{ "variant": "q6", "recheck": false, "model_load_seconds": 0.626447749993531, "conditioning_seconds": 0.12570883400621824, "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": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.5.mlp.layers.2", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.5.mlp.layers.0", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.5.attn.out", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.5.attn.value", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.5.attn.key", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.5.attn.query", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.4.mlp.layers.2", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.4.mlp.layers.0", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.4.attn.out", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.4.attn.value", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.4.attn.key", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.4.attn.query", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.3.mlp.layers.2", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.3.mlp.layers.0", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.3.attn.out", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.3.attn.value", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.3.attn.key", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.3.attn.query", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.2.mlp.layers.2", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.2.mlp.layers.0", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.2.attn.out", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.2.attn.value", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.2.attn.key", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.2.attn.query", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.1.mlp.layers.2", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.1.mlp.layers.0", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.1.attn.out", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.1.attn.value", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.1.attn.key", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.1.attn.query", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.0.mlp.layers.2", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.0.mlp.layers.0", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.0.attn.out", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.0.attn.value", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.0.attn.key", "bits": 6, "group_size": 64 }, { "name": "_s3_tokenizer.encoder.blocks.0.attn.query", "bits": 6, "group_size": 64 }, { "name": "ve.proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.mel2wav.f0_predictor.classifier", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_3.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_3.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_3.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_3.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_3.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_3.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_2.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_2.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_2.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_2.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_2.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_2.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_1.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_1.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_1.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_1.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_1.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_1.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_0.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_0.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_0.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_0.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_0.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.transformer_0.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.up_blocks_0.resnet.mlp_linear", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_3.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_3.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_3.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_3.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_3.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_3.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_2.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_2.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_2.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_2.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_2.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_2.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_1.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_1.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_1.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_1.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_1.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_1.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_0.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_0.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_0.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_0.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_0.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.transformer_0.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_11.resnet.mlp_linear", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_3.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_3.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_3.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_3.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_3.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_3.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_2.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_2.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_2.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_2.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_2.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_2.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_1.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_1.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_1.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_1.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_1.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_1.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_0.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_0.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_0.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_0.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_0.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.transformer_0.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_10.resnet.mlp_linear", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_3.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_3.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_3.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_3.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_3.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_3.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_2.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_2.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_2.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_2.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_2.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_2.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_1.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_1.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_1.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_1.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_1.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_1.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_0.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_0.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_0.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_0.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_0.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.transformer_0.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_9.resnet.mlp_linear", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_3.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_3.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_3.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_3.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_3.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_3.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_2.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_2.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_2.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_2.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_2.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_2.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_1.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_1.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_1.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_1.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_1.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_1.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_0.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_0.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_0.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_0.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_0.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.transformer_0.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_8.resnet.mlp_linear", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_3.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_3.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_3.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_3.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_3.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_3.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_2.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_2.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_2.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_2.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_2.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_2.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_1.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_1.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_1.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_1.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_1.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_1.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_0.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_0.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_0.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_0.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_0.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.transformer_0.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_7.resnet.mlp_linear", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_3.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_3.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_3.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_3.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_3.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_3.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_2.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_2.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_2.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_2.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_2.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_2.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_1.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_1.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_1.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_1.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_1.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_1.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_0.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_0.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_0.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_0.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_0.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.transformer_0.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_6.resnet.mlp_linear", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_3.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_3.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_3.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_3.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_3.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_3.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_2.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_2.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_2.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_2.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_2.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_2.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_1.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_1.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_1.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_1.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_1.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_1.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_0.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_0.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_0.