{ "schema": 2, "status": "PASS", "method": "joint qkv/z AWQ smoothing plus zero-centered RMSNorm-aware BF16 a/b input-column compensation", "zero_centered_rmsnorm": true, "formula": "W_ab_compensated = W_ab_source * (1 + norm_source) / (1 + norm_awq)", "linear_attention_layers": 48, "compensated_tensor_count": 96, "touched_shards": [ "model-00001-of-00003.safetensors" ], "smoothing_scale": { "min": 0.24673694372177124, "median": 0.45125672221183777, "p99": 0.978907585144043, "max": 4.0 }, "effective_relative_l2": { "min": 0.0016395641723647714, "median": 0.0016574671026319265, "p95": 0.0016669777687638998, "max": 0.0016791383968666196 }, "tensors": [ { "name": "model.language_model.layers.0.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "ff9e6d0a05fc85c2a61e5f73200d6778c84d76df3df2ce6c80e1e9a4c99c0ce1", "compensated_sha256": "3d4189ae9528e2aee43792fbca9ff95e16c2261deca51345fe8e02fcdfe1d6ee", "effective_relative_l2": 0.0016527726547792554, "effective_max_abs": 0.000270843505859375 }, { "name": "model.language_model.layers.0.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "9694f11c0393c563dcbc8c1ba120c2eaebc36b200d6d19bf8761034acde4ac33", "compensated_sha256": "0922cc2ec95a868ba2dca469b117e3ba8e58d6abf2a3607f7216df08716dd75c", "effective_relative_l2": 0.0016629273304715753, "effective_max_abs": 0.00013256072998046875 }, { "name": "model.language_model.layers.1.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "a88f63a36240c5021a860cc07a89b2431bd134c7691e7130034837a6ce1efebd", "compensated_sha256": "735ea7c7aaf30f5e71ec15bbb007bbf65b4c3583dec1555ba70837764284dcbb", "effective_relative_l2": 0.0016579352086409926, "effective_max_abs": 0.0006999969482421875 }, { "name": "model.language_model.layers.1.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "4347a0a2072b375862a8661178e7177b51b8f503656e00d835ffe76bcff13713", "compensated_sha256": "b9227fb5873bde8c22f126282ed2b6cf1b9a6d7a28d09c2c2c9d4f5d4e20a867", "effective_relative_l2": 0.0016556739574298263, "effective_max_abs": 0.0007171630859375 }, { "name": "model.language_model.layers.2.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "0404ee88abb353c7af96e0ba7f053ed2aa235aed4c964cc1198810fc934a9a50", "compensated_sha256": "13619ec2ec778feba63e57ee973a2d9d5385d567590bfe803e4f6a431f230d9e", "effective_relative_l2": 0.001666152966208756, "effective_max_abs": 0.0007781982421875 }, { "name": "model.language_model.layers.2.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "7661b6df75565169daa1fa083e768e8b54bee8b041005868d009416247e8aff6", "compensated_sha256": "5230ee710a3e4058730f5a27a416af503100bac1d6440a6d036920d6e47e0d1e", "effective_relative_l2": 0.0016654946375638247, "effective_max_abs": 0.0008983612060546875 }, { "name": "model.language_model.layers.4.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "5bd456397980252e77eca90d6972236812e77fef153375b2f1d61fe12936c95a", "compensated_sha256": "9fe2db950e54e04b0395cb32cbb51ad37800a375c68b897e6373eef79ac990e1", "effective_relative_l2": 0.0016495444579049945, "effective_max_abs": 0.0007572174072265625 }, { "name": "model.language_model.layers.4.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "e1d8a7f55cafc92896ce201c096d21832cd1cac88c1be75980598f5d78ab82c0", "compensated_sha256": "e311748b209a5660ce9c940281213a54d96ede707e88954a64ba71534aa02c9a", "effective_relative_l2": 0.0016652691410854459, "effective_max_abs": 0.0008831024169921875 }, { "name": "model.language_model.layers.5.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "b0a7e343a52c783be5bf51e888fbe10f0dd818812556fa35c680161005c87946", "compensated_sha256": "7a48469123cf9a07aee73042580a6efc39942da26f12a5fdbd6748bf19d83797", "effective_relative_l2": 0.0016498331679031253, "effective_max_abs": 0.0006866455078125 }, { "name": "model.language_model.layers.5.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "ff174a7556fc768d52246e82b6281145fb1e1c290d0cee7c8ff2fab36d268d6d", "compensated_sha256": "47033adab6cbcca0592d0199387d046e93ce3b00d79dd48fc9ce089fa1437be0", "effective_relative_l2": 0.0016484289662912488, "effective_max_abs": 0.000640869140625 }, { "name": "model.language_model.layers.6.