144,):1c5f0823a382f9e67db57825af37d5fe3b7fe99701055a34341b6d14475bb56e",
"28/linear_attn.norm.weight": "float32(128,):e3a297511ae57cb27a196918fa18cd1637a56ae43458208e8d511f11e050d787",
"28/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):aee4dea11e210df5db340029f3151c74cf1c5326cfde712b9406d192be466c2a",
"28/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):83d7320ef09756a4fcd68cbc62d99da003548f800d062590b252fb0a87f3a7a8",
"28/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):d19fb660928bd6cbd7e62675c882957ab32a8250028f3f30397436a5d8f5431b",
"28/linear_attn.out_proj.weight/lut": "float16(16, 1):56652c3711deb6eab16c9c9436aef4bfc213ca3c24649efc0b1fdb62c27f4f46",
"28/linear_attn.out_proj.weight/scale": "float16(5120,):960ba83e764ba612381094d28e166707b57e0a869590e71f81a88f53b71c67ce",
"28/mix.rotation": "int64(2,):3d32f57dda3ee681d204d87851f0229fff1c6862b198857607e7f9a6ea04efa4",
"28/mlp.down_proj.weight/idx": "uint8(2560, 17408):4843f5a66225cbb85314a4f1c5bfc793bcf014637df5ce2113c26b747a75ff27",
"28/mlp.down_proj.weight/lut": "float16(64, 2):4878e7df23e0c59c19b08d6870d79fd273d5c2010749a3da4519c5c403d04ae8",
"28/mlp.down_proj.weight/scale": "float16(5120,):2547fda7322c8cd44fa0aa741115a0be3f22a2b7bc614df21ff002a1561601b0",
"28/mlp.gate_proj.weight/idx": "uint8(8704, 5120):275c423152b98ad3c67001a5a7149c440be4f54c08684e3bef1b68840ea81227",
"28/mlp.gate_proj.weight/lut": "float16(64, 2):118316354dc4879d19d0f989d9bbd5548f4b1ac2a9c3d365a1fd8cc548dd4703",
"28/mlp.gate_proj.weight/scale": "float16(17408,):bfbb830621a49b8adf3e947c1436ec142d2c8dc4f90d7d0f4e30c4f3a02507fe",
"28/mlp.rotation": "int64(2,):691adb6cbf52b75299d7f9a06a09898305e7ef2553b7b36ffa90288658061fa2",
"28/mlp.up_proj.weight/idx": "uint8(8704, 5120):3e047b79c77cb52e7a7e0490dce7bf28f52bc463614b40fd3fda4cabd85d736a",
"28/mlp.up_proj.weight/lut": "float16(64, 2):7fd08705b654cbb8e0a69ac121f25f6bbc307f416b89675203de7f593a0af015",
"28/mlp.up_proj.weight/scale": "float16(17408,):d815433ab32f4fd60ba1037d82b2db43e94bd481dcfe1f7a78616587e512a957",
"28/post_attention_layernorm.weight": "float32(5120,):e29f7776e2cd430698eb60d707787123aee59e288b730b9d77348da96d481935",
"29/input_layernorm.weight": "float32(5120,):49c94fad8926f7e5873b3990d6d32bc77419ccab4bdcb1d74d2461b5d4b949c0",
"29/linear_attn.A_log": "float32(48,):405411af2c5a5149b90a29818685d15b3127c20e55fde38e9ff74f1d028bda60",
"29/linear_attn.conv1d.weight": "float32(10240, 1, 4):cd66b7bf77951db3ec20145189bf18b776efa327cbdcc830f9120e862edf72bc",
"29/linear_attn.dt_bias": "float32(48,):33485f94ce3cd122a902c7e3fc1e8f67fe06c4d34376f80763739af0acdbe9c2",
"29/linear_attn.in_proj_a.weight": "float32(48, 5120):5f196791c417f7c22ab291d87974c9ca68182b34bf25d0f87ea2f70e48275b94",
"29/linear_attn.in_proj_b.weight": "float32(48, 5120):04cf50980b12af7a6c431c64f10c969d7226003b61c38fac0c911c1b7c3db52f",
"29/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):d95167d536ad996c870018edea81b99fa43fd247f7898e8c14f7d5557c999dc0",
"29/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):ab3501989fcd65e6749052befd8ae17468ddf8e3751062060f82239d41481096",
"29/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):1427678c3ee9bf8322b73b795e4681ade26a2a1054c786927e9ea306552059eb",
"29/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):a869ca43fec011fe09c5e0870d01a4cbf24c47b9300ffd01a452480b8a53b13a",
"29/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):38853d61fa772191e7ffce22c12f24e2ff1eed947a9c3bda4d06a5e79ab160c0",
"29/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):fc5849df87afbeda7be530108e88d73117db1192542259c4020e6e2112441e12",
"29/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):cf07716b695cfc370a8f94f0194c83afd35ff531ce0dcae76e76c61eb3bed941",
"29/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):a04dfae03caf7d7d26375a7446d92e72fa13e30471f7c7fc5865623012783d12",
"29/linear_attn.in_proj_z.weight/lut": "float16(16, 1):3fea5d76d7eabe7d76310f92fa22d0a32c716660bb17a9ffb569b0c1b63d7123",
"29/linear_attn.in_proj_z.weight/scale": "float16(6144,):cd63ae456207736843aae15d4ca5d0870cc61c9f60c5ec358e61d3f229776142",
"29/linear_attn.norm.weight": "float32(128,):0c400dbf53d1c12f05d43d88a3b06b1399febe5303d277215152878a7a326e30",
"29/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):a70523175ff2217a08a62b66d17f742eb42c9385ac4f0a5eb39f5f6114a62588",
"29/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):b91f75fba5d8fc57437289990cc5551370a135a08207456b05bf88c0dcc31cb8",
"29/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):2e1b857f85387944ebfcbbeb53cf44d9ee44adad3a717b958975433f5cf36ad9",
"29/linear_attn.out_proj.weight/lut": "float16(16, 1):7fe9fb5b0118547839ae0b3b65539cfc330f6d59ca1e6a8c8f254d004932c620",
"29/linear_attn.out_proj.weight/scale": "float16(5120,):3312709b3d8458f77596f335a0be06e657ce6c9a4c6d75d8d054c984322e81f3",
"29/mix.rotation": "int64(2,):b89c18e21d6ff552a5a5959c71a6d808a5d3d28c37d818bf8588c6caddbeda7c",
"29/mlp.down_proj.weight/idx": "uint8(2560, 17408):9f299f8d874e16f49703cbbf44f210337e59f37d928b54491f5e70fdbd594b19",
"29/mlp.down_proj.weight/lut": "float16(64, 2):ed524b6ed1b5e6674b718ad85c7bd8fe7d0adce078105bd4a819912370e5ee72",
"29/mlp.down_proj.weight/scale": "float16(5120,):068aaff3364e8f940534ec1a112ee1b138cd8123e61d2daf77bba414b288c439",
"29/mlp.gate_proj.weight/idx": "uint8(8704, 5120):5a9f759114d72c910b37cc02b86d5c4fa24257fee3e8020c823e600210506b11",
"29/mlp.gate_proj.weight/lut": "float16(64, 2):47bed486a7c55d3bd6b9268162c1540622e5d1da59f28874bb5ace635a20b3a2",
"29/mlp.gate_proj.weight/scale": "float16(17408,):3d3cabcb9430b10e91c830cdea53ba9eeab1760090feed3b964ef50d8521bcf9",
"29/mlp.rotation": "int64(2,):4b5371f5fd95adead17b7964de7af91c37b873e37188e691db98cd313bf0d214",
"29/mlp.up_proj.weight/idx": "uint8(8704, 5120):71cecedf24c51c853b77d6f0a6671f7fb8f42b3ce61c9d3f2c8747d7f441fb87",
"29/mlp.up_proj.weight/lut": "float16(64, 2):5b0bde3f45dca5e4d6f049a7b60472c34f03d037c42737b2eb3d89200bba8831",
"29/mlp.up_proj.weight/scale": "float16(17408,):3d5a3a5d39789bad52b9a5069d78b5563f97e36ec473049ed1ee5a6b04724f6b",
"29/post_attention_layernorm.weight": "float32(5120,):045125c9bbe3378aa79c5e0bd1b6844f2adfd3150b062c2deda48d450311eb57",
"3/input_layernorm.weight": "float32(5120,):b6c915d6f0dd2d239aa02e9cde78c57807ef9f2c84852efba491f89443444731",
"3/mix.rotation": "int64(2,):19b1dce82326293dfce62a883ee7ca6e83d1d1cd8aa3584cd9beb17347db5279",
"3/mlp.down_proj.weight/idx": "uint8(2560, 17408):647494d8b41a7932ecebf7a5acf76c16de9e3ecd078b976442d5d11438b41ff2",
"3/mlp.down_proj.weight/lut": "float16(64, 2):64a4cfdfaed8abb10f657db1c11ee352500414861ef18ade6ffcf4ae764fe0a8",
"3/mlp.down_proj.weight/scale": "float16(5120,):c2cede7e63ffb3b9e8dc311c94516248f825485d3c59f362db7108450213e0e3",
"3/mlp.gate_proj.weight/idx": "uint8(8704, 5120):8e09817761fbb6485dac1fdb8167cf348f3d70f1034714c4647a552d266478f5",
"3/mlp.gate_proj.weight/lut": "float16(64, 2):43d9c69614d32dc1abf7fbc237b79a63c2d690885633c02112d108e995005e20",
"3/mlp.gate_proj.weight/scale": "float16(17408,):fe5388e36e64b74b2ce22e5b1dbfdb8eee9637960a96ee406bcaacc49c8b9cc1",
"3/mlp.rotation": "int64(2,):a91c9fef41d2f77c12b8b0b2c09f7cfea89ea95c2ab40cd20e23f19ee12d7380",
"3/mlp.up_proj.weight/idx": "uint8(8704, 5120):d13243396eccc12e6651a25c850991f05676e4276d209e56983984e0dbf4bee4",
"3/mlp.up_proj.weight/lut": "float16(64, 2):ce23eb54f68b9444b35c15adae00d843023ecf46d045ca555adbbb6b11bff80a",
"3/mlp.up_proj.weight/scale": "float16(17408,):ada8712781906155b002bb23e06a2a10eb3ea87fcea6fd6c87dd41c5baad17da",
"3/post_attention_layernorm.weight": "float32(5120,):4c8a4bd5b9052e8935e2cdc31b38b90ac0df0a7ade55263bd0b67888efe666fa",
"3/self_attn.k_norm.weight": "float32(256,):52fad02d9b006c770b852be613d063655bbde0fcac3ceb0ccee66a43ccf9559d",
"3/self_attn.k_proj.weight/int8": "int8(1024, 5120):ae00738b4eeb5444a0fb7c15651673fcfddb5d378d221c4840868339b959aa7f",
"3/self_attn.k_proj.weight/lr_a": "float16(1024, 64):dcf5949f64bef4149532cbf33b9d3eb4e20137b6c651eb7b668ff770764a207b",
"3/self_attn.k_proj.weight/lr_b": "float16(64, 5120):3f9d6e70359eb89e44f94bafeb779e599b3ff89f51981e8ac01182cd8f47ad44",
"3/self_attn.k_proj.weight/scale": "float16(1024,):825c3b3e3d3883f832be40f422fac4838319d10e37ef0f9ec5348c759dc4e1e0",
"3/self_attn.o_proj.weight/idx": "uint8(2560, 6144):e834cc001df754520dda81ba7ca32b139cf943575c57e3e1f275e2438522bcaa",
"3/self_attn.o_proj.weight/lr_a": "float16(5120, 64):a5fca313e4add16c30e05bd52f8d91515cd1d61b8395773c951a4f328159f854",
"3/self_attn.o_proj.weight/lr_b": "float16(64, 6144):204149cf4c82472a9398eba0a6677001012b11d87bc4a171fba3ab813b802e32",
"3/self_attn.o_proj.weight/lut": "float16(16, 2):6889e4d3113c0aa8b895b27a40590311b5a491d6dce8ada7a4d878f2517cd4c8",
"3/self_attn.o_proj.weight/scale": "float16(5120,):7a9cd54ac39df3739b1d807d267a4dcea6ef1f7616f92cdd077871d03b302610",
"3/self_attn.q_norm.weight": "float32(256,):e49e77664d2933745cbe29c25524a3e87728e4093daa1363621799bacf0d0324",
"3/self_attn.q_proj.weight/idx": "uint8(6144, 5120):70aa6cd1fc631ce5661763c3c126f1b6829c7abb80c87ef49b5c12293de6a0a6",
"3/self_attn.q_proj.weight/lr_a": "float16(12288, 64):50f317e2ae39b09b0dd81e2810a25aac693df9b4f9efc4f62598711bb1cebe77",
"3/self_attn.q_proj.weight/lr_b": "float16(64, 5120):9d46966e7c4d6a760689d80272c690cf032ff1464137dda844cecfbb5c38d5b9",
"3/self_attn.q_proj.weight/lut": "float16(16, 2):43a7ff0f7518fe515a609c979fbef834bc6eb816b2b5aaf947b22b12deb99203",
"3/self_attn.q_proj.weight/scale": "float16(12288,):dfaac7bbdef2ce3965dafefe143365a05cd74ded735cc7fcfc24e60f45e36fb5",
"3/self_attn.v_proj.weight/int8": "int8(1024, 5120):5af7118bea6c1fea8ade76420f26a857a12120185c4db3d3086204f18baf58a2",
"3/self_attn.v_proj.weight/lr_a": "float16(1024, 64):2d452df521a36bd213d37d1724197626562c6a236663204aeca021e7e6f4d4fe",
"3/self_attn.v_proj.weight/lr_b": "float16(64, 5120):27610c143e6ffd04409ea2c792b6906946ab614c425f3e131e7ef288ad5fb1fc",
"3/self_attn.v_proj.weight/scale": "float16(1024,):75f95d1e6bf21ffdbb72ab5c32a8f4b430a56c0a13838f0e3d9903f1054769d1",
"30/input_layernorm.weight": "float32(5120,):ef96a792ae2d95eda997ef27a78354e6777f4298acb869fb69fecf54128e6b7e",
"30/linear_attn.A_log": "float32(48,):5b46c63c7fec40dd40794f8e4a99d85ebe47d6936769a6981b5979cbdce0ea56",
"30/linear_attn.conv1d.weight": "float32(10240, 1, 4):2d2e6c484375e690871e65a841b8c13b56270631a9c0fd3aebf46f92302d129e",
"30/linear_attn.dt_bias": "float32(48,):1c4ed6236b1c778d8d911d66691179892a3802c4935fd5e2ce6c95fab5be9981",
"30/linear_attn.in_proj_a.weight": "float32(48, 5120):993e7816eb3b9fa1c573794355f7e21945257cdefa7a5e5117191b94ed32f7f2",
"30/linear_attn.in_proj_b.weight": "float32(48, 5120):0d69f04be0d47ea2a58747cba501b55354ce7ff4466b67b3848fafc787c1db86",
"30/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):129a27c02d8b8d1fa2d558ddcc60e49b81f918f106ba61e19da7253570525890",
"30/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):85cdb0a04cebf442c3b4936f6292639632ff535e5722315a875c9e850fcda33f",
"30/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):0244e27dcfe8466bf5253d5bdd29cfe5302fe8a1430dead714a53850cc0fde21",
"30/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):5100bb0e7ea4ec85f7830f42aa86c1ec5fba7ba66b355cb34adfd19398c2e8f9",
"30/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):11259ecaf99356ba5424961c49853ba6d3009a905b689ed02149518ebe4257bd",
"30/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):e871b1ae13092720fa279eb93f48f45984207e27dbda7c1d2b07f106d8b46375",
"30/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):2b9c1c9dc48f4aeb69cea3b48d18e287f6df0a5841b1aabb6ca06747f39c967e",
"30/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):0076e0cfc9430b949aa33444aa9aceb99ce3a5dcf9550315c4756f72e1da0dcf",
"30/linear_attn.in_proj_z.weight/lut": "float16(16, 1):fad723ead84cb9fe20aae42ffd7c2b97f91f82afd4eeeee6e65e5b489afd2c63",
"30/linear_attn.in_proj_z.weight/scale": "float16(6144,):7a47cfb1a5616895c45fb2688a77da0ea20a2d073e1a4cc8136c09466fda2a5c",
"30/linear_attn.norm.weight": "float32(128,):b6de949813faa0f5660753d2b2aa8b4fdf5c02708c1336f80471be35a637d20d",
"30/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):e02da27d4983df0bd27d65a6981c3b55507164d736cda449b2a0bb5b6af3c4f8",
"30/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):19dbbeae7b760e87721e439b73aca9c54244223a548e208408f6b1daa0f928cf",
"30/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):15ca23726caf1e199b4bb8c161e3b5ca9de2ae0f66fec6d2834c9fe30c8509d5",
"30/linear_attn.out_proj.weight/lut": "float16(16, 1):d5423cac72ff4673c739eeb29b36e568cf404c28edbfbf11ac0013d555e0807e",
"30/linear_attn.out_proj.weight/scale": "float16(5120,):4ad9d3788eddecc416ac6981ed1f348a29fb7a8ed6232471b1a1d20054a3a465",
"30/mix.rotation": "int64(2,):d2d5d2fb1779fb7c5f7e0f7042767c1e10e4df388abc9cb4ea48a8ef3a5c87ef",
"30/mlp.down_proj.weight/idx": "uint8(2560, 17408):1d0356f0178144b6c88c9d824b520bb8ea8c9120fbfb7b35248eedb17ff8421a",
"30/mlp.down_proj.weight/lut": "float16(64, 2):1f678965ac134326445e44e01fb7df393f3931e29380672fefb9cd37cf0a1d55",
"30/mlp.down_proj.weight/scale": "float16(5120,):85f32ebd2513b2e6fe0736b459addb5cc7f5c90b1a2278f6de17fb6d26a13643",
"30/mlp.gate_proj.weight/idx": "uint8(8704, 5120):74ccd57bf9b82d5a77623c896d8284dc874e106098d7941793486b48dfcf5ca3",
"30/mlp.gate_proj.weight/lut": "float16(64, 2):be8d8559a6e1a258d444d29828c1efba958fa733a157ac7040fadd1da53201e0",
"30/mlp.gate_proj.weight/scale": "float16(17408,):3a9df8180d844628f4726fd958bc64dcdb47830b9b202bc8b7b72c52b994c62e",
"30/mlp.rotation": "int64(2,):b624847329a938b165f96766d96860fa761d6828adb74f28f9811a0a9a484d65",
"30/mlp.up_proj.weight/idx": "uint8(8704, 5120):1ee01d8d3d0a7a6fbe0901e759491e3bdbb913a3dda10a78e8ee414bcdc4c515",
"30/mlp.up_proj.weight/lut": "float16(64, 2):2b74a2b65397ac3de4dd6e980abf14031a932a3be6110e6f336ebae7134d9564",
"30/mlp.up_proj.weight/scale": "float16(17408,):f86f40c5c301175389d7554a02b6bdefb7d480b4fa65cd276bf357a7d74ccf43",
"30/post_attention_layernorm.weight": "float32(5120,):71775c6a465032ff2f3aa7b6673f2d9c247f39d0d556347d31cd20fd159924ee",
"31/input_layernorm.weight": "float32(5120,):4f28f1b820f4de7912b183a73071d4641e762b42a16a35358cde793f9d1fb2af",
"31/mix.rotation": "int64(2,):865162daafce1f22a90881f722b13944978b87235f4795677e9a35b528da84ac",
"31/mlp.down_proj.weight/idx": "uint8(2560, 17408):908d093095f31693c20bc01b63645ea677ac3ac40fe3394e3d289fff12f77ed1",
"31/mlp.down_proj.weight/lut": "float16(64, 2):232842a38cdcb0c1389ee9f8f684d65c4a5b0207957dff11f9164b41f4484f7b",
"31/mlp.down_proj.weight/scale": "float16(5120,):a62cd0a1c46a3e603bbd46098dd21e275b8793474faef8ae260d7d2cfdd77737",
"31/mlp.gate_proj.weight/idx": "uint8(8704, 5120):e27ce717135e09c05a427740278aa5c502e4069e2050468752d947a27e920808",
"31/mlp.gate_proj.weight/lut": "float16(64, 2):9c04b06c682cc44a70af02722d39857df389fa3ff29066ed6f2eb1ea2ad5ea4b",
"31/mlp.gate_proj.weight/scale": "float16(17408,):4512df5f746c5e5c3f6ed39aa4dd0ae3bde05b5a51fd9731cfc66961b8f4adeb",
"31/mlp.rotation": "int64(2,):74f34b57e8ab40d9c95badaf31e038028a79e3e2f2c218b9c61779af004efb60",
"31/mlp.up_proj.weight/idx": "uint8(8704, 5120):612b5ff545fe4b5e42f9c4009376e07f38b7a75ed927526b62dc1287676166fa",
"31/mlp.up_proj.weight/lut": "float16(64, 2):e7628dc0b64fdf5c0e6128f592212ff674beaf82fbc84a33958f4fb7b5781c23",
"31/mlp.up_proj.weight/scale": "float16(17408,):f76b8f336857595626ca35d0fec40cf2c8958b678c723c2dba680bcad0f475c1",
"31/post_attention_layernorm.weight": "float32(5120,):aafc18c9cb09fb42acf9e44dc1d7004eaa1dba87e57ca6a56f11c9c5f6bb96a5",
"31/self_attn.k_norm.weight": "float32(256,):68e8dcf3ae7e896953a6a9192eced656a867806cddcd4bb6297cd26dd57eef5b",
"31/self_attn.k_proj.weight/int8": "int8(1024, 5120):8460aaf59797f142b031585b98a2975ce8db1d547aefe69db52e93966262c06d",
"31/self_attn.k_proj.weight/lr_a": "float16(1024, 64):264814c25294b2ec982292f3390e455eff5a233c0cefca09b82326944be7decd",
"31/self_attn.k_proj.weight/lr_b": "float16(64, 5120):1338b2bb072eac82b5f1c3e94370262ac32573b266933bf59b783273afa48321",
"31/self_attn.k_proj.weight/scale": "float16(1024,):473a8f9c22b2cdd750f5fead46f12fef76c8e2929abbfa6ff1d81ea0bbd689bc",
"31/self_attn.o_proj.weight/idx": "uint8(5120, 6144):1097a5a08c3561050418146ac13897741f5e36367c360d10333d1a19b4ec2e27",
"31/self_attn.o_proj.weight/lr_a": "float16(5120, 64):4232b3cdd3bd8544f4b4e53223aa9ff52a54c0dada53ad3f18216aa60e6a339b",
"31/self_attn.o_proj.weight/lr_b": "float16(64, 6144):3e8e75dcc613de81e760a8428ecaadfc308005761d4da1e556e647704cc64175",
"31/self_attn.o_proj.weight/lut": "float16(16, 1):e16e304257910bed37beb1a5dcf992a347a2bd3314acd4336c3d63f072a9cf2c",
"31/self_attn.o_proj.weight/scale": "float16(5120,):1b01cec90a7236fef4f39eaaf6a93fd8d1d9a3af1aa358aa8e92dcedcdc1007c",
"31/self_attn.q_norm.weight": "float32(256,):405acfff5033d70a85203b29206b3e40762701f3a687e252ec98741b785cf087",
"31/self_attn.q_proj.weight/idx": "uint8(12288, 5120):e1c9107020af0aba67d18c4f6add0708e9fa5f36dc811f746b32e8330c90fac0",
"31/self_attn.q_proj.weight/lr_a": "float16(12288, 64):1ee72b088eac78d97b084a4b4524fd317c58b4a8499c8b0e5396dd9981f988ed",
"31/self_attn.q_proj.weight/lr_b": "float16(64, 5120):bdcc0bb4885a69ec15454da73fc8d26c21298f273a2df4ca3b9ad9f3cf5c5851",
"31/self_attn.q_proj.weight/lut": "float16(16, 1):9924cb8e5bf686fc4d41b16121c228a72f6a33469b2feb3f32f358fe5543f822",
"31/self_attn.q_proj.weight/scale": "float16(12288,):df1292fd2726544357fa3bd6d723c90c9c9e219c7fa868f4e0e559f0f125bf42",
"31/self_attn.v_proj.weight/int8": "int8(1024, 5120):255c16c151abc7d3edd10c7f078ed8aa83fab0d0ba5d5f99b65d67b7ac6ae237",
"31/self_attn.v_proj.weight/lr_a": "float16(1024, 64):87fc8c47cfd298ce5a1e8c2813dcfe3aea562e1c80e5c17d6249b133acf66991",
"31/self_attn.v_proj.weight/lr_b": "float16(64, 5120):35b08a85cf4dc780411a33121e6f853e5cbbb30606be926166b5d89a1757e379",
"31/self_attn.v_proj.weight/scale": "float16(1024,):a4c072a31552ee756c252f215518347525bb7b4ca716e8672754528e24314b1c",
"32/input_layernorm.weight": "float32(5120,):ef36be87d8699493216f99b057126cc955f43b023fd6cb320abf0329f1a2445c",
"32/linear_attn.A_log": "float32(48,):75507ab3bc2da7e2b2f1f8dc4cd163885f594b4a0ebd049db473e9c6fa10ea8c",
"32/linear_attn.conv1d.weight": "float32(10240, 1, 4):ffe5213c226aa1f686641006ebc6561297ae96a5bbc37673568875fb32ab1a79",
"32/linear_attn.dt_bias": "float32(48,):6265adbf8f829e6569001977f8c967fe89dcbd27e9e95b052a86a33f8a382667",
"32/linear_attn.in_proj_a.weight": "float32(48, 5120):395a47fb3c0234d0115ce944f030f74495ee9e68f26b95750dcafad1fe706d76",
"32/linear_attn.in_proj_b.weight": "float32(48, 5120):c872b9d4fec0a9db1ed417672f8f9dcd98008c9d5038a2efd354b0efbbf11cac",
"32/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):8f50c1c6a8cbd2fc1cea7eab8d1047e5c220a44c6ea9c1af41b15e84b53ac2bc",
"32/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):dd27ccd13695be73bde7478a404a7478b15276be37ac20d4be9433dbbf11c1ed",
"32/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):4b291664c51700722b12c79924ce8e48ce6d253d229457c918fdf450e5d47e17",
"32/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):60a26b8c3ab72d328644e280343cbb9b2031d027f18c99673bb990604446f221",
"32/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):306dfc052f6eefc029e3e185debb7e20ec0d6f7771d8128489d7a84e9f9deb93",
"32/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):81a00b71dd8db8f1ff30cef99fd4f9bc92aebf4ff2a4af734387618f56269538",
"32/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):b7e8d20d70e5cda59a19d1399cb8cfd277ad4113bef58686bf05de41df2322d2",
"32/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):f312481249f2acae9bb219a5ac685dc9a46f4d7233a3fb325359b6a334a91cdc",
"32/linear_attn.in_proj_z.weight/lut": "float16(16, 1):7a5c50617c227d00201b88c9d06879a43ceb7f6453dad1d17c5167984fcf6f3d",
"32/linear_attn.in_proj_z.weight/scale": "float16(6144,):0e14be27927336e51a878e9865dc42b7ff351a4e535def469b30bddd1ce37d75",
"32/linear_attn.norm.weight": "float32(128,):757077e482f04a2078f2ee8241fefe7bde7670a1c36780fb27123206fd0b8d3d",
"32/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):7ec3ec570ad9cc76578775f3497ad4ceaa806ceeeefb00c4e2c74b670d722069",
"32/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):f73d451fd7b84b0ea8d78af0ece9d2734fcf10a37e6c4aa0240629408193d94c",
"32/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):7a6801eaf98e7cddaae25e4ff31d92a61d7070261ef7dc6bb8195136600bc280",
"32/linear_attn.out_proj.weight/lut": "float16(16, 1):7f086d3d6b153828242976239b6cf23d73aad0ff964e4a04b10f902d16a3def4",
"32/linear_attn.out_proj.weight/scale": "float16(5120,):37df6b7aae5beb626b32653627a308cbaf29bd81270b23f04589f27ec278bb6d",
"32/mix.rotation": "int64(2,):3b78b3e79fea91dc0a8b79f6115967c268406d76f9eee15ff98f84171ec7d32f",
"32/mlp.down_proj.weight/idx": "uint8(2560, 17408):e03ef721535be2c194d2e6a4e11d142af4a3332957694ee9094e91134943f795",
"32/mlp.down_proj.weight/lut": "float16(64, 2):110f632a424ee36eb40ae33209e956f42019a243fe1695157d5b725763083b03",
"32/mlp.down_proj.weight/scale": "float16(5120,):4ea90e7bd50ab77182d95d41f9bd659af715262ff3ec7030b38b491c30c7885a",
"32/mlp.gate_proj.weight/idx": "uint8(8704, 5120):4aeb0f69e4023bc17ff658faeece1a85ad1f81da4d1b3eb545566ca1f144719c",