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_0.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_0.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.transformer_0.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_5.resnet.mlp_linear", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_3.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_3.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_3.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_3.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_3.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_3.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_2.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_2.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_2.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_2.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_2.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_2.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_1.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_1.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_1.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_1.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_1.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_1.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_0.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_0.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_0.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_0.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_0.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.transformer_0.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_4.resnet.mlp_linear", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_3.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_3.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_3.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_3.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_3.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_3.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_2.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_2.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_2.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_2.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_2.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_2.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_1.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_1.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_1.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_1.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_1.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_1.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_0.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_0.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_0.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_0.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_0.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.transformer_0.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_3.resnet.mlp_linear", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_3.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_3.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_3.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_3.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_3.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_3.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_2.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_2.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_2.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_2.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_2.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_2.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_1.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_1.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_1.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_1.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_1.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_1.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_0.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_0.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_0.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_0.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_0.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.transformer_0.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_2.resnet.mlp_linear", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_3.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_3.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_3.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_3.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_3.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_3.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_2.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_2.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_2.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_2.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_2.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_2.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_1.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_1.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_1.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_1.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_1.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_1.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_0.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_0.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_0.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_0.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_0.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.transformer_0.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_1.resnet.mlp_linear", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_3.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_3.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_3.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_3.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_3.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_3.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_2.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_2.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_2.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_2.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_2.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_2.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_1.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_1.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_1.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_1.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_1.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_1.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_0.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_0.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_0.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_0.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_0.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.transformer_0.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.mid_blocks_0.resnet.mlp_linear", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_3.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_3.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_3.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_3.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_3.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_3.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_2.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_2.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_2.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_2.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_2.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_2.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_1.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_1.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_1.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_1.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_1.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_1.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_0.ff.layers.1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_0.ff.layers.0", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_0.attn.out_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_0.attn.value_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_0.attn.key_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.transformer_0.attn.query_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.down_blocks_0.resnet.mlp_linear", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.time_mlp.linear_2", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.decoder.estimator.time_mlp.linear_1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder_proj", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_3.feed_forward.w_2", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_3.feed_forward.w_1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_3.self_attn.linear_pos", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_3.self_attn.linear_out", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_3.self_attn.linear_v", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_3.self_attn.linear_k", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_3.self_attn.linear_q", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_2.feed_forward.w_2", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_2.feed_forward.w_1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_2.self_attn.linear_pos", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_2.self_attn.linear_out", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_2.self_attn.linear_v", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_2.self_attn.linear_k", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_2.self_attn.linear_q", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_1.feed_forward.w_2", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_1.feed_forward.w_1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_1.self_attn.linear_pos", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_1.self_attn.linear_out", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_1.self_attn.linear_v", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_1.self_attn.linear_k", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_1.self_attn.linear_q", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_0.feed_forward.w_2", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_0.feed_forward.w_1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_0.self_attn.linear_pos", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_0.self_attn.linear_out", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_0.self_attn.linear_v", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_0.self_attn.linear_k", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_encoders_0.self_attn.linear_q", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.up_embed.linear", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_5.feed_forward.w_2", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_5.feed_forward.w_1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_5.self_attn.linear_pos", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_5.self_attn.linear_out", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_5.self_attn.linear_v", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_5.self_attn.linear_k", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_5.self_attn.linear_q", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_4.feed_forward.w_2", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_4.feed_forward.w_1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_4.self_attn.linear_pos", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_4.self_attn.linear_out", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_4.self_attn.linear_v", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_4.self_attn.linear_k", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_4.self_attn.linear_q", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_3.feed_forward.w_2", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_3.feed_forward.w_1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_3.self_attn.linear_pos", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_3.self_attn.linear_out", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_3.self_attn.linear_v", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_3.self_attn.linear_k", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_3.self_attn.linear_q", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_2.feed_forward.w_2", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_2.feed_forward.w_1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_2.self_attn.linear_pos", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_2.self_attn.linear_out", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_2.self_attn.linear_v", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_2.self_attn.linear_k", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_2.self_attn.linear_q", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_1.feed_forward.w_2", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_1.feed_forward.w_1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_1.self_attn.linear_pos", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_1.self_attn.linear_out", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_1.self_attn.linear_v", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_1.self_attn.linear_k", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_1.self_attn.linear_q", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_0.feed_forward.w_2", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_0.feed_forward.w_1", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_0.self_attn.linear_pos", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_0.self_attn.linear_out", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_0.self_attn.linear_v", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_0.self_attn.linear_k", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.encoders_0.self_attn.linear_q", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.encoder.embed.linear", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.spk_embed_affine_layer", "bits": 6, "group_size": 64 }, { "name": "s3gen.flow.input_embedding", "bits": 6, "group_size": 64 }, { "name": "t3.speech_head", "bits": 6, "group_size": 64 }, { "name": "t3.text_head", "bits": 6, "group_size": 64 }, { "name": "t3.speech_pos_emb.emb", "bits": 6, "group_size": 64 }, { "name": "t3.text_pos_emb.emb", "bits": 6, "group_size": 64 }, { "name": "t3.speech_emb", "bits": 6, "group_size": 64 }, { "name": "t3.text_emb", "bits": 6, "group_size": 64 }, { "name": "t3.cond_enc.perceiver.attn.proj_out", "bits": 6, "group_size": 64 }, { "name": "t3.cond_enc.perceiver.attn.to_v", "bits": 6, "group_size": 64 }, { "name": "t3.cond_enc.perceiver.attn.to_k", "bits": 6, "group_size": 64 }, { "name": "t3.cond_enc.perceiver.attn.to_q", "bits": 6, "group_size": 64 }, { "name": "t3.cond_enc.spkr_enc", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.29.