{ "variant": "q8", "recheck": false, "model_load_seconds": 0.7792904170055408, "conditioning_seconds": 0.13881229200342204, "reference_sha256": "ace8518ab1d3b45e702a4c54700cab378a3d6e8172dfa8cd3dd51d8e176cfac3", "parameters": { "exaggeration": 0.5, "cfg_weight": 0.5, "temperature": 0.8, "repetition_penalty": 1.2, "min_p": 0.05, "top_p": 0.95, "max_new_tokens": 400, "verbose": false }, "runtime_quantized_matrices": [ { "name": "_s3_tokenizer.quantizer.fsq_codebook.project_down", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.5.mlp.layers.2", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.5.mlp.layers.0", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.5.attn.out", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.5.attn.value", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.5.attn.key", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.5.attn.query", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.4.mlp.layers.2", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.4.mlp.layers.0", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.4.attn.out", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.4.attn.value", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.4.attn.key", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.4.attn.query", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.3.mlp.layers.2", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.3.mlp.layers.0", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.3.attn.out", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.3.attn.value", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.3.attn.key", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.3.attn.query", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.2.mlp.layers.2", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.2.mlp.layers.0", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.2.attn.out", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.2.attn.value", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.2.attn.key", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.2.attn.query", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.1.mlp.layers.2", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.1.mlp.layers.0", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.1.attn.out", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.1.attn.value", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.1.attn.key", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.1.attn.query", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.0.mlp.layers.2", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.0.mlp.layers.0", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.0.attn.out", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.0.attn.value", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.0.attn.key", "bits": 8, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.0.attn.query", "bits": 8, "group_size": 64 }, { "name": "ve.proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.mel2wav.f0_predictor.classifier", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_3.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_3.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_3.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_3.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_3.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_3.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_2.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_2.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_2.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_2.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_2.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_2.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_1.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_1.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_1.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_1.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_1.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_1.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_0.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_0.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_0.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_0.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_0.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_0.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.resnet.mlp_linear", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_3.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_3.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_3.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_3.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_3.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_3.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_2.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_2.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_2.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_2.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_2.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_2.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_1.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_1.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_1.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_1.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_1.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_1.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_0.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_0.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_0.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_0.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_0.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_0.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.resnet.mlp_linear", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_3.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_3.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_3.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_3.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_3.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_3.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_2.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_2.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_2.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_2.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_2.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_2.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_1.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_1.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_1.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_1.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_1.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_1.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_0.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_0.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_0.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_0.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_0.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_0.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.resnet.mlp_linear", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_3.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_3.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_3.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_3.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_3.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_3.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_2.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_2.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_2.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_2.