"32/mlp.gate_proj.weight/lut": "float16(64, 2):7db014a8fe62f760068f4821dd9519b2fc8c1e3134292c370acbaa2717c35add",
"32/mlp.gate_proj.weight/scale": "float16(17408,):55ce8a0dc3814bc08c3f556cd82ac87eb2ed177399af1d2f8174cf917c7e6871",
"32/mlp.rotation": "int64(2,):ee80c6d3409ae56c557f975e123ab78297943914fb4b006c218cd47d4ff0bd81",
"32/mlp.up_proj.weight/idx": "uint8(8704, 5120):6da02df2c1cfe385ffc0476e6ff4df3d5d1d041e035728de0232fb52bfa061df",
"32/mlp.up_proj.weight/lut": "float16(64, 2):41096e249218381dc268ee18b165bc5721a963866a4ed2b67e0a5dd6d107d9ca",
"32/mlp.up_proj.weight/scale": "float16(17408,):572bcfaa0f8fb118a30207d389543d7fc474468d8404c24818ade6defe85ad4d",
"32/post_attention_layernorm.weight": "float32(5120,):967e49e98a86cb5a9550497c28de6587524c71139697a359ffd6a3bac60737bf",
"33/input_layernorm.weight": "float32(5120,):4c21e29283dd3856c313c7709d94f705d587e48d105bdc4d453042c06ef514ab",
"33/linear_attn.A_log": "float32(48,):4979172efb58e868ac7918f6b5e4227258b2c11426b6fb4f0a438fc079cb68fa",
"33/linear_attn.conv1d.weight": "float32(10240, 1, 4):6603ba7caf299caa2cc08e3498ffcf3ced992754b68f55da418a6fed4ca9d803",
"33/linear_attn.dt_bias": "float32(48,):d963162a8f295e9db1d81acb30e769c537d22c6868952b4b9a585afc42d9a8c3",
"33/linear_attn.in_proj_a.weight": "float32(48, 5120):920f1c756b0d2850798e65c0e6076c76d324b387a1b4ef0ba84d6c1c705d20ea",
"33/linear_attn.in_proj_b.weight": "float32(48, 5120):b42e5e1d7c6d27b60dfa14a97715cfbb53a3e8db853b3e39bc4ca99e8ff71c7e",
"33/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):24f9fcd268d7dcee2113d6bda043f92b6f59537c064aa8a7d54949fa35009371",
"33/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):9c0ef719cf60762cd2c628bc689d203c3a5bcb08c35b933ae11f7c254200cc9f",
"33/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):2a2d9945aab222037ccb669cf5cc0fc75f17086d22f600b0f1e4be62588b8c8a",
"33/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):7f15007b88a271f08e3dbebb3eb345872c5e83eff5dff33f107a877fe318108b",
"33/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):c69e98a13240d5f1ce2d519ee0e8b6591f8a6b43b8f4d1e55e741da5214871ad",
"33/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):ae2fb920c3800bed4b201a7899765783bda159afbfe67b77408bece9a88be911",
"33/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):c67038657db9388a72a7ce58a0fb247b19087a523c7d6f8eb1953c5c3d1c2cdf",
"33/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):2d354a39b021a9baa53103a77cd45cc3bd89cc135af7095d29b58cb65ee649c3",
"33/linear_attn.in_proj_z.weight/lut": "float16(16, 1):5a33e9fcd188af39414c03225a36463b27d08fb48aec5a790b8c3443824acae6",
"33/linear_attn.in_proj_z.weight/scale": "float16(6144,):8e0d41190966ec7df4b33f1dacd10528fd252f58def79a8929db7f9609998a22",
"33/linear_attn.norm.weight": "float32(128,):4ff3f47c6f2d8cbbd2b76a1240d358483014c5729f1a0087d70d23a627029855",
"33/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):b7947c99805b042118668ef4fc8e5ff1d9ae6984d0c6832034de652e78c667e5",
"33/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):ea39a8400661704b1d2778753b50119b696448a991d81fe700d479ae37bdd6c1",
"33/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):62c020402fbeac3e91fef12ca1b9ee8c828e4616d5fb08563decc91286ee3163",
"33/linear_attn.out_proj.weight/lut": "float16(16, 1):223c9c04fb6620339ea795dbf475e902c261b8bff2085b23c727540e9c74419f",
"33/linear_attn.out_proj.weight/scale": "float16(5120,):84eaa42f9181e4abc0c1658b98cbf7f287caa6cb16e212c2dfb54fc29decdf3f",
"33/mix.rotation": "int64(2,):0bfdc775f978a1824df82423c20012f53eefd3bac3738a8af92a1a427451e9e5",
"33/mlp.down_proj.weight/idx": "uint8(2560, 17408):62ec4bf9c9226750ba4c9b5454958be8fcc27c6bb920d778a9280cd3c8a212fe",
"33/mlp.down_proj.weight/lut": "float16(64, 2):006e892bd0a082c2b29a609316fa657f56f961ad6e6d722d6547193af213fbfd",
"33/mlp.down_proj.weight/scale": "float16(5120,):982a9ad42aa2a4ce4d14ccd2e7220f12eb35129998578ec9a058a17a5d0a37a6",
"33/mlp.gate_proj.weight/idx": "uint8(8704, 5120):db5c77c3b8a24fa3b10ab46d5111d92ae90b0f3af0ab539cee9e4609743b67b4",
"33/mlp.gate_proj.weight/lut": "float16(64, 2):fc7cec553f3ba7caa17abc7415d6f00d90a99bffd4dc1ebacac3afafae080b53",
"33/mlp.gate_proj.weight/scale": "float16(17408,):98ca3a4c346c9c334175f8979004992febd4d36edabd1bf5cd4b7d11ee7c3211",
"33/mlp.rotation": "int64(2,):52524de3eaa815b9c68f112404f9ffb6544655c7e3df4174688ec093631349d5",
"33/mlp.up_proj.weight/idx": "uint8(8704, 5120):8abe7a06b01d551fad278c4e57884c79e0a9b2e0cdab25ba81a8ae2e46a429e4",
"33/mlp.up_proj.weight/lut": "float16(64, 2):5e2570b052a7622a1817c31987344984641956e5c4d0e589cd53b803d95a2dab",
"33/mlp.up_proj.weight/scale": "float16(17408,):b234bda6f2e43051dcbd4acde5ad3a2bc30a9a4a3b31a36ba7ba06a5605b4cf7",
"33/post_attention_layernorm.weight": "float32(5120,):1109b805681624326da257743105decb3190347379735897a7ae29cb5462c13a",
"34/input_layernorm.weight": "float32(5120,):bf4a74a4493dda62593653e1fffc84dd898a20f57704cbde85ec96634e468c67",
"34/linear_attn.A_log": "float32(48,):da9a7b1636376183f2b82ce69067c7a69448ee0a70f3c0dac0182babea67b0fc",
"34/linear_attn.conv1d.weight": "float32(10240, 1, 4):973ee304301603b12327a19091ee58ccfee2932f41dd8edf13790c3799419923",
"34/linear_attn.dt_bias": "float32(48,):768bb1458a25ddbaacccb4d5138e15c591c416081d46d2c1f07a0bf8a345dfa6",
"34/linear_attn.in_proj_a.weight": "float32(48, 5120):5586891171ccf3825b79f119653eb262d33b696c37bf3422aa92a150ac6903a1",
"34/linear_attn.in_proj_b.weight": "float32(48, 5120):fed9c8339efa2bfcbd9842c3dc0973a7e62c8b0292f36b5d012b3d761ece23ac",
"34/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):e84d91db176e9877a13b921b7d6b1cdfb3f90be9b5130d6b20eec09897630f6c",
"34/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):8cf5596471df9f220474465976b3785d5d64e59eeaf774efe12359aa2f9aad90",
"34/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):a324ce09ea6a036681778cdfb5b609c593ab78406658a39b5dd190a816e3f13f",
"34/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):d3233ae46eb4d2b4bff7b61b1c59ab7d3e285769769810446636f236375b720b",
"34/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):0c8e96372716f64ac996e2e64bab944873404ce24c7e7f468b0e5847b72be92b",
"34/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):56d820685b08bc0c71d6bb1a7fe9c8ef371c04623be5f6cf66526c3490f10eba",
"34/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):f41c46670c0a48b692be907821de892c9ea94307f7e9346bfb5c8084feadfb32",
"34/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):a3bdc829f6103f4303dcfefcc4c73d78cf0b10fa9d9c74baef8697c374b232ea",
"34/linear_attn.in_proj_z.weight/lut": "float16(16, 1):880b54774acb352c08b8dc101d74a6dc4aa6ac1a02f4fd119e59c1e48feac818",
"34/linear_attn.in_proj_z.weight/scale": "float16(6144,):2b6fde80bd20664fcd0df5a0b02344ba830e306743bfe2d7db18cc4c7119c8d9",
"34/linear_attn.norm.weight": "float32(128,):48c2efb358b0b04fe6e0ffa5c76fac63d2f1d727a6d3b81a15d184ba7138fc47",
"34/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):5400dc862899b3f4c62351998923d82f177a36209d70bfd5df1b3beed3a16756",
"34/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):af510f29196c6e9e080ef8e8f8144b7f25d5070990b0b51a0e17aa9dd0ebb797",
"34/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):d5289cec027b8ffa2ec548e9fa44ee334abc137de9340738c1576aedaf97da9b",
"34/linear_attn.out_proj.weight/lut": "float16(16, 1):79a1d1642fb35c0fcf128ee9e98531d73e68a9cb21c25ce82153ec09daae8d7f",
"34/linear_attn.out_proj.weight/scale": "float16(5120,):114fddc94f854cc2f71bbfa04e3735ae6619996f3c4b684d47d890c73bd10947",
"34/mix.rotation": "int64(2,):894a337532333e4397c612d50d592e75bb5e492c17bd5ee8fd37d1c03e6f207d",
"34/mlp.down_proj.weight/idx": "uint8(2560, 17408):744aedc9b0567752bd9539931fa7dd9a909ae54a7c841bee2f0b0b7d15edb100",
"34/mlp.down_proj.weight/lut": "float16(64, 2):7d4638cdb6c7fcd355bad2bc8c8912e9d7cc05579214a0bca189b37cc368706b",
"34/mlp.down_proj.weight/scale": "float16(5120,):8067d9803a64cd765524faa77360ee94c35e9aa3899cc749c636181cd0c0e128",
"34/mlp.gate_proj.weight/idx": "uint8(8704, 5120):10c6ba7f3c5ec59f0481042ac4f76f3f5b4a90c94f6fe840dbd9f4b25102c234",
"34/mlp.gate_proj.weight/lut": "float16(64, 2):389b8e14fcd89f037c9e8ebe631b429af08797671fb3a68d0559e54ccf01cd1d",
"34/mlp.gate_proj.weight/scale": "float16(17408,):91c6044c0aaa06411d346e314ab4ddea6f98f87fd1cbe90639e9174bce59e83c",
"34/mlp.rotation": "int64(2,):19bb88ba100e5fd3f6630a27ae0f65cb7f6f433d23f3e5f82bc0004ef8ab919b",
"34/mlp.up_proj.weight/idx": "uint8(8704, 5120):9c9078019ac6d39a1ec76005d44115293f540e0c6bd9c63712554204937473da",
"34/mlp.up_proj.weight/lut": "float16(64, 2):aef14f62c31fa236ae6da3ff36e15a5186aa4a0bb2d8ea6f6fab2d2edd7e46d5",
"34/mlp.up_proj.weight/scale": "float16(17408,):84a8e6c2324a439b999bf952e301668545d77b6fa4cc6c0553dcfe1f687327fd",
"34/post_attention_layernorm.weight": "float32(5120,):b9b837f80393a30382bda5703d9d681e5289067b1f86295bc0727449271787ad",
"35/input_layernorm.weight": "float32(5120,):f65cac373dcc3971ca9abb97e74ef022b0cf2be1ecfd1811b98dad5cba033b68",
"35/mix.rotation": "int64(2,):9e6796eee6afe097a04bfe0744cfa7c53c571588decf201418421039457c5087",
"35/mlp.down_proj.weight/idx": "uint8(2560, 17408):38dab5806643bf4f7404ded7b44b5f17080dedc4a1bf30e7a047ac065117f1f3",
"35/mlp.down_proj.weight/lut": "float16(64, 2):5ce6c55a19cf88aa600a3c0d7f66f789ffb36814c885c07a83f4d6b6e4ef9fe3",
"35/mlp.down_proj.weight/scale": "float16(5120,):4f7304b3bcd3f4cf8dd04fb1333f086f2dc544a0a6d055fe4d253ccf72973e8e",
"35/mlp.gate_proj.weight/idx": "uint8(8704, 5120):6c1cbcae9dd30c62730c9dfc5a8a2839acc9d28b69fbf23ab6b943e9db76a950",
"35/mlp.gate_proj.weight/lut": "float16(64, 2):96539ccfedb73d4dfabda4ff26056e8a087049902df6c631a7cb3a8583baaaf6",
"35/mlp.gate_proj.weight/scale": "float16(17408,):aaf20ef5deac090eabd2bbc6c089eb8b254982128ba73fedf33ad3da3b0da07e",
"35/mlp.rotation": "int64(2,):994ec8f69de5dc527f06f84636d81e1f86aaf442dbf6f4fec8922a0be0c17bf4",
"35/mlp.up_proj.weight/idx": "uint8(8704, 5120):014def924e6ddce278f2c23144f33bd3f9c55295eaa3f874da17dd4409413860",
"35/mlp.up_proj.weight/lut": "float16(64, 2):769ffeb0d79b85e4da9a3edd1c68750cebdfaf0620488975b2aa826ed143d042",
"35/mlp.up_proj.weight/scale": "float16(17408,):38798f4ad2b3ca5d1a7b61db61a2d21ad25f7336de318aa7c52e8b0e9c93db08",
"35/post_attention_layernorm.weight": "float32(5120,):8fd67043491920f8bdf8439ee395aaa24bc816660d85f68970bee031bf283a3e",
"35/self_attn.k_norm.weight": "float32(256,):d8396a52c2ae070f57a3f321d304b3c541cfecb29e90a8552655236522d92b96",
"35/self_attn.k_proj.weight/int8": "int8(1024, 5120):3fccf333fe10595a3f879a225701c6169be79bc5fa053f43d7a58422de2ae1ae",
"35/self_attn.k_proj.weight/lr_a": "float16(1024, 64):e80ab456b93b83f3b794dd41d488487c4d2de3e387bc025a4d5159d72e9c253d",
"35/self_attn.k_proj.weight/lr_b": "float16(64, 5120):913d35ca0170248bce497b20379b8df5323cb65f60ebe07fa30ed5b0c04bfc49",
"35/self_attn.k_proj.weight/scale": "float16(1024,):a4b4ea37c1fd1f666f9519a089289f381f9b63169bccd82232f60d71ad5ba2ab",
"35/self_attn.o_proj.weight/idx": "uint8(5120, 6144):5733277293770b1a9a0692daf9f170b6d3e811604c3ff32bdc90404c603d64b8",
"35/self_attn.o_proj.weight/lr_a": "float16(5120, 64):16ed8e7c61653ea2516459e31d0ba63a4ccd3cf6d8f24e95699226cfbfc0b93e",
"35/self_attn.o_proj.weight/lr_b": "float16(64, 6144):4d4d4e78b3dd1ecb4a04f1c47155d92046e03f556be98d49ac37e614fbcccaa4",
"35/self_attn.o_proj.weight/lut": "float16(16, 1):dddc690d720a244f3c2a92b09ab64ec1fc9c30d2e68e2d1d53ad1a279f948b1f",
"35/self_attn.o_proj.weight/scale": "float16(5120,):bc6edb7b831c379a34d090ef3ea7162eba6bcbdb47e30c615e5eff7394133180",
"35/self_attn.q_norm.weight": "float32(256,):2652eb317156a156bac54375243131116daf9ca488e4da270d7d2485e14cd32f",
"35/self_attn.q_proj.weight/idx": "uint8(12288, 5120):2055319e166898ec1438867ce662485c13f5455c768d3cecdb866b333791fd42",
"35/self_attn.q_proj.weight/lr_a": "float16(12288, 64):a1936a4d84e895e5a77d5b4410b84f52fb4baf771e9f85cced0b52e40d88d1e2",
"35/self_attn.q_proj.weight/lr_b": "float16(64, 5120):b23609e12c23e0003eb379795d0d730274eab4c83718d6ff2ea5cef7723c4b3d",
"35/self_attn.q_proj.weight/lut": "float16(16, 1):7f7b8a21529a4e4d6dc3e9cc077fd4543b6d5bc99d0d9093eecb77341aac2921",
"35/self_attn.q_proj.weight/scale": "float16(12288,):6758d991c02f091d271bdd33cceea6238c68c76b2d91be95b89d1ff98f01390c",
"35/self_attn.v_proj.weight/int8": "int8(1024, 5120):17c7057b4fda2b717612e19867db37d1e5333eb6049cc416bc1932ae00fbd430",
"35/self_attn.v_proj.weight/lr_a": "float16(1024, 64):7e316ea7c53ff7398715da324b19730e9979c6d419fa85dbb3fec2f0774b2411",
"35/self_attn.v_proj.weight/lr_b": "float16(64, 5120):6c59a3f6d069cc5a9553991c10c24c9af5af52b444027bb25eb88c5e2bc07532",
"35/self_attn.v_proj.weight/scale": "float16(1024,):1ef9c4a3e2b83428e992cc54210b8bbf0062f5fde71cc3cf97e17c07c6d210a0",
"36/input_layernorm.weight": "float32(5120,):9bd2697de3b7a4393970b59ff5bdc925fd0457d0cb987dd5917206022aaba4a5",
"36/linear_attn.A_log": "float32(48,):1938542b53135dbeb3a7bd12ad30e61929b512505d0b71608ad89d00b36d4b48",
"36/linear_attn.conv1d.weight": "float32(10240, 1, 4):553f188ce3b9b0e38425a6a2ff568188dcef48ac9716b10c36110b7e253b40b5",
"36/linear_attn.dt_bias": "float32(48,):d179d5722009095f8c0e0c5a17bd3304f1a9fa002c2a5eb89da0ea0c77ed3c54",
"36/linear_attn.in_proj_a.weight": "float32(48, 5120):80c9f4b7d65348a30379b9a12a0a1e7dee4c1ebf7fa43ada06a6042bffceff89",
"36/linear_attn.in_proj_b.weight": "float32(48, 5120):4b95c1124d1caec944e74f2f1213fa54483fcea942f3a7905f9e1c1d62cb7373",
"36/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):e82a1f5ed1cf219b4e10f39d4a8607e8c505149872cc570ad2c77fa2cbfc6f93",
"36/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):28f7ae92f5a4a8f8facfd9361e4939016e6a72680df6dccae5e3eee757b0cb2b",
"36/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):0ff6d32576bf185a466f13a38ac099e5f4d65f0f3480fed383d758c2fe243f70",
"36/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):207f2991bc745d3678c88da4ee1607299bf1a64b1eee7f7041b41c6477d2ac64",
"36/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):664f754682c19d85852ad6f1f1e5b861721283419351dc6266d7abe1145bbd54",
"36/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):831545ea8169f2b32e7e4d84329aa8a1ae4a5841907626773d448b88f962b947",
"36/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):e3d64a8edac09e0360b56beae040d023be7148e1b0190bb1a4765fb740717fda",
"36/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):fcb7e5528de40ff2c87a7ca14f37d2bb78e0d14e9eed1ffef319b67ae7f36e0a",
"36/linear_attn.in_proj_z.weight/lut": "float16(16, 1):92d33ba6421d108b00dd227f9aa46b0a9dd900dabf273f222a0b83692497d472",
"36/linear_attn.in_proj_z.weight/scale": "float16(6144,):0df270f39c5dc72c149d8bdb54b2a6acee781008a9c705a6778f01f5d6f35d36",
"36/linear_attn.norm.weight": "float32(128,):5c8e45187325e700886f22ebd36a6025ad35718b2c4ab58687eef99575ab25e0",
"36/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):03e09520f826eee5f7e6e59f68748e9ce3fb36bdddc511e5c96be82ca5d49b70",
"36/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):fc3df26731221465091f2151ebe30b2c8f7dea7596a53a0d8ff727f9c0fc15ae",
"36/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):b291dba17d19c1ced596baf7a5ccd94ec2158d8dfa99c4dede4acb411bf4b1f6",
"36/linear_attn.out_proj.weight/lut": "float16(16, 1):a7b5dfaa2b07050dadb16d8801276c06148400518ba89e57dd2806f288bc3ecb",
"36/linear_attn.out_proj.weight/scale": "float16(5120,):4ff6378d825182620a8872762741ceab949c676a2035edb8cf76a908eaaf60b3",
"36/mix.rotation": "int64(2,):73d7f9d6a5f50ab1ecf2cae774f9bbf653084091a5924265768ffd0faa288e35",
"36/mlp.down_proj.weight/idx": "uint8(2560, 17408):af2815e64c5bb784a0d20ff7b5de5f96c33a48b538b923200de57f4e2470059d",
"36/mlp.down_proj.weight/lut": "float16(64, 2):2b1fd280474787987e6a05a46b4033470308cab19f22e01f845b580cc43c57a0",
"36/mlp.down_proj.weight/scale": "float16(5120,):3ff27570025f6b364311e7fbb0fa072cce63208602c7680d04817f7b84e2e9c0",
"36/mlp.gate_proj.weight/idx": "uint8(8704, 5120):bbbd6e2f17f07d87eed36ab299cc0242ab19849464fdca902adfcf18c390867c",
"36/mlp.gate_proj.weight/lut": "float16(64, 2):220760c017f22c83ba11bc75cc356f685c419c25cb8541a8224195189348eea0",
"36/mlp.gate_proj.weight/scale": "float16(17408,):391aaae1f5b25df4542077e79af588c81e9874e5e6722adaa87bce2ab9f7c704",
"36/mlp.rotation": "int64(2,):9935f5d35d82bd737de6e1e1bde19c5a874a438bb54aad8d55ccaca86364b6ea",
"36/mlp.up_proj.weight/idx": "uint8(8704, 5120):fbb671c84a8088ff371d20eaf5ca12e79b8ee7e0f135bd67757af41401975f30",
"36/mlp.up_proj.weight/lut": "float16(64, 2):0a7a72fb87cae8bb216e2e84ff709a488c3de90425865ce3c451f3b965dfce8e",
"36/mlp.up_proj.weight/scale": "float16(17408,):39d4db9fe31277ebd3e10fd89e6aa971a59cac407c51ff27a80bde98fe385833",
"36/post_attention_layernorm.weight": "float32(5120,):b74d13d8b6833166513c909620ac7246fa9a13f0fa34fb4dbdbe383f90101755",
"37/input_layernorm.weight": "float32(5120,):8dd9dc862fe640016663726d0f9c74494b7e0096e92349a4f0a7c8c83ab8e999",
"37/linear_attn.A_log": "float32(48,):3b09942347c3a477b3301c9b7cf849585e4454b4f426c103f69fca41c332be6b",
"37/linear_attn.conv1d.weight": "float32(10240, 1, 4):277b7556853d2aed6ce386265183e379aee3932f759b1eb16b5ae0593aaea959",
"37/linear_attn.dt_bias": "float32(48,):1bf415b25310415e72436a3f5a7ac139c31ba36dda04ae9f8a73115f73e196a1",
"37/linear_attn.in_proj_a.weight": "float32(48, 5120):e8a81976fe3d53c3379c12106494d84d42edfa791dfbf018706d626066687fad",
"37/linear_attn.in_proj_b.weight": "float32(48, 5120):c9bf2f7a8b7f7407bfdf026740644b100be98c3ef63828ac2be071a5f8e4c6e0",
"37/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):86d54088526288a17346fd85284fcf0ed838d837cb91aea4b72dbd46773082a1",
"37/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):9b0fa7e891fdce63fd934aee19ea58be387ec72aef930eaf6899033f704ffd7d",
"37/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):92bcc273d67d422ca5fa9c47793176b9ad94ab4d03d0853b6149c27037bddfd4",
"37/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):781654bd1c1082fad0bb90ce012acd750a7d96b6ec482268bc5094f318a8e1ae",
"37/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):1e84fbdefb65c4b4ce0adcd9ba78aa494f93502c558feba2b76edd2e7b2ce4dd",
"37/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):fb803c60042e89db2b70fd46864add0cfb76113347e9d0c1bbd578f7b7e06121",
"37/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):dec577bdd3addcdfbb0a2ba2c24e50af4747b9c0010b1aa020f44d522cb93d27",
"37/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):93f33b4ac83d4ae95b49ad5fad53b0074841810ca27e89d949a284b15bca1cb3",
"37/linear_attn.in_proj_z.weight/lut": "float16(16, 1):d9c8f439a78ff9f24b5169acdeb9ab742e78fc60f933d28d77b6b12f6dfb6691",
"37/linear_attn.in_proj_z.weight/scale": "float16(6144,):572894b0be97024b3fb1ca69494f27c0a7377bd0361eee21e8f63f5ced32ef5a",
"37/linear_attn.norm.weight": "float32(128,):a84a51f082c360ce89c666be5a4bf42b27429a6247c24c9dce2561822df1cc9d",
"37/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):4ab81c3c2e8e8805a9c3cd29d9e98473565d7e909bc077422c5ad3f591cdc386",
"37/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):59ca61054bca6b047ee07ef890388c4ec63d435e78d721d11cd86eb95adafac4",
"37/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):a03caf433a62eab0dd6ddbddde8dc68ae61f01f88f5359526ccb968b389f8c3d",
"37/linear_attn.out_proj.weight/lut": "float16(16, 1):fdeb445c497c8de533461169b7417ec81412854051243a8066a05e7b6b96be0d",
"37/linear_attn.out_proj.weight/scale": "float16(5120,):de27d8b9c4883ab326a28e496bcf7c099b69c8670ce4c8c4b16a4b56554430a7",
"37/mix.rotation": "int64(2,):4e23011c1795361a0ad3a86a1689b5df15723239d89c63f2fefbc0454bc36df9",
"37/mlp.down_proj.weight/idx": "uint8(2560, 17408):161ff4ba6e37c50b06617866d7381d857000a0c801a04f563b9661d10db01256",
"37/mlp.down_proj.weight/lut": "float16(64, 2):d85373037ea3746c8fefb89745831d0b154b7a2156ffd3cfdf02a921fc93b616",
"37/mlp.down_proj.weight/scale": "float16(5120,):6e72b27d6755cc19027aa3e838368a14292e69873ad8c76eea14e0d92ee44e96",
"37/mlp.gate_proj.weight/idx": "uint8(8704, 5120):404b44f1b6417a6ea24e7752e4afa0cd492ab5f17ac3e1d50f9fe514f39fbf1d",
"37/mlp.gate_proj.weight/lut": "float16(64, 2):ece6cf18a678f02328efebb6a09e136c52af2f245fdb6a1dcff80054f3157e59",
"37/mlp.gate_proj.weight/scale": "float16(17408,):e6c9beddf088b0d68a201f31630afb95f51a46fa58c4af41fd12515c3c32099c",
"37/mlp.rotation": "int64(2,):94b526c0e96bcd2fb5888edf5d10164d8739629e5327446becef0259d5b2012a",
"37/mlp.up_proj.weight/idx": "uint8(8704, 5120):73c9c5d10fe560656636c18a7b69be9d0e8882c5e54eb7c462568d3e493ed9eb",
"37/mlp.up_proj.weight/lut": "float16(64, 2):5486a339d638835e870e5acb1675225ec42fc6344c2d7636e0f9d2af66121ddd",
"37/mlp.up_proj.weight/scale": "float16(17408,):eef6895e3aba3f4f105ff10dd7ede66b7ee42e34228396591d9bd6d93c15db16",
"37/post_attention_layernorm.weight": "float32(5120,):887a6aa1de26116beb5da48ea92e8d3e6b1fe8e04d9b0eca9f57f04cbea40bb5",
"38/input_layernorm.weight": "float32(5120,):f8fccc24b41ef51328949c2baa7379cbedd8763e5c2cf1bb86c371ba8d685528",
"38/linear_attn.A_log": "float32(48,):17b36340c26b6ca77cf26fb1ea383f8ede71ade17835d76b73712652ffd97791",
"38/linear_attn.conv1d.weight": "float32(10240, 1, 4):d52f0cc83862bc776b1590827c6239d349247b7664020cf9c94dbe07aa9f3d7d",
"38/linear_attn.dt_bias": "float32(48,):b609c8ac97549038b216c9817446f99a2bca632f6abe971f07f406085f6cb82c",