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "a63a587197f916fd23a81300ff06ab89da7553b86bae68ca7e178c213af62d6c", "compensated_sha256": "87786a88ab69bf1ae24e3b2598fcab76e84444e016a75976d44c115ba00ed5ee", "effective_relative_l2": 0.0016593246255069971, "effective_max_abs": 0.00067138671875 }, { "name": "model.language_model.layers.6.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "f49a069d9a3f3ac3a9f264c0c82a63467b728954a76c71bfe697f52c71e53a1a", "compensated_sha256": "9ff72cbc1df689fc15d6869945ab146968804a39d73ef06e3e638d818fa5cd76", "effective_relative_l2": 0.0016482515493407845, "effective_max_abs": 0.00063323974609375 }, { "name": "model.language_model.layers.8.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "e765660f9b905ede2a752fc05019cfee1a8cc80fd07fccac8515981934c996c8", "compensated_sha256": "bb1b11845baa8a53c287ff835a3a301571ea785141d357ce2a9fc8b47bdde435", "effective_relative_l2": 0.0016640282701700926, "effective_max_abs": 0.000804901123046875 }, { "name": "model.language_model.layers.8.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "e8327e7163c1a2c9db724430e60f9f886cfe09b4364cdfe7dcb34e323f89eba9", "compensated_sha256": "69cd79f9ef07632705646f721400f24dedfa6db3b9f0176389e76e7d46ff8e10", "effective_relative_l2": 0.0016489507397636771, "effective_max_abs": 0.0006256103515625 }, { "name": "model.language_model.layers.9.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "038f9fcb34d6e8b32945bb6a241a89ebc9e2b7764bc5d35fc5c13d39bbf8d56a", "compensated_sha256": "f3ca509abcda3fa3d35def929e83eeaa7c4fb5a4432ce8f6165ac4734a04240d", "effective_relative_l2": 0.0016600192757323384, "effective_max_abs": 0.00064849853515625 }, { "name": "model.language_model.layers.9.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "f944ab1a5a90487248d61b531edabeb9078788237c5f1180f6234509bee3c8c8", "compensated_sha256": "82ffcef39fc91b6df56eff60514ed774b8980262931d7d0e220d499cc0150807", "effective_relative_l2": 0.0016664986032992601, "effective_max_abs": 0.000782012939453125 }, { "name": "model.language_model.layers.10.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "2d06a15ec0205eca194bbf86b0c44ff5fb7f02ac76cd61df9c0170aaae479532", "compensated_sha256": "8f243756bce0c5d2818ff09d90677ed570281e840fc4455ee1e7c34c5216db6a", "effective_relative_l2": 0.0016589862061664462, "effective_max_abs": 0.0005507469177246094 }, { "name": "model.language_model.layers.10.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "4fc50f06d27bc7ef690540912d79442257da6041fc73ac9c69577db8c8c21c9a", "compensated_sha256": "0cdbf21bee289c948efae8aefc357e7331c9dd6a7ee1862c2ab89fc3692ad808", "effective_relative_l2": 0.0016565158730372787, "effective_max_abs": 0.0006608963012695312 }, { "name": "model.language_model.layers.12.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "4935b986d70470e1b82f7290606eae4c6b240576498ca42f354b1cd208337f9c", "compensated_sha256": "0e0d1c47c445ced243cd9a1e1f00acddc54d123a63fb03e68f944eb68e400de9", "effective_relative_l2": 0.0016622868133708835, "effective_max_abs": 0.0008182525634765625 }, { "name": "model.language_model.layers.12.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "113c8788e67c5fa7e7590bad137273aaa85ffa6bafa575ea564ddca3546ae17e", "compensated_sha256": "a406a30b08a6ba0c631312389771d6fb95fe2468e384614a64824415fa3d26e0", "effective_relative_l2": 0.0016659669345244765, "effective_max_abs": 0.0008678436279296875 }, { "name": "model.language_model.layers.13.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "4ac682e81aa1e16ef2a91b6f687fc410aefc15e2617dbbac6c1b77b94b4b9186", "compensated_sha256": "36ecd0828b244279f33118893b90ee12940170441d4fd289bf474509ea14437a", "effective_relative_l2": 0.0016534066526219249, "effective_max_abs": 0.0006771087646484375 }, { "name": "model.language_model.layers.13.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "f0cd3ea1add7d352681f7d24d871b3524ef2c45f1749753719fb5453563080e5", "compensated_sha256": "56d291ab15efb36dfbed01fd1a89ff739a59c18cdbf64846f8d8f880a6931c55", "effective_relative_l2": 0.001658450928516686, "effective_max_abs": 0.000701904296875 }, { "name": "model.language_model.layers.14.