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_2.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_2.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_1.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_1.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_1.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_1.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_1.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_1.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_0.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_0.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_0.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_0.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_0.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_0.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.resnet.mlp_linear", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_3.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_3.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_3.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_3.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_3.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_3.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_2.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_2.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_2.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_2.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_2.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_2.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_1.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_1.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_1.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_1.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_1.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_1.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_0.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_0.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_0.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_0.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_0.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_0.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.resnet.mlp_linear", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_3.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_3.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_3.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_3.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_3.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_3.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_2.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_2.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_2.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_2.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_2.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_2.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_1.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_1.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_1.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_1.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_1.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_1.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_0.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_0.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_0.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_0.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_0.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_0.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.resnet.mlp_linear", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_3.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_3.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_3.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_3.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_3.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_3.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_2.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_2.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_2.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_2.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_2.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_2.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_1.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_1.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_1.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_1.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_1.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_1.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_0.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_0.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_0.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_0.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_0.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_0.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.resnet.mlp_linear", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_3.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_3.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_3.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_3.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_3.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_3.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_2.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_2.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_2.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_2.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_2.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_2.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_1.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_1.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_1.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_1.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_1.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_1.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_0.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_0.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_0.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_0.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_0.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_0.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.resnet.mlp_linear", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_3.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_3.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_3.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_3.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_3.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_3.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_2.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_2.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_2.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_2.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_2.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_2.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_1.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_1.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_1.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_1.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_1.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_1.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_0.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_0.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_0.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_0.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_0.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_0.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.resnet.mlp_linear", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_3.