"38/linear_attn.in_proj_a.weight": "float32(48, 5120):730e27baf442840d38ee1f8a9d0e996d08a778a817b391764b23ca8c29fd0006",
"38/linear_attn.in_proj_b.weight": "float32(48, 5120):f45ed72bdba6f40afe6c00e9a2b556154aee7161a22e5bf96c0e878623ae57f5",
"38/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):1ef4cc4e6479738eb407d2dec89b17d25a3c9c1799ac42d4760ebc7b475910f7",
"38/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):05dc704270f014562ef8413498942d2946453e879e7ae965065551d6d99a7527",
"38/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):004d2d7c2b927d5806558b73f6e70cfbc970d18d74a9fb64ee74cdde05addba8",
"38/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):1c8fd0d28294bb0baf0feb552f34bb26e1e491c7c02cb37f9d9430fa4f7fece5",
"38/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):9b47ca7eac29d72013ec4cce73c9eeac99846f53cf9cc34c1841cd12e9e9c863",
"38/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):6fa3ef5ac3e5642daa7c176c576026d76fb98c3d49b39d74bed45c3e946dc2dc",
"38/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):752b0465dc1929fca8db6d3fa7201fbeef327b782e6878d78faa3e06a1dcd61e",
"38/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):c0b229cefd74ce2c4c7ea36e9212e7f35642d3f5268e97243c8ef5cae8c4071f",
"38/linear_attn.in_proj_z.weight/lut": "float16(16, 1):939d8fafbc044fbf67f190056c42dc7b09a3006d1b68c499c0b6892256ed0550",
"38/linear_attn.in_proj_z.weight/scale": "float16(6144,):86eb1baa2625e8810e2840faf59577c6a77dda1f4809465d096c815cbe3b6ab0",
"38/linear_attn.norm.weight": "float32(128,):4c63c8a5618ecc1f14fb03d2900917d6516d406ed558367364437a5d32cbca5d",
"38/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):f6262b3d88bb02de436d9f6050a53dcbffc6d15b1035beec2b6b7c5d96ee15ba",
"38/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):b7b4e5d61e1eb031dc1b72720ec2331a228c953a6f4d7e90b3954fe89b095a6b",
"38/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):37708546dbda5ea13ec30b7dc5f7d77568fcb2f06855ada3836e872513f1549e",
"38/linear_attn.out_proj.weight/lut": "float16(16, 1):7b6fd51d3d0064ab322a45b8be5e9613ac760d077048594c958e3741b7f41734",
"38/linear_attn.out_proj.weight/scale": "float16(5120,):33579f5253b4f8f30e4f96523ca638885820d3c506d6df886769b14a4a55741b",
"38/mix.rotation": "int64(2,):ff1bd3806dcdbbc5a8c8f39f6169a8fb5a8f5fb2b38b6503cfb241eba5dc2bed",
"38/mlp.down_proj.weight/idx": "uint8(2560, 17408):27f784951b9d25fb29e5dfee4f5de010761072a9599cfa31b9360d8327697c38",
"38/mlp.down_proj.weight/lut": "float16(64, 2):553ec03d39a80f5b0212ae9265dba588a28e338f6f33007d7868c759980be3f0",
"38/mlp.down_proj.weight/scale": "float16(5120,):16f2e6cf84a4603b409344d723f445ee6e930d4bd6c3af1b70c42f00914e54a5",
"38/mlp.gate_proj.weight/idx": "uint8(8704, 5120):2cbd05f959bb09ddaae7a3a29464f469697f62cd4117b4a53f930a02ef4ad30d",
"38/mlp.gate_proj.weight/lut": "float16(64, 2):e5c02b341d9ce9607583f1dcc8ac2c1ea091f9ef847eef076b532a7620bd44ea",
"38/mlp.gate_proj.weight/scale": "float16(17408,):a2e668c1f05d0ec67197d2c4f7012c271f10a5cca435ba9a8c2b6fb72c4628e6",
"38/mlp.rotation": "int64(2,):98283292401772de40e5115adab4aa43ad31851da9bcb6e35c1ae86b788c1823",
"38/mlp.up_proj.weight/idx": "uint8(8704, 5120):137c2e6d185fb2f2a5c1a8d81357b4bab710bb67e67bc4eaa4d575985f39a0c0",
"38/mlp.up_proj.weight/lut": "float16(64, 2):658bf4b4481bc539b2972d7198f7a004a487aa2d2e5078dbd3e5e1df64d892db",
"38/mlp.up_proj.weight/scale": "float16(17408,):40cebef79ef4cd5afcf93038d6b30245f5968c65517ecefb19aac0dcc15cc09c",
"38/post_attention_layernorm.weight": "float32(5120,):a74d6a29ea9534bbec63dd7b40d2ae21294e83060a6f34a9221ed33b3c2cc5b4",
"39/input_layernorm.weight": "float32(5120,):c1345cd6d57d73a97a8909a1b5216bfadb0ad97dcb5a19c1ca45c93f46702479",
"39/mix.rotation": "int64(2,):42006562117d745d80ed8b29470195227b6aaa8b43edb3dd98f8445c6cd9c642",
"39/mlp.down_proj.weight/idx": "uint8(2560, 17408):6ebdcab7e943d734d5930b4abde6f993eb6d8871b8783ac7ccdc3a9c20a00dc3",
"39/mlp.down_proj.weight/lut": "float16(64, 2):f2462da0892b4236913251009cdf5d8540995e055d166cb8fbca1b2a7b32ff68",
"39/mlp.down_proj.weight/scale": "float16(5120,):4e0b3dde5b491ff9e4943f8810c2c90aef36d6ad0780a4dc6844498663cf96e2",
"39/mlp.gate_proj.weight/idx": "uint8(8704, 5120):cbee8aefe87a89565280acb472e73fe6566861d3856eb365cb7200132a96cac3",
"39/mlp.gate_proj.weight/lut": "float16(64, 2):64ef9e857c9816aff3acd53e709a6be66eb8622f2fcbd6f70906fb71d9b06847",
"39/mlp.gate_proj.weight/scale": "float16(17408,):74fc7fac8fec3999433856a73818e9a770de96004f4ae492a0cb05dae3749b9a",
"39/mlp.rotation": "int64(2,):6d85b5d73546e823a25f84b6f4bdfe0f90be9814f7c7dc9e8ca412637dfae4f7",
"39/mlp.up_proj.weight/idx": "uint8(8704, 5120):bbbfb518fc81d7af13e7b11609a8ddc51ef06b4aa112db7eb486959c996bfeaf",
"39/mlp.up_proj.weight/lut": "float16(64, 2):80cddb131d2735d84d1ab7d73ecae467d0e40b20cc2979b047f2fe6272c783ec",
"39/mlp.up_proj.weight/scale": "float16(17408,):4f8cb8111819d54aeecc2216eda0e18baf980453aef2934328ad49ae30de3916",
"39/post_attention_layernorm.weight": "float32(5120,):03c39dfac3ffe95fe6c09729fe16ff93a9649acea340c9fb184a389918c387f0",
"39/self_attn.k_norm.weight": "float32(256,):ec341a2ec7fbf7144b63bf3063e21310526304981cd78d61ec094141794db684",
"39/self_attn.k_proj.weight/int8": "int8(1024, 5120):5c6ad94af43c3b28c3d8807fa326106099a49918aac404f365f3e6bef11ae2c0",
"39/self_attn.k_proj.weight/lr_a": "float16(1024, 64):45036f412c0d68793932b52c52163756b6895699589bae191deeb4fae6a2e871",
"39/self_attn.k_proj.weight/lr_b": "float16(64, 5120):4c8509d2f06c3eb99a260fe3625a7704524abc18754b721ab92ba17f998fb81d",
"39/self_attn.k_proj.weight/scale": "float16(1024,):5c4e2ea19519ad8dd4600d4f8caf954d6601edb56b5200e59a18a1ec8bd269e2",
"39/self_attn.o_proj.weight/idx": "uint8(5120, 6144):15332c7e6d3d61159b2605ce4d9718136606b8ce31d8576e06409544702322d9",
"39/self_attn.o_proj.weight/lr_a": "float16(5120, 64):582b2824e77d3b355362e9f971df3e76ba72be01f36384d0efadc7f00d0c994e",
"39/self_attn.o_proj.weight/lr_b": "float16(64, 6144):974fdbbc8af1585afcba1b5f273b3c67caf396254cfe459f4c47a9f1317c69cd",
"39/self_attn.o_proj.weight/lut": "float16(16, 1):b636275727271572345370bf07a2fa0e3ba3399b347155a70b69ab72ec1d4752",
"39/self_attn.o_proj.weight/scale": "float16(5120,):34946c48f8fe62eead33a074b16514002ed900c9ebd3a69f12a0041f7e80fa56",
"39/self_attn.q_norm.weight": "float32(256,):7ee3bd8df8fa1f1bba577dd21535626003cddf48217a5424673b6212529ebc80",
"39/self_attn.q_proj.weight/idx": "uint8(12288, 5120):18a7db8486a6c513ca0444f6778859a7a7aa6f46df29c07d253ef2162dedb12f",
"39/self_attn.q_proj.weight/lr_a": "float16(12288, 64):89c1cf8d40442862fd78c7a3c8814be10929f2349f2331666f3ae76b9757d39e",
"39/self_attn.q_proj.weight/lr_b": "float16(64, 5120):73dc5305eee33a97c0b3912e698ef5cf3692ce8c129651b16ce4c82161ac655d",
"39/self_attn.q_proj.weight/lut": "float16(16, 1):bd7ff39eece1b310001defc83c720a0c480ef36076ded828f7ba56c794f4962b",
"39/self_attn.q_proj.weight/scale": "float16(12288,):70831b1d6e1ec98e27cca86f30fd0ac85b6c5ddaedd837f18d55ea2a40f016a6",
"39/self_attn.v_proj.weight/int8": "int8(1024, 5120):791227963f3d4d77ece7783f63733ed4dc083bc12840f077829052f7f93c646c",
"39/self_attn.v_proj.weight/lr_a": "float16(1024, 64):b39c3d0711396e1ae28aa9283defa0ccf0ac655ca3cb31452c082829638928fe",
"39/self_attn.v_proj.weight/lr_b": "float16(64, 5120):135a505a7f50c96925fc2bd7782142453b29a92ef44cf35432d663377515f002",
"39/self_attn.v_proj.weight/scale": "float16(1024,):fc7d81e5b3d51c04b0ae6082f28518efe0b94a963b0c790eceb0dee03a054411",
"4/input_layernorm.weight": "float32(5120,):9af2418aa17a1fde3135bf00c0a4be94757011be4a20f9691ced37cc2a84a4b6",
"4/linear_attn.A_log": "float32(48,):8e1570e4b1c358bdc8815c8369bc389247f525e3cef79a58798f35d6442e9fe0",
"4/linear_attn.conv1d.weight": "float32(10240, 1, 4):3ec88f4c112ff501046db18275ddd2ceb804a5623cee854f9559a7035135b0c1",
"4/linear_attn.dt_bias": "float32(48,):83c7e8f4b52d6cbe52bf56ea40ead2a3a64b1497bb45b4c2e8784d6f2ca3f426",
"4/linear_attn.in_proj_a.weight": "float32(48, 5120):7a4d9069b33be70724bbf71803c4bd436c38e851ecabd32e6c1078a1eb2c91fb",
"4/linear_attn.in_proj_b.weight": "float32(48, 5120):3148596aa45c3ac2e9280413708688c3072811a25bd587f5be2a965d4874fd21",
"4/linear_attn.in_proj_qkv.weight/idx": "uint8(5120, 5120):8160aa9f4a71bfa77b08ec0050975422eaf3d705ce4e62709b41c0f2ddd60dfb",
"4/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):8cc76c1a75fdb0892af450f3c7997e681eaf24ec62644d5e1fbba1c551d3536c",
"4/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):4eefde67a5179bddd7494f919bec51db3fdf70c95307f30927da4adcc4426415",
"4/linear_attn.in_proj_qkv.weight/lut": "float16(16, 2):47b6ec9b698ee9e360137b26f290aaf8c681d17400ad3269f4c34979348d8abe",
"4/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):c771268023b5b9f8f32115c4bd1e88fc40b7094daac2ee20b5000592c9a3fc13",
"4/linear_attn.in_proj_z.weight/idx": "uint8(3072, 5120):badbcbd1d30f527a9e6dfaa22ae439b27f409de7039cc66305a1d26b60eaaea7",
"4/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):aae549a506050068263272874c98754fd34f65b2ae39a51363ed599ccd2ea237",
"4/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):2a5e0948709e294cd0d886c52b96ed0b156d3d207a5b460b930e1552382b2b6f",
"4/linear_attn.in_proj_z.weight/lut": "float16(16, 2):c087aa3f989f98a79a58e0b030d5b2c8def83f51272448c4d462dd3d5c081b6a",
"4/linear_attn.in_proj_z.weight/scale": "float16(6144,):988091f1371cc6592f02150c4b924a08875e39420bfcd9d6a2740a7b75a5feee",
"4/linear_attn.norm.weight": "float32(128,):733007a2e58c2a03fec95ca83432c0ab7d3a2ebf69f7768ca333c407fea874fb",
"4/linear_attn.out_proj.weight/idx": "uint8(2560, 6144):3f8790642196881acbd6ef6a5564ed253e91986bb4dd4f1ad8a5e0787cd03334",
"4/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):995d7a7e81630aaa9482d82c68beeb009040f11288a5af172fa9a1d5c91beef8",
"4/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):f8c8fd636c36c28af47c5f1476c0b12de733e276d57805679a50a11275fdac24",
"4/linear_attn.out_proj.weight/lut": "float16(16, 2):154bae5619d13f0b584a014e15e1d4d40d2c6e8a3ce0b2bfe25188d29468cdf2",
"4/linear_attn.out_proj.weight/scale": "float16(5120,):a7a28c97b1179907c86d756aa86d67795a9d7096df155e0fd97e5cae45abcdb8",
"4/mix.rotation": "int64(2,):be3993026832899de26427cf6bd29118858a859b67cc926ec8efb0563e8bf394",
"4/mlp.down_proj.weight/idx": "uint8(2560, 17408):90b81468fd51e302345439c69ea22aa25d859df0f0df6f9d07580062ea0ea2f2",
"4/mlp.down_proj.weight/lut": "float16(64, 2):f3ec1f3f331e36fdeb78e6a2dc3977a3d9accde0c62134d4a3d18e955fc0e1c8",
"4/mlp.down_proj.weight/scale": "float16(5120,):5032b13e5583b32467ea37ce426140aa29fa1b36fef40825dc052591a6cfec34",
"4/mlp.gate_proj.weight/idx": "uint8(8704, 5120):3dda92d613ee4411bc23d431f125053722f67acc2918302b239ecf9c35867e65",
"4/mlp.gate_proj.weight/lut": "float16(64, 2):a6b298db52a6ae3049d945796fd1373f9fb0fd9651c16713114d8a59a0d0ce87",
"4/mlp.gate_proj.weight/scale": "float16(17408,):f501b031a717152ed4b93a8bba863fb34467951e319bfff79457415f87d6124b",
"4/mlp.rotation": "int64(2,):581257451b2acd3bfc132647572ad90eb6301153e531718561e6a6860b329447",
"4/mlp.up_proj.weight/idx": "uint8(8704, 5120):a5a864f40f37e7ff0594ecafb7f3419effb9762e8ee6f8337d006bfc54cf1366",
"4/mlp.up_proj.weight/lut": "float16(64, 2):ca59ee2df3b9425e3420a288ab35a8be10452e5c26b23e00c0c88b404fe5cce6",
"4/mlp.up_proj.weight/scale": "float16(17408,):437ceb85251c29ef234785b4c03c9d773a6ddd8cced72ff9ad253f5c8d61f1f1",
"4/post_attention_layernorm.weight": "float32(5120,):f6d53494a16bacd141937833cabe79c6190373d888ac8e565262c52f832c9ed4",
"40/input_layernorm.weight": "float32(5120,):e4653db5fc3f63e5e6f48de55460290fe4a3ee6fcf8e498005e85cc0f865d118",
"40/linear_attn.A_log": "float32(48,):d171a2f74870ea52b4fd2a3137a17ae3d69d46fecefdeb1e05f0814477610082",
"40/linear_attn.conv1d.weight": "float32(10240, 1, 4):c9ce9479b24cc19c98d9136c0c0ce7896d89697db8c07c0bb0564408051e99d0",
"40/linear_attn.dt_bias": "float32(48,):1550be0d3aaddadf5c41588b02b3e479d96b7d634d6e9d433486def17864535e",
"40/linear_attn.in_proj_a.weight": "float32(48, 5120):5648d537ff0b2542a3ad72e2caa2bce9f31e743743464973cfeb5b69b210bcda",
"40/linear_attn.in_proj_b.weight": "float32(48, 5120):18b343ed09c4da596eca2e209182f3774e71f946d0d5ae594d2b782b3c2f3ac6",
"40/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):cf4fe29fa03217288ee4af977e87cf3b40d7ccea1e877ac93de00c3a3834cf89",
"40/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):b8332875d4b7a460559d93f17097dab89223480dfb22280a88171fed309f381c",
"40/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):a1c63691a1eefb7067f923f32e4dfe1b02c54db61058ef1d2e9d132d84546ac4",
"40/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):776b6af5ebd997c2bac4af9a1e6e7f0a4a6ab690f09998bbdd40dd715e09cc90",
"40/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):82a56eab779593408c9818e72f6f5cada55f9a574574b2895f008c9e34c8e6fe",
"40/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):c0ca0517d479629205aa6ed0da1478097efb12d002d842f002f288e322441466",
"40/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):d56b892f2dd29b819ecf18fe4db6051f093091534444beb9f5ad6ee43ff2b801",
"40/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):3c1fbd8a0e86baeb2c706455efbf07838bad47944aba8cb743383e975083d66a",
"40/linear_attn.in_proj_z.weight/lut": "float16(16, 1):0600fdc113468f01f4465e08f06933cb71df0c913246e54a1c81f95bd9160301",
"40/linear_attn.in_proj_z.weight/scale": "float16(6144,):a1ed69626fb891b3119f0b7c958163ff047ddf446ae01f4ac1d65a94a6c06d1c",
"40/linear_attn.norm.weight": "float32(128,):705e6f08f89110cd987ace3e3967332d725987e533e495542e5e5f5c02b88cbf",
"40/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):78984c2ee67b9ae2466c8ee1f00f0d0bab32c22f12f1e4fa27eeb437dd128cbb",
"40/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):827a921d1e993d9a637b99764366149b59b12cf056065262f856a435d7e5c8de",
"40/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):793551f66c083031bed78baa35e62ba4820ff8cdcf1b5ab594ed96528fdc0ecd",
"40/linear_attn.out_proj.weight/lut": "float16(16, 1):e0b68a5afc302d14de365f32cbc17023476f4ec6b80c1666ecc6027e6fd2c9a4",
"40/linear_attn.out_proj.weight/scale": "float16(5120,):0e2608adc1d85382d27d6850c4beba4adb4e1d9086315c53273c8d77985cbed5",
"40/mix.rotation": "int64(2,):32786a2939a41ca268fcd4304c413c5ed4f1ef5bc6b4f45419438620ae79e603",
"40/mlp.down_proj.weight/idx": "uint8(2560, 17408):15ea6f6050fb2f34373f0dc90100ecf2ba424cbe527173dceb7dbcf4039316f6",
"40/mlp.down_proj.weight/lut": "float16(64, 2):84d3ad5f12c380b8e07a7c68cfcc0d9490d6e7a587cb2102ab0eea1782feac04",
"40/mlp.down_proj.weight/scale": "float16(5120,):7b6b406ca67742457bf5180bdcf547f88360c50aa184e831e85f537b97d279cb",
"40/mlp.gate_proj.weight/idx": "uint8(8704, 5120):6949e404a0b522a240eab42ee5acc05709cb9cd1e1039b460b5672871d02e1e3",
"40/mlp.gate_proj.weight/lut": "float16(64, 2):d454851ff060b8fba5be2efb6a6dda8ff172312ab0b66e8d0ea0c02dd5949a81",
"40/mlp.gate_proj.weight/scale": "float16(17408,):fc6f7911003465496fbf8b15f1e2e1730d09cc87914c12e0e0a912a20f7b43b2",
"40/mlp.rotation": "int64(2,):95b5881d1b4e167b052c2eaa0b116f865bb8ff1aa2ae8b4b1d9ad8d1abdfc6b5",
"40/mlp.up_proj.weight/idx": "uint8(8704, 5120):23028bdba5d24b5382fd27ce58225d24b9e810078b4dbfb628896fb9d3a59c51",
"40/mlp.up_proj.weight/lut": "float16(64, 2):dad9a864b2c9699d6c435597d3fa386927a28f68f31a1dbd4585b96bbf4480fb",
"40/mlp.up_proj.weight/scale": "float16(17408,):4057cac5299687bce364e213e6ed01804c2b75560dcf4b241acd293248f6280a",
"40/post_attention_layernorm.weight": "float32(5120,):be375363bf94be016ad2a1425dbae302c8b3f9882636856e2ce4bc005e094a06",
"41/input_layernorm.weight": "float32(5120,):a4675f6fc8468ce45283380761ba9fbabd5a3b8aeb70319208f954ce9066e5c2",
"41/linear_attn.A_log": "float32(48,):8d32fabe373336f78ea91b2cde4a794ffebf1bb80f41ed263d46c59293a01dac",
"41/linear_attn.conv1d.weight": "float32(10240, 1, 4):5f55df32d37292bcb4883a9fa7e4089c4e14d89785eba90fb6c0aeda4f6b3a44",
"41/linear_attn.dt_bias": "float32(48,):876bc0e679c7f538a6718ef422857db5fe8f87f5ac1c6c3f5b2c7bc9a435a9c9",
"41/linear_attn.in_proj_a.weight": "float32(48, 5120):0fd9ddd38a91760a31169b8a80c73abb4232523b0000856f672065dc99e766de",
"41/linear_attn.in_proj_b.weight": "float32(48, 5120):43c75c4cbdd313ca5221b78098ef3d37b07c20a853bb253d155e945b16659f13",
"41/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):595a4c8cdb9dbedcaf9bd2608d3efe4bb3725c0d5b0353ada3949d820cee04c6",
"41/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):862026115fceef127f570ed5125b9d49cc1c87c840e47c0c0b7c81920b7a9a9b",
"41/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):f6f40bf808b927b3633fef3d9e43022191fab823e77cad89413bc8f7d23df982",
"41/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):9206115abd1283e666b250c8bffc7fe3f2a2a7327c48f5f0c92c74d9764b319c",
"41/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):af3071556890d403d208528f069dcd24591490d2e1974e4e9884ad65184a5950",
"41/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):c46faf1e39718425f68267f5408311ee5cb1d34da127e33ddc4ade8cf748e2f5",
"41/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):c376ba3a89651abc86def336c0d840dfc6652f5ca8c760e961f8a5899ef70a41",
"41/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):b7a891868af72c9ded25ed951d1a2545246ad6ca1239a3d0303d6f95c8949b78",
"41/linear_attn.in_proj_z.weight/lut": "float16(16, 1):28b15c4697afbd002e8381a223641dea3af6859a09f9cdce87a307f9d281d3d5",
"41/linear_attn.in_proj_z.weight/scale": "float16(6144,):202e5e3b1ccb79b218bb41cc7f80cb163459f0b04b2d626c841513d81f733dcc",
"41/linear_attn.norm.weight": "float32(128,):c7670a531f6af02c5eaff3ecb3c1282241f309d7f0a9d196c496ddd88a68a227",
"41/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):5bbad25e27e349d27453c5825d8a4c09139aa470fcf49891282237dca292a850",
"41/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):28958ad39cf78b215d720f6f424c5165da6ab7f1ef6565d30f8d3f212e1fef21",
"41/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):f83fc6faca8a88a394549fadc95b06f94f4cec0a0119d5c0d1d42e43e95a8459",
"41/linear_attn.out_proj.weight/lut": "float16(16, 1):b0206f1c1fd0642f62b9e23d46e02649d15e52b8055fbce228976785196a4e7e",
"41/linear_attn.out_proj.weight/scale": "float16(5120,):1d484cf4caf2d65e0c0c7bf03df9e3802e103f414f4c54b8bfa7ab8c3c529fc9",
"41/mix.rotation": "int64(2,):9305c8386aca65086c4d0cbbde5dabe0b4e23690642e9ab89c56a67b6c8d84fc",
"41/mlp.down_proj.weight/idx": "uint8(2560, 17408):3107750277f330048d11de6b45638a88f87d47c4d776922c08f6754de74c57db",
"41/mlp.down_proj.weight/lut": "float16(64, 2):f63260a2f4724c72e7e010b81d79ec73c94f0e31583cf9b47243555658edd0c5",
"41/mlp.down_proj.weight/scale": "float16(5120,):d5e7f28e308947cc4d6592b1621d28d0fb02c9987e462229fa69d178b18369ae",
"41/mlp.gate_proj.weight/idx": "uint8(8704, 5120):e2fbc5e747acfdb20a1523e012c7bc21efac17b6a94579a75766c3883f0bf448",
"41/mlp.gate_proj.weight/lut": "float16(64, 2):52c1745023b70d0fb38622fc0b9c8c2cae1338f2ecaf86f50a02125da5f995ad",
"41/mlp.gate_proj.weight/scale": "float16(17408,):bbc51489d85cadfee9b2a7fa68c3a1e82917a54d05064891de4308acbd2dfa3a",
"41/mlp.rotation": "int64(2,):0e54dd14bc8b1a9d79197cb595bb15917deec48ce0be4466e16fec81a0f9c2b9",
"41/mlp.up_proj.weight/idx": "uint8(8704, 5120):d00f73070bc6dc216e52fe1f8fc65e60407a11c76fcf876b061d5fd230f9fd6f",
"41/mlp.up_proj.weight/lut": "float16(64, 2):f90e8340297966aa3657582d0d06c984a95a9f621e5662db82c7c55bcbd327fe",
"41/mlp.up_proj.weight/scale": "float16(17408,):fdb4f9cc105884c20ab001e8e95decbf3d75a1e1fca1ac599f974c9525aa7280",
"41/post_attention_layernorm.weight": "float32(5120,):c6450ac2bf597aea5a3a87a1509dec81830bce31293bb357339849cebb2b37d8",
"42/input_layernorm.weight": "float32(5120,):a05f0f52f5e9ca8f3a312dd44df1406fffaf1395f97ed5962e61790e2cf782d6",
"42/linear_attn.A_log": "float32(48,):b56b299ecea47e0a5dd0c0cab34d8e31c665dcbc5b5279f170f005b6e1245fe2",
"42/linear_attn.conv1d.weight": "float32(10240, 1, 4):a949303a57f74a6481b3dc13366983541c8b73f0f342deb051334d95e25403d3",
"42/linear_attn.dt_bias": "float32(48,):94d2bef74a30693c997a8800dc4dd1d25ed85d2a71840b664a5b257c27d1ca2c",
"42/linear_attn.in_proj_a.weight": "float32(48, 5120):2cf4656bd259e8de946098c91b59e0545ea7da012642c75448d510cfc049753a",
"42/linear_attn.in_proj_b.weight": "float32(48, 5120):211688e28de5c6fa4856304d91d1e9468bb0ccad834fb1efb97ed947733a5c1e",
"42/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):07b8e293732e0535d5ed59e25c155199b69c15d80d577f7c406850dd764a24b9",
"42/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):d010df2bb0a37e5e93be397a22ae2e01fee7eec92ede6c312db961cc327f985b",
"42/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):d020395114df0c05d65f38e8d4bc596ef3e6e78320cf4e5775e1fe3506a3b798",
"42/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):b65ebdc4e2afd7ec560754de66a63f629465753500737833be7b72de7a961949",
"42/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):602117e0b1d2a17498a0297aaf269a7366fe82cb4bf38b7348dc96b47b96412a",
"42/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):42a25f19ad04f8927821a231f4c3579ac716f837a682b703926bc5e8bd2f2b90",
"42/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):b65f237e9489d4e1a2aab3f36c83e78d681ef4644f026352ac9495a1e9d329fb",
"42/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):b2336cf8aa1df44e64841b95cc8e8053e06ab434d200143c95adafe69f423b2f",