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "778441693a16f933285b3a5810cebcac745c52fa7fe7f0bba2e1ad4bf7f8f0be", "compensated_sha256": "8ffb9a692cd0f351e706f1ddf9007da57beb751ed9bde514f56d501fa40b8728", "effective_relative_l2": 0.0016591643216088414, "effective_max_abs": 0.00049591064453125 }, { "name": "model.language_model.layers.14.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "5f36b29a94e80ff25c8eccf59ecf572e3e578273728d499a63b1e7cee1780532", "compensated_sha256": "0d4a03cb0778cd4aa788ee3700619e6fe6f820819318bd24a47e7b8643943c19", "effective_relative_l2": 0.0016573481261730194, "effective_max_abs": 0.000762939453125 }, { "name": "model.language_model.layers.16.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "ed9df994841fb05179a24309c553e30ed09195985425466b687735292be3d8be", "compensated_sha256": "775208a3f3710c000951bc8a8916c14be9e3870260a342cac8195b1d1a0ad6fd", "effective_relative_l2": 0.0016599814407527447, "effective_max_abs": 0.0007162094116210938 }, { "name": "model.language_model.layers.16.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "3fc86b38160b6cde9eea4c2c8fb47e81a006dd330a827d4b3f67c6f136e611da", "compensated_sha256": "f5247c3a2726805628f8a164bb57da021f220b495274ce712e1e8bc24b9f3dd6", "effective_relative_l2": 0.0016585588455200195, "effective_max_abs": 0.000476837158203125 }, { "name": "model.language_model.layers.17.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "c91a5a84dfb1688fc9cdba1f214dc5c7de892622019274b6c620cbb19c62d72e", "compensated_sha256": "c26bd92cf3ce5d40a18f14fa32cf301703beb55cfb6a1e830653967dc84f06c3", "effective_relative_l2": 0.00165650830604136, "effective_max_abs": 0.0006809234619140625 }, { "name": "model.language_model.layers.17.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "9e2860e988b8f01f6dc6ab19ec048b0ed70640a23899c5151c86b07540fa0343", "compensated_sha256": "86f765e054d3b37dd03a9b8fc5e801c9a0d7f6295aab8a433a14f2e7837bdc3c", "effective_relative_l2": 0.0016671593766659498, "effective_max_abs": 0.000614166259765625 }, { "name": "model.language_model.layers.18.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "a0cbd4560af3137ca009998669fea66483f83ef042bc2f04ae51e6efdc36e239", "compensated_sha256": "eb446bbde724d3c410c1ac3cbde954c9003caa0549b11a64b908bf761d2c1eb3", "effective_relative_l2": 0.0016590068116784096, "effective_max_abs": 0.00045013427734375 }, { "name": "model.language_model.layers.18.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "ba0e7f2e0148d7531ca184508910b6e8fba2b8ea18f90120d0326bb89b20242c", "compensated_sha256": "990f0c25618a42c62068e0606951e0d2f14dcfba6fb39b54a6290cab2f4958ec", "effective_relative_l2": 0.001657023560255766, "effective_max_abs": 0.0007276535034179688 }, { "name": "model.language_model.layers.20.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "b8b86988aed15e54babb9902665d75143a3f44b319af16f7232d36773bc308a2", "compensated_sha256": "0754f0902e2ea7e3be416bb487543469d36034fba2dc395ac5c8a99b11e0c5eb", "effective_relative_l2": 0.0016576852649450302, "effective_max_abs": 0.00054931640625 }, { "name": "model.language_model.layers.20.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "43e268d39de515b9d24732ce98db3ea18d62aa4870f40fc5fc57397f6a7b3d5e", "compensated_sha256": "5636251e9d8519ce45fc58da38c5e8a8c3ab93167433aa71048c1e1a753bc659", "effective_relative_l2": 0.0016573143657296896, "effective_max_abs": 0.000453948974609375 }, { "name": "model.language_model.layers.21.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "3a7313ab3e50b2ff6841ab6a11d0c6e0479b333f1ddc7f536a606667f859b7fd", "compensated_sha256": "4b8723e0f2aa88d1e4b93f80a52fd7b7449a8929dc537c3071f3761a9a792054", "effective_relative_l2": 0.0016563779208809137, "effective_max_abs": 0.00046539306640625 }, { "name": "model.language_model.layers.21.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "38237df20272a609142defa62295420f97cbca7580499ddfe7c3e5852cdca67f", "compensated_sha256": "0b448c993f35342d27f807f516724f498596e331afc8603cea0ff4e5e74aefb0", "effective_relative_l2": 0.0016621362883597612, "effective_max_abs": 0.0005035400390625 }, { "name": "model.language_model.layers.22.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "9c4fb068be917e84fbadb9c3a06f92ff85ec644d9d901cd6ad3adaa061037c20", "compensated_sha256": "9f43f945337b36a4e2f16efdcac2190806661633043f8304bc9c7888f3b5c141", "effective_relative_l2": 0.0016567081911489367, "effective_max_abs": 0.00054931640625 }, { "name": "model.language_model.layers.22.