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_3.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_3.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_3.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_3.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_3.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_2.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_2.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_2.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_2.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_2.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_2.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_1.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_1.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_1.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_1.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_1.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_1.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_0.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_0.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_0.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_0.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_0.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_0.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.resnet.mlp_linear", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_3.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_3.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_3.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_3.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_3.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_3.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_2.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_2.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_2.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_2.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_2.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_2.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_1.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_1.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_1.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_1.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_1.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_1.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_0.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_0.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_0.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_0.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_0.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_0.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.resnet.mlp_linear", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_3.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_3.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_3.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_3.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_3.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_3.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_2.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_2.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_2.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_2.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_2.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_2.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_1.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_1.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_1.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_1.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_1.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_1.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_0.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_0.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_0.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_0.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_0.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_0.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.resnet.mlp_linear", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_3.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_3.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_3.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_3.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_3.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_3.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_2.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_2.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_2.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_2.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_2.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_2.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_1.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_1.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_1.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_1.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_1.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_1.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_0.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_0.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_0.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_0.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_0.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_0.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.resnet.mlp_linear", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_3.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_3.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_3.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_3.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_3.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_3.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_2.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_2.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_2.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_2.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_2.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_2.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_1.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_1.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_1.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_1.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_1.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_1.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_0.ff.layers.1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_0.ff.layers.0", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_0.attn.out_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_0.attn.value_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_0.attn.key_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_0.attn.query_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.resnet.mlp_linear", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.time_mlp.linear_2", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.time_mlp.linear_1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder_proj", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_3.feed_forward.w_2", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_3.feed_forward.w_1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_3.self_attn.linear_pos", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_3.self_attn.linear_out", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_3.self_attn.linear_v", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_3.self_attn.linear_k", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_3.self_attn.linear_q", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_2.feed_forward.w_2", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_2.feed_forward.w_1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_2.self_attn.linear_pos", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_2.self_attn.linear_out", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_2.self_attn.linear_v", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_2.self_attn.linear_k", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_2.self_attn.linear_q", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_1.feed_forward.w_2", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_1.feed_forward.w_1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_1.self_attn.linear_pos", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_1.self_attn.linear_out", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_1.self_attn.linear_v", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_1.self_attn.linear_k", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_1.self_attn.linear_q", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_0.feed_forward.w_2", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_0.feed_forward.w_1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_0.self_attn.linear_pos", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_0.self_attn.linear_out", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_0.self_attn.linear_v", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_0.self_attn.linear_k", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_0.self_attn.linear_q", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_embed.linear", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_5.feed_forward.w_2", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_5.feed_forward.w_1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_5.self_attn.linear_pos", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_5.self_attn.linear_out", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_5.self_attn.linear_v", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_5.self_attn.linear_k", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_5.self_attn.linear_q", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_4.feed_forward.w_2", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_4.feed_forward.w_1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_4.self_attn.linear_pos", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_4.self_attn.linear_out", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_4.self_attn.linear_v", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_4.self_attn.linear_k", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_4.self_attn.linear_q", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_3.feed_forward.w_2", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_3.feed_forward.w_1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_3.self_attn.linear_pos", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_3.self_attn.linear_out", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_3.self_attn.linear_v", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_3.self_attn.linear_k", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_3.self_attn.linear_q", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_2.feed_forward.w_2", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_2.feed_forward.w_1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_2.self_attn.linear_pos", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_2.self_attn.linear_out", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_2.self_attn.linear_v", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_2.self_attn.linear_k", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_2.self_attn.linear_q", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_1.feed_forward.w_2", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_1.feed_forward.w_1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_1.self_attn.linear_pos", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_1.self_attn.linear_out", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_1.self_attn.linear_v", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_1.self_attn.linear_k", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_1.self_attn.linear_q", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_0.feed_forward.w_2", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_0.feed_forward.w_1", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_0.self_attn.linear_pos", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_0.self_attn.linear_out", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_0.self_attn.linear_v", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_0.self_attn.linear_k", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_0.self_attn.linear_q", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.encoder.embed.linear", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.spk_embed_affine_layer", "bits": 8, "group_size": 64 }, { "name": "s3gen.flow.input_embedding", "bits": 8, "group_size": 64 }, { "name": "t3.speech_head", "bits": 8, "group_size": 64 }, { "name": "t3.text_head", "bits": 8, "group_size": 64 }, { "name": "t3.speech_pos_emb.emb", "bits": 8, "group_size": 64 }, { "name": "t3.text_pos_emb.emb", "bits": 8, "group_size": 64 }, { "name": "t3.speech_emb", "bits": 8, "group_size": 64 }, { "name": "t3.text_emb", "bits": 8, "group_size": 64 }, { "name": "t3.cond_enc.perceiver.attn.proj_out", "bits": 8, "group_size": 64 }, { "name": "t3.cond_enc.perceiver.attn.to_v", "bits": 8, "group_size": 64 }, { "name": "t3.cond_enc.perceiver.attn.to_k", "bits": 8, "group_size": 64 }, { "name": "t3.cond_enc.perceiver.attn.to_q", "bits": 8, "group_size": 64 }, { "name": "t3.cond_enc.spkr_enc", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.29.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.29.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.29.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.29.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.29.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.29.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.29.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.28.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.28.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.28.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.28.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.28.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.28.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.28.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.27.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.27.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.27.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.27.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.27.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.27.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.27.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.26.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.26.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.26.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.26.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.26.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.26.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.26.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.25.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.25.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.25.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.25.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.25.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.25.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.25.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.24.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.24.