"42/linear_attn.in_proj_z.weight/lut": "float16(16, 1):610a2375cc9fc31801dca638b88945257a8877f66a3640cc7920f786e0fa16be",
"42/linear_attn.in_proj_z.weight/scale": "float16(6144,):ab1a7d25afd03c3dbac8a33c5caad866b525503ae1fb0da627bfaa58e8c22164",
"42/linear_attn.norm.weight": "float32(128,):ac41d0f1398f4822a629e234af1cb71c10b958312a8b67fd16248d1465f3f616",
"42/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):18446018e4c2f54b0d08c66c570ade0c1216b4fb15322fc2e5dab5e13edce4a2",
"42/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):b4186f1b26570f2dde9849f32285d5713909b13a94d3b6718d78821038291a03",
"42/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):9d8a451291071f870741e753f5da09be1fc0305e7a10abdd8b1a11fdc09640a3",
"42/linear_attn.out_proj.weight/lut": "float16(16, 1):8b6c19f8c616c21ec912856d96db3fb62030a60405c0fd7d9d7f2ccb92310ff2",
"42/linear_attn.out_proj.weight/scale": "float16(5120,):b3211fc31cd20254be13fe08ce68eb58142b4cb5a5cbf67e52b1d2f31f0b3f8e",
"42/mix.rotation": "int64(2,):581cb034525f98ed1df45256f18d1d15394c3493c48ffcab68e8cca4344de9ef",
"42/mlp.down_proj.weight/idx": "uint8(2560, 17408):d562ecd7f8966b29c1a58bf5591d5994220e3325623cc9c85c1f4769a4f66d1f",
"42/mlp.down_proj.weight/lut": "float16(64, 2):382ec00d5816856870441f959a277108e01297cc15ec307a6ddd059fabdcd421",
"42/mlp.down_proj.weight/scale": "float16(5120,):df81f5ccbd0670bd41480da96244e7cc6dc8fa9f696f1ad401e3570aac6e86a6",
"42/mlp.gate_proj.weight/idx": "uint8(8704, 5120):082b80c3124364a912c679f4560a5590d070230abfcbe0104da7b5c5407960ec",
"42/mlp.gate_proj.weight/lut": "float16(64, 2):e5f118560d959ed39ee5cab14b0f7aba4fe670076e93cd6353a4cc636d0bb858",
"42/mlp.gate_proj.weight/scale": "float16(17408,):f84692db068820d9008266fbce7bc50bf78542b785a3888aec172b6b86e36be3",
"42/mlp.rotation": "int64(2,):a6bc09033cebf8fa6617d1826c12111e65048284d707b40430a077aafec1db94",
"42/mlp.up_proj.weight/idx": "uint8(8704, 5120):ba19afadc2078c7adca101766ef7ccf1f2bceeb7090d1f739dbd3c57b2ad5cdc",
"42/mlp.up_proj.weight/lut": "float16(64, 2):fb4c1b33236afd20b6c915ea1ed3f56ba41080052e680539c0261be67b38cb32",
"42/mlp.up_proj.weight/scale": "float16(17408,):1356fa7f034d94f00345d7b729b66278ff12a0b58e26d6e7b9b1e5ec5836d1cf",
"42/post_attention_layernorm.weight": "float32(5120,):65ca7687e213ee18538efde92e56d7a9f2e67c0a3bd67cd6922f7f298bea15c9",
"43/input_layernorm.weight": "float32(5120,):db2ddb54629a54591ca3d519f85d0c45a99991c782be392129e094da140bdaca",
"43/mix.rotation": "int64(2,):791034b03e8d2537aafd8acd6b970821208e00a6d4cba4725773381f673cc980",
"43/mlp.down_proj.weight/idx": "uint8(2560, 17408):8ab2bc56f9f07c983a2f85d0f91b2d13315dc30ac93d719a273fb06e36a1924a",
"43/mlp.down_proj.weight/lut": "float16(64, 2):0254d9bc60e7d343fec29351a66b4bd8eeb5dd82bcfa51e5064f0f2994f9ea48",
"43/mlp.down_proj.weight/scale": "float16(5120,):ce04cd127ff2c20818f08b81a5afdfdef73d71985820eb9e0d58792b24f6c35b",
"43/mlp.gate_proj.weight/idx": "uint8(8704, 5120):70a27efd75178f029bfc51f76172c07cf7a2fb7ecbf057d7facb2be3b2f1b390",
"43/mlp.gate_proj.weight/lut": "float16(64, 2):41105939042d2c51e1fd400c0f72265dc61df3258d0a07da38a819ffcd1d8bbb",
"43/mlp.gate_proj.weight/scale": "float16(17408,):f18fc756879ddd552925fc00862e8d8b03193fd092b12dd6f10885c9f706a189",
"43/mlp.rotation": "int64(2,):68a0781be35338d095336cae01681a0aa1e05eacc9d6af3b6d81d27258a8f38b",
"43/mlp.up_proj.weight/idx": "uint8(8704, 5120):bbeb0e6adef77068b41585330c330d9aea24071141a1c67242a498defa7b739b",
"43/mlp.up_proj.weight/lut": "float16(64, 2):75d3ccd35e1e45de5b89e36030f82591aa62aa6b1ee30e05339010c4aec22d81",
"43/mlp.up_proj.weight/scale": "float16(17408,):6e224c85dbae3da79e0e1abd42236480886735a928c616a2f1ac0bef27d05fa9",
"43/post_attention_layernorm.weight": "float32(5120,):b35c74e2a15812f47621661acf8f477c17bd702315481caf503d19ee6ae7fc6a",
"43/self_attn.k_norm.weight": "float32(256,):f3c166df47a9a9f0080ba2c12fe523eb0062f09505ab867c1ccc68fb6c2d747d",
"43/self_attn.k_proj.weight/int8": "int8(1024, 5120):79a2aab2c786064167aaf894557b56cdbb612a4081dad872ed8eac44905dd3db",
"43/self_attn.k_proj.weight/lr_a": "float16(1024, 64):0a5d9be28eb6fe43354800ccb836210bc5d6e5cc144c187695e8b790114791e5",
"43/self_attn.k_proj.weight/lr_b": "float16(64, 5120):ca3ac7aefeca7219d03ff4d196631559e50ee016e0dabccaf3ab0b0cb89875d0",
"43/self_attn.k_proj.weight/scale": "float16(1024,):9bc8eb2a98749904c3b07603cff26dcb15672755286f183e88bb82264c4008bb",
"43/self_attn.o_proj.weight/idx": "uint8(5120, 6144):a634cb322b80e3a4afe1af53af6d5ae91d35678b59e690bae9da8932697f0abd",
"43/self_attn.o_proj.weight/lr_a": "float16(5120, 64):824460c1defe964e98b44937244dfcef1f6418505d4897afa6bc1aff97e5a1c8",
"43/self_attn.o_proj.weight/lr_b": "float16(64, 6144):9d7e0cdc5208b9ab56bc0563e3a2749a60bd163322175e8029c8c4c26c67db3e",
"43/self_attn.o_proj.weight/lut": "float16(16, 1):82fa680765b66586540473cdeee4e44def6e6891daa056461a5afc013d75972f",
"43/self_attn.o_proj.weight/scale": "float16(5120,):80c6361447976336f3324a8c570675f33224b230fcb612cd36f744bf7ea43dbc",
"43/self_attn.q_norm.weight": "float32(256,):cf89612d97ca3e8fc4cc46d1c64bc4490ecb1c8ca34672afa9eeb34349b33ddd",
"43/self_attn.q_proj.weight/idx": "uint8(12288, 5120):4310df6736f5ae5e7b85c75111e4646aea83bd30b7cdf3219172c4add79356da",
"43/self_attn.q_proj.weight/lr_a": "float16(12288, 64):2a0e283e951480eb228076ef452cbde33de4f56efdf054d4a2427c71c3bc0815",
"43/self_attn.q_proj.weight/lr_b": "float16(64, 5120):c29799d1d34df40bfc41c4cd24ff2701542cb2ee952174a54f3428039bad58bc",
"43/self_attn.q_proj.weight/lut": "float16(16, 1):3a76eef8ba14f288310fbb0d8a99bcd964f590749816a885871314d840a2948a",
"43/self_attn.q_proj.weight/scale": "float16(12288,):f1c1973b1c8568aa6c57348e2b88ab9ffe09673a47c53f9b6b2529f35069f1e6",
"43/self_attn.v_proj.weight/int8": "int8(1024, 5120):c8ff0c46eded97733c7ee772d37b57f6286f7521e5373c443e737a9d7dfe02ae",
"43/self_attn.v_proj.weight/lr_a": "float16(1024, 64):abdf543ecdb00dd56305c8d3965a76a9e35aa88d0da4ebe86f2515464e9a34f1",
"43/self_attn.v_proj.weight/lr_b": "float16(64, 5120):00059a17db7e9ef394af8d09fa67e9ddc26cdced971db96baa5d0628908b3d55",
"43/self_attn.v_proj.weight/scale": "float16(1024,):6e35235c676a232197093086df539103e002f271b9e89cfcd0a83c4aa5bf85fb",
"44/input_layernorm.weight": "float32(5120,):10dfea94eb0283d1a6257ac0a910730c5d5a011a8df697eb424d46d2ea243a8d",
"44/linear_attn.A_log": "float32(48,):fade71765b1a26ebf2e339cfb89f5d39448f63bb57d518b2f2f9e11c4b77f6fd",
"44/linear_attn.conv1d.weight": "float32(10240, 1, 4):3ee4ff7e6fbbcdebd36bcea7a28f1932a2de38af7e12f299a8067faef0e35460",
"44/linear_attn.dt_bias": "float32(48,):29e006a35a15b1572bbfd1d8b7a3ab24c32873ca26a3e7fb34884a4aefa21e03",
"44/linear_attn.in_proj_a.weight": "float32(48, 5120):d10430c1d90425f79cd3acda36618fbcbf65881d4c32532a287260c9506b9f42",
"44/linear_attn.in_proj_b.weight": "float32(48, 5120):a3ddb7e1618459588519eff5cb3a411e17fdf8e8a56cc8f60fb1e373ca4872b4",
"44/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):4cce7ce7fbe5796e8d62d79f3a053c66e23f771afa7976f70790850945805544",
"44/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):f3e51ed038ca3d718a3cee133d9ee2b5ea49b06cf60f14f08d52f5865422de8d",
"44/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):2db628837158ac63b95e5f31c632a4fdac537e4131b84c933c244cef5f27ed01",
"44/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):c4b45f32b64de24a049db8d2c1cecfdbf65a8b27cd6228d135661adbbad7f1a7",
"44/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):4b0b4088e468e9bc14b128fba5193d9f4a77068be0c32fcefcce5c07000375fd",
"44/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):68328a138e83a2d120cbbdefd0bb53230aafe62fea5a200356608a6f6aeee397",
"44/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):f73463d1b5a03dbaabf9029f20aa56498fd861a3bbb8b79019a2d4b036c1ec41",
"44/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):f06fe9a456fafa9ea04b021ee2b2836f877201f9037cb7bf262db2d1a8f6cfac",
"44/linear_attn.in_proj_z.weight/lut": "float16(16, 1):193a84b6cb98a7bcdc9808bef77dc1615e50a3b2f6f7de8b341a88fc8e8d8d2a",
"44/linear_attn.in_proj_z.weight/scale": "float16(6144,):2afc8f6cf17be744bfdb75b976cde100261058c7711f89144ed9229656fdc906",
"44/linear_attn.norm.weight": "float32(128,):eb05891c80d44363c226d01c054deb504e23c51b302071287e3986afaa4422ec",
"44/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):9b56d7ed2bcf2bbe2cdec1f2da8a67be496e1fce5980375e129378f2567d7d92",
"44/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):395e7a1fdf94cbea2904219825f088dc13ca9c1b2f708577df7e2f21484bcb19",
"44/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):ce6daae9fb974cd413a91a6887951118c54c653e6263c1f1777573311d10b50f",
"44/linear_attn.out_proj.weight/lut": "float16(16, 1):216ce149ab37d21a40fd8cc5ce00ba9c5e0dbaf57966a6c852d349825d5de77b",
"44/linear_attn.out_proj.weight/scale": "float16(5120,):ad50270d6ab464d4dcd6c8f10f69ee0b1e9b5506f7d48f87a1eca5ddfa54c0a3",
"44/mix.rotation": "int64(2,):ba32747805f833177e5b5914fab2277637d7f716a0f24d538ee1ba1791fcdad4",
"44/mlp.down_proj.weight/idx": "uint8(2560, 17408):ab1b525364371a97b4290775e7dc948a23d28264a1f6becb0e8c3ba7cd1be399",
"44/mlp.down_proj.weight/lut": "float16(64, 2):7acc16c668c91bb9a6a22be0e95fac34bdf0820c5e4b5ce29a878a70610b9aab",
"44/mlp.down_proj.weight/scale": "float16(5120,):7db30301c4fdd1ac2a36ed33b96fbf6849ebb777bab7d73874e6cb06394e61ae",
"44/mlp.gate_proj.weight/idx": "uint8(8704, 5120):24ecb5142804558f0af2c586f3e22bab23b07295450eea822d753f4e2af7117a",
"44/mlp.gate_proj.weight/lut": "float16(64, 2):517aee5fa69af2cb00a1a5cbd03e8b5e08f546dfa4d54c95e66b8f20c625d2b6",
"44/mlp.gate_proj.weight/scale": "float16(17408,):c90d6fec449df7a2902d925fbe6634ee6c643ef83d4993ab644d4d06edf2b6c2",
"44/mlp.rotation": "int64(2,):df64ff10bbaa78adb4d31ff13d793c35241dc02968a2656fe5362fb9f90c08d6",
"44/mlp.up_proj.weight/idx": "uint8(8704, 5120):4482bdea9237b2c2262292a54616551e20c09f07315d50e5fe457cf377e15034",
"44/mlp.up_proj.weight/lut": "float16(64, 2):2e1b210780c7de6c39103b04d0002846739a0b989c17291ff4362a7f0ef0745d",
"44/mlp.up_proj.weight/scale": "float16(17408,):1ca7856a9bb86bcb0a3bb3a71967ac3c1f0e300f115d98fb8a7b1444ad50a910",
"44/post_attention_layernorm.weight": "float32(5120,):9220401d46318e09c798921cb413ead66d3a9b8796d90b03b7d06a43f25239bd",
"45/input_layernorm.weight": "float32(5120,):884a63f48fbff27083a326f744c9823fb8fe6d1db6eba0a9ad704a9c2941caef",
"45/linear_attn.A_log": "float32(48,):68e2316dc74427780a0ef1fafb4847e0c12cbe2a911bb7dd93b617789be1f35f",
"45/linear_attn.conv1d.weight": "float32(10240, 1, 4):cef3c2c5b72f576bed61310e15940a5473ad3690b1793cd7b3c7b2fcd04f93e5",
"45/linear_attn.dt_bias": "float32(48,):e273fd6c649961763ec63278fb8a728fe151874dddbd953d2de2221338c10a60",
"45/linear_attn.in_proj_a.weight": "float32(48, 5120):6f0400690dafc69de77ed19ba0f2d35db356eb5c7e521688fbc9d32d9abee129",
"45/linear_attn.in_proj_b.weight": "float32(48, 5120):17c3d5a3fc59781c50d02b47f0ba103264710d0a44038516fbcb276bc67dc983",
"45/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):cb58a2b64bffac866703a3f0c73f59af8d3bc93d25473726ac06c2f2db22b30b",
"45/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):f70375051cc98d56d16da5f445ecfec0f6658f575ceb1c6e74023012b53905f1",
"45/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):57dde02d9095ab27b5926e7b662686ff72d6b01d00a5aad42290315f2aa01fe9",
"45/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):5cd0caab2d64d6194f219c331804dbb53d0e5dd66a0dda5746d3cec4e8e43aae",
"45/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):3a6f295ae3e752f94006d706af69fc49929ee314d3f1d73af68f7316cb935e2a",
"45/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):322c2687cb77fdf7981ee130c31c4bb68b29fc72f73a6391c802b9068362bfe4",
"45/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):0793f975182ab99be9615b011103957bc01af7ea37ab99eae147ac6b44304d4f",
"45/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):caf3ee2fa264a4f89fc3f0945ce5958ff03207687fdd1d36dcfffc507763e6c5",
"45/linear_attn.in_proj_z.weight/lut": "float16(16, 1):2119cb9050a4660cca15a8e27dfc13a78453f5a6be757f93f3cd25943e13405e",
"45/linear_attn.in_proj_z.weight/scale": "float16(6144,):b0a4af6aaa2610760cbb66adab932b5e4df2555bbf758f8d592da5c7bb45d1d1",
"45/linear_attn.norm.weight": "float32(128,):6d95323776bba1314f04e5308130f8174a3904156c1bbb52419e830d858170dd",
"45/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):6a4a39dad43180bbebc7ace8639b65921379ad7b01181a8974d0e50c96a559fd",
"45/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):ddb24aa97d271c43a53c379e85ae0c15f3549345c09e511489a32bc99d86d7c0",
"45/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):5ab68a7d00076029b733b3621fac2170b009cb9fabf4f43675fc0008fd32f5a5",
"45/linear_attn.out_proj.weight/lut": "float16(16, 1):c00b7cb090018b946cf1fcacb6efcff1851078eebc7570f4bd302a33462da933",
"45/linear_attn.out_proj.weight/scale": "float16(5120,):519c957e83c4ed9cde6ebcb3a268260ddd730f664505c76fc878fd164c612734",
"45/mix.rotation": "int64(2,):7a0f85f651e3aab187bc5a4a1a1f3a9b399bc92c10eb03cfcca4dfe7652d7b13",
"45/mlp.down_proj.weight/idx": "uint8(2560, 17408):9c2ceaaf465f68e22126f2eb98d52c852afd6229d5ff7f25d343d2c41ce5f7de",
"45/mlp.down_proj.weight/lut": "float16(64, 2):5f3104fdb8eec4041d3d0e6834e0b8e366d3a31940a451c163d54c325f0d05d8",
"45/mlp.down_proj.weight/scale": "float16(5120,):be499cdd945f82798106454749e22ef1165b6ef3cc9de17687a72fc6dc2c6c12",
"45/mlp.gate_proj.weight/idx": "uint8(8704, 5120):c7bd075ccb66c73b43e2332198db4cfcfc67c3d66f5e0a5865dd028fa4a095a4",
"45/mlp.gate_proj.weight/lut": "float16(64, 2):956d7cd16a5a2212f8f325c509fde9e8ad8ae31f9db3b07dccf465a54807d8a3",
"45/mlp.gate_proj.weight/scale": "float16(17408,):3fc894a574439310ff688c90b9f62cf84b60be59fe4ddf809bb571f9079e9824",
"45/mlp.rotation": "int64(2,):149c5fcec1a11fe4646515b9d89291df305955d869667da8e5e66bfefd8050b0",
"45/mlp.up_proj.weight/idx": "uint8(8704, 5120):be6abda54688398378fb88cd04dad4c14f79c878fda031d65599f70f8c27b0bc",
"45/mlp.up_proj.weight/lut": "float16(64, 2):40da8b7476d5b31d933cef5c1e7e2680f526707400916dd6316781e2308ba5f7",
"45/mlp.up_proj.weight/scale": "float16(17408,):05078c0218686b308ff1171d0cab17c89dda44b632b699c9db7b9925ec52718d",
"45/post_attention_layernorm.weight": "float32(5120,):02390ae3476a92768ac1b296d6249d32ddd64c1a3a66f43df6cf5e1f0cd9ee33",
"46/input_layernorm.weight": "float32(5120,):5b459d933d24390a39d1b2a4bbbbb153ea09077ab8ab6d8cd7452d9a0ab46a3e",
"46/linear_attn.A_log": "float32(48,):86f3fc5efc00fab7af026e95a94c7d21fa301eb85bd5c031eea62141b3034c50",
"46/linear_attn.conv1d.weight": "float32(10240, 1, 4):fe71c8123182e0cb79ffd52b2fcec874199336cdfb6ef9f6c24e9ee1e404306f",
"46/linear_attn.dt_bias": "float32(48,):2a71f1f0e43fc8ac6a0873cb648769fb03918692117160935da3001b36d626ec",
"46/linear_attn.in_proj_a.weight": "float32(48, 5120):c56ec4dc22e168e9b9d92c091ee53509b24c20548b77debf4c6b93457dd55a3d",
"46/linear_attn.in_proj_b.weight": "float32(48, 5120):175b73efca2f078b0753dc88012ce04641a907aef04070b5004e741b2359c6c6",
"46/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):283600bca603d330bd48caa0191afecdc48c2c18e66d56cedff83add2e331834",
"46/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):264f2b8ed8aa8e71e8d670ca146683f7f691a49e51aec0e835da535c34245be8",
"46/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):27aeb289b18540afd7cf1be82baa89c24179f9365ac29a5e7cdb001e3b2ce85b",
"46/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):3d1e5e764219d8a47b91c18a3ca1de394ccf65a77d0ba999e4afcaf56065f053",
"46/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):3c71c97e8dc0b6d093ac2d38d6519d4847ff0ae50c753568e809728115594657",
"46/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):d5eab421a25adc4a31846b6fa26316800f4d2148d63db76d357f1a70ba2e38af",
"46/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):0c87a690ca379137af09fa95273f02ff7415f7c97f6bca22ba4e988b657e86bf",
"46/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):aa832f35e86d58fd74b2fdc8bb89c31485a3b594541d32fe54ae8fd90bd5bd60",
"46/linear_attn.in_proj_z.weight/lut": "float16(16, 1):6ad1c459da90e602c24df42158b77098e3e873a7908773f3b8c828e7cb0d79e5",
"46/linear_attn.in_proj_z.weight/scale": "float16(6144,):81dc5dd5adb6814b03bd3a64b633b39e53709aae2f04069c873acc95c90d2dc9",
"46/linear_attn.norm.weight": "float32(128,):c1b5c5f4fe45f833d3f5127488d234308ade18bffa5404de352028c8290b29c6",
"46/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):19bcd7460fa32fbbf5ec20a4ca1a934bfc687772d449b66eea1816f27bcb67fa",
"46/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):0684549c4ba694af0727833ebc788709696e500991fc8dc99fc0c8258ba2b897",
"46/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):2bcd9ef2496ea233ba0c562a516dd9f9e571131d33c454b018da3df7ad883f09",
"46/linear_attn.out_proj.weight/lut": "float16(16, 1):882bea1fde2d8ead4cb83ed67b2d3f61466470338a5325ce435a8bb97a5ee567",
"46/linear_attn.out_proj.weight/scale": "float16(5120,):b409ab0620dcf88620231dd99599fa887a79829f03dac8ddb652bcdf071d3c5c",
"46/mix.rotation": "int64(2,):67c33b29ccb8e5c631a9024e7d77dc6bc1d82c93db9489465e95af63e33bdd73",
"46/mlp.down_proj.weight/idx": "uint8(2560, 17408):7385f4ec00527fd32c00f8f776185fa5b8d786bba4cc6a74421a24be29e5417e",
"46/mlp.down_proj.weight/lut": "float16(64, 2):a1e5109147d0f9eff447439c168b9249271d0a35c528fe9e17f3212da8cfde4f",
"46/mlp.down_proj.weight/scale": "float16(5120,):86b5332b0635d8dd3f9bd9373fcb1b9a3ec3c748a02a9072cfb1a9536e532ee8",
"46/mlp.gate_proj.weight/idx": "uint8(8704, 5120):35460aa77eb957c2a43186e4d174c5a773c6685a27beca2cc6995c367c7643eb",
"46/mlp.gate_proj.weight/lut": "float16(64, 2):31b10f2415988716abcf6775311e8588e17ecd2ee3f07720d454d5324f51c713",
"46/mlp.gate_proj.weight/scale": "float16(17408,):1fd41d23fdd48b98973cd64fe531a6a3e78b9a4f866cc5bfa4d09be5518be587",
"46/mlp.rotation": "int64(2,):4d3812b7347c1bfa364147f79154a625ce1c41227d7504d04eb676427ee987de",
"46/mlp.up_proj.weight/idx": "uint8(8704, 5120):a8fea58eb396d1e751af2156b650decd212cec34edcbef3fa04e35913d9527eb",
"46/mlp.up_proj.weight/lut": "float16(64, 2):fa4e8c3695cf101b1d720beff3d1afec5187413fee7e7c11d8895e7f1561d1b9",
"46/mlp.up_proj.weight/scale": "float16(17408,):51c8e51b3c6ba748db9566b92e88d229a21647407d5874bbe8a488883c01cf9b",
"46/post_attention_layernorm.weight": "float32(5120,):d7207f5e3170c145c0932f4cc2ce451060b7bea7fb6b4d98dd1177f6546c905a",
"47/input_layernorm.weight": "float32(5120,):283fcd1efebe1c0d73e274ddd637f4e04f233a8d1500951e8b9fd23f26dc6423",
"47/mix.rotation": "int64(2,):e12dfc6add1577fe2944850797f5e321ba6b58d6f389e4d8116b6831632352c8",
"47/mlp.down_proj.weight/idx": "uint8(2560, 17408):ecf5205949cd7123bb963c0c4cdef298f4db5a51867d8646a45466ee3b18fbee",
"47/mlp.down_proj.weight/lut": "float16(64, 2):729d7c89848e6ce5e75993ff507bf0be375df9a6d30d87d619bc4c6df64a0d56",
"47/mlp.down_proj.weight/scale": "float16(5120,):e4f3a556c543b16eb0b174065606556c33e2d3da90cb28a5304d63793df15f33",
"47/mlp.gate_proj.weight/idx": "uint8(8704, 5120):0ea6232bd00c14e991d8560687e15ca0b3ce5c21125b91734ee8a5aa32ffa95a",
"47/mlp.gate_proj.weight/lut": "float16(64, 2):ca86b84149c3b7e0106e324e644768e9f336fee377c92d1267745d5c4fee72b4",
"47/mlp.gate_proj.weight/scale": "float16(17408,):0f038b774a6b828aacb632cc733c1b1d820662d6a62865a47ff58fe70620103c",
"47/mlp.rotation": "int64(2,):11611950d04032b87113e0573b4e4443ffbb952d2e3e85127fc25e69275b1085",
"47/mlp.up_proj.weight/idx": "uint8(8704, 5120):d8edc3a69c670eee4c81ff297cf7cb3a0f7be33253f44f642b03cf15d470c965",
"47/mlp.up_proj.weight/lut": "float16(64, 2):42780ac692b6e00d0a83bd1fab4b9f4f5b33bb181866caef8f0e71dd963fe5e3",
"47/mlp.up_proj.weight/scale": "float16(17408,):b88e5be381684d26fbb1236ff2f3c329b81b671d455cef4f0c1e5659323cca57",
"47/post_attention_layernorm.weight": "float32(5120,):64c94437ef456b5b4b06c16fa179d1d2e26182390df38253e4ffa0c1c528a8db",
"47/self_attn.k_norm.weight": "float32(256,):adca19bedb3c4ea3da58c75878a525de59cb7310c1f9ef0633098505e031ad4e",
"47/self_attn.k_proj.weight/int8": "int8(1024, 5120):e8cb95fa8a7136e8b17857d5d033ff5a5004e0fa5f06c3002af3e848a19f96b6",
"47/self_attn.k_proj.weight/lr_a": "float16(1024, 64):4aadd49e280b2a8eef9b48918051b88c4590e7787f330d19b49bbf35247a8563",
"47/self_attn.k_proj.weight/lr_b": "float16(64, 5120):10445ca3f064ab06dd94f2db30436886a349e27929c2b9e1518a78867db60696",
"47/self_attn.k_proj.weight/scale": "float16(1024,):8f584a235060400876dbcc5c07528d8ddf3cfdf8b45a506cf028e171835d618c",
"47/self_attn.o_proj.weight/idx": "uint8(5120, 6144):8047454c145f923f2a0df89ca5a00050ecbc3919081992c37821392981d902a2",
"47/self_attn.o_proj.weight/lr_a": "float16(5120, 64):a7ad4dce91c86835a552c2fad4c2a9e26315fa4fde2c094f20b348b8a59a4b52",
"47/self_attn.o_proj.weight/lr_b": "float16(64, 6144):947582dba6e264fade24b8adce84b8ae75c8613b262937e9b8885a016160c044",
"47/self_attn.o_proj.weight/lut": "float16(16, 1):deca59c417ff09f79d127e53ba6c146820c14b53e747bd2fe14c2e2aff54d7c4",
"47/self_attn.o_proj.weight/scale": "float16(5120,):39b40a0e8fc370b7f326467cf5e5c7fa6f0bad5e6aaf13eab92c875d75e03eef",
"47/self_attn.q_norm.weight": "float32(256,):adfd7b8ddc3271c9333b958745c97048d11acf6ab0d66325830aa0ba8855e4ae",
"47/self_attn.q_proj.weight/idx": "uint8(12288, 5120):104c9aab7eebdde3599c8e12b57857f59412aae601d58a5709c06a47fe0ebec6",