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "51bbc9043e1072a661f7f2e8a2d7b82f5d9a0205b482f98752d6398afaee1dcd", "compensated_sha256": "c65e42329a4a8cfdbf7d542a7c72b3fb83249c035f3534b928d300a3207a5cf1", "effective_relative_l2": 0.0016601731767877936, "effective_max_abs": 0.000560760498046875 }, { "name": "model.language_model.layers.24.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "4e4dcc8ffe1dfee5481c6982defb34a9ce89eedbc7f185475d5ba383cf573acb", "compensated_sha256": "5a71ff34b4a51fb644a196765a3c2d10b27d8e4c81a8261249689e137cb3b67f", "effective_relative_l2": 0.0016621012473478913, "effective_max_abs": 0.0006656646728515625 }, { "name": "model.language_model.layers.24.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "428b899782d500663b5073c47a0e53182dc1ab1fb2296d937f580d02514f149e", "compensated_sha256": "ba730d89633e5b60b119abc82beb9254600ed7d288abce5d03803646f3d771d0", "effective_relative_l2": 0.0016723843291401863, "effective_max_abs": 0.000713348388671875 }, { "name": "model.language_model.layers.25.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "9f955eb55cc879cecbf2b3fbc1af1be053d5393b891540b2d976ceb4ba026891", "compensated_sha256": "54c10bc3c3a04a21bae38cb0706575ecf317c85ea632e4c415fe6b7cbda55894", "effective_relative_l2": 0.0016506081447005272, "effective_max_abs": 0.0004520416259765625 }, { "name": "model.language_model.layers.25.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "b5e38d769ebfe889a17155290c9961c55ce14413d2465b06d07fafacf5b0ca4c", "compensated_sha256": "a975ad63b21666fbcbb4015095c17a2e63854836a62783c4a3dc79889611dea3", "effective_relative_l2": 0.0016601216048002243, "effective_max_abs": 0.0005550384521484375 }, { "name": "model.language_model.layers.26.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "0526b8536ede64c97cd70122a1e2bc06313d16870b95f2b975202fcd9dca45f8", "compensated_sha256": "85b5c04335db7c3ccfc4736ad7e33ee91bdeda44d2351e078d0c38242699c652", "effective_relative_l2": 0.0016463639913126826, "effective_max_abs": 0.000408172607421875 }, { "name": "model.language_model.layers.26.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "0236acae7de6c5f47991bb0ec17da8547b75689903ecc810659105c508d79227", "compensated_sha256": "57fe6c2beb4dab54e856bcfcbde65dd0918bc632ada31ba74a4eb8922848c6a9", "effective_relative_l2": 0.001655510044656694, "effective_max_abs": 0.000522613525390625 }, { "name": "model.language_model.layers.28.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "a909aa4303c724542ff788c41fd25a5cb7e04d946c52956c89198eb27a9f1339", "compensated_sha256": "b7a4b2917e7943f38fc5e0ac7939638a984b0307036c3f6c057a6b950adc88b1", "effective_relative_l2": 0.0016550678992643952, "effective_max_abs": 0.00054931640625 }, { "name": "model.language_model.layers.28.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "34132f5ce08cf2f2e059e2eca6a879e1b08f914b1e919f12016c0eadd02377b8", "compensated_sha256": "babfafee3e214c02e2627c23a0eafb13c6eb86e9becad690b9b0668314b24cd4", "effective_relative_l2": 0.0016409943345934153, "effective_max_abs": 0.0004119873046875 }, { "name": "model.language_model.layers.29.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "c73ca5c0b1825360b17781a17932e54a9c4c5cbf0bdfb50bb8ba21f2f7db9097", "compensated_sha256": "842d72e7399685364262daa95e996f773fd6a4d7abef7f969c8c007a97a507a1", "effective_relative_l2": 0.0016648827586323023, "effective_max_abs": 0.000457763671875 }, { "name": "model.language_model.layers.29.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "bfc2eb569f764262009de616ceede437ac6bc0a5eb96b76984912f5eb9f585ec", "compensated_sha256": "e45c7f342949f899836a29189e2c5ea31cccaec6401ab0ec9abbfc0201bca505", "effective_relative_l2": 0.0016640661051496863, "effective_max_abs": 0.0008745193481445312 }, { "name": "model.language_model.layers.30.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "ab078c0ace3c05acd7c5fdfd8ded09e57ba3641d5e588e828d153e109eedf294", "compensated_sha256": "27d69076241a7ac94078bab04a12856d00b63980872b286aaa11c052a46d0c28", "effective_relative_l2": 0.0016587129794061184, "effective_max_abs": 0.00044918060302734375 }, { "name": "model.language_model.layers.30.