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.29.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.29.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.29.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.29.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.29.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.29.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.28.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.28.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.28.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.28.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.28.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.28.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.28.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.27.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.27.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.27.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.27.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.27.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.27.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.27.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.26.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.26.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.26.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.26.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.26.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.26.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.26.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.25.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.25.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.25.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.25.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.25.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.25.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.25.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.24.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.24.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.24.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.24.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.24.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.24.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.24.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.23.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.23.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.23.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.23.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.23.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.23.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.23.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.22.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.22.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.22.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.22.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.22.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.22.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.22.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.21.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.21.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.21.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.21.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.21.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.21.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.21.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.20.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.20.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.20.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.20.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.20.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.20.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.20.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.19.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.19.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.19.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.19.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.19.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.19.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.19.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.18.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.18.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.18.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.18.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.18.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.18.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.18.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.17.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.17.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.17.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.17.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.17.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.17.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.17.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.16.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.16.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.16.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.16.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.16.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.16.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.16.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.15.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.15.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.15.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.15.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.15.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.15.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.15.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.14.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.14.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.14.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.14.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.14.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.14.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.14.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.13.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.13.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.13.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.13.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.13.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.13.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.13.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.12.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.12.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.12.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.12.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.12.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.12.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.12.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.11.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.11.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.11.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.11.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.11.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.11.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.11.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.10.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.10.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.10.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.10.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.10.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.10.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.10.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.9.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.9.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.9.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.9.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.9.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.9.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.9.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.8.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.8.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.8.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.8.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.8.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.8.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.8.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.7.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.7.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.7.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.7.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.7.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.7.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.7.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.6.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.6.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.6.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.6.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.6.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.6.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.6.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.5.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.5.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.5.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.5.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.5.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.5.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.5.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.4.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.4.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.4.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.4.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.4.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.4.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.4.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.3.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.3.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.3.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.3.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.3.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.3.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.3.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.2.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.2.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.2.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.2.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.2.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.2.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.2.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.1.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.1.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.1.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.1.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.1.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.1.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.1.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.0.mlp.up_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.0.mlp.down_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.0.mlp.gate_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.0.self_attn.o_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.0.self_attn.v_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.0.self_attn.k_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.layers.0.self_attn.q_proj", "bits": 6, "group_size": 64 }, { "name": "t3.tfmr.model.embed_tokens", "bits": 6, "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": "q6-warm-0.wav", "sha256": "4d2ba5abb77f6206e76128aba2cae457b368fd0bc487a45591c68d44e9902531", "float_sha256": "887ae4b3e52014fdfad15c32ef9a4c1576f87975cacce52525b765d9d232e161", "generation_seconds": 5.849353499987046, "audio_seconds": 8.16, "rtf": 0.7168325367631184, "mlx_peak_bytes": 2357940084