"47/self_attn.q_proj.weight/lr_a": "float16(12288, 64):0c32d5f7df559bc7a6cb0a736561adfbfd0f916d70513ae8fa0952dc20fe94d2",
"47/self_attn.q_proj.weight/lr_b": "float16(64, 5120):2daef9d1364953fc661e25051ef215d37df1febd7de6a85bb34720ac272c649a",
"47/self_attn.q_proj.weight/lut": "float16(16, 1):91bac828c6586ae38118461592497a96eadc2a3adaed4c45c58d884c860f30fb",
"47/self_attn.q_proj.weight/scale": "float16(12288,):0f498692889b97116efdd977a89a0aa5c7a191d37c3f62dc35c29ae4604da18f",
"47/self_attn.v_proj.weight/int8": "int8(1024, 5120):d2c678f641eac7aaa2afa8a788c42d9e008754be8a9424113cb8fde46bb29216",
"47/self_attn.v_proj.weight/lr_a": "float16(1024, 64):273346e5a9b382a0d0322a0e6b9b35660d7c2bfbbe9911e6d124ae4fbb3f00bc",
"47/self_attn.v_proj.weight/lr_b": "float16(64, 5120):992a6ddaf3de12375f45764f2a16b4e5bd46802477d7a7d4aae5ac203ca55eb4",
"47/self_attn.v_proj.weight/scale": "float16(1024,):7a5cf99d0be85de9563f802dc600804dcff6b6ce68a6bedf7034db851184ce4f",
"48/input_layernorm.weight": "float32(5120,):5d52e28aa05f219b9de5b6c599ceb394d16df68a795466423c6ed7e0ec105e06",
"48/linear_attn.A_log": "float32(48,):a947a43c6120db6e572688df92ed498c5cd3c1392bc4dc28781db3fcdf0b7273",
"48/linear_attn.conv1d.weight": "float32(10240, 1, 4):66dfeb40b76dfc3940c2e35d92025330bfdb9e59b3d9377c26587d318c0e9334",
"48/linear_attn.dt_bias": "float32(48,):d3f72acfde9921395cede9f974da816298ba526e8b35520b46ca16d9d9bc9da3",
"48/linear_attn.in_proj_a.weight": "float32(48, 5120):3342c637f27f53e3cff88c4ddb824e0180407cdc045efb9ae1dfca787eb1cada",
"48/linear_attn.in_proj_b.weight": "float32(48, 5120):8f9fbd1f95d56177b7d0fa7216016c3593006d2844901faac387084e4a8af95f",
"48/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):5fd955d8e4ac51122d8226f91313f862f3ede3db3d0b5be55b10446d87c93756",
"48/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):a43c4fb205812f21382b23209828c92c9c88ef18b6f8b77d9f848d83b114cfb2",
"48/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):d79cc3e215046c13bc6df8f44561ddefd532a7faceeaac4d6e1bf3d99f23fdf6",
"48/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):d51fc0608fbc0478bc9cd8b99352b0f9f7dce0cc7d5324b24af04836265a2619",
"48/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):388644b3a7b233448a02b56715dfc7c5ada7086f21a43c31ff28cf1720a70ed3",
"48/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):86dd2c5f86eb599fd1779832b5d23bbc5f5c2fc7edab73fba08db38c63fdce2d",
"48/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):b1513ba4b9c3956d49c2470fe77005a24ee3d3fea5ae7e4a9cc7c23506e0ba73",
"48/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):d5c662e44dc18b0946ba9ae75a2374aacc4e4ca7dbaf1d681e981d8c3a8c5513",
"48/linear_attn.in_proj_z.weight/lut": "float16(16, 1):d30cd689459867a48a60f44a0bbd09a674e74091d7872bd8f5b51539649acfb1",
"48/linear_attn.in_proj_z.weight/scale": "float16(6144,):7dfa9108d0b9f2a6bbf9c2a624cf560f7867f03c609575284a90149814ffbee7",
"48/linear_attn.norm.weight": "float32(128,):f10e123465f485674e2090bec319e155ea2409892dd2bbdbd0df79a5317460e3",
"48/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):381fc06dfc63b58d36a0a83d01ce970de8da362cc04c0b07598d9e84d7c49060",
"48/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):950c3b391a5a5b0f53d077a37ed93cfc9c3a087118276be4365b8c941e3ba93b",
"48/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):79d8a0a3f74f2ec5e36a24549b5e75d2a9f2cd3a070131d5b13da51f5aff97a6",
"48/linear_attn.out_proj.weight/lut": "float16(16, 1):f6fd1f5e307877bfeef03e6b14d91a2627f75d2a54a1ec1fad2e3fec5ad0e76a",
"48/linear_attn.out_proj.weight/scale": "float16(5120,):7b64711b51349f0cd3768c072740ab4eb8988f03d0b7ff3c7987c3d9db0f127d",
"48/mix.rotation": "int64(2,):e5ab0a5927e680301d51ec4f1b274fa9d508384dc2a1f3ae8af34b89247ee2ae",
"48/mlp.down_proj.weight/idx": "uint8(2560, 17408):95fa61f3357dfaac030d36d8c79de0c396739b71c40d0cdfe0a1eb7c3ddae199",
"48/mlp.down_proj.weight/lut": "float16(64, 2):3db642c8ee4430b8f117f35293c383fe55856b3123ab8b6bc0a980dc1f5eac69",
"48/mlp.down_proj.weight/scale": "float16(5120,):377a19c1491c0d62a8d3bde03395a06b9a07a214e4062e206bc83759bbbe66fd",
"48/mlp.gate_proj.weight/idx": "uint8(8704, 5120):4d4bfff6f31507639d1211d219661b08b028bb69768f7c4412139ae477e26ac8",
"48/mlp.gate_proj.weight/lut": "float16(64, 2):588be4cdb0e4384b072c0a2e9d6ed43b41ab2f9d220b2017599148fb9edf6869",
"48/mlp.gate_proj.weight/scale": "float16(17408,):5af9047a9f78b9283aea5534edd66acc92c0f03fa783eda58638c22036f1ed24",
"48/mlp.rotation": "int64(2,):db447a7c44da487ac93755f0084bfa9c01fbcd4e29eed8f2bda6cbdd4f7cea95",
"48/mlp.up_proj.weight/idx": "uint8(8704, 5120):013e159aeebac29521ae899af31078b5ca88de41cd2308e994eed9df2c4bd688",
"48/mlp.up_proj.weight/lut": "float16(64, 2):3abc13c6600a27ee1d487fef7cd53e4c89d10bf758765a181602d0a1c00a4dd5",
"48/mlp.up_proj.weight/scale": "float16(17408,):9c9ecab1ea17926f512605a12ce1b560d64bd5ca2f7e453eb1c66b677d612b70",
"48/post_attention_layernorm.weight": "float32(5120,):ab25e8f8442e8b81b5a84641aa7bbde9c0c6acfac682793d7840679f52a2e09f",
"49/input_layernorm.weight": "float32(5120,):765d1951b06a5389352b7231bf4818b860000ce4e918b949a562c489ece6ca95",
"49/linear_attn.A_log": "float32(48,):774fe716b4a319a81217698292fcfe8b93958d649be61355b091384fe7b9ab8a",
"49/linear_attn.conv1d.weight": "float32(10240, 1, 4):7dbff126a4ea9f04d72785e25332cadd607f1b25695b7aa5a6f65436e228a133",
"49/linear_attn.dt_bias": "float32(48,):1273e2474f0ef6f688a308248780b43f59bc24163f36d4fa5f7e31de79256095",
"49/linear_attn.in_proj_a.weight": "float32(48, 5120):dbcbee0d1227d1441f41d05d01064b6756187b5b9694ce6ff05a170cdcb16aa0",
"49/linear_attn.in_proj_b.weight": "float32(48, 5120):2faa6bd689d8bb67a76e057c35996b70d407eeca94eb59a93f335c3f89747ea7",
"49/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):36c2e9d79f20955822924b11a77b5fbec11e0d32a8ef2bbce6f8999b619bb846",
"49/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):a67f4b11dc644e5044698c070a058d1ce49b7539166280de81fba5c7dd0a7548",
"49/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):3dc282f0ce55742ebe3e4e45266739a69b38dee6f176e5376b5636ff16a373ce",
"49/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):bcfc373d0faba5043d2629726fe918638db772c315b753591f8e9edfd37bec55",
"49/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):7d7f9a1ca04fe53198a9a38e197743fd4b04afbea7b9429a642654aceb2ce360",
"49/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):a99ddfad218803756a59ef9093a0a806e381d0d811cca6a03e8966e3bcb6e614",
"49/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):321ac22b8ecb0c1ba64a889ed8b7c61febebb391cf1facdbda02b3ba0ee2e513",
"49/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):1de6e241133bf4cde3db4b0d266647e5a2fce3253fd94953ead53ad5d40fefd2",
"49/linear_attn.in_proj_z.weight/lut": "float16(16, 1):0f7281543af495397903ff4c4050e04a5c0e5013c8e30a449c0765e655d525bf",
"49/linear_attn.in_proj_z.weight/scale": "float16(6144,):133ef4a7cc9c026c39210e6214fd20325081f53960e04dce4c72248caef12ba5",
"49/linear_attn.norm.weight": "float32(128,):34d6ed78d9c37bf3d7ab94b4c2edd728fab51b80502703f86e4d77aaddd043cc",
"49/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):5c059651c7e9a241327192db4b27d79d62fdd57e06a699c8185c8a5c76513e73",
"49/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):326fe3ae06a71d4b2e44552f83f03d7b6889ebe2caec86d61cfd96c17d69416a",
"49/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):608683175a3229636e14c0ed2dd5593b5b8d6aa1b2b53cd7d266be5beac86935",
"49/linear_attn.out_proj.weight/lut": "float16(16, 1):51acc79c30ac6dd374005ad31e44505dab857ef3a44ee179978bc67562e6342f",
"49/linear_attn.out_proj.weight/scale": "float16(5120,):d538de89a83de46ae4b7ce9d0236d39bfd1622b1ad0f667f8b101a71d36fee9b",
"49/mix.rotation": "int64(2,):18a79ea84489b44d4397ea4b2e9f2d22f97ff36f4f90477924069be5373b3ff6",
"49/mlp.down_proj.weight/idx": "uint8(2560, 17408):c88e870debff82ed9935db8eac479f136e3de5abbb603d0470258f880264ea8c",
"49/mlp.down_proj.weight/lut": "float16(64, 2):2f1f291002562f39a96cb42d72b789606b529e9c865d7bd6b4b89221d59a2839",
"49/mlp.down_proj.weight/scale": "float16(5120,):4f513c46a01010f04f7f020802c897502089a059ee5ec82624e56edc594f01ef",
"49/mlp.gate_proj.weight/idx": "uint8(8704, 5120):229b329f058dbec7ba2b6810e475f0d6835ba6bb81a86c37466b754eb861903d",
"49/mlp.gate_proj.weight/lut": "float16(64, 2):ff6e5f1f1038658cd32e65ab10343d7f308186660d9f97fb9c87e1cfe2c14939",
"49/mlp.gate_proj.weight/scale": "float16(17408,):25e1d2e796e7c4691a0d90b64cddaf571897e65cd9c14309405e384515493bbd",
"49/mlp.rotation": "int64(2,):e6200628855e59250a21f8f9a88ac6986516bc83daad25009b29a00934e2c77b",
"49/mlp.up_proj.weight/idx": "uint8(8704, 5120):5e03eb68c7779ddf6325dd56621ed8c6dc14cd811d7f070d9542081ad039eda5",
"49/mlp.up_proj.weight/lut": "float16(64, 2):d8e02b7a2ee26275cde0b03f8e5d9dd78c9f5c7c373913bbf1b81dc6c27c1ce8",
"49/mlp.up_proj.weight/scale": "float16(17408,):7791e2a6deb4e08776eac1446e82af2a4907f1986e06be26d9497c3deec6b67e",
"49/post_attention_layernorm.weight": "float32(5120,):fc7dcef5c4d6a895da6d1ce9e361dbed1df76111acb99897ee8448c0391fba8d",
"5/input_layernorm.weight": "float32(5120,):a8432466ce906ab107b6a2e8a29fcabed5904c1f55b6466d90f1892a4bf6dbb1",
"5/linear_attn.A_log": "float32(48,):414d035af6ba70769d717ae705e7736efce87ba59a0f42b48c6577c1674b19f6",
"5/linear_attn.conv1d.weight": "float32(10240, 1, 4):0be9d7861df2edf7742980fa29f991aa4076690cd3718d08fe811537a36b0b7d",
"5/linear_attn.dt_bias": "float32(48,):22186c77585f050a71757669cfe02f08d17e518e1e92ff29d8674e6ed68a14f1",
"5/linear_attn.in_proj_a.weight": "float32(48, 5120):696dd92ae43777e99f7f4aaa4f5645bd9bf069e80f3fbfaf7f958cbbefecd902",
"5/linear_attn.in_proj_b.weight": "float32(48, 5120):d8eee079085fe6e1f99f7cd9a3e95893c8d87da12099272b4d94d76d51c28511",
"5/linear_attn.in_proj_qkv.weight/idx": "uint8(5120, 5120):ff3fee10d9c77b9bf2162dcd6bbb3a5c840b9106a90f96f016532b67f9b526ad",
"5/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):a33898df8ce9131eb332ac4b1b44835e20bff074ce5c628bafe74a0a00f492b5",
"5/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):537aab603b744ddc975db1d21f19268bca50b41d8e00004bc5303ecda3f4cb30",
"5/linear_attn.in_proj_qkv.weight/lut": "float16(16, 2):b54dc5772bc461a69b19dc4fc9b4f96d78dfdf81aaae26026633fcf3dccf484c",
"5/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):ca9e0632682c9b29258caa01558f2df5195116a4ae92322b4ff1503e0bb79162",
"5/linear_attn.in_proj_z.weight/idx": "uint8(3072, 5120):93ed4819650a02e8dec7bffd1535d4c180bd755e35b8d94a4e0bbd47a910194a",
"5/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):b21f29552383ae298a987600c095b69f7200fbe6925950a36cad06b1a253b466",
"5/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):b69a5161a8a3302bac7340d7b76f6c17a383a97f02d5bb2a54f07f713fe7eaee",
"5/linear_attn.in_proj_z.weight/lut": "float16(16, 2):ba17820c29f4e49d74d164915832bae8c7da1aca6303e1f6ae2f85c44ec238bc",
"5/linear_attn.in_proj_z.weight/scale": "float16(6144,):775ae3c9ccc48d57e5cde06fea7b4553a63b82603c6e33b5193b828844fd16e1",
"5/linear_attn.norm.weight": "float32(128,):49e9db7d50c07a35a1239b4a66dc0cd2bd18768c8dc0daa7fdea792310ba7fd1",
"5/linear_attn.out_proj.weight/idx": "uint8(2560, 6144):f8d212f9711dba8552cc9279888056a6f94f67decb94335f33a56cd03c350e1b",
"5/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):ffa12ad11c3c109941f9c23106f521b6aa5cc5f00d618e7bfd655bdca3742cd4",
"5/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):9b654438c8d4fbb375b9016a2731e12867f312514ca84ebe5f500ae3cc489ac3",
"5/linear_attn.out_proj.weight/lut": "float16(16, 2):d6973b926dcc8c23277c0dc72302bf2a0b163d5e99c9c306cdc9e361b82b19a3",
"5/linear_attn.out_proj.weight/scale": "float16(5120,):840a9f36ef65b9cd662b543b160d2560b817e260881372a2a9f5014f8e000d79",
"5/mix.rotation": "int64(2,):793447e5e3e89e7687ac8acf434c72dee42ef51c96a3bb8ded93f5bf0ca0656e",
"5/mlp.down_proj.weight/idx": "uint8(2560, 17408):74e54a37ab2139c5135e72ffbdc3880b064ad99a659705039089b41b9c093f53",
"5/mlp.down_proj.weight/lut": "float16(64, 2):e8b440e12c6f64378e6ed929624b4f4b2eeccea2e54c877f86b82e73516c9cc1",
"5/mlp.down_proj.weight/scale": "float16(5120,):062ff944e4e0a45545b220c7ceb8bf622dcc79f2d02b05e91fcbf3fffff6aec3",
"5/mlp.gate_proj.weight/idx": "uint8(8704, 5120):f19edc070c51a18a5908775adfd944b07033e98053073dc105affb035fb2f04b",
"5/mlp.gate_proj.weight/lut": "float16(64, 2):36d4a78e8355f7bbbe81d41cac8c5c2bec54fcfb50489720043a6e963726b8a6",
"5/mlp.gate_proj.weight/scale": "float16(17408,):759d18c75f4bde765c671a568587a36e661f1dd3a819beb54e790668355eaa18",
"5/mlp.rotation": "int64(2,):798fd29fcfb5aac78d22603bc2e4beb424cb7d6a99643fc28c1859ed02c8b781",
"5/mlp.up_proj.weight/idx": "uint8(8704, 5120):fb169035dd981a0a51da3e58607e7daf0e45dae45cb27a446c9c46c9a36c04df",
"5/mlp.up_proj.weight/lut": "float16(64, 2):9595ffaab37de22f9d472f68d830b15da9c985884546b02c48e7befb22c52ff4",
"5/mlp.up_proj.weight/scale": "float16(17408,):e8e90d77a7da11d38aa101d7328f7e47033505a5c469cc39b3e28d8113765e96",
"5/post_attention_layernorm.weight": "float32(5120,):65ec7099d9aa56a95164d2169b74d0d09d6aaca09134a20de29dc61d2bc7de9a",
"50/input_layernorm.weight": "float32(5120,):051ea96f373b24cbd0f8af9ef70e161fcf8caa9af2bc962728cc63d266ea5fd1",
"50/linear_attn.A_log": "float32(48,):bdc48512fac63a7a75a618dcde534c0bb5aa45605326c0a893afc834d0be4ddd",
"50/linear_attn.conv1d.weight": "float32(10240, 1, 4):c7bad2e8ad4175c8439a8c26c24521c32cc467239c585d7adb91d340c8c6a161",
"50/linear_attn.dt_bias": "float32(48,):73b3d59242cb4965a02ca365f0c880c1e0ed8be79d6aafda3d8de2d4d0214acc",
"50/linear_attn.in_proj_a.weight": "float32(48, 5120):599c245e9f2474714a2fb1f5ea898f4e98d08c398576f87a748e7c390e602a45",
"50/linear_attn.in_proj_b.weight": "float32(48, 5120):9d82799d3b2bbdc4d90291180b828aa2ac9c3a95352d700f436f57fbfc18f797",
"50/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):480ccc2f79e9f89020a34979da347bc947c6bcbe101f7283e7837ea44fccebc4",
"50/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):99089e5e8b9deeb2d644fe7feec070a63814c75031e9bf0099c927a1a9f178fe",
"50/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):63f34466ff452bc2b7cd1b3425f19bed62d1ae2ba26413d80948001a96e870b4",
"50/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):45fcce17f8d3e5340c892a181f8bf551e992e0dbdcdce9fd6fecf9504ce0cbd0",
"50/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):8d71fa0dd99f166e2311456ea930fdca588834be8f25745e4349384156596674",
"50/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):a336c078a39be2fbcbe36b3c6fbd5737426dcca25178d9650407a9f36eae388b",
"50/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):914202bee75ed1954f776770f0b17032c6ac754f93aacad49c598c881b21e6e2",
"50/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):1d74c609e6453005723c2fae12cdddb80af8b293e80c6addcda2a54bb24fdadf",
"50/linear_attn.in_proj_z.weight/lut": "float16(16, 1):73832017a2dce0a37d56a19536615cbd6ac39b5d2f9098e366f6e1fbe7115a94",
"50/linear_attn.in_proj_z.weight/scale": "float16(6144,):00a5008d77c00c06768c79941b2e714af9603f3153d3eaee50569b2b2b0335d4",
"50/linear_attn.norm.weight": "float32(128,):99c5df80424c5fe5626f4e01ba799baac9cc7c3198f78521641ecc32345f10da",
"50/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):d5c0fe19e7367b9c9993a8334db025ca168a4471ffe53a55be770378ca54bb06",
"50/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):b4bd2939914a95a798005c33f7eb633adf031a732d60fb35810892494d7b3b2f",
"50/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):f93a7bafa0b3ac07f8ca6f1acccb3170502ddb2898a6fefb95fa7b22fa5cc4fa",
"50/linear_attn.out_proj.weight/lut": "float16(16, 1):fe3bc4898fe88e35d9b934b1b60bd2a25ac4571b6c170a97a3e01e7f65c0d178",
"50/linear_attn.out_proj.weight/scale": "float16(5120,):66a70a48d42aa5139ea486d51c87ef1111c02a8bd55cecf8978db57b53e9d9bd",
"50/mix.rotation": "int64(2,):5643b26ac3901349c317abdf4602ca1331080b358fcab9c1211c61355e77b9f8",
"50/mlp.down_proj.weight/idx": "uint8(2560, 17408):06ba1c4c5f0305655ceab1c0b7570816b0e5ec4d1da574cb76f49ec2554f33cf",
"50/mlp.down_proj.weight/lut": "float16(64, 2):52094bf25ada878cc666efdbcb1f8bc4363c3a1c3cbd6cdfc7af5b37c7100d97",
"50/mlp.down_proj.weight/scale": "float16(5120,):ee36d28eb1640370f8a23504cf72f477426dd327a14ae0ee76c0c351e9ad9e21",
"50/mlp.gate_proj.weight/idx": "uint8(8704, 5120):d694349f6d04c8acdcc91f71506d57eb421b78b0f7ac6c39fae1cedee484c34f",
"50/mlp.gate_proj.weight/lut": "float16(64, 2):2ee8c5f41502a5d9cd5496e80dd0550496c8ace54b2852e5beb6dd9c57a65264",
"50/mlp.gate_proj.weight/scale": "float16(17408,):b02ec6dadec785e65c0e450115e6306bed1589489dfd19b68838269b518c8538",
"50/mlp.rotation": "int64(2,):47a3d50aadd1025021bcd52efcb6caa72debe2b976d58ed37212109e6b94b3df",
"50/mlp.up_proj.weight/idx": "uint8(8704, 5120):245dc0d5ff56faee1ba6ae6e51b43d1b83483168f28999dd15a4b0ebbae6cf28",
"50/mlp.up_proj.weight/lut": "float16(64, 2):0f0a09738a298ca3d20b6359ef7b150fa1ceb3e0497118d63855eaba1a520c58",
"50/mlp.up_proj.weight/scale": "float16(17408,):8902a0afb3c5bfd3b7f5eceec156369a7d75561386d2289e86a23f44c17b72ac",
"50/post_attention_layernorm.weight": "float32(5120,):ded2cafc8a5310a1c4063b1c7e5fbc0fb602eb0bbec8e29a1669161869583eac",
"51/input_layernorm.weight": "float32(5120,):e4704fb2868b44037ba763cfc76fc87c96c9f05e942ce90a25a0997878436f29",
"51/mix.rotation": "int64(2,):0c1155a1145be01b5f25d3cf25449d258a6a5cd566b5788a2044236b3be5bcf8",
"51/mlp.down_proj.weight/idx": "uint8(2560, 17408):a2ec137017657430c64d37b504c6dc21e342bbe6a40b6c03de397dffaffcde5c",
"51/mlp.down_proj.weight/lut": "float16(64, 2):1ad8deafd0049aef3a0e060b8f2ecc2beac1f7cae176f044abb103361c4de955",
"51/mlp.down_proj.weight/scale": "float16(5120,):a906068db33d5a280def6032602dab6b8ab4017306fc01cbc3656158f0f58286",
"51/mlp.gate_proj.weight/idx": "uint8(8704, 5120):8e61493dbe3fe07054390bc46a29365fecfe09739751b5373202ccb12ef369d9",
"51/mlp.gate_proj.weight/lut": "float16(64, 2):ead38a973c470ea534aa54aed2933e3c8b8caacc83bd5717a72758105661c996",
"51/mlp.gate_proj.weight/scale": "float16(17408,):390afea38b82148f0423acf99f28370f3cfdd3645415d5f1665f695af225e7e1",
"51/mlp.rotation": "int64(2,):7b17fbbfd099f62c19cf49b5ae4a255427b1eeef9ff3de19d3e758e1758d73b5",
"51/mlp.up_proj.weight/idx": "uint8(8704, 5120):86f248f27e201d50bebc09636ef5df02135f33f4a383b9053ea40316e035ba07",
"51/mlp.up_proj.weight/lut": "float16(64, 2):bccf97ade03cc2007b3d51650e05582e403ec3dcaa4d8721a080ee42ef986523",
"51/mlp.up_proj.weight/scale": "float16(17408,):6c91e82a7e75c7af54b4ff2633eed4ebdfa14523a0de219822eeb3da16b8c843",
"51/post_attention_layernorm.weight": "float32(5120,):e712aaee32dfe14df04b857ade61a5686eed01450d0b9cd6122747057d1566f2",
"51/self_attn.k_norm.weight": "float32(256,):745e82eeccfe48e3c12ef539565e8747062e778a1e53ef1460ab5db265600ce7",
"51/self_attn.k_proj.weight/int8": "int8(1024, 5120):76223c13fc7f04eb5263a855cb471a8b475049cf08363b5bc8d02ddde1227dc5",
"51/self_attn.k_proj.weight/lr_a": "float16(1024, 64):4afd4a7bcc055683448317cb5ecb139464e9d2abff4d073ab1166e61e779ef3c",
"51/self_attn.k_proj.weight/lr_b": "float16(64, 5120):f63afa5e3a25e22028d1ee5fcec1a908c8210689c71e3fb18d4e758c9ebbbac1",
"51/self_attn.k_proj.weight/scale": "float16(1024,):f9dab2622f0100016f5f92fe3e6015e5bca58ef6db2fc59a0a78bc4cfb647758",
"51/self_attn.o_proj.weight/idx": "uint8(5120, 6144):94edbc4c624f7108885fe00a1f5102db4c0c1aa8ba77cbe28f3dd3b85d95ff60",
"51/self_attn.o_proj.weight/lr_a": "float16(5120, 64):0f9c8fa59a31ac952eb4c18e54f17464bf3c34b28967dea85541c8bf6f49620f",
"51/self_attn.o_proj.weight/lr_b": "float16(64, 6144):492fc804ff6c5528339296f87669667d3919a9e67483aa5c3767e65e88e69624",
"51/self_attn.o_proj.weight/lut": "float16(16, 1):6b4d0be3812566e104d72471ac8eef5c6dec9a8ef9d8d979dac4a4b02d3610c6",
"51/self_attn.o_proj.weight/scale": "float16(5120,):b5fba50fb618c3473c2dbec2537b6352b9c8cb1d3db334d9f955c5036bc0b732",
"51/self_attn.q_norm.weight": "float32(256,):6ca26ca6d48c4fb7b14ab25044969c12919c3264dabf095271af16b792a808a6",
"51/self_attn.q_proj.weight/idx": "uint8(12288, 5120):062e2d71449e01834cb35f602e3db256a9061734c13a3cfb019a0a2a29a066bd",
"51/self_attn.q_proj.weight/lr_a": "float16(12288, 64):992b11cedbaae259bd14b7068491c6bb6e43d13737217a55a1ecb0a802bbf7df",
"51/self_attn.q_proj.weight/lr_b": "float16(64, 5120):dc1b6f970dfe9474d8e7502cab97167914c19a344d0bb68bb60fa590211f4312",
"51/self_attn.q_proj.weight/lut": "float16(16, 1):dc162163731391c0b885040704ce7ffb0de90ab62e9328e4d5d85671f9dbd317",
"51/self_attn.q_proj.weight/scale": "float16(12288,):7d3e3f394614647f64240f3a96bc0099594b2ba389f26f0c82b2919659600a07",
"51/self_attn.v_proj.weight/int8": "int8(1024, 5120):465095c7b650931c513b784a050e8b0ac82ec64c0fa1e55d1c1fef6ffe5e2515",
"51/self_attn.v_proj.weight/lr_a": "float16(1024, 64):997fa2a96372e478959c75f00e9c350bc92477b1e1c039bb66bb754c656d84c0",
"51/self_attn.v_proj.weight/lr_b": "float16(64, 5120):60b3f4e8658fcfaa270958d57c5939effa1ceb5337a0bdb6162d704f36f1971c",
"51/self_attn.v_proj.weight/scale": "float16(1024,):c88a999ca3b8d66f7541b625c90b75348a683575ebcb8b0ab3f5eb942f073ee7",
"52/input_layernorm.weight": "float32(5120,):5ac57d77840ae3d9bba057086c9b9ee81f9273d36d191cf888539c7aef5c3793",
"52/linear_attn.A_log": "float32(48,):32b19762fdc79e47f99794d4ea5f389bf50e09fd0ac20c4d7443a0514eeeda37",