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "0a122a423da6c57c93156bae1cd64a3c444aa3db0567f329c79ef4f2db6491a2", "compensated_sha256": "273dd4c5b8bffacc67e1bb9341c8fd4d317128eea8192f31c58538fc14b977c0", "effective_relative_l2": 0.0016516799805685878, "effective_max_abs": 0.0004978179931640625 }, { "name": "model.language_model.layers.32.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "f60de07c692273b97dea3764c1f40f1c14ec6f00147f45d76d4670d63aa6ec31", "compensated_sha256": "66835dcc6d9b1f5642e941e9243c53941c6ba165f46ab2afcc673c823a2e1690", "effective_relative_l2": 0.001655478379689157, "effective_max_abs": 0.0004620552062988281 }, { "name": "model.language_model.layers.32.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "bb13758d74e20307f561e5ac0b5b20b12b4cf24555abc172499e95501dd1be9a", "compensated_sha256": "340e551ad2e3f30e38e8320eaffa3e6a913306e86a36610b094ce08516ec8de1", "effective_relative_l2": 0.0016557917697355151, "effective_max_abs": 0.0006103515625 }, { "name": "model.language_model.layers.33.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "b02d23c3b8a496c49bf1a6f8764f1d0c0f0ff47c4c5aac585c72068642c2c49e", "compensated_sha256": "9d890ab97e65fac1ea0a716a264b80c747d4605753028b9c2dcc77ec7f5b396e", "effective_relative_l2": 0.0016603744588792324, "effective_max_abs": 0.000423431396484375 }, { "name": "model.language_model.layers.33.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "df9b695c14712c3d2c78a531c1f223124f77f540a9d1b5b13f386872f8853c6b", "compensated_sha256": "20a761aa258afab644a41b1dfeb73a4f5de3d2bfafd97269bb1799e64f97045f", "effective_relative_l2": 0.0016669172327965498, "effective_max_abs": 0.0008077621459960938 }, { "name": "model.language_model.layers.34.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "80009614ced097840783c0e145bc50527163ebcdf2351de420e9981bf5007f44", "compensated_sha256": "000b72a9982ed36db023079ae5aeddce36f7ce1765743d5ace906b9242a406ea", "effective_relative_l2": 0.001661101938225329, "effective_max_abs": 0.000446319580078125 }, { "name": "model.language_model.layers.34.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "c9d871f7067fbcd0c1a5a25c9848deeb9248062ba4ff1a5c65afabcaf8cd5ecf", "compensated_sha256": "88709588b089ba751589146dd09b55a27161448aaad7e98798d3b0309f900f7c", "effective_relative_l2": 0.0016517220064997673, "effective_max_abs": 0.0005064010620117188 }, { "name": "model.language_model.layers.36.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "da19d8cb49f146838f887e409fd2cc19ebfa0f67acdaae26197569e07b72e996", "compensated_sha256": "f43e4f9e26d4e8ee2db691158edf88481b47f7e9c2ad922ae28d21f7beb9005a", "effective_relative_l2": 0.0016524834791198373, "effective_max_abs": 0.0008678436279296875 }, { "name": "model.language_model.layers.36.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "96ba3e40b9a6f36fae9f4acde64f4dbfdfffeaa73dba2797c5b5782ddd6bca92", "compensated_sha256": "715722927138541e234a72c8811c9379de7fc2d935c337459a290dde8211f57b", "effective_relative_l2": 0.0016606975113973022, "effective_max_abs": 0.0007171630859375 }, { "name": "model.language_model.layers.37.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "3579ec881e290e9b914efb218cedf54a1aef62bf82cee454492d0dbe72adf7dd", "compensated_sha256": "b4c17b61fbf0d9c0566d9a5aaf5acefd944914a139a9b1a0a8e821c8fd8c0de2", "effective_relative_l2": 0.0016583043616265059, "effective_max_abs": 0.0005340576171875 }, { "name": "model.language_model.layers.37.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "e1748b0330f440ee372180eb8cba3b39947e118f0768b8d40f4a7f67bd9a3d75", "compensated_sha256": "61eeb433f12df4ae8b793ac2110b923535ca20f13ce0b812c85086a06d650d3c", "effective_relative_l2": 0.0016632085898891091, "effective_max_abs": 0.0004596710205078125 }, { "name": "model.language_model.layers.38.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "18b3fd19f6c4c92e548cdd23c12c7fb841e205b60ec8b62ea0ab392ead39e7c9", "compensated_sha256": "f8973b26a4dbe8e92a71c99ba66fe9bfdcfe9badf4bf337b45b3e2a3399808b7", "effective_relative_l2": 0.0016466279048472643, "effective_max_abs": 0.00091552734375 }, { "name": "model.language_model.layers.38.