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.24.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.24.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.24.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.24.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.24.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.23.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.23.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.23.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.23.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.23.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.23.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.23.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.22.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.22.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.22.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.22.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.22.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.22.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.22.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.21.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.21.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.21.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.21.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.21.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.21.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.21.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.20.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.20.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.20.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.20.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.20.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.20.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.20.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.19.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.19.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.19.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.19.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.19.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.19.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.19.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.18.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.18.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.18.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.18.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.18.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.18.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.18.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.17.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.17.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.17.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.17.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.17.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.17.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.17.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.16.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.16.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.16.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.16.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.16.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.16.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.16.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.15.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.15.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.15.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.15.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.15.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.15.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.15.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.14.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.14.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.14.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.14.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.14.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.14.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.14.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.13.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.13.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.13.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.13.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.13.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.13.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.13.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.12.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.12.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.12.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.12.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.12.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.12.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.12.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.11.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.11.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.11.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.11.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.11.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.11.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.11.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.10.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.10.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.10.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.10.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.10.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.10.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.10.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.9.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.9.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.9.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.9.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.9.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.9.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.9.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.8.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.8.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.8.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.8.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.8.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.8.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.8.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.7.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.7.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.7.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.7.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.7.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.7.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.7.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.6.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.6.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.6.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.6.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.6.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.6.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.6.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.5.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.5.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.5.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.5.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.5.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.5.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.5.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.4.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.4.