"52/linear_attn.conv1d.weight": "float32(10240, 1, 4):4e42e77edd9a7ef8c86c05733d50876aed39b3cd80c978e7015be2e5c0588df2",
"52/linear_attn.dt_bias": "float32(48,):b9609fafa1fd7632a76773a125ce93bc2d8166b6b4c2e7ba27f50248588cfb33",
"52/linear_attn.in_proj_a.weight": "float32(48, 5120):5f8c2d7f67772dcedff758a4da1da4b4aadc32b52a5b977c92578040300a9552",
"52/linear_attn.in_proj_b.weight": "float32(48, 5120):528ce35d448624aa893900183610b7efc2356ef974369ed683f2984f7a17df03",
"52/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):8b3b7407bae93a69b50270c3fd090a5b0e2adaba50bc5d2e86a4774e7182c205",
"52/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):045c071d66f64f692bc202c4c4eb933ca35f4a810ab5405041760e6f371a5bb5",
"52/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):ffb56aaa5615fcf006190feb77e5f314fdfec984eab2a16e9f81a986814843e9",
"52/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):d1db5f9ac383e9d9d6980fb771e69fc939f3bc5e77914ef73bfc4816ad782c3a",
"52/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):4751b4d89935bb132a4bf4c8449e142fc81d4a7214fed63f0b2c7a417d578c89",
"52/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):12d2252bb40f7a20ce1393990d75a2bccf0c2e4cb37454bfefa9e049fda45887",
"52/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):701c9622de54f74308326078634aabaad186971d8fac05cd3a04194d9ce4e13d",
"52/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):7935213c19437c523ae13acb8df7cdd62f268ff32c3c746feec6fb2f83b4b7fb",
"52/linear_attn.in_proj_z.weight/lut": "float16(16, 1):bdec3cd270de77347a7b36dd2b59894aa2d9fe4a088288c26e0b768ba8460054",
"52/linear_attn.in_proj_z.weight/scale": "float16(6144,):3f824a712f14399a525a9429789651acacbdc57c926205ec7e4a26bd13f3c58c",
"52/linear_attn.norm.weight": "float32(128,):151818738f67ee69d17cf8c080c1cd74b2507de450018df6c8aa36046e13c28c",
"52/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):a607c80c04ca35815d1e9592def664645caa13302bdbbcaa4c99f3f8ab453428",
"52/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):1cdb126b55fe349d26bc0e1d23efd3eda347cc317e54a9c8ade19cf421c8e801",
"52/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):4dd4b79af1dc4428a4bdbb37c3552f6e46f5fc48b8fcd92334d2095b12a224e1",
"52/linear_attn.out_proj.weight/lut": "float16(16, 1):0445c9a5aa69b0b504e1d39b6ad93ede421eff6b731c76d9c5718622555fe10a",
"52/linear_attn.out_proj.weight/scale": "float16(5120,):d0bd909a45b3a09b1376ece747eebdcfc2f2b934287fe19229df364b0c1a6589",
"52/mix.rotation": "int64(2,):abf51ac577c4ca3c3e1ebc9c53b87f9cbfe9fab8cd64a7d5c03f10bd09b8d933",
"52/mlp.down_proj.weight/idx": "uint8(2560, 17408):b0f623283ebec75c90c3674fb39ebbfd46f19c8072d889f38f6f81b217033a55",
"52/mlp.down_proj.weight/lut": "float16(64, 2):583e838fa23327752eacb40d63fa2c5913711f483dd1f4d7aad29c93beb0085b",
"52/mlp.down_proj.weight/scale": "float16(5120,):f462e4b1191adbe064998fd5ee7aec3a5d52fd0f0670bdfe528973ae6d31d009",
"52/mlp.gate_proj.weight/idx": "uint8(8704, 5120):3c7ff5ad4afbde50de9bf5e5bf70e5a216ca6f82c715484b54859f64f42a79f2",
"52/mlp.gate_proj.weight/lut": "float16(64, 2):2d14d3b2ce2489533788f586a0d093f807cfe4b3dea4bd51395241cfd5744b52",
"52/mlp.gate_proj.weight/scale": "float16(17408,):04e56e7485fedfb732671c467ba528520716218e371bc4cc2e567081d74bf803",
"52/mlp.rotation": "int64(2,):464a412070ae802d586d01163f4e07d723bfa54b219d0bc17dd13af1ac797e44",
"52/mlp.up_proj.weight/idx": "uint8(8704, 5120):f3cf2eaaa805952a8e8a9f0778a2cfa9ba1caaebbd456cfb7797d86292dd4c2c",
"52/mlp.up_proj.weight/lut": "float16(64, 2):7e7ce67c9337e0ea76e80cbad9454d8b58ef528ba1d321b5dfb2a178c04a61fd",
"52/mlp.up_proj.weight/scale": "float16(17408,):5f74abf27e1239667d56c78a523dcde7ff9ff7b01584b6deaad3428cc8feb0cf",
"52/post_attention_layernorm.weight": "float32(5120,):0d10dd7eb820272ee2fd522161808a628209793f8d4d9971506d403c6d4cd154",
"53/input_layernorm.weight": "float32(5120,):1863b9cf8e9e721090d12807597d60f70d4e990eae788609e02eaa8d860cea3e",
"53/linear_attn.A_log": "float32(48,):a58c10fda667db08e3ff431be911e585b0760eb4136fa80e6169ac3548625410",
"53/linear_attn.conv1d.weight": "float32(10240, 1, 4):33550ac24c51a38d01474b9a1ff5d0e383cb11a637e2fd79e1a1efd47f73c61b",
"53/linear_attn.dt_bias": "float32(48,):f65b2761e03c9bae75ae3d4fde7aa129d50f18f6f71bdb38e5cbca14232aa8e3",
"53/linear_attn.in_proj_a.weight": "float32(48, 5120):9b64a05a8ded533ab49935cb8185782ada708ca5817114151a8196fbf8a2fc2a",
"53/linear_attn.in_proj_b.weight": "float32(48, 5120):a68a10c5c58ef5747d8b4fddbff50f1764a0549c815da706233787791a6bf6a8",
"53/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):b2d195edad29e421ad832f72d6480ffdbc2eaa815bb30b591b3645d8c683d46d",
"53/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):5c953415344b2557f90873d88344f7603f03a747cf175b3f7d0e7fa928c194e1",
"53/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):a5f17d9d138d683a606e5c760ca19960535c65dcfe72a71b5f52f8455cf352a7",
"53/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):fb117a54b6b3a5a74db511007938ca37026c2b775adb89be84031f860e8cfd55",
"53/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):03b712cab4fbcee24b8be2e72d8abb1193307711cfca4919a145a8a59b63d818",
"53/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):250d9f66b2f19d75a98f66db28cf41d7f93374506c4bda27174634146bc02124",
"53/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):64738dd2a07f482b01da4643ff5770224437f847690f540ac8e250b81b9e5779",
"53/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):57ca96b067ffa593fa6fc582c9cdb45b665855c2a4fba66b7389d87c03441ed6",
"53/linear_attn.in_proj_z.weight/lut": "float16(16, 1):5b475237a1882d040039c0022d71b2c6ed58f54b9acc736f3a995a7821c77d75",
"53/linear_attn.in_proj_z.weight/scale": "float16(6144,):852dc7aae9642665e4fdacd26b3ffca08571c74ba2008f308bd2384ee251227b",
"53/linear_attn.norm.weight": "float32(128,):039ac6c1f658759fabd51ae3a28da42c2f7152ea6c73a9b0c6083e21588c1932",
"53/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):8c5e740fce4990eb0d8c2641e5f866ea9d08bc5e12e1827dc4b8a843f4eb6a3c",
"53/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):fe574a1b30373981a31a5226e25bf720cd4c727549732fc2f924de4830d94055",
"53/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):a17ba89bd8d1c433b1645b055e801bc7f5a0a0f81554ac7a791c20427ba1a39d",
"53/linear_attn.out_proj.weight/lut": "float16(16, 1):dde66b1705559b442a1cc0aa0a7d59d09a389951fda1171ac125a2b25870e471",
"53/linear_attn.out_proj.weight/scale": "float16(5120,):1d940ac0fe3f26e80bb3c82af85bcc343819aa1f3e5e0f9b9b1345ccbb904735",
"53/mix.rotation": "int64(2,):5228b7a699b17bb54e7cd14e62251bd9edcce2091ffd75f4c261ba466c00f12a",
"53/mlp.down_proj.weight/idx": "uint8(2560, 17408):6a91384e98e14438319b597b949bf7091cce3c743ec4d7ccc563231aab7be428",
"53/mlp.down_proj.weight/lut": "float16(64, 2):db42df87ad81e400f0a23dc58c9d5176184290a5bcb4048a187fb7c8137b5241",
"53/mlp.down_proj.weight/scale": "float16(5120,):da41d2b0ad0ad297369fbcbbac638b83dbc9cec6f89789a716eaf2fcf21f4104",
"53/mlp.gate_proj.weight/idx": "uint8(8704, 5120):9e962b47ce817758b906ab5962aaaf0e893394ca8f7cf4373ef39e3cebd8aaf9",
"53/mlp.gate_proj.weight/lut": "float16(64, 2):4179d122efba50fef8c28b23ca7e29f69afb03347c373f0db052fe822ec7da41",
"53/mlp.gate_proj.weight/scale": "float16(17408,):0f4d27decd7ed1f64c1a357fd95ca50a8a41c27d961c86dbc813e40c4aae5821",
"53/mlp.rotation": "int64(2,):f7bed9d7dcda1d0470ab75f6d3a34680e09da4973c108cbd5bf15a38b614fccc",
"53/mlp.up_proj.weight/idx": "uint8(8704, 5120):a7219c469e3aef0dc71d9b000d5ff496f4e14544835b7ef494826b6193fb6282",
"53/mlp.up_proj.weight/lut": "float16(64, 2):853b08014c9f990612b1094411a5284c50151a7cd3da395f5d47901ec053f782",
"53/mlp.up_proj.weight/scale": "float16(17408,):da416d6599272380a1bb1f57eef9f0efa7245a8ef6890073581d26cdaf99f4dc",
"53/post_attention_layernorm.weight": "float32(5120,):6164afac184205dd6ed17d1fcd53bf9316d4ce56256c13f62995a0f978cbfd31",
"54/input_layernorm.weight": "float32(5120,):bc68483d6d2dc5c01e78853e408f22bd5d6d9f20bfaffb72c2aad16c98fb551b",
"54/linear_attn.A_log": "float32(48,):5dc52b53d8cced6381a0b037fd90129b853bf1078e08a48f8c396cb972d0fe65",
"54/linear_attn.conv1d.weight": "float32(10240, 1, 4):eef67543b5a85a4bb3f14991b150af73b825b07cf3b8dcd8c27dba4e43a14eb0",
"54/linear_attn.dt_bias": "float32(48,):675e440459feedc368aa98db92351d3eb3101cc40effc2d8873d25ccc0e5e961",
"54/linear_attn.in_proj_a.weight": "float32(48, 5120):5dd969b5c0157ace13574f2e7c76ac7d07a842e18d4141e4f5c3f2c31ab42ed5",
"54/linear_attn.in_proj_b.weight": "float32(48, 5120):5e8d5ac202503a37a6de01addd6afd8cab920b5af27d32d4aef3f17235a94de9",
"54/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):94824c566f7fdfd6fde3f6e975289f85df6abc085c3f35188fa8ca4982c6bb7b",
"54/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):e17d21ba663f7b1269c7a92b7205e786d4ad03601b87e164df72b647730f929a",
"54/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):e7ba601467108ef00bb291cd5c5ff67e8c897af3e9d4047c497594cf61921342",
"54/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):cca9358673a386cbbdfb0c5fd4e96933628c95e21f61aff5ca5f55c3b3260d88",
"54/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):f11cd431691754020248c584818831aa67152382b99224445d1cac0b094c4953",
"54/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):a2a688177caa78962719f232c90e4ca574ff5a375c004043770a9a476bb88e0a",
"54/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):98edab694198dbcc6eade1c65e01d5401d50b97d78689a646e44496842ebb5db",
"54/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):7e00b9af64dd501bb3f5ce0b627a045b681359c32f2b59ed8082b05fdccad78f",
"54/linear_attn.in_proj_z.weight/lut": "float16(16, 1):22c5ca9a02636a7ed4c74039a2bb05eceff935ab92284f70d67b3995339f28f6",
"54/linear_attn.in_proj_z.weight/scale": "float16(6144,):263cfb46d6dead8c05c7bcbbcc8c2e394c256106705376498105c784e04edc99",
"54/linear_attn.norm.weight": "float32(128,):14fa574029f0eaec2574766f0fc3eeba0feb22f6d90fdc0df929cf0dda1b3fc7",
"54/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):61305417959cc957081f9f7b731f623c8d1fd3fcd9a99901341a655ab74ab69f",
"54/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):73029f250644f885eec4348aefbd3c73443737bc591770550c2cd49436e728aa",
"54/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):ffa97336bbcf29ae249858cfce923d00d072f37a462a9a2f0ace98c03c223f7e",
"54/linear_attn.out_proj.weight/lut": "float16(16, 1):0a9607df7a500468b56db3c099c698f4fc13d545c063da17cdc82f74a8e1879a",
"54/linear_attn.out_proj.weight/scale": "float16(5120,):be5f721fd98830056d6bc07b2823a3944a6794d5c2d0b3c9168c69fe4369ba5d",
"54/mix.rotation": "int64(2,):5ffc22212a6f2e9ab1ab125ae602a245b39213d6a5c320594f9b6cdc4305d7cb",
"54/mlp.down_proj.weight/idx": "uint8(2560, 17408):d2cbc30227bd5147ac9f3666c14f1df62b4c68e45c05e286fe18a61cf9755621",
"54/mlp.down_proj.weight/lut": "float16(64, 2):7f0cd963142f4848cc2480aaa4b7260099c0e223baa4195e5992f90896b4c1c9",
"54/mlp.down_proj.weight/scale": "float16(5120,):0c491b73e86a8f65182d2bd33a645773502d40e423ef58a1ff91a1f6fb64e103",
"54/mlp.gate_proj.weight/idx": "uint8(8704, 5120):93b7891127aabe9602c43b4a635d53ecf23c64e0c0deda404401c90d8ccec4a8",
"54/mlp.gate_proj.weight/lut": "float16(64, 2):55363d66c582531ffed829e971387936933c2fb983a1c0fae71a804c440fe908",
"54/mlp.gate_proj.weight/scale": "float16(17408,):08d00529bc8fa550ecb44722a521c26b3d887226fc4f17beda22d3e86e781a62",
"54/mlp.rotation": "int64(2,):fa3ed95c2a97a7992a785da8bc8fd2a5a4b30dd8e6b78baad858f377d0acc498",
"54/mlp.up_proj.weight/idx": "uint8(8704, 5120):338a777ffa7e0819e7711a34055469ac5c4cfeacc74b0a46a41e0bbab2a2ed5e",
"54/mlp.up_proj.weight/lut": "float16(64, 2):c44b8fcf0ef0e0080d3af2b6d636aa292428327aa5cdbd900a0f192a2c928c94",
"54/mlp.up_proj.weight/scale": "float16(17408,):558806d564404e55b1059e7dfb5ef9a4d47bf58b0f43efdabe5b2b87fc9d8caf",
"54/post_attention_layernorm.weight": "float32(5120,):164fe8fc02cf87ff0be51b9e36b8a0685e8b2ef89fc5f85b18235412206f1ee3",
"55/input_layernorm.weight": "float32(5120,):9d7ff37696187bda32728908783de6ac024a7155073b6e31f6e6f8f05e158c1f",
"55/mix.rotation": "int64(2,):d6b9b5c319d3ad260552dfd017601764e09b453def6113c0fa336a5bec537db2",
"55/mlp.down_proj.weight/idx": "uint8(2560, 17408):279ef2aabe180fa86bcc852272c7af561d678248e651ec8f931d67fbe8e714e9",
"55/mlp.down_proj.weight/lut": "float16(64, 2):598141f7352ece3a973824cdf95a13974d049cec6fa584976600e0de1c7f3ec1",
"55/mlp.down_proj.weight/scale": "float16(5120,):3c9ca4cd11c2a8538eaf620c3b8b8eee658fcdef5f8d0df93c348c8a8aedffd1",
"55/mlp.gate_proj.weight/idx": "uint8(8704, 5120):9553051acc5133d634ebce85dd530369ad01b8eaea75c9201a39ed37a88d6e6f",
"55/mlp.gate_proj.weight/lut": "float16(64, 2):70d7d0617706a94c7e6ba12c1aab77e59ca3710f76bd645644c388b306af1f56",
"55/mlp.gate_proj.weight/scale": "float16(17408,):a0117a404a92ec11b3666ae541315dba857e3e72c75cbd0e34aaaa21e2f7d0e2",
"55/mlp.rotation": "int64(2,):b1a86156afb7c404b12f271c3113f52ff717fcb381e3242350d8bc0d7742345f",
"55/mlp.up_proj.weight/idx": "uint8(8704, 5120):0f719db52180480391ed4060793771f9c45fc1ce61b6f3ad9de89352a62cbe35",
"55/mlp.up_proj.weight/lut": "float16(64, 2):457f4cbc329ddb22c540369b92297714f0a825fc8d12a3745e3f75b5245dec94",
"55/mlp.up_proj.weight/scale": "float16(17408,):a011b62aeee8d70f5739ae87b6ec4696c6b74f2e3644ec7d7f1b745f3d85b2b0",
"55/post_attention_layernorm.weight": "float32(5120,):b43e2191c8b7ccf870a1e3a87cde67afecf78aa83776e066bdd069513df4a4ba",
"55/self_attn.k_norm.weight": "float32(256,):3e15beb3810e3c11228a6052034e614c3d04a66aff0b52fd4ad7135fcbf9f4ba",
"55/self_attn.k_proj.weight/int8": "int8(1024, 5120):9355c20ce448d582de111cd70e9e832f8d726f398d84b4e856e581439e6ff085",
"55/self_attn.k_proj.weight/lr_a": "float16(1024, 64):b043bcc5a0f63ff9425a6acfdd4ece417f08e2dfd860ca5151cd933904bf0a9c",
"55/self_attn.k_proj.weight/lr_b": "float16(64, 5120):07af548d61e8e2709eb660ee78d5efe89fb01654dbd86a07c46636e9ed706a1f",
"55/self_attn.k_proj.weight/scale": "float16(1024,):5546560e2f386c41d16d08016e658b916317d7cc47033c2710084a5cea3e55ce",
"55/self_attn.o_proj.weight/idx": "uint8(5120, 6144):3da86325758ec2335c0cc7e64a909e03a39eda78f526bc70d55a7e43e6998554",
"55/self_attn.o_proj.weight/lr_a": "float16(5120, 64):56ac3d767c658102f6e5bec127bd2c5da709b512f2174a4f3c8725cfa25441bc",
"55/self_attn.o_proj.weight/lr_b": "float16(64, 6144):3f677dbf526e99978800c63267d9023800ec5c844fccd43eda71ad1ece43d783",
"55/self_attn.o_proj.weight/lut": "float16(16, 1):af5c9331573e09f89f74a8ee39dbe22bf7ae35505e373ebb36aeb323f35c39dc",
"55/self_attn.o_proj.weight/scale": "float16(5120,):a062fd25d39e30241142b94a184fc747bd427759c15eb9b3cf6f4629f1d0e199",
"55/self_attn.q_norm.weight": "float32(256,):a47be09c5d01d5404f244280290a693541822a30ca27b42862e4e8637cce2d54",
"55/self_attn.q_proj.weight/idx": "uint8(12288, 5120):b452a4d38f2eb817b69a0211f3807d05446523426407bb0c3a092b816b060c01",
"55/self_attn.q_proj.weight/lr_a": "float16(12288, 64):fedd19a4ebbd7603975fc79a16f098de0da73fccfc26e3e55a047c6c038efdd1",
"55/self_attn.q_proj.weight/lr_b": "float16(64, 5120):147aa4b1d222e03c03d5cb7847fa7df9d90b3a24b2dbeb0b551887a09fa1ec0b",
"55/self_attn.q_proj.weight/lut": "float16(16, 1):ea1b0ae19f7a17fe3dea8b42afebd268db9356d98702fd5fc601f55c2b67554e",
"55/self_attn.q_proj.weight/scale": "float16(12288,):f564fbc4e07d05c03d5d4ee7f901ed3b718c48e98dcb185cf231275d9916c34d",
"55/self_attn.v_proj.weight/int8": "int8(1024, 5120):24dad72da76d81f73f30bcfca5bfdc2d6413c0c18c8d8ff6259d44177129a398",
"55/self_attn.v_proj.weight/lr_a": "float16(1024, 64):e830b464b4ca0314067a59682127c5e8a440824c888a0d5edbbdab4e9534f714",
"55/self_attn.v_proj.weight/lr_b": "float16(64, 5120):dbb73dfb3972af7e62b8bb17670f953040a20897e3b0c77602137d427171bb31",
"55/self_attn.v_proj.weight/scale": "float16(1024,):bbdb42cc8d4bc7173a5d33b793852e8f671d75f2851dc8dadc3d1e4781a2297d",
"56/input_layernorm.weight": "float32(5120,):4c2b3b7c9caf5126564a0ccd36d5d722540b284a36249b57adf761013985ab99",
"56/linear_attn.A_log": "float32(48,):cefe8d92f05d1a0aebfdb163d1fafb00705387acb5c3a70572292559137d882d",
"56/linear_attn.conv1d.weight": "float32(10240, 1, 4):277e493c538558a7d279165a02325455846c65bea9aa5cad71a776490a72f343",
"56/linear_attn.dt_bias": "float32(48,):10c29a5a2de7c79303841f436fe2fe5b172cce810c3594b33ba0548d5cfcb959",
"56/linear_attn.in_proj_a.weight": "float32(48, 5120):d3310497f14cdbd6fd87113174bdbeae219e27e2ca7b401493f884ed7b5d92d4",
"56/linear_attn.in_proj_b.weight": "float32(48, 5120):be99933a0387362c5891de9f71225dd693fbbd03b58587bc372c6b8f9d45709c",
"56/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):567c4d62364e960466a16612221bc97801a0324a7480581702b4e1571af2f877",
"56/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):6736c972711c0928bd646f622a80985be3b7e9f79fa1e833230799912d6f97c0",
"56/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):c78b1194ad9179d596cbff71adf2b5c666bd8fd39074494a2de7c3073dc2ec4c",
"56/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):95888173af05740b352063a7bd777513a5e8bc017b9e4c28dd89467f43925a94",
"56/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):8a76fb3c0208f52cba2494416925e821155c40343df40a4dfc5591d5866d6e58",
"56/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):27d3314af7cf6c1c48c149c3ab5c1716461e498aa73d80e9c6c518247d1c4412",
"56/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):1fa8100ac0a87c64c3555782ce0b522ad239ecbe4e180289a6e4165dccc09fa2",
"56/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):c3f2e1a0b99820a9c9daeded68b2a13bf561bb93e2373f858a1e7cd4b789d02a",
"56/linear_attn.in_proj_z.weight/lut": "float16(16, 1):c0995e4b555c740334e10f1d69e01fb857e2bfc811d0453be2203b484ac5e40a",
"56/linear_attn.in_proj_z.weight/scale": "float16(6144,):d1874fb2dd6939dbf436b229d5e7c44ecec8a67e3f2f85eeadd8b1e8991a91b7",
"56/linear_attn.norm.weight": "float32(128,):a7e9d81801cee992dc8cfebba30647c940f90606ad7dee8eeafe598b0e748347",
"56/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):7bd0e3790f5a858daa9152049ab5df40a08fa9c301a9d0a286ef6dbad7de3429",
"56/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):1744be57ef743054240ad2ec6fa4715a29389835cd64df5b5739f64e8d124937",
"56/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):017da0ca2c3efdbadbaaf1ffa2a2ee4d35dbd28a58582858a2a2406ccfe0aedb",
"56/linear_attn.out_proj.weight/lut": "float16(16, 1):48a36fc21a2ae88c19677e509f8f9c5e3a7ed5764be4c953dd4d833ed1996f16",
"56/linear_attn.out_proj.weight/scale": "float16(5120,):fcf18171096548cd80f3ee70ef4b0d71636e8ab843fd84949475fedfb7142fd9",
"56/mix.rotation": "int64(2,):b7cf43e87a84e5c529bf8db1a6cb1dfc1e259dea17b9d967f2f9483590e2900c",
"56/mlp.down_proj.weight/idx": "uint8(2560, 17408):0969708598b393677c34ef857e0813fb2f4a16c547e78d40d21344042711ef56",
"56/mlp.down_proj.weight/lut": "float16(64, 2):f05410d4a7c42df3349da7b2c140353e26fa50ef9bfa0dcbc43cbb1c984516d8",
"56/mlp.down_proj.weight/scale": "float16(5120,):7c685ea32fbbb243d6fdd4f32e3f1703ab10dd9b3e08d5ea6d66195e2c6fc80f",
"56/mlp.gate_proj.weight/idx": "uint8(8704, 5120):5423c009c58d6537e3b88b32f520815f3b88c90926b1d61f6f30f2f37ac723a9",
"56/mlp.gate_proj.weight/lut": "float16(64, 2):d0693983c1f81e0668d8950d3659bf52f0fe0a282f2cd24cf486a3d0fa9d4bee",
"56/mlp.gate_proj.weight/scale": "float16(17408,):332d587575b238d15654d57dd025660b2779fc4786cf5734ca0718e8405be575",
"56/mlp.rotation": "int64(2,):c1d7d79f1d6f6dfb243132b6998a34ec3d5278ec576a793a739ee9072cf3e6cc",
"56/mlp.up_proj.weight/idx": "uint8(8704, 5120):7909b5a13c5c2feb3e59b881bfdb0c78ae539a11182fcd6a55d143de6a10a641",
"56/mlp.up_proj.weight/lut": "float16(64, 2):cb507bf366b02f849c7dfb0e97c66c292a698aab22015a4d41296ff968d9737a",
"56/mlp.up_proj.weight/scale": "float16(17408,):081b23cd6eb1b6c8c342f69da4263954b1fbaf3eb70077ed72d90b026285ce48",
"56/post_attention_layernorm.weight": "float32(5120,):b7a48751802fd089dcd32971e76f8bc6c72629907f52037911cbba0dd3f5a96d",
"57/input_layernorm.weight": "float32(5120,):18169b63bd70a64afe7be449abca6d77bbda8f1fd2d7dc90cd7c12c8b3b472b9",
"57/linear_attn.A_log": "float32(48,):72efad97093a2885d8f3e8d057198eac2ccab0904bc3b23f5efac596e9e6bc67",
"57/linear_attn.conv1d.weight": "float32(10240, 1, 4):e61943ec01dad5451d2851e5ec012c818c0665abb5ce0aaeb82dc9bac14547e8",
"57/linear_attn.dt_bias": "float32(48,):2b379b8b987758f45545f7c380d442930a2ede02fa044dc1feb5c9bee6d94c0f",
"57/linear_attn.in_proj_a.weight": "float32(48, 5120):e3c63c80aebd1d51b6099d1c24ea4857cb127353c065b1d9f9e05b1dab74a10d",
"57/linear_attn.in_proj_b.weight": "float32(48, 5120):2285542adcaeace23a83b838f000b70848e8158109480207df4e94730c25826e",
"57/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):1fcf0bc2e7833af8bfa28bd0fe23799e4aba408b43c8aab79da19786a945c9f7",
"57/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):5fd150727e8a4732ec56ca40d36b06e807bfe93920694951ad07e4395554a197",
"57/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):a3184fbb6e9db63791b6e27a8fae3a431efcd0e71730876d7243d3f6d3cafdf9",
"57/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):ddb991fb0a6f1f0b2d3ec6b0e9ebcc39928fb3361e80601618463d78a3aa6bce",
"57/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):5743fd33324b60d45e5809fc82cca79c2ce4bd511827245132af34196a186fff",
"57/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):4061f8368735d708279e1d746f2a7cff9d6f5d2fccbd84f0909ac900aa3cac57",