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "15f48cec07e085f11b3acebce468d35a2745a1f9a9492f02d23416d0e9434f0a", "compensated_sha256": "cb639afacd0f6be98f41647a3c88c2f6af03ea78293da7616e16b4a2b7a733cb", "effective_relative_l2": 0.0016614120686426759, "effective_max_abs": 0.00043487548828125 }, { "name": "model.language_model.layers.40.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "c64b1cf9bc52b296011f9d9aad7ed69baa79823d10bffb8a56971d83b1453384", "compensated_sha256": "6d5e300b8e9f918055303dc15aa33ba4c85cd15d51de12c8a3d9e356d4d7de6a", "effective_relative_l2": 0.0016576560446992517, "effective_max_abs": 0.0006237030029296875 }, { "name": "model.language_model.layers.40.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "fa951fa1c6a02e8979cf96b381a2ab893686b2c63b802eed24e812fd09b37696", "compensated_sha256": "19fbb87debb0fb4889413cf713d8cf0a7bba89c33beba2ac2eb97bd471688968", "effective_relative_l2": 0.0016551772132515907, "effective_max_abs": 0.000576019287109375 }, { "name": "model.language_model.layers.41.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "3a3660aaa311d1da1f60eacf28626bb23e00730073aab2bbe791071af8a9bdc4", "compensated_sha256": "db0850265e31362bf48346503005e9e38b80853c9bca51e434e8ae1c6e6cdcd1", "effective_relative_l2": 0.0016511942958459258, "effective_max_abs": 0.00048065185546875 }, { "name": "model.language_model.layers.41.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "2ca456353551bb9e689632b5e966d39848f88c9194a3a5d6c718bfa236d909d0", "compensated_sha256": "d41f058f23ec9d7e26b660ab5da0e3d4fc17d72788bde43288cdf1be30a0b5c5", "effective_relative_l2": 0.001652382779866457, "effective_max_abs": 0.00052642822265625 }, { "name": "model.language_model.layers.42.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "55315a0b8cf40cb69a86d18550d7a582f6536f2afd7b1ea3e538564afa715264", "compensated_sha256": "a8f1e25613bcbe6f0c90414843c4fe25c020d2dae6efb51acd6e6474b74b4e39", "effective_relative_l2": 0.0016532354056835175, "effective_max_abs": 0.0005035400390625 }, { "name": "model.language_model.layers.42.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "43fabda9f2522c485e366f4fa7cac3e7a997dab6d134acbe471e1d0c647d9ced", "compensated_sha256": "8853a9cdc62a565af69fbd754e66f4818ebcd072953a8d834fd613ed6997ebae", "effective_relative_l2": 0.0016565934056416154, "effective_max_abs": 0.0005297660827636719 }, { "name": "model.language_model.layers.44.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "dd7fdde3d44e8738929b5a66f2921a81a6012c24410f724d2abb86a1375784d8", "compensated_sha256": "3f785aa17a16e4d3deda13f3d11410118e18589286007ffe9f7ad89a671bab32", "effective_relative_l2": 0.001660814043134451, "effective_max_abs": 0.0005207061767578125 }, { "name": "model.language_model.layers.44.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "4fab02a7c181040dbcecc6d80124359f908c07e2707849d9913547808716cc2f", "compensated_sha256": "9d1f47943765abdb6e46a500472687b256b0c96e03bae27b0d35596259634d76", "effective_relative_l2": 0.0016532079316675663, "effective_max_abs": 0.00029754638671875 }, { "name": "model.language_model.layers.45.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "0e6e7cdea8c705a2b434652b87c9053e368b865459628194ddd07e125e5b121e", "compensated_sha256": "db39baba7a6e1dcd0e5647eea35e82f3c266cd8af23dd4ee3bd288b111738d86", "effective_relative_l2": 0.0016497626202180982, "effective_max_abs": 0.0004730224609375 }, { "name": "model.language_model.layers.45.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "ce45c92b29c6760d71fb6cc598ef20ffd3d29b6440caaacee6a9df2cdb200da9", "compensated_sha256": "718db6a4906232e38b6ffa8a8d29c0be3f31e2c83baa39b420444560839fe655", "effective_relative_l2": 0.0016556757036596537, "effective_max_abs": 0.0004367828369140625 }, { "name": "model.language_model.layers.46.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "093862569b1a3ec4eb3b8f2564ccf707604a1c93ab581039b8f25025e403090a", "compensated_sha256": "5a6805305af053a1ad2377ee1c843d74c32ee263081fe4b07a097b0b75ea2b94", "effective_relative_l2": 0.0016574367182329297, "effective_max_abs": 0.0005340576171875 }, { "name": "model.language_model.layers.46.