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.4.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.4.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.4.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.4.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.4.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.3.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.3.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.3.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.3.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.3.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.3.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.3.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.2.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.2.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.2.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.2.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.2.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.2.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.2.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.1.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.1.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.1.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.1.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.1.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.1.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.1.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.0.mlp.up_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.0.mlp.down_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.0.mlp.gate_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.0.self_attn.o_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.0.self_attn.v_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.0.self_attn.k_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.layers.0.self_attn.q_proj", "bits": 8, "group_size": 64 }, { "name": "t3.tfmr.model.embed_tokens", "bits": 8, "group_size": 64 } ], "versions": { "mlx": "0.32.3", "mlx-audio": "0.5.7", "numpy": "2.4.6", "soundfile": "0.14.0", "huggingface-hub": "1.33.0" }, "warm_repetitions": [ { "file": "q8-warm-0.wav", "sha256": "c4032ddb8ce1bac397d456b101177a9a4d26dae3fbfd14094807161b9d231f51", "float_sha256": "297a84b7ae411e6bb1c1512e8e4e7957f7eddb495cbd4ea417ec2f427048fe64", "generation_seconds": 5.244187625008635, "audio_seconds": 8.56, "rtf": 0.6126387412393265, "mlx_peak_bytes": 2584147472, "rms": 0.07231333926376518, "peak": 0.6776053309440613, "clipped_fraction": 0.0, "repeat": 0, "first_generation": true, "text": "This is the oMLX Chatterbox Turbo checkpoint. It was quantized from the original fp16 model.", "lang": "en", "seed": 20261009 }, { "file": "q8-warm-1.wav", "sha256": "9fffada6fa80b3ba88efd317737b69727f561cda2df6debd9f10a9f55b0d2188", "float_sha256": "2c0c675d6005ca88fd71924666aa02153f72087e4e0f15d5145d8fd5bc8a1be6", "generation_seconds": 5.2366827499936335, "audio_seconds": 8.56, "rtf": 0.6117620035039291, "mlx_peak_bytes": 2584147472, "rms": 0.07231333926127466, "peak": 0.6776053309440613, "clipped_fraction": 0.0, "repeat": 1, "first_generation": false, "text": "This is the oMLX Chatterbox Turbo checkpoint. It was quantized from the original fp16 model.", "lang": "en", "seed": 20261009 }, { "file": "q8-warm-2.wav", "sha256": "2210c60a6faca29debd4970a575991180ba889282ed73a20abb3489c7226c86f", "float_sha256": "7fd3cf6d5344373f1622d6e8a83aafb8672edb08448c4447cf7274cc743fc376", "generation_seconds": 5.531241208998836, "audio_seconds": 8.56, "rtf": 0.6461730384344434, "mlx_peak_bytes": 2584147472, "rms": 0.0723133392729002, "peak": 0.6776053309440613, "clipped_fraction": 0.0, "repeat": 2, "first_generation": false, "text": "This is the oMLX Chatterbox Turbo checkpoint. It was quantized from the original fp16 model.", "lang": "en", "seed": 20261009 }, { "file": "q8-warm-3.wav", "sha256": "843c6fd0204ec3997e3a2560dba2a7d068766f5bfff67928590b23c286fc7b86", "float_sha256": "57c598984ba9780520dbc346ffcf1e0a06d169ef2e3fe232895302adb245217a", "generation_seconds": 5.837617708006292, "audio_seconds": 8.56, "rtf": 0.6819646855147536, "mlx_peak_bytes": 2584163856, "rms": 0.07231333925440961, "peak": 0.6776053309440613, "clipped_fraction": 0.0, "repeat": 3, "first_generation": false, "text": "This is the oMLX Chatterbox Turbo checkpoint. It was quantized from the original fp16 model.", "lang": "en", "seed": 20261009 }, { "file": "q8-warm-4.wav", "sha256": "414ec303a5e9c67008ea3e95cde19bf9103af2b3ee23ed30b01997dd31e4386a", "float_sha256": "20e8c11341147438b4e9b18b453a5ac5ba5377b69cb571b671ad757a46521bd5", "generation_seconds": 5.777289957986795, "audio_seconds": 8.56, "rtf": 0.6749170511666817, "mlx_peak_bytes": 2584147472, "rms": 0.07231333926728001, "peak": 0.6776053309440613, "clipped_fraction": 0.0, "repeat": 4, "first_generation": false, "text": "This is the oMLX Chatterbox Turbo checkpoint. It was quantized from the original fp16 model.", "lang": "en", "seed": 20261009 }, { "file": "q8-warm-5.wav", "sha256": "e1b2426208925a13fbc30ddd5b6ec1a0259e363ee3c6a2a3f32e5d6de748f37a", "float_sha256": "74ee8f3bbb10414d6cdfddde1a5fedd676019ff4ddcc09f0102cfd9ec25b1ef9", "generation_seconds": 5.568548291004845, "audio_seconds": 8.56, "rtf": 0.650531342407108, "mlx_peak_bytes": 2584147472, "rms": 0.07231333925429473, "peak": 0.6776053309440613, "clipped_fraction": 0.0, "repeat": 5, "first_generation": false, "text": "This is the oMLX Chatterbox Turbo checkpoint. It was quantized from the original fp16 model.", "lang": "en", "seed": 20261009 } ], "end_to_end": [ { "file": "q8-e2e-0-20261009.wav", "sha256": "3f72b995abe6adc9149d825a9a7848d7d747e49e6ff7a50a3744b9feaca14eb0", "float_sha256": "7da46c3baa32c82b0d223abc586b2a4cf1da97609302a3c967608cad3f8a93a3", "generation_seconds": 5.755737208004575, "audio_seconds": 8.56, "rtf": 0.6723992065425906, "mlx_peak_bytes": 2584313088, "rms": 0.07231333925480417, "peak": 0.6776053309440613, "clipped_fraction": 0.0, "prompt": 0, "lang": "en", "text": "This is the oMLX Chatterbox Turbo checkpoint. It was quantized from the original fp16 model.", "seed": 20261009, "speech_tokens": 214, "speech_token_sha256": "b6d7830bfab8076e3566e972301f6a27ca3e4b454f8d59dbc0b7a8f050f1e028", "timing_includes_token_capture": true }, { "file": "q8-e2e-0-42.wav", "sha256": "ebd25ac870a34421aad3a6e186710e70e0e682b8733b59ed6e8340f856bc44d2", "float_sha256": "91ab7f232ba780b5cf8b4692944af2a794646db52b018c52f51128dd17508987", "generation_seconds": 4.5601065829978324, "audio_seconds": 7.08, "rtf": 0.6440828507059085, "mlx_peak_bytes": 2494332676, "rms": 0.0832787560041797, "peak": 0.8690729737281799, "clipped_fraction": 0.0, "prompt": 0, "lang": "en", "text": "This is the oMLX Chatterbox Turbo checkpoint. It was quantized from the original fp16 model.", "seed": 42, "speech_tokens": 177, "speech_token_sha256": "0659bbbfd0b9af83f87dc2eff63dab907136d51961aca2b7bd258b93ce35a43c", "timing_includes_token_capture": true }, { "file": "q8-e2e-1-20261009.wav", "sha256": "57ea7ba9c10f47c3437dad5d507335eb20317cf250de2c5a17b9d54db7461dc3", "float_sha256": "422a6bf0b13f58f95fa59e2025e9194b5aaa37b9da421f1382e722e421834d3c", "generation_seconds": 3.987167083003442, "audio_seconds": 5.4, "rtf": 0.738364274630267, "mlx_peak_bytes": 2289525580, "rms": 0.08092633722917864, "peak": 0.6097702980041504, "clipped_fraction": 0.0, "prompt": 1, "lang": "en", "text": "Hello mate, the training checkpoint is saved, and this is Chatterbox speaking locally on the Mac.", "seed": 20261009, "speech_tokens": 135, "speech_token_sha256": "24c63a3e429f62d20ad37e66a6b098ecb610d03b4a920fa225f8ba23c7b1cfe1", "timing_includes_token_capture": true }, { "file": "q8-e2e-1-42.wav", "sha256": "aa4ee565c9531c076e03d82adbf6a282cdcadb6c2e58f5fb4a572e4cb3928122", "float_sha256": "9ec7200541dd4ce1b22ae89d82582a68eda580f21ec2e16622fcaf6942f943eb", "generation_seconds": 