"57/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):f9a1c1d95eae15d152eaa1b7d81c6935a4d37c38517b721aaabb02f94900d84d",
"57/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):6a009b18d456c5da472027c224ec51d42babdb0d9ad8dc9af4709f88193d0e9d",
"57/linear_attn.in_proj_z.weight/lut": "float16(16, 1):ea23dbfdba08be9e156791c837c63a31e0ba2b8efca0cddff3c2253bdb51972a",
"57/linear_attn.in_proj_z.weight/scale": "float16(6144,):8dfe3726c9529ce0382da033e7898342135cb24f64b28f59f4558388d1e99445",
"57/linear_attn.norm.weight": "float32(128,):fe7e2f286f3ddfdf38a4aae8036697c2c69e9f32f88f98c71da9a3181c3222e9",
"57/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):80b5cb2a72e5da30acc6be2ed0fd8292eda27658929452e71b1cf64cc732e2f9",
"57/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):3d201335ca8364bb6da770452cd2f92decde47ad7039faaa0cb0193dc2004835",
"57/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):f91bb3f079b4ce2495ebc3723d18f1128248ce1c9396d51177d1d63fdf5c0d3b",
"57/linear_attn.out_proj.weight/lut": "float16(16, 1):0d1cb881cada2383aa9d80dd4aa559b8cb2c138c7a6bfbff689f236e246c2149",
"57/linear_attn.out_proj.weight/scale": "float16(5120,):26038aaf96098ae1324748d2311d3a09be191ba51a38da2f505278c3b41cdcb6",
"57/mix.rotation": "int64(2,):17071d86ee194a0cea60ce2da4acd996e50039522a66b13a93e5eac2e162ffb0",
"57/mlp.down_proj.weight/idx": "uint8(2560, 17408):0e3eddb7a3d3243a49bf44de72edec8827c2ae26c44b5a89deaf588bd682d9b6",
"57/mlp.down_proj.weight/lut": "float16(64, 2):00479629333428307d73f17aa93238ef98d0499a94b3f2359b0b757eb436ade0",
"57/mlp.down_proj.weight/scale": "float16(5120,):e000200f726b5e19cae7c8341bfe17571b704fcca8605dd9394a2c0de144c28a",
"57/mlp.gate_proj.weight/idx": "uint8(8704, 5120):528b2f4555aef6f09524483b43e6b70803d56f960dfb0eac886b66cbcffd6966",
"57/mlp.gate_proj.weight/lut": "float16(64, 2):e4f13ef9c64109344364cff9026b5038c67886db485cf11f022b6e3fad43ccd7",
"57/mlp.gate_proj.weight/scale": "float16(17408,):57fd335b2a25ffa0eaf60fff3a7aeeec5f5e5c431252f4cc516a108f5b764c88",
"57/mlp.rotation": "int64(2,):6981cabc20178e647cc6fd459252bf26fa44ca7c5701fac03d2c10a612d2a7b0",
"57/mlp.up_proj.weight/idx": "uint8(8704, 5120):e161f710e652893861527921661fde892aee36fb6602def2046269ee8364be19",
"57/mlp.up_proj.weight/lut": "float16(64, 2):2805aea661f1ce6edbc4836051429ac4a630b6d112f27bcbe4f698d5f9cd44e3",
"57/mlp.up_proj.weight/scale": "float16(17408,):8ad27284ef104990e35df677eac2d676ef20c12cbf21ede8a55accea8e339b1f",
"57/post_attention_layernorm.weight": "float32(5120,):e108849773828ef52b019655fa6531a40beab87477260335f9d339adfa43ae52",
"58/input_layernorm.weight": "float32(5120,):10dbd84b2d52dd791f9f1b3d6efde042112679a1dd726f0d52dfb856125e740e",
"58/linear_attn.A_log": "float32(48,):495474a405d913371db1d3236d17cea9366e2c200906ade56c7e6f96f9936cda",
"58/linear_attn.conv1d.weight": "float32(10240, 1, 4):008d23022b82eb8df7a7a79b63921b3a8f103701e8c457c232ebef8423df2892",
"58/linear_attn.dt_bias": "float32(48,):48c50af765795808083916b05bc03e197d61fa9987ae380dd4dd96b720a86414",
"58/linear_attn.in_proj_a.weight": "float32(48, 5120):9e6cebbc60a0a8c403d130341a3b1efa8fcbaf67a7772bbe4a4f4a6686fd98d2",
"58/linear_attn.in_proj_b.weight": "float32(48, 5120):9a79f37f063eaeb78fb5db86e52512beb14c3bced18a0c3fd7f64a71402b3443",
"58/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):4592936ee1da542159ff994a169bad966bc07adcdf304ea649e916ad6fc1a5fc",
"58/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):989e39026ee3ece62e51b9a55466831c982b657f5c8bcd3fc7bc237ef90eef57",
"58/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):c925d8818df4b28bed55a8ddf66580d946d7057ecf3e70bd7708ee5c8fe67911",
"58/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):2f7abb5858bb7973e9e7b59d0073bb75074be14d7fd6c25a787820aa55173a64",
"58/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):3b357afcd536f12bdf6c3f1bb4e4f430de63de88f906e6769e946b39f5bdd067",
"58/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):fe1a1fff17f33a9d9c3f9458f671cc41f53645a3804944318535bfd238efe252",
"58/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):8596486fc71fa60913ed679918d9d06f62d77fbd1bec66761c0dd6f992a3e25e",
"58/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):9369af1a4b8d50f810ea1446eec51c345678890f0bcb4d7e02c2a96275639a84",
"58/linear_attn.in_proj_z.weight/lut": "float16(16, 1):b3e65e0e0c7f264b543277808cae23f9346253c51d1b6a52e5c161b0875f92b5",
"58/linear_attn.in_proj_z.weight/scale": "float16(6144,):84018a0fceda4736bfa51370cef4ee97cad822e43d8b4bb565a7a392e13b328a",
"58/linear_attn.norm.weight": "float32(128,):2032d8473e50ce521f8593936984ebca67b2f364028ab4488d8fa203b38c25af",
"58/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):f4a23465d290928144273c25319213f2f5adcae7af89d97b9f4afb850a09a064",
"58/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):ecdf2e22f093ef217e389a35a9a92f63f041363cdb81ffbdf92d6b55df2c7646",
"58/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):7e61673ff88a78fcc1e7a0e34a393b9413ac0dd676ed5db208cf8d6a1c8d758e",
"58/linear_attn.out_proj.weight/lut": "float16(16, 1):a28d0f6fc2f95ff01436a0eecebb1ab4e3bd214f086103247240ca1d94bf34f8",
"58/linear_attn.out_proj.weight/scale": "float16(5120,):9bfdfe4db3c5e3ede9113435d788e3a76dfdaa35d3d60a5fa906934a26281f50",
"58/mix.rotation": "int64(2,):a94f81d03e24fda668c34478b57fb3270c481d48808d8a380dc6261d34460b46",
"58/mlp.down_proj.weight/idx": "uint8(2560, 17408):60f4a541a323f70ec31395548766dfc53fdd3de57a1618169d21a6c03b86f204",
"58/mlp.down_proj.weight/lut": "float16(64, 2):55f35d54c0b027d8e8e4bc0b897bc46db4d55e9e2facc9369f202e7e3238cb2d",
"58/mlp.down_proj.weight/scale": "float16(5120,):eb3fe07b08301bb775746de35bf29a6d88637a3e1e39d9854f18d15d2cb63e4a",
"58/mlp.gate_proj.weight/idx": "uint8(8704, 5120):cd72b5e9ba5f5121747e12e0b5478b0be94c705d6eba8180e50472dc2f0c017c",
"58/mlp.gate_proj.weight/lut": "float16(64, 2):6d5857ad6681f4fd8d5798daba0de7defcee999d1945f973f6835eeef7ab65fe",
"58/mlp.gate_proj.weight/scale": "float16(17408,):1c12bf21b476c7e380e5182167d17fbb898c361417894cf3788fe27e514d49dc",
"58/mlp.rotation": "int64(2,):0abf3f8bf9afa83e743d5b31fedab020d55ab51b516a5da68a8320e915e492e7",
"58/mlp.up_proj.weight/idx": "uint8(8704, 5120):bda3421d9cc9478fc1c40c034295673d7313d8ffbe60b36a441b93bc5115e960",
"58/mlp.up_proj.weight/lut": "float16(64, 2):28685f263a92bb2448b28f217b66b77d9f6c6f7db7633254f375da1f7bb69389",
"58/mlp.up_proj.weight/scale": "float16(17408,):e0f49d899370a781c90c0343ba9d0a64c63802e89f4cd0d882035b3605270ca8",
"58/post_attention_layernorm.weight": "float32(5120,):fbe5348c346649690fe4f67a81e9292f174f5a9e003d787f27baac62265e504d",
"59/input_layernorm.weight": "float32(5120,):dbb4c4080c3679de9db324c01ae61ce6506e824b0f5c624910b47f7041f2c640",
"59/mix.rotation": "int64(2,):4fb8eb2755499ede96d41d370dc43ff2705fcf95e6ca5394ae97d680adf19e1d",
"59/mlp.down_proj.weight/idx": "uint8(2560, 17408):d1ca1f2f06fca7cf765424de84864ab6f5adde8fcf6d80f6508786161f24639e",
"59/mlp.down_proj.weight/lut": "float16(64, 2):68f2fd2b2f63ca1f9f832df1618a7517f5b3201b28adaf3159dc27ff330d88d2",
"59/mlp.down_proj.weight/scale": "float16(5120,):93e2e9dea7a304470ed96d4e3b541cd31eeb6b972f312d0386003339d5994c16",
"59/mlp.gate_proj.weight/idx": "uint8(8704, 5120):32f9f1bc888876ac93c5866b1511daeb12e47c674673eae5e2d07a1f6020087e",
"59/mlp.gate_proj.weight/lut": "float16(64, 2):25b06623ef2f001451d29c4fa1b68ae0f3ad89f02066dd42065fd4cd0a827d7f",
"59/mlp.gate_proj.weight/scale": "float16(17408,):818343e8e1a34caf880106086c4f6768b14cc888abbfa30eef4eaa31ce821787",
"59/mlp.rotation": "int64(2,):0799d7a28d2d392b30362526cd7235cae28f5cae743d1020d0eb6fd2a64cf00d",
"59/mlp.up_proj.weight/idx": "uint8(8704, 5120):48884966b0f90479371989e45f96bcca77fb614671b6035f7e9ecd576963a426",
"59/mlp.up_proj.weight/lut": "float16(64, 2):361bbb6f62250bb1fbb0bcffefa17167bcedfcd5935d4ea72eb7aae46c85e47b",
"59/mlp.up_proj.weight/scale": "float16(17408,):2e00b0ad23bbd5ac4512fb2da01bb6f049afb7e7b9a683c2ec98bc3f9c4ed822",
"59/post_attention_layernorm.weight": "float32(5120,):f35262f79d3e1980e38729f5f4aef6a29c328f8b8c3e40bf95b8ad083bc6b9cf",
"59/self_attn.k_norm.weight": "float32(256,):9c8485dbe2470dbff4a90a6ee67ca45c5befdc5366cb6865193238f0fd2ac2e5",
"59/self_attn.k_proj.weight/int8": "int8(1024, 5120):c1d47d463b6c8119faf61711fec6b684a938483790bdcd07048a4d8c546fd22b",
"59/self_attn.k_proj.weight/lr_a": "float16(1024, 64):436fdade5e2863ea3808aceefc64db6e8c644c536b91eaafd67e6e44b25ad3e7",
"59/self_attn.k_proj.weight/lr_b": "float16(64, 5120):be04684c5f650bfab7ad2e063f7cce072f755648ba7190523a8b99faa87122da",
"59/self_attn.k_proj.weight/scale": "float16(1024,):d765925414928cad73ac33b1b5552f8a4911501c8db2345b10ae8d6e6e046486",
"59/self_attn.o_proj.weight/idx": "uint8(5120, 6144):0ba34384542168e5b970f687eb5bbc407763c6d8226c512b1791bef51e5cc105",
"59/self_attn.o_proj.weight/lr_a": "float16(5120, 64):23218953802ae41f080671ee57a4d6e8d858e34558a6f9d836777f1e29cdf8ec",
"59/self_attn.o_proj.weight/lr_b": "float16(64, 6144):f30bae9f1e0065cb03ff6ee4cc8d38bac77f13d726fc5cde5e9507dc031d8c40",
"59/self_attn.o_proj.weight/lut": "float16(16, 1):64c29aa52fa34a5e8ca86b703436e77da090153152fd9bb0c134d0a391af65c0",
"59/self_attn.o_proj.weight/scale": "float16(5120,):6d8e222a38cf975e57e6d0477338aa500e137639c82d8b677bd86436f1d3d89d",
"59/self_attn.q_norm.weight": "float32(256,):aac386d7fa8858c065271081dda287c3d07184933b5761a60f5c0842b2b39867",
"59/self_attn.q_proj.weight/idx": "uint8(12288, 5120):1569303b2db4eb072cca542fe732225967341f4758e998a6d61c93b3e6c11823",
"59/self_attn.q_proj.weight/lr_a": "float16(12288, 64):169d1e4d6d7097df1bc9f10b56ff62eee854431da603bd0f16ff107a80e73e5a",
"59/self_attn.q_proj.weight/lr_b": "float16(64, 5120):9c8cd4b53a90c33d45ea5d583001191e0e9b8067faf71499469e60c3a8c45251",
"59/self_attn.q_proj.weight/lut": "float16(16, 1):c73338dc8a0796a8cf8fe62d8cca683c1012c8870397fc352ec24ca6bfb9e0f6",
"59/self_attn.q_proj.weight/scale": "float16(12288,):93930bc5c2ae226db2259f1157eabc6cb454a842839c89f6c04aa0590b136544",
"59/self_attn.v_proj.weight/int8": "int8(1024, 5120):63f985f3d2eadaedc84862acd7943b1df4693d9c0b298aa3097e9683c94cfbf3",
"59/self_attn.v_proj.weight/lr_a": "float16(1024, 64):0f8b79c76e74b07be7474b2c25cc61f44eb434d176439bf98dec390cf8c10cd6",
"59/self_attn.v_proj.weight/lr_b": "float16(64, 5120):7c9590809b6ab59684dab90d4b8c8faf6d281f10813e2cdadc6f7671004b5f6e",
"59/self_attn.v_proj.weight/scale": "float16(1024,):76e2412f9d762fb44344ec2efa083c1221bc1083f6632586645f785b8fdd10fe",
"6/input_layernorm.weight": "float32(5120,):33dca254bbc9cfb9d8d4c1041410728894e0cb3f6b9de5bd012d12f35f0c9afb",
"6/linear_attn.A_log": "float32(48,):f19fbc4294342ac4383b6ed4a2d2aa8ab82562ec4e0012fa167a5d0a0e3ce620",
"6/linear_attn.conv1d.weight": "float32(10240, 1, 4):6cf670d6d444ca0da1d01f9f877a45e30c31592ce61bdf740c1ca71cfcf59c7f",
"6/linear_attn.dt_bias": "float32(48,):7d2dd7efc8306ab54f325560e4354ff8058d319faab63516b6d6f6b62a641dc5",
"6/linear_attn.in_proj_a.weight": "float32(48, 5120):32fcf339acb66707b87b2e34cffc18814dc215689c7b448a90f732cd83425d95",
"6/linear_attn.in_proj_b.weight": "float32(48, 5120):6d94df1f44aa91a31790208f5db7225287d520905222a3512875bf94db8c4f68",
"6/linear_attn.in_proj_qkv.weight/idx": "uint8(5120, 5120):168d054e2251e57f847dc9c7a16b981de5c0224cdd12a0c48e51c094b3247b74",
"6/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):492bae0724bf7da74f3b0b008f99b2dbc607b0327dea559109d02b29b33b4432",
"6/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):d94f8ffe8f2928a8636983fc3feed5a5068b9b789752d8018bafe0f296eb97db",
"6/linear_attn.in_proj_qkv.weight/lut": "float16(16, 2):e140bb21b4701005f8e8d9131ded4cc50e926d1a5d2d3ac74ac3bfb27f10e2ba",
"6/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):c78ade47ed3aefd18fe501b855db1c296f6756eab12b9f52d6d189547b90de80",
"6/linear_attn.in_proj_z.weight/idx": "uint8(3072, 5120):713143efb728a0ab29f3339ea939ab0807a1314737406466341cc2f29beb2e9d",
"6/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):5b25ec4c9f912892a187b83cddec5c2743e1d3f466bcbc5d275eeb3e3f2e36c3",
"6/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):09683d03a770e9c1c400540a52a252dd92b7c0f58bc8c56eb6ee658588312efb",
"6/linear_attn.in_proj_z.weight/lut": "float16(16, 2):9eba3e00dfa47de1af6af4f05a8b26be5347473a85025e429a352a43c7fdab60",
"6/linear_attn.in_proj_z.weight/scale": "float16(6144,):27cdf31581ad2db65360b9ed947a637f4658753ff0af8f86b1fe4788a07e850d",
"6/linear_attn.norm.weight": "float32(128,):337de9c86e938a87b08baf66c3ad5bfce5a550db12ee0ad1a1f896b6055cd134",
"6/linear_attn.out_proj.weight/idx": "uint8(2560, 6144):106aee02c17b6447f20718d3126cb1917d7a6d83716e70e1dd5c0e0ee72ca903",
"6/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):dd4563cee89dbe07a7eba167a6494cc8a62a47dcc083262c237dc458ca175a0d",
"6/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):1e7125d0bb4ca329d6a5abddb1ffe2d16bab5d0165d4e62975775f67dd1d3ccd",
"6/linear_attn.out_proj.weight/lut": "float16(16, 2):ca229b9db0f663d26076341c3ca0cb1a26f9eb1b793c0e3f393fea93bccb3320",
"6/linear_attn.out_proj.weight/scale": "float16(5120,):9106706771e98e7a2c6211c348da1b67ef151859e84b37556047839469334421",
"6/mix.rotation": "int64(2,):7cff86a18f7f6c73b44952d54ea1fd66b99383b340aa5fbef60a98d7f5c9d20b",
"6/mlp.down_proj.weight/idx": "uint8(2560, 17408):63bed618b04a40688dcb6702ef74e4b884ec8b64b45c9fe824b03553be7184fd",
"6/mlp.down_proj.weight/lut": "float16(64, 2):8ed4ce420fbb069e5359ad40e93bfc7514f09d55a17ce6f8dba0af8a72b08347",
"6/mlp.down_proj.weight/scale": "float16(5120,):f9d775bff6f44dc79669378cea67bf8997d47b6a5b199e08dfef6166ffea68a9",
"6/mlp.gate_proj.weight/idx": "uint8(8704, 5120):a4f049d3ca6d6b8ccf0313fa28ff58e02b0af3fa2ca54e4f895929f1bda74e2b",
"6/mlp.gate_proj.weight/lut": "float16(64, 2):97466f13bcd9391128c1eca4f81179a89df2dca1235aee654724583d6f98c167",
"6/mlp.gate_proj.weight/scale": "float16(17408,):b6b1d13478c5324307d0c9e7e2a8f1c02ef5a17c9b69929f67c016cee9c8674a",
"6/mlp.rotation": "int64(2,):4c8f1a3232614e4bf093961ba6c2b331055f408a613596ef8492e7f7182fae53",
"6/mlp.up_proj.weight/idx": "uint8(8704, 5120):77f65c070f7f530dfffd2e784c7972a462826b38c0894f55557a1c0d101349d1",
"6/mlp.up_proj.weight/lut": "float16(64, 2):327d0df4882436d603d25d76d753c839830871e4ae3e372dbbd7abc14c8d9751",
"6/mlp.up_proj.weight/scale": "float16(17408,):7511ee17d89a735f1d7025b6eacf69567a64c635c22710ecd7a83e1dc1aa54f4",
"6/post_attention_layernorm.weight": "float32(5120,):8126b72ac736397b6de1d6407bc1d32e0eea6edbe32d69ea45bbb7e85e96a3f9",
"60/input_layernorm.weight": "float32(5120,):d73e9c2f1e9aa495a11f25cfd4de2189eb29faaff5bdbf5bc61afaf2f6c7ff95",
"60/linear_attn.A_log": "float32(48,):82bb8fbecef3a83a3e0234399f527839d220209fa9c4e31e4b19268950159afb",
"60/linear_attn.conv1d.weight": "float32(10240, 1, 4):012910958f41d5b55cf9fe2f39e9ab8a39f7ec21d0315cc862c4d32898ff5007",
"60/linear_attn.dt_bias": "float32(48,):7a682d8e56616a6ba10caac8f681b6b439e1baaf95c3f115cb603646578e561b",
"60/linear_attn.in_proj_a.weight": "float32(48, 5120):dd033c94cf20e445e28c7e55773208fd5d471362b72125f02a7f2105898adfc8",
"60/linear_attn.in_proj_b.weight": "float32(48, 5120):3edfaf897f8ae2b14b5244668dcf9173fa7177ee74a528fb6ce3eec6cfca56e1",
"60/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):f38163ecb95269af74249d192cbe3fc4e9b0b303c109f2b8c0f26ccd9d6c0c7e",
"60/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):295236c9ca46e0b1e63c87f89f862d81a5614a59a68aaafc3c5d476b5a230d90",
"60/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):4df74a68afe5da328f04dbb38bbc8a69d8f2648a4c6e8a30a15405744a145c49",
"60/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):f92a8e79ab454836ea41da6df8490d34c73decb8c46749c96e9f4736c82bf246",
"60/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):5c4cff7024bc8cd3c1cf780c3e8e2a9302bc6f30d56bee2cc63aca713a8c9cb9",
"60/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):9e69cad78df7fe088419a2da7f9a9f17a5ead2cf2fb10a9b776388a0d118a9d1",
"60/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):e157f792b8bf65d4378c2e7021108c6cb74f6eaaa0ef1f3c2b6d9e3ca5aaad5c",
"60/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):4b8876b4811efd9d47d9e525a19be74dfa1dfcd20648674dbb4d8083b0cadb9d",
"60/linear_attn.in_proj_z.weight/lut": "float16(16, 1):71608f00b931a6fa32ca80e40ede3786d6c75b0d741b5e019a2b6601a05cc6e6",
"60/linear_attn.in_proj_z.weight/scale": "float16(6144,):2582f7af4308a2952065b0c333d484b4176a95273dffe399ae536ce3a9740c49",
"60/linear_attn.norm.weight": "float32(128,):bea90e2369227f8b29878629fe3eeb5d1be42f62002c9a12f9584c4bc94f10f3",
"60/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):b9c26985806895fef97653f6bc039942d32535dca9ac94c3e17da6650e081349",
"60/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):a007ab4518a71a1316ded0b0203b8cd669673ba3884bd5093f5eb3b0c6482a36",
"60/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):542f7e013b7658cec6804afce8035e948214536680bc8cb752c8a8187de9e023",
"60/linear_attn.out_proj.weight/lut": "float16(16, 1):2d1d142bc38cdae679761102bd0246ffea8e13f122b7d37c9ffdd2e912de19b8",
"60/linear_attn.out_proj.weight/scale": "float16(5120,):63ca2014569251932c61db91d41e277636a4a674c2b387a516af153fc1d230a7",
"60/mix.rotation": "int64(2,):6eac9d1ae8889e9f03710917ceb960763a2f34d1ca3feb48bfa48422c76e663d",
"60/mlp.down_proj.weight/idx": "uint8(2560, 17408):5919287b551fd52b92b3e212be0824461bbee69cdae2e70a78c372e1787b9d25",
"60/mlp.down_proj.weight/lut": "float16(64, 2):846ff75e4c28106cc16c6459643538eaeccfd2f6fbbc453e6994df8fcdac5dba",
"60/mlp.down_proj.weight/scale": "float16(5120,):0d5d7b12aa077ea1e99d105a93cce8b82d0deef4f4e1ca9704289c04e6c0679f",
"60/mlp.gate_proj.weight/idx": "uint8(8704, 5120):23cd53ec19889b5138ebc1e48995a2d26a2b268c9533a65e5f42ccdcf632ddc0",
"60/mlp.gate_proj.weight/lut": "float16(64, 2):95e74b78fa32ae2acd5cfe40a214f751f217049123828bdf2c00da46d4c49be5",
"60/mlp.gate_proj.weight/scale": "float16(17408,):f0382e8b7e5b9bd8eb8d3ce0e1676106070e6cb5c35ee5dd893714edc072bd93",
"60/mlp.rotation": "int64(2,):ed5266b883c245981bd55c042a4a3ab48578069a3a6bd7b62d55b9e12029bfed",
"60/mlp.up_proj.weight/idx": "uint8(8704, 5120):da94f7521e0c0897d20795788d19494712185cd20360e347207b7ee7cb0cdbc0",
"60/mlp.up_proj.weight/lut": "float16(64, 2):fd33b1e6d220c9c7ace40fa9e5c7c702dd30524afcf83f8252da02127fb4db3a",
"60/mlp.up_proj.weight/scale": "float16(17408,):09bd77414117c6b3bbc5fcffe084b85885702147d8c36aa7c695c81ec346c119",
"60/post_attention_layernorm.weight": "float32(5120,):7ef8b0ac8d283a3b3469a52774ba9240fdccfaef3d39567780b54f801c70acda",
"61/input_layernorm.weight": "float32(5120,):86c0d1eddbec8e870cd589a056f9dc434d237cb781a3268368b44f801b9d99e9",
"61/linear_attn.A_log": "float32(48,):d080f2651b2cf29a33056f252a92f14b6372091846e245d322e71550178ef687",
"61/linear_attn.conv1d.weight": "float32(10240, 1, 4):a2c8efb9432d0848adf796f1634bc7b8a4523e29538a1b4e5d38dc39a40648a5",
"61/linear_attn.dt_bias": "float32(48,):749f1a225bb9432a72861da467559d7bffce520f05f29dc63a7a8aedfbd4e7ba",
"61/linear_attn.in_proj_a.weight": "float32(48, 5120):ddaca7533107940c8dcecb93374f7479712f280f66492a08de68adbd6c47da8b",
"61/linear_attn.in_proj_b.weight": "float32(48, 5120):720e6ed998455cbaae12c471a6355e5a0e50b4e4772bc38e42b80a617a2fc9f4",
"61/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):5c36b9ec75548af560b8d2fb612c5d67f0891e5e1dc23a616ffad65642a9f1dc",
"61/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):40b606c98dceb5a4c62163fd3e3fc19e81f5330d1afa9a881292bb2fa59e99ff",
"61/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):0df1c8748613f0a261cbc611bba2781812802d63223dc9b13764ccc204bcf7af",
"61/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):45d1f3a7860c634472fcc13ff47a23cf1b76af2441999cb811dfbff0fdd85988",
"61/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):cc30130cff99fc696f7da4fec2881b819e576785e5d700c3eb97851bccea8d34",
"61/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):4e2911d61873ed0e98a7b58cb1f5fbb77788cc4464f0dbda49b627a667fa2e87",
"61/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):9fbbd919bc0eaf9c648c615b1a1e25d1a45e4f46a3e7e0299f92e0b3958123b9",
"61/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):a20a326d37b6bcea19157d8f415e3a425bdb4441b4afd83ffb59f3f572c245dd",
"61/linear_attn.in_proj_z.weight/lut": "float16(16, 1):38bec36fcdf13c401348207e816d4446f88dbfe07447e2a3145f58a1946fbdec",
"61/linear_attn.in_proj_z.weight/scale": "float16(6144,):19b84d72c73ab875f35f12c320463d1639288f1fe508aa0e1043d9219d039992",
"61/linear_attn.norm.weight": "float32(128,):3f169623f54fc7c15d14201a79f9b9c34344749db3771961678cc820fd3313fa",
"61/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):4952f600fc2f9072f222f83a9d59b71aace41bbd937bd7dfe1a4712960732115",
"61/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):14c900aa86ef96335164352a0d7fba910cf1a2c60cee80be4e97105de68fd33d",
"61/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):72b1fe0eb4e6cfee76457267313c0ec68db325ee707b05147ea835e64a5fa85f",
"61/linear_attn.out_proj.weight/lut": "float16(16, 1):b39da6becda20ee388ffde2fef43f8c4c1d92e6629b4039e5c5aabafac045397",
"61/linear_attn.out_proj.weight/scale": "float16(5120,):13c3687402cf8b85cd7bfd81bf27c2349e798854add2bab81b4e91682c5d99a3",