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "617e384f577a1e57fb30dfe5bc8c95025e4ee18a20a9c1be132145168f3ba77a", "compensated_sha256": "e19846ebb9981a2419de621ee4b297e6b6f3356ee7699227cd230510ed22c3ee", "effective_relative_l2": 0.0016574974870309234, "effective_max_abs": 0.0004730224609375 }, { "name": "model.language_model.layers.48.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "20e859c8961e3895c849689f8f546227e39551db7c2ffd0cf2574ab4f0e9c6a4", "compensated_sha256": "760366b83087d494a1a8ba1e00168aa082bbe072e214b645cca56c7d968c9824", "effective_relative_l2": 0.0016541295917704701, "effective_max_abs": 0.000518798828125 }, { "name": "model.language_model.layers.48.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "e7e8685e0809ff95e475aacdd23d1e85b3cc0c3601626fa7da7339b39746daaf", "compensated_sha256": "dfd5db315d56570e2b2ad25db4616f86de61f0f4ff47cc10bada2a1cf7cd7470", "effective_relative_l2": 0.0016606285935267806, "effective_max_abs": 0.00032520294189453125 }, { "name": "model.language_model.layers.49.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "415328fd36da7dcfd45ba9079e6f002c269889d7744d7fddefc9f25667578d66", "compensated_sha256": "78e29c3da1bb87eebd61acd7f2699db0e7afe50f286a37247b24b59a19de31ab", "effective_relative_l2": 0.0016595997149124742, "effective_max_abs": 0.000553131103515625 }, { "name": "model.language_model.layers.49.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "7939674bac5b10033aab38256799c914862af99befe0ef010ba02af4e7d152b7", "compensated_sha256": "cddabd46958b47dd833d10722b0a50ff0de719bc6594c1157e98a03ffa60810f", "effective_relative_l2": 0.0016558790812268853, "effective_max_abs": 0.0003223419189453125 }, { "name": "model.language_model.layers.50.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "58abaf84701b2770fe42b85e6026d890fedd8171d49f580536ae7405372815f9", "compensated_sha256": "d73085bcc6eb95e4028aaeacc7718a29b6cc26179753c0873717c6dc9bd86752", "effective_relative_l2": 0.0016587554709985852, "effective_max_abs": 0.0005550384521484375 }, { "name": "model.language_model.layers.50.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "bfff4647297905777800210a59f11b9e60bd5ed8f6c6dc0b76b977d1b9397e64", "compensated_sha256": "43130013d6e91ee9905d0d5d14c06525ee561bc52fb85f8f7939a6b3e770bfaf", "effective_relative_l2": 0.0016569271683692932, "effective_max_abs": 0.0003485679626464844 }, { "name": "model.language_model.layers.52.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "835d231c1301853a28d9433df451ee50c29f3ba13f7fbf864cda0d41186c45d3", "compensated_sha256": "3d948d52c7f33c37395f6da612e3dc26d8ea0577e77eeede4dcb61a616555ce9", "effective_relative_l2": 0.0016395641723647714, "effective_max_abs": 0.00091552734375 }, { "name": "model.language_model.layers.52.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "27b55a5632583de237e39fa4a8c3be5c6031409185772737d00cb66d5746c0fb", "compensated_sha256": "bef7f89e89f5fbdbc0a6c83249206574955422b2efa47c56715554227cefd272", "effective_relative_l2": 0.0016649364260956645, "effective_max_abs": 0.0005035400390625 }, { "name": "model.language_model.layers.53.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "f5cae03b260fc21a9b402b63fc7c8cb1b981e030d239910beff635ccb39373fe", "compensated_sha256": "c28b82464a886b1bb41037a3a50718c8a31235f91f22cf55081bae0620aa4722", "effective_relative_l2": 0.0016542052617296576, "effective_max_abs": 0.00063323974609375 }, { "name": "model.language_model.layers.53.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "89f6a33b315a9c16b9ee2301ffcafe9ebcca6ff3876e041485557e36aa21696b", "compensated_sha256": "1dd8ee969dc79b527ad6a4b77326bae20b320e09ed831ade247c39c6d1cc7b8d", "effective_relative_l2": 0.0016618178924545646, "effective_max_abs": 0.00040531158447265625 }, { "name": "model.language_model.layers.54.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "1dcec66d34d7387c09116122e9fec06eba77c42a1698acb103a9caa89ac4f73e", "compensated_sha256": "5e026eeae1cee4f96d178f9931e2be3f2d89c3fa390e391ebf104ce6845924c9", "effective_relative_l2": 0.0016490858979523182, "effective_max_abs": 0.00072479248046875 }, { "name": "model.language_model.layers.54.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "6fd35161d7c8c933d3a3f41be0d874dde03f0be6a6a238acc6f57930b988862e", "compensated_sha256": "f5b5b797a2711b2a5e052cbbe7bee32fe5bef2c9f0d1eb96116e561f0f3d424a", "effective_relative_l2": 0.0016489217523485422, "effective_max_abs": 0.000392913818359375 }, { "name": "model.language_model.layers.56.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "a65f798a2cfee68c1db0c04b616dc524bcec3b9d7004fcdb8800bae262939430", "compensated_sha256": "01196c6742807150dfdf337e7365ecd1c8dfb77ee78214ddb6ebb83e681f79bf", "effective_relative_l2": 0.0016540050273761153, "effective_max_abs": 0.00152587890625 }, { "name": "model.language_model.layers.56.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "acc4f46da2377fb92b5fd712ebc7663c2a246a6d384b71cb26a9d49a5349ff67", "compensated_sha256": "569dc2f7e3e85c1b481b2379f566f10e99801310e0d6586560eeb0ed4b596182", "effective_relative_l2": 0.001661153743043542, "effective_max_abs": 0.000762939453125 }, { "name": "model.language_model.layers.57.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "9b27b0668a6cce520aa4821919c6c81b40d85c977fab835a64886de2cd5bb315", "compensated_sha256": "d4ca9b4cdd29c5d9e28743cb2f247548eabaa891b9150de065c90b87dd83f779", "effective_relative_l2": 0.0016791383968666196, "effective_max_abs": 0.00167083740234375 }, { "name": "model.language_model.layers.57.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "9fde1c28c422d2832c79f638577a429f79e78980974ffc19f1cc90e96ddbdc16", "compensated_sha256": "1206379590b4435d642f64d372f4675d76a94385c351a29094251c64b34bd75d", "effective_relative_l2": 0.0016480133635923266, "effective_max_abs": 0.000835418701171875 }, { "name": "model.language_model.layers.58.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "92a09954927d2fcb6a963607055fc6f17820aeb7752c51298260f26c1d4074f7", "compensated_sha256": "0f0e79880d4408246e4a885aeb05e5a78a6a7712ba3d23e0ed4f3222d4ef61b6", "effective_relative_l2": 0.0016433141427114606, "effective_max_abs": 0.0016632080078125 }, { "name": "model.language_model.layers.58.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "63df0df27835f4b583e682458e26375e6479959f7f72c2008dabfdbb731946f5", "compensated_sha256": "840885badb2472434adb3440d29f881318c990fd6bd0087e4cefc67776b99cb0", "effective_relative_l2": 0.001650591497309506, "effective_max_abs": 0.0004138946533203125 }, { "name": "model.language_model.layers.60.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "b45350124711ed56c7968289ecea9861206aacdbce071111ec91e123cf613efb", "compensated_sha256": "871e983b2acadf60feabefccf26eb62b5ae332ae4786250821f6429c5b1b3c1c", "effective_relative_l2": 0.0016410085372626781, "effective_max_abs": 0.0013275146484375 }, { "name": "model.language_model.layers.60.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "910032012679faeca66f49ecd8449e9861030c1f26abe1428f5669d061cb9032", "compensated_sha256": "ab54e4f260127deaae59db7a422b7732ecd919d115cddbb29c1bf95e2813bcfa", "effective_relative_l2": 0.0016631311737000942, "effective_max_abs": 0.00066375732421875 }, { "name": "model.language_model.layers.61.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "e0b83e862639e3fffbfe05739bed4ad2d9e5bb904f4cd68593a6e9fdd1eb8ee4", "compensated_sha256": "2a04d7b029928ee45a3670596aaeb0434e73165024868d8a82dd4781fa7162af", "effective_relative_l2": 0.0016657045343890786, "effective_max_abs": 0.0014495849609375 }, { "name": "model.language_model.layers.61.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "3f38339ee1cbb263102ae746bab72b98616129afca5d83db3da89568480e5393", "compensated_sha256": "95083bb1e7d969e9198922360aca5c3352fe0dd670d2f6612fb5483abd9d8c7f", "effective_relative_l2": 0.0016511136200278997, "effective_max_abs": 0.000499725341796875 }, { "name": "model.language_model.layers.62.linear_attn.in_proj_a.weight", "shape": [ 48, 5120 ], "source_sha256": "cb298268ae905423fca30b57aadb96cbb3dfd2a5577064c7818607b6722a1623", "compensated_sha256": "450c7b6ac1585ac514dae601c48a929235a31dd3a4aea5ae6c93693539c61b76", "effective_relative_l2": 0.001670765457674861, "effective_max_abs": 0.00146484375 }, { "name": "model.language_model.layers.62.linear_attn.in_proj_b.weight", "shape": [ 48, 5120 ], "source_sha256": "fe9131571c35b8a8c555d7d73eb051994b75ec08c5f458d90dba7f2207ae0675", "compensated_sha256": "3a190fbaf08bff039a2733258bb2d52567dd5452663231a3bbe92d770953f52e", "effective_relative_l2": 0.001669173827394843, "effective_max_abs": 0.000698089599609375 } ] }