6.684393834002549, "audio_seconds": 8.64, "rtf": 0.773656693750295, "mlx_peak_bytes": 2587742244, "rms": 0.06537511927990118, "peak": 0.5386520028114319, "clipped_fraction": 0.0, "prompt": 1, "lang": "en", "text": "Hello mate, the training checkpoint is saved, and this is Chatterbox speaking locally on the Mac.", "seed": 42, "speech_tokens": 216, "speech_token_sha256": "b3f8d4a17a12955f081aa76e57da8d3d19fe2abb1e9018973324309570243e2b", "timing_includes_token_capture": true }, { "file": "q8-e2e-2-20261009.wav", "sha256": "30d78223f557d5c5e60f6bd1bc35b6985b19a4c8b9bb97d341e85058d3d9f00c", "float_sha256": "52c6905dfb5b9f3093c344b9bf7c13d1b208e9bf91b07b03bfd24309f223e284", "generation_seconds": 4.970803208008874, "audio_seconds": 5.76, "rtf": 0.8629866680570962, "mlx_peak_bytes": 2331437308, "rms": 0.07985603556164887, "peak": 0.6071439981460571, "clipped_fraction": 0.0, "prompt": 2, "lang": "en", "text": "Please check the appointment on Thursday at three thirty. Bring the blue folder, your keys, and a bottle of water.", "seed": 20261009, "speech_tokens": 144, "speech_token_sha256": "964459d6c5c17234678bd34edaa8f8c6f8a04de9511cb499ac9ab50bfe288b6c", "timing_includes_token_capture": true }, { "file": "q8-e2e-2-42.wav", "sha256": "c895856dd4b458ba226f85d643cbefef4a1f030b89f52d773dd24ad3652b85cb", "float_sha256": "0189d7d7f58e9f7041a0f0b75d8bd7b0424da936cf4a276dd253d679f0c41bb5", "generation_seconds": 5.049558790997253, "audio_seconds": 6.28, "rtf": 0.8040698711779065, "mlx_peak_bytes": 2395432084, "rms": 0.07344408685271332, "peak": 0.5616801381111145, "clipped_fraction": 0.0, "prompt": 2, "lang": "en", "text": "Please check the appointment on Thursday at three thirty. Bring the blue folder, your keys, and a bottle of water.", "seed": 42, "speech_tokens": 157, "speech_token_sha256": "e276c73ee43684a0ad50eab46393818a37e9411cd8c56e89c337d070e8ec1c73", "timing_includes_token_capture": true }, { "file": "q8-e2e-3-20261009.wav", "sha256": "11e54624c5967474f6b364fcf3d330c6a6837c9aeef2b9dfd44982653aa95130", "float_sha256": "4b6af38109ed874030d2f914395b7f48761ae5c48c30a20cf4c1ee2f46ab465c", "generation_seconds": 5.4503103330062, "audio_seconds": 6.52, "rtf": 0.8359371676389878, "mlx_peak_bytes": 2424587636, "rms": 0.08913555176421256, "peak": 0.672126293182373, "clipped_fraction": 0.0, "prompt": 3, "lang": "es", "text": "Este es el checkpoint de Chatterbox Turbo de oMLX. Fue cuantizado a partir del modelo original fp16.", "seed": 20261009, "speech_tokens": 163, "speech_token_sha256": "def5e709d490090263c659aadb51f5abeee32e8170ce1c971bfbb34d6d8f64e4", "timing_includes_token_capture": true }, { "file": "q8-e2e-3-42.wav", "sha256": "433b09414cbe52bbf424ded41c5b9780f1eff4191cc87c6d7876d200e62cda35", "float_sha256": "cf2ae59cbc44fb244efe50721b5ef35c54de4056804ac90e0a0d431319ba1920", "generation_seconds": 6.983207249999396, "audio_seconds": 8.52, "rtf": 0.8196252640844361, "mlx_peak_bytes": 2580249852, "rms": 0.08073411920874711, "peak": 0.6531962752342224, "clipped_fraction": 0.0, "prompt": 3, "lang": "es", "text": "Este es el checkpoint de Chatterbox Turbo de oMLX. Fue cuantizado a partir del modelo original fp16.", "seed": 42, "speech_tokens": 213, "speech_token_sha256": "721e568ab5c68d79b33cc2a2c4c6d3ea731dbdf860fb57c5d5a7b3fe86f3e9b1", "timing_includes_token_capture": true }, { "file": "q8-e2e-4-20261009.wav", "sha256": "7ce05bfb1cdcd1a34645ceed9e60082974f52ab20c4882300fa077e30e2dd73e", "float_sha256": "989455a723e3841401ad2d1fce85569fdf5de4165204ae374f93ad2d4411fb63", "generation_seconds": 5.2006425830040826, "audio_seconds": 5.96, "rtf": 0.8725910374167923, "mlx_peak_bytes": 2356035048, "rms": 0.09536592761282509, "peak": 0.7870880961418152, "clipped_fraction": 0.0, "prompt": 4, "lang": "es", "text": "Hola, amigo, el checkpoint de entrenamiento está guardado y este es Chatterbox hablando localmente en la Mac.", "seed": 20261009, "speech_tokens": 149, "speech_token_sha256": "f7271047ba79ddce67a828b9ee18a42daf1f1e1983c004b0dabdf1577fd2265a", "timing_includes_token_capture": true }, { "file": "q8-e2e-4-42.wav", "sha256": "bd201b9948f5e1567f4b38afaaf85ae6beaffd940380cec1878f9969df9c6351", "float_sha256": "f39f10c01969e31d2de63678474d59f61b89249b190240ddd330b50078cfe534", "generation_seconds": 5.054334249987733, "audio_seconds": 6.28, "rtf": 0.8048302945840339, "mlx_peak_bytes": 2395432076, "rms": 0.08816368297440172, "peak": 0.7103111743927002, "clipped_fraction": 0.0, "prompt": 4, "lang": "es", "text": "Hola, amigo, el checkpoint de entrenamiento está guardado y este es Chatterbox hablando localmente en la Mac.", "seed": 42, "speech_tokens": 157, "speech_token_sha256": "3c83a99208d7e8c50b1e2ef714bdb5823af2d0e986d4a20fda990f6d3b71c0d5", "timing_includes_token_capture": true }, { "file": "q8-e2e-5-20261009.wav", "sha256": "0c202f152a7d0423730f039d5bc56b1f1471ab9734f1ed00441ff0926b38702b", "float_sha256": "911d223bfcea3df8dfc71e434aa0d0aee9c236c26818bf1b7530ec4255c81f4a", "generation_seconds": 6.266688083007466, "audio_seconds": 8.32, "rtf": 0.7532077022845511, "mlx_peak_bytes": 2563570220, "rms": 0.07896153290862118, "peak": 0.8897371888160706, "clipped_fraction": 0.0, "prompt": 5, "lang": "es", "text": "Por favor, revisa la cita del jueves a las tres y media. Trae la carpeta azul, tus llaves y una botella de agua.", "seed": 20261009, "speech_tokens": 208, "speech_token_sha256": "1244a41e23b6f03738308ec4cae8f05043fd367bf76741eb6c0ac1fe27df7536", "timing_includes_token_capture": true }, { "file": "q8-e2e-5-42.wav", "sha256": "f2c8cb00f7e990b710ca930d50ad0fd824bea88a8b30fea7c1f67f8bd13b6222", "float_sha256": "c9e5c39b6ed2cec631756b25c23abbdcd618d6615b3debe741229a7130c083cb", "generation_seconds": 5.922888875007629, "audio_seconds": 7.12, "rtf": 0.8318664150291614, "mlx_peak_bytes": 2498716992, "rms": 0.0930324119523856, "peak": 0.9518831372261047, "clipped_fraction": 0.0, "prompt": 5, "lang": "es", "text": "Por favor, revisa la cita del jueves a las tres y media. Trae la carpeta azul, tus llaves y una botella de agua.", "seed": 42, "speech_tokens": 178, "speech_token_sha256": "83be09476bf4bd23215abb6e27edcf9369efe12b16184a64c4374f1fa8622524", "timing_includes_token_capture": true } ], "controlled_decoder": [ { "file": "q8-decoder-0-20261009.wav", "sha256": "8afe66d2cb59f7d7ef383feb946b91486f31a5c8312dccbcaba1a22c4db0c14b", "float_sha256": "3cf7291e9b5fd8f6a3d501412c964b7fe187751c61785c154f0f372080a31f91", "generation_seconds": 4.057174083995051, "audio_seconds": 8.48, "rtf": 0.47844034009375597, "mlx_peak_bytes": 2580298868, "rms": 0.06935898892124007, "peak": 0.6296741962432861, "clipped_fraction": 0.0, "prompt": 0, "lang": "en", "text": "This is the oMLX Chatterbox Turbo checkpoint. It was quantized from the original fp16 model.", "seed": 20261009, "decoder_seed": 9182, "speech_tokens": 212, "source_token_sha256": "1225f7a4fdc94204bc692e58fa0ea124684670fe74211411d007ad53eb4e4477", "reference_shape_matched": true }, { "file": "q8-decoder-0-42.wav", "sha256": "c6b030be318cf08d13fc5560b4e4dddb089487cc40e7bfea5ca0543c23abb929", "float_sha256": "4301027118c4f4f78999444996f312664eafe4d06eb9b98b4dbd74556d00ca19", "generation_seconds": 4.470503707998432, "audio_seconds": 7.84, "rtf": 0.5702173096936776, "mlx_peak_bytes": 2582280156, "rms": 0.06988808329992759, "peak": 0.5527527332305908, "clipped_fraction": 0.0, "prompt": 0, "lang": "en", "text": "This is the oMLX Chatterbox Turbo checkpoint. It was quantized from the original fp16 model.", "seed": 42, "decoder_seed": 9182, "speech_tokens": 196, "source_token_sha256": "c1c2de8f00a7122c5becec088297c2236ea36df1c2c2c744447d65123736f0ff", "reference_shape_matched": true }, { "file": "q8-decoder-1-20261009.wav", "sha256": "a24c11edf5c9f09e50510f7771b89272e10b4949a7dc5209540bbbacea482832", "float_sha256": "f541cba05a24755f8dc2a7742ebe37d2a2aa9ab2bc8ed4fa0423fda548223aeb", "generation_seconds": 3.480579625000246, "audio_seconds": 5.8, "rtf": 0.60009993534487, "mlx_peak_bytes": 2336464720, "rms": 0.08003359140624661, "peak": 0.6371778249740601, "clipped_fraction": 0.0, "prompt": 1, "lang": "en", "text": "Hello mate, the training checkpoint is saved, and this is Chatterbox speaking locally on the Mac.", "seed": 20261009, "decoder_seed": 9182, "speech_tokens": 145, "source_token_sha256": "5cfc24860df526a94926fb5c5e241ad4ebf77207f4f4f9c1008c90f6f7a15df9", "reference_shape_matched": true }, { "file": "q8-decoder-1-42.wav", "sha256": "6c854f57b64cd817d591ee83d7d5d2b7c182bfa64d4696db8968cf2468cd3acc", "float_sha256": "b9cd4656246b6b5dc433ecebd8886e1424f09c14e6b994a9a2d8f32ed1d6732b", "generation_seconds": 2.7550749170040945, "audio_seconds": 5.6, "rtf": 0.4919776637507312, "mlx_peak_bytes": 2312148373, "rms": 0.08175650593203415, "peak": 0.5611691474914551, "clipped_fraction": 0.0, "prompt": 1, "lang": "en", "text": "Hello mate, the training checkpoint is saved, and this is Chatterbox speaking locally on the Mac.", "seed": 42, "decoder_seed": 9182, "speech_tokens": 140, "source_token_sha256": "fef7f0074299cbb05c480ebb87e82d76eb57335e55e6820a40d0d524b6675496", "reference_shape_matched": true }, { "file": "q8-decoder-2-20261009.wav", "sha256": "c2d9435a5af213fee12a2d1c5921faa6da7ec2780561036bc2fa0739c64619b9", "float_sha256": "5d5934c98a73cf7eeec6360dc67252470cf56654528585d7f51b9a592ffb0c20", "generation_seconds": 2.742295000003651, "audio_seconds": 5.72, "rtf": 0.4794222027978411, "mlx_peak_bytes": 2327811836, "rms": 0.07537337436473573, "peak": 0.5510967969894409, "clipped_fraction": 0.0, "prompt": 2, "lang": "en", "text": "Please check the appointment on Thursday at three thirty. Bring the blue folder, your keys, and a bottle of water.", "seed": 20261009, "decoder_seed": 9182, "speech_tokens": 143, "source_token_sha256": "0162792dd3cccdf3b23f649b2f317e777d5431be8b43c4b898534c3915ba5c30", "reference_shape_matched": true }, { "file": "q8-decoder-2-42.wav", "sha256": "2697612009efc3e8738f23c54b70a140ef2dfddd02dcbb9b10f42e785ddfdd31", "float_sha256": "f6707ec499a904dd2835416503c4871b8925a7616f4c4e50689c8058204a253f", "generation_seconds": 2.543105167002068, "audio_seconds": 5.32, "rtf": 0.4780272870304639, "mlx_peak_bytes": 2279121516, "rms": 0.07561528716104214, "peak": 0.5643180012702942, "clipped_fraction": 0.0, "prompt": 2, "lang": "en", "text": "Please check the appointment on Thursday at three thirty. Bring the blue folder, your keys, and a bottle of water.", "seed": 42, "decoder_seed": 9182, "speech_tokens": 133, "source_token_sha256": "13365eea735200dba4ce5afe9e981a1dd747ae44e67036cbe8704b23cc68834c", "reference_shape_matched": true }, { "file": "q8-decoder-3-20261009.wav", "sha256": "3416d2337251a45ee7afe0fae6b2e55ea2eeab1b2d4de88cc91a486374cb6568", "float_sha256": "b0ca70accc9263280b49d458d7958dff73682664e5dbd4d67a6ce71c609ba0fb", "generation_seconds": 2.9213512090063887, "audio_seconds": 6.52, "rtf": 0.448060001381348, "mlx_peak_bytes": 2424572972, "rms": 0.09031595118138053, "peak": 0.6536054015159607, "clipped_fraction": 0.0, "prompt": 3, "lang": "es", "text": "Este es el checkpoint de Chatterbox Turbo de oMLX. Fue cuantizado a partir del modelo original fp16.", "seed": 20261009, "decoder_seed": 9182, "speech_tokens": 163, "source_token_sha256": "912cde02e0ae66e34fefdca7455d29fe0a3a88094574e2fea2f3af2d35252004", "reference_shape_matched": true }, { "file": "q8-decoder-3-42.wav", "sha256": "de6a4a77c36f5ee3a6e9a34722ee9be8cbb690ff55e2bcacb32795bb0f6536fa", "float_sha256": "60c91e3b82e845dd327bf72e1801cea366d96f63f180654f156e26db3017bd47", "generation_seconds": 3.3051404590078164, "audio_seconds": 6.4, "rtf": 0.5164281967199713, "mlx_peak_bytes": 2408363204, "rms": 0.08381916768441286, "peak": 0.6601104140281677, "clipped_fraction": 0.0, "prompt": 3, "lang": "es", "text": "Este es el checkpoint de Chatterbox Turbo de oMLX. Fue cuantizado a partir del modelo original fp16.", "seed": 42, "decoder_seed": 9182, "speech_tokens": 160, "source_token_sha256": "510a36cc71321f20b5a06826b4884130d4adc8618585757ffdc2e74ec5899b9e", "reference_shape_matched": true }, { "file": "q8-decoder-4-20261009.wav", "sha256": "dc351a4975af20f55c09fa804ce7492c55aefd11e61d004aadc6cac83a553909", "float_sha256": "65821d3fb5e435ddcd1f5b3c1a945dbb14c2de0fa1b50fec9d9d56df53fcbe90", "generation_seconds": 3.5853033750026952, "audio_seconds": 6.08, "rtf": 0.5896880550991275, "mlx_peak_bytes": 2370562292, "rms": 0.08822706929269773, "peak": 0.8327633142471313, "clipped_fraction": 0.0, "prompt": 4, "lang": "es", "text": "Hola, amigo, el checkpoint de entrenamiento está guardado y este es Chatterbox hablando localmente en la Mac.", "seed": 20261009, "decoder_seed": 9182, "speech_tokens": 152, "source_token_sha256": "bbee0e53132295f6c223b590727068ebf1b10ddc1a98d20c9889f806c232e742", "reference_shape_matched": true }, { "file": "q8-decoder-4-42.wav", "sha256": "08d92cd982f9e487add096b72e6174f8f176bec91bcb771b397a9b5bd14bb9d3", "float_sha256": "74a10a51d6202b95e00bf6ac63c4c9e75346323f1365a6d020c08aa9ebc23af4", "generation_seconds": 3.4176434580003843, "audio_seconds": 6.16, "rtf": 0.554812249675387, "mlx_peak_bytes": 2379688356, "rms": 0.08303270510416638, "peak": 0.7862178683280945, "clipped_fraction": 0.0, "prompt": 4, "lang": "es", "text": "Hola, amigo, el checkpoint de entrenamiento está guardado y este es Chatterbox hablando localmente en la Mac.", "seed": 42, "decoder_seed": 9182, "speech_tokens": 154, "source_token_sha256": "a1be4dc194c481b4c1dd62337f92e6bd9769fa6465784acc4cfa4f018131704e", "reference_shape_matched": true }, { "file": "q8-decoder-5-20261009.wav", "sha256": "4e6405ad35e6dd87bd92312c85ede079f3ed3f0603be1f7831ab6a9a52cc6b81", "float_sha256": "494b5e1ef3e9e4cc8143b3300ce22eb5105080efe0a432b16741ab823f548bb9", "generation_seconds": 3.3086566669953754, "audio_seconds": 6.08, "rtf": 0.5441869518084499, "mlx_peak_bytes": 2370562293, "rms": 0.09916244046463293, "peak": 0.7611903548240662, "clipped_fraction": 0.0, "prompt": 5, "lang": "es", "text": "Por favor, revisa la cita del jueves a las tres y media. Trae la carpeta azul, tus llaves y una botella de agua.", "seed": 20261009, "decoder_seed": 9182, "speech_tokens": 152, "source_token_sha256": "aa5fff77bc601f519ec9ba885eb0a4918aad4ef242fd076670f4ffef9fd50847", "reference_shape_matched": true }, { "file": "q8-decoder-5-42.wav", "sha256": "145cd99c7676d9fdc7230c124fe3534b115793909dcaab67480202ff45a29eff", "float_sha256": "1413f96316a3481f34b7897dd6f7923c5ba7315d4fb10c93b720877c358fcfab", "generation_seconds": 3.850878625002224, "audio_seconds": 8.28, "rtf": 0.4650819595413314, "mlx_peak_bytes": 2559588992, "rms": 0.08344251813468516, "peak": 0.9723390936851501, "clipped_fraction": 0.0, "prompt": 5, "lang": "es", "text": "Por favor, revisa la cita del jueves a las tres y media. Trae la carpeta azul, tus llaves y una botella de agua.", "seed": 42, "decoder_seed": 9182, "speech_tokens": 207, "source_token_sha256": "0012d919bf529baf19b526f9ee344ad618f3963beb3bb616efac8b261443cd59", "reference_shape_matched": true } ], "full_process_peak_mlx_bytes": 2587742244, "process_peak_rss_bytes": 1621524480, "warm_median_seconds": 5.568548291004845, "warm_min_seconds": 5.2366827499936335, "warm_max_seconds": 5.837617708006292, "warm_median_rtf": 0.650531342407108, "warm_same_audio": false, "warm_profile_peak_mlx_bytes": 2584163856, "shared_decoder_conditionals_sha256": "209d21a9230f596ffc854ad305919a3fc33a4bdc746e60baf67239ca9a17043f", "decoder_flow_signature": "(token: mlx.core.array, token_len: mlx.core.array, prompt_token: mlx.core.array, prompt_token_len: mlx.core.array, prompt_feat: mlx.core.array, prompt_feat_len: mlx.core.array, embedding: mlx.core.array, finalize: bool, n_timesteps: Optional[int] = None, streaming: bool = False)", "decoder_method": "Deployed S3Token2Wav.__call__, finalize=True, original FP32 tokens and FP32 decoder conditionals; fixed MLX RNG seed." }