"61/mix.rotation": "int64(2,):9f311a8de90701eaa59d1f6f8d19c977ae262dd10d63f0a2ac21ad1681e49a3e",
"61/mlp.down_proj.weight/idx": "uint8(2560, 17408):616bf303dcdb136914c0d8e330396d28d13982464920f11bba78fa0eefb98cb4",
"61/mlp.down_proj.weight/lut": "float16(64, 2):d6fa4ac6fd0f15d60702ca53f329c19861f00bb923691ccaa89684cd308efd29",
"61/mlp.down_proj.weight/scale": "float16(5120,):d58e874de4885e3a1c8288e4c80098c8bfacc05a500107c9c6ec358c22530fdc",
"61/mlp.gate_proj.weight/idx": "uint8(8704, 5120):5bf03f0efe42aa48ed061160cdaef8b6861e91ecc93e4096404a119e7e4b37c5",
"61/mlp.gate_proj.weight/lut": "float16(64, 2):87202754ebc9e529efbf89ebd8049a523dd20729ab631337e68b902243609f7b",
"61/mlp.gate_proj.weight/scale": "float16(17408,):b9e7509b9fc055b92d6706fc82a7a6ed5ada5f248d938a18ef9fe1cdb014a395",
"61/mlp.rotation": "int64(2,):9a008bc094a74c3f26519ec59d98cd643b14da42c9e77bfaa965abc1d312556b",
"61/mlp.up_proj.weight/idx": "uint8(8704, 5120):2f1d9ee873f1073bc78e3e0709ffdfb775d999bc5ed58f2c5d513394fef9c06d",
"61/mlp.up_proj.weight/lut": "float16(64, 2):bd852242eb7df859d8101ec53fa373fdd58e5cd21c43bb03df08925a9cdf9562",
"61/mlp.up_proj.weight/scale": "float16(17408,):9952257e7178c626cc53164bc3fd85f7dc834bbc10ffcbf2993d2a067d3f6396",
"61/post_attention_layernorm.weight": "float32(5120,):9f61bc1627fa274cbcf979c2acf52d52dbe2bb47c0c6c28ad907d3687bfae270",
"62/input_layernorm.weight": "float32(5120,):7e8f14811509997c12fbe053109ac405be3613451340d97a82dd67f4a1c4b705",
"62/linear_attn.A_log": "float32(48,):5997c4f1ae205b583a7620c9bbed3d84189c28cca0c1783577d823b42302958c",
"62/linear_attn.conv1d.weight": "float32(10240, 1, 4):17a28b357ba81d21b50a2b5f78816e3e685d5a81d269978774acdd6305fcef8e",
"62/linear_attn.dt_bias": "float32(48,):66db6c88916c0a7a88de1cb1860a3a4dfcc135e3274fc84129efeb3c936492c3",
"62/linear_attn.in_proj_a.weight": "float32(48, 5120):2fe2c7a83b55ace900844311c9f59c4374eeaa6550074aaee2459f7abbc91bf7",
"62/linear_attn.in_proj_b.weight": "float32(48, 5120):b1e66de4cb0141ece71cb99fbf9cc3d54a268547b7c7cf5e196c4cca13e057ca",
"62/linear_attn.in_proj_qkv.weight/idx": "uint8(10240, 5120):93cc26ac22785000995f7d83d2a86cb09a996214519953f4bbbb3a5d54f28ca7",
"62/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):bfdba963bea108435f2b39abebfbe60535008cc06e492cbf25f5cede79611af9",
"62/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):f1c92d1a99de7ef92b1862cca01a815252438bfb4710a5b3190eb7aaa875dedf",
"62/linear_attn.in_proj_qkv.weight/lut": "float16(16, 1):4830800e738501495fcc65bdd586647de07c6cf48c8d84b31addd9b400284587",
"62/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):2c9ae4dd8a310f823ece16f0a28be4a6d7e06c8eadf10b21f397e2a98ada3399",
"62/linear_attn.in_proj_z.weight/idx": "uint8(6144, 5120):385f9e2b051f8e8ed996228f97ca2ed4e748ab69a5b5a53c0d6ea43d3796851b",
"62/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):0afe62149f1d13cb6bd426ba70d3be60dafc0b293415c57542af2feba2b2462e",
"62/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):a3871a5fb1dc2d8345b3e7fb3a1c8e74172823dd902f6def0f882ef67fc44240",
"62/linear_attn.in_proj_z.weight/lut": "float16(16, 1):0ecfd084616b8df7ac4ca06438616b711ca4a544e8f635c77786cb5ccd717151",
"62/linear_attn.in_proj_z.weight/scale": "float16(6144,):e556cf270702df061b34454c6b31788cdb7279391792df0c9c2f21836fe3ff56",
"62/linear_attn.norm.weight": "float32(128,):e0d98987760867713536f1c1b393b3204dbda5b12138e617bc782115d3e20cc7",
"62/linear_attn.out_proj.weight/idx": "uint8(5120, 6144):be26084cdba9afe82d8196164b86d85289e37641d5e277b198f4779ecd925e42",
"62/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):c3ea539517afe83506e542c6e736045653a647a14ac9172419224e1c13f174fa",
"62/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):bcc8060f13a3d738e0aed3b58e371631ddb1f75d0a5f1cd1d96ecee41dbb876c",
"62/linear_attn.out_proj.weight/lut": "float16(16, 1):98dc9151602794bfc8ecfcf539a373024a15acb2a5f5c785f4a76cfe362d1f7a",
"62/linear_attn.out_proj.weight/scale": "float16(5120,):8d6376a03c6b16b2dcf293a12a18cf854a87d2ea9af92129962bee16f9a90fa2",
"62/mix.rotation": "int64(2,):bf0944c1cd3105501cc8d6f8b27e14511ea5e0a1ed12796e8ef263c8906382d8",
"62/mlp.down_proj.weight/idx": "uint8(2560, 17408):39c1fe5e0e57001c7327674ef8274942ed234b35657d367cf4b32940a26ffb9c",
"62/mlp.down_proj.weight/lut": "float16(64, 2):23b9b5a2f04f58bdc734e5e6e7c863da81ebc42e41066f218a05ea03b4c4ab0c",
"62/mlp.down_proj.weight/scale": "float16(5120,):b7619df731296fe3d37ef9838abb8c4b40dc871a3959c6aad1d0dd9f777de10e",
"62/mlp.gate_proj.weight/idx": "uint8(8704, 5120):2e33cb27348d791d84303a9e4c8abced90873a753cc3a732c089d924292630cc",
"62/mlp.gate_proj.weight/lut": "float16(64, 2):8e23487ebededd79986b9fcddd6576ecfa94b4812c4bdfce645ffac9c6ad16df",
"62/mlp.gate_proj.weight/scale": "float16(17408,):80a4c8f7ba2ea88781dd7fb0e7782e241d793f8f3656b80ecfb01af608dc2336",
"62/mlp.rotation": "int64(2,):2d47b26d8b24496101a9b4ba9d186bd6790b910d8f93f3f364e5864646921b85",
"62/mlp.up_proj.weight/idx": "uint8(8704, 5120):064d2a1d202bc54f7821c072f7e637be046810381534f28bed6e68f3e08b079c",
"62/mlp.up_proj.weight/lut": "float16(64, 2):fc6256d084efdfa0e83883fc23b058e6a8a06f5c7dc797af755cdef20a9c8a27",
"62/mlp.up_proj.weight/scale": "float16(17408,):21b6bebe00b33a249791a67d08187d10c0c76f4879f85cc4b71c9bbedd2940d9",
"62/post_attention_layernorm.weight": "float32(5120,):2a2df71c8b88b44512e1834c52fa70c77805f0201b2c622246031f1bfc4ca5a6",
"63/input_layernorm.weight": "float32(5120,):f8d416392940882bbf0351b592dc6badadcef0289f3e83ed91ed36145f0ee147",
"63/mix.rotation": "int64(2,):40e9e7705ffb5e598f531a031c8eaafb79e35d905f9dfbe2442ac36a9c711713",
"63/mlp.down_proj.weight/idx": "uint8(2560, 17408):2b34647fddd6dd19e65869eeb03e74878829fd9cb1b53d344848c53c8322ff81",
"63/mlp.down_proj.weight/lut": "float16(64, 2):766c9ca195584a6ddacd1012f4728ba67931e2c04bfd84958f584123c5da2402",
"63/mlp.down_proj.weight/scale": "float16(5120,):639a91957dba796168738e7d23103740c4da1f47038726042a93bfe93b5b4816",
"63/mlp.gate_proj.weight/idx": "uint8(8704, 5120):6b7f69e8cf1e1353d512798f0a0c09ad02438439ed84e0328d5d46b8b912664e",
"63/mlp.gate_proj.weight/lut": "float16(64, 2):95963fdafe03d2843000018031701822e65e5bd1b10dd89f98fbdeaa2fbc508e",
"63/mlp.gate_proj.weight/scale": "float16(17408,):a441aae67df213a779c792d3265864df46a4044838c96d7d58e1e787563e4999",
"63/mlp.rotation": "int64(2,):5b36460602608d07bea49f2c7df493432bb4c84361529fcf7feea57ff1e2f90c",
"63/mlp.up_proj.weight/idx": "uint8(8704, 5120):c78e887d114d5d79570dbb7a0cb971494b22f8178eb9172da4a72a25e69832b9",
"63/mlp.up_proj.weight/lut": "float16(64, 2):cd20288a3c028f342f0b2daa83300349168185a1024c021f3d0745f3fc08917d",
"63/mlp.up_proj.weight/scale": "float16(17408,):3baa2607780b99d19bed8216bf06acecac84c8eded7ccb7bf490a47bb79c5b56",
"63/post_attention_layernorm.weight": "float32(5120,):0226cd0246a3115daf8339cbefbf5d3f3bb4e38fac7f63df8f97dd730ccb4ff1",
"63/self_attn.k_norm.weight": "float32(256,):95b70e85bb9960fe1c683e3f93799d68858d8f596977db1223275a487da880e1",
"63/self_attn.k_proj.weight/int8": "int8(1024, 5120):d49ea95fa56af5859a08e9d3acde6f9d837e850ba2aa907494e3530ecba52c26",
"63/self_attn.k_proj.weight/lr_a": "float16(1024, 64):c70ab42ce525a088e0dec204bd69c3172c9e279c7533e6164db6820cbf597869",
"63/self_attn.k_proj.weight/lr_b": "float16(64, 5120):dfb3836985f4925d531534e3faf1fc091085251cde8bbabbf66feb6e88a03809",
"63/self_attn.k_proj.weight/scale": "float16(1024,):58c409914fb32eb07ab155942e6c9b73a49916d192be8dc46872da0f37ce8c8d",
"63/self_attn.o_proj.weight/idx": "uint8(5120, 6144):9df9de8e861e3b4cd70985a3b8503941b9354ac316cd661dbcc251accfb3d06d",
"63/self_attn.o_proj.weight/lr_a": "float16(5120, 64):4c1dcd6885ea877b0de7dbd6087e1072b41628fc445b706f83d569fb757fba20",
"63/self_attn.o_proj.weight/lr_b": "float16(64, 6144):042b52039fad41f8d86f7b9f10aaa1a14c0c44ce46efb513a1c7229c3bdfc088",
"63/self_attn.o_proj.weight/lut": "float16(16, 1):cd57b03bbdce7406d0dc6b839e615630d9824d78e7fddfeabbe286e9a2bed3ca",
"63/self_attn.o_proj.weight/scale": "float16(5120,):f6a4d4d073afbc07fcf549532cf704da0a0a9fed88bac7bbd8eb209592fccf9b",
"63/self_attn.q_norm.weight": "float32(256,):1377d34a008d2038fbfee97ce4be9ae5ab7f4110de40b80f54cc431c5a374c0e",
"63/self_attn.q_proj.weight/idx": "uint8(12288, 5120):279d4a186d72d21372fddb69ff0a037b3373121f4be9a2d040f80e7ff85c857c",
"63/self_attn.q_proj.weight/lr_a": "float16(12288, 64):ab09acb509a52cf30fb20de6f8d6bfb10fd2f53efa55fe7de43ac4867533bbfb",
"63/self_attn.q_proj.weight/lr_b": "float16(64, 5120):ddae6990ed3600ad69cea705d885f8ad6a8215d1ac291c45a0bfcffcc5a91672",
"63/self_attn.q_proj.weight/lut": "float16(16, 1):1d982c9026fb072449f1558cf8015b7654df812a318c0f68ccf262070cbe6209",
"63/self_attn.q_proj.weight/scale": "float16(12288,):69ad0687e3489a540e1bfcffb524fc3db8d1eb66bdb7cb120a73a6460a00bd73",
"63/self_attn.v_proj.weight/int8": "int8(1024, 5120):df7589ce6560da2522a91a3cc2013ed57b214ba0651f5868b69e616eaef2f282",
"63/self_attn.v_proj.weight/lr_a": "float16(1024, 64):561232eb9d67a4ad23392ff9f7a6d38e4c421327482b9f70999dbc797cb83e3a",
"63/self_attn.v_proj.weight/lr_b": "float16(64, 5120):f04b6c00301ba2e9b4b1475207ed495fcbe24732e90dc587cbb6e1e3e98cd9d9",
"63/self_attn.v_proj.weight/scale": "float16(1024,):8a1ca59598ecb0515e3bb65509fe5bf7b5b3509c053750e8439a0f30bcb656d6",
"7/input_layernorm.weight": "float32(5120,):6bf28cfe70e319b95dc1d2a2acf15eef61c187edecc1a7e986ecca89dad5f423",
"7/mix.rotation": "int64(2,):c45fc6f3c28388d13af6590cf79c4ad40870b1a57cc7e8902125ec520efc0de7",
"7/mlp.down_proj.weight/idx": "uint8(2560, 17408):aa914f5d013d16055c1477242f018987ddb9c737c8bdd253bf8c5fe001b03c50",
"7/mlp.down_proj.weight/lut": "float16(64, 2):6b4ba618bd03b5e0084139f78bbb8c18b3702e6a27aedcad6ad7c137709f3362",
"7/mlp.down_proj.weight/scale": "float16(5120,):d450d8ca0c42c87c48ca493df14d1397d22e1e3173891f742a59023788693ab0",
"7/mlp.gate_proj.weight/idx": "uint8(8704, 5120):2d8e0f23449397d316e4c5ba7ecb9f28a44ed52fd1389e18c8671ea2307f04b4",
"7/mlp.gate_proj.weight/lut": "float16(64, 2):a2858695284daa544896c35f80100deb4a7bcf1ffe393fb50298e24219a08e32",
"7/mlp.gate_proj.weight/scale": "float16(17408,):a6098b2f6067b0a3e1f7f33ef0cc1f8b7f1f3d3510afbac3eacbaa6c4711f385",
"7/mlp.rotation": "int64(2,):31025c622f7aa5c196ff9cc90b622d13554b859b44b7978d6b5e08c0c82fda6c",
"7/mlp.up_proj.weight/idx": "uint8(8704, 5120):f3fab6af856d68c383e7ebfdc73c451bf8cfcb0cb6a6f32144f7a9e8921034c2",
"7/mlp.up_proj.weight/lut": "float16(64, 2):e5929ce7ab7f16af2b9e04f473e065e645506912657ed80bfe8636c179e9130b",
"7/mlp.up_proj.weight/scale": "float16(17408,):ae42d3789039a3c5d76ed02d293fcb1182075031236676b2e5582dc4ecca9f02",
"7/post_attention_layernorm.weight": "float32(5120,):dc03e0bba9d2892efcf27cd5e4550b2238cb5dde9baeaf18f7a26d57909baa8e",
"7/self_attn.k_norm.weight": "float32(256,):500a3116ea947860786f2198daea1f755276519e37de110fe5182acd894f7a1e",
"7/self_attn.k_proj.weight/int8": "int8(1024, 5120):98e149ed55b05336839016acfaad0e86047d50d1b6b7567eb2e4f7b64d3d016c",
"7/self_attn.k_proj.weight/lr_a": "float16(1024, 64):afb6674ccef08938d980f826f9dd422cc4a17867b9abcaef1cec02370299c87b",
"7/self_attn.k_proj.weight/lr_b": "float16(64, 5120):c65c674ff217cfe95f37196304f7b4caae63ea99e7e015ed1201d9d650207799",
"7/self_attn.k_proj.weight/scale": "float16(1024,):50d18e96aa760098d5f0000046cd87cd236522dae0ffa46ae11a1aae6f10dcd8",
"7/self_attn.o_proj.weight/idx": "uint8(2560, 6144):84165efc6f0e793240d14653382edfe707c3105dfe6cc3fb2cb04c811d387597",
"7/self_attn.o_proj.weight/lr_a": "float16(5120, 64):4f33c7156716fe38ac21f188572e6d71c992c08db714880de9b3293287b90924",
"7/self_attn.o_proj.weight/lr_b": "float16(64, 6144):1ecfc216667ff93f81b9007b164e77ffd244f96d479acd53d0f69bf79b80d148",
"7/self_attn.o_proj.weight/lut": "float16(16, 2):41ac83e2887f97c428bce68cb94efff1baf769f08f3e3edd17b151e78dd73ac8",
"7/self_attn.o_proj.weight/scale": "float16(5120,):aba864174318543b46efd806bb2e2eb82f413a92b3b3f36b53e4db9950edc68c",
"7/self_attn.q_norm.weight": "float32(256,):adfd1d32aa1f452af11f2fc8369b8a794d9ca7a2793287bcdd1d4223b65e835d",
"7/self_attn.q_proj.weight/idx": "uint8(6144, 5120):0d4e1d56c68be7e931a70a4bf57c67268f58fb8b301544cd1003bd7d401f4976",
"7/self_attn.q_proj.weight/lr_a": "float16(12288, 64):570cbcca5119cc2b2856bcdd597d4cdc2a0828861426497aad65078e45666eca",
"7/self_attn.q_proj.weight/lr_b": "float16(64, 5120):46485942155e49288f7944f1d407142f3f83d5310e58449840f1ce6004553e96",
"7/self_attn.q_proj.weight/lut": "float16(16, 2):4aaf3ded524279350a1ab9a3b4bccded6d37594558d001abc9e92fa2ed578a33",
"7/self_attn.q_proj.weight/scale": "float16(12288,):1b4f85fabb83bb7d2f490bbd1e2997084d75fd5e8f95161f0252b93e02bc6245",
"7/self_attn.v_proj.weight/int8": "int8(1024, 5120):4b76bf61f4d9ba3f103320ff0d9b5706fd786cc20ccae2d1d27f4242830b2b92",
"7/self_attn.v_proj.weight/lr_a": "float16(1024, 64):997dab221dac9fa92b255aee9e18d7cc6f73af4223cf0fe9ca30fb5ecc2f1a9d",
"7/self_attn.v_proj.weight/lr_b": "float16(64, 5120):bc956e3fd3ebf85f33405962cde1cac9517b2c3b72d6516c2c491f71273f73fb",
"7/self_attn.v_proj.weight/scale": "float16(1024,):db6c5bd32174d9aefd3a603d51f943ceecfb67868b3242d5e8ba5d0eb1ba8098",
"8/input_layernorm.weight": "float32(5120,):5992bdbb0bbc81483eadd2767b4141772acb11e1720f300b381f0b06e052f96e",
"8/linear_attn.A_log": "float32(48,):db4d1945caf7cc3ea7a74a784c1032e2d6a9e13c1963f0d9faf7cb2240e3d2b7",
"8/linear_attn.conv1d.weight": "float32(10240, 1, 4):0f080e9e13dc087d50345c85d907d916480e538cb20805259f889bb161e4d70d",
"8/linear_attn.dt_bias": "float32(48,):a54d79528f2cb1d21f57e7499b0dd05862ae81171807146fff71cc6b17f885fd",
"8/linear_attn.in_proj_a.weight": "float32(48, 5120):5850abec49c08242529f12efe600b01db36dc6c4cd579bc77d32b5a57c147112",
"8/linear_attn.in_proj_b.weight": "float32(48, 5120):6746a4aee04a2a0ba26f3d347d5c00500c551b7d61fd330602e0fa40e29567ce",
"8/linear_attn.in_proj_qkv.weight/idx": "uint8(5120, 5120):a50a7856b4445365819a897deabc75329638e2c4d7cdcefaed5e9e81f79b0802",
"8/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):dc610c0b205247a6a2244f36682d67a8ea69d819c3db1f985f7d47365ad69178",
"8/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):b05281b840d59b8c8b6518adecd6b4cfb7798ae3aac2845ab643853b6f8e947b",
"8/linear_attn.in_proj_qkv.weight/lut": "float16(16, 2):74f74a2b53b85deacda651502220dc3abed266d961d5b887676a095e0add5428",
"8/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):16223a67fed03a1d52e2e841570872fd0703895e5dd0d1d016861825066899bb",
"8/linear_attn.in_proj_z.weight/idx": "uint8(3072, 5120):7ad33aec1f05745acffc0293ad197265e0dc3d7fa0d9f8ac8590b37a39c889ac",
"8/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):4b0e57062d5e10af58e2948d4ab62b74e6e66ecfa9a5eb6d4aea43e9aba5f1bf",
"8/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):fe682523e5064c3b832170f6ccca12481b102ebe829bdb60437dc476e290bf22",
"8/linear_attn.in_proj_z.weight/lut": "float16(16, 2):6a876981056b18f53e6d82bfbd87d5365d6d09fca36174b491d7e5372876be37",
"8/linear_attn.in_proj_z.weight/scale": "float16(6144,):cf73e482e9d4a7b7c027bf2e8de20156b3a36671a9db2d64638aa2ef585bfa0f",
"8/linear_attn.norm.weight": "float32(128,):a1ec028c0c8e6b766f5d236c27ef1a16e0bd7f8ba2336ce40571d657ca7c9439",
"8/linear_attn.out_proj.weight/idx": "uint8(2560, 6144):40791a7babc4917ad84450ec34ee35eb796d77e66e4dab415feca9140bded8a7",
"8/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):16d9eb825d166c810619f54f6faa05c98d5f9c5d07d558b6f1dafa00bbab1cf5",
"8/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):4f582e8c8f6567af50ce8d4a7473e70ad0139697019577c1edddc2713dcb25b5",
"8/linear_attn.out_proj.weight/lut": "float16(16, 2):4cd293d591fa23df810af1dfb1a03f3a75eae94484a9eff2160d7ea32a8430ba",
"8/linear_attn.out_proj.weight/scale": "float16(5120,):502b6e104f5b125cfd1a0b550bbd56ffd6c365bafd4fd26f3f0c160963b53a2b",
"8/mix.rotation": "int64(2,):183ebe0f96237c3c74fba18117e26add5604223db12bd3329222f482db56993e",
"8/mlp.down_proj.weight/idx": "uint8(2560, 17408):14aa8a9b214375bb8a2aac23cc01d4e33e6ec96cde7b4ba2403728eed9216384",
"8/mlp.down_proj.weight/lut": "float16(64, 2):d9ba0d105950f4764efc0191f1393c0adcf721c8803d081becbc20c0401bbdca",
"8/mlp.down_proj.weight/scale": "float16(5120,):16e506be14c2833336f643b56e4733449c3cadfd699afb3e76c5c9b2bca4c1dc",
"8/mlp.gate_proj.weight/idx": "uint8(8704, 5120):eded686fe99c0940e2b56ed481fb7c80a5d5d336ade334dc69bfbc5051c4d024",
"8/mlp.gate_proj.weight/lut": "float16(64, 2):a2055874815cd914c7d80c5f92312ae498b8ea2fa112e5522787761e0479ee70",
"8/mlp.gate_proj.weight/scale": "float16(17408,):36157378ced77687c6d67e1a66fdbf17b1ba43699ca6aefdea72c4f3014e086d",
"8/mlp.rotation": "int64(2,):8a6a584ca65ffe74490a1852365fe1aace7f34c35ff8d01da8ac353ef0d71756",
"8/mlp.up_proj.weight/idx": "uint8(8704, 5120):b28b0143d0ee396d6f61e51cd7ffb05548da21d3d12e03fd138be0a6b3f7a2db",
"8/mlp.up_proj.weight/lut": "float16(64, 2):1378e66678b923060fea04885ed8d30902fe1479afb847195682cb3928ce3092",
"8/mlp.up_proj.weight/scale": "float16(17408,):5944b67ddd63b71d614cff684ce69aab1bebb3cf77137117472b5fe2e85d494e",
"8/post_attention_layernorm.weight": "float32(5120,):85bb645389a6e62bacbf34cb64b530469b0f1e28e471e2e488b4ad868f20cc40",
"9/input_layernorm.weight": "float32(5120,):911826d9f28fce283ff800f27bb487ef46f367db54c2103975e6b924823d02f3",
"9/linear_attn.A_log": "float32(48,):df08bde08aeba35ac06d013d5600df62b18def76a96798d75ce40264316f80cc",
"9/linear_attn.conv1d.weight": "float32(10240, 1, 4):3e9d3bb50d56e3504aa056dc0f3e684759770af4d4ff65bfc9f36568b779c657",
"9/linear_attn.dt_bias": "float32(48,):8142ce5c2fd0798106d77db6408a5bb7e92be8ed1d6bb26487f5eeb13673879d",
"9/linear_attn.in_proj_a.weight": "float32(48, 5120):ebb7825dfd57ffbcc02715cbc49ae3a967672ab0663c78c28d196ee30b6080db",
"9/linear_attn.in_proj_b.weight": "float32(48, 5120):9a3142a91d0ac326ea758d51b96d1b5da5c5d729aa43715f29bb2e5031319eda",
"9/linear_attn.in_proj_qkv.weight/idx": "uint8(5120, 5120):3fac4915597bf94f710acfbf3de6fcf7c44d3c26f31e8f3e07d2a49ed3d1997d",
"9/linear_attn.in_proj_qkv.weight/lr_a": "float16(10240, 64):f44b5189edde43445b04e1760116f69c409ca348ac718cfcb1a7422d5e663c2d",
"9/linear_attn.in_proj_qkv.weight/lr_b": "float16(64, 5120):8130df185d5b348dcba1a8aed238f028216a78b056503d41a5ae51f37548451c",
"9/linear_attn.in_proj_qkv.weight/lut": "float16(16, 2):1780e345045f2cc6c7393b4dfeb9f12f4750b641f768dacbd868e65c088b2f15",
"9/linear_attn.in_proj_qkv.weight/scale": "float16(10240,):162739b5fff672fe36dd5bf23b0e9ace981a5d03ed8194019d19e3bc91a18ccf",
"9/linear_attn.in_proj_z.weight/idx": "uint8(3072, 5120):cccc8ae52de43ccc69e6c624d5592d0ea8fdd076d7bec8c3f43cdc8f5d90bac6",
"9/linear_attn.in_proj_z.weight/lr_a": "float16(6144, 64):0e7182d8b2ffc9779174e817c334b66583d404d4c9acca3433ee654ee10ffa7e",
"9/linear_attn.in_proj_z.weight/lr_b": "float16(64, 5120):d85f466f66e4aa8a24f0a0691c9a53ca4e2f6af6c8162a9f55096e9989148d44",
"9/linear_attn.in_proj_z.weight/lut": "float16(16, 2):394194232b060084df62597d84674be59c75c40202e107fad892d487c9a89091",
"9/linear_attn.in_proj_z.weight/scale": "float16(6144,):aba6b3c30823855f8b81a5f5ad56c29b8cfd90a9bb4c4e9aad3e8adccb42160d",
"9/linear_attn.norm.weight": "float32(128,):f112e884424b437b13dae9325ea1c588939e61ba6c511b6e5626cb0e034d0471",
"9/linear_attn.out_proj.weight/idx": "uint8(2560, 6144):cd435cc3965f61aed7a3b3f67974b5cb78d2ada283fdd8156956ca72af81a7bb",
"9/linear_attn.out_proj.weight/lr_a": "float16(5120, 64):ebc659c3ef1bab5cff8071e535cb398631e8ba630697ab8023779e1587106547",
"9/linear_attn.out_proj.weight/lr_b": "float16(64, 6144):3b245589c89957b5f9c4ca80f490d44b06ea86163af952024b30299357437a36",
"9/linear_attn.out_proj.weight/lut": "float16(16, 2):654f0771ab406e76fd6f69b8a17ccbde70d6489d51c0f07caff72c37a4b7b104",
"9/linear_attn.out_proj.weight/scale": "float16(5120,):c805c085c27675f0f1df1cee1e678fc13c0cf26b93e4cf7a2b4a7cf146ea5db0",
"9/mix.rotation": "int64(2,):0b64019124cee154c9ff082f1f64a641903ecf47302cad6df6e60ca90476ee4b",
"9/mlp.down_proj.weight/idx": "uint8(2560, 17408):ddeae337effa2f860b370d34cb243ed70dbf22c510d55c3235c914c3e19d9b78",
"9/mlp.down_proj.weight/lut": "float16(64, 2):9d8af8b32d2c3f6a36a3914ff0c6c5ee8c96fdf49c726ab09461d26d127178b9",
"9/mlp.down_proj.weight/scale": "float16(5120,):ecd74c00c92243cd991e5ac48c31a0321bceaeace55b2db180c2d8867c674797",
"9/mlp.gate_proj.weight/idx": "uint8(8704, 5120):c7f6e4297a23d2861faf95c723893e82249e8c54986dd18c175cc1b85ee1b186",
"9/mlp.gate_proj.weight/lut": "float16(64, 2):1f90633efc0b79c6b78794c11ceb3d9726b1e5da15f2837aa8e05c7100c7c7ae",
"9/mlp.gate_proj.weight/scale": "float16(17408,):5ceb22c3ddce033a39d770b09a6f304dd85d8807eac496031a97d9b2f810d333",
"9/mlp.rotation": "int64(2,):4ae1a0485f33eab8f5e2c983fc0f1c507c932b355bde5ccd726328bd88161288",
"9/mlp.up_proj.weight/idx": "uint8(8704, 5120):a8b8f9e7dd6cdf163250a6a72144300f0e8f56bc96775973721c151b493eaf0d",
"9/mlp.up_proj.weight/lut": "float16(64, 2):b2cd98895530d58f7d19af38a0a5f939c605bf74eed5ad116944cbae13be3641",
"9/mlp.up_proj.weight/scale": "float16(17408,):8444718dffe36fdbf85e7fa280fdbd3f5927cdae97ab4a83ee2ddc8c0aa08240",
"9/post_attention_layernorm.weight": "float32(5120,):727788f55faf2933cb84b5b2d7e8d89036a13eee35479a380e50d8ec91b50546",
"norm": "226b0749e33401bd9116c7690a6b38791a3e8864403b1f03f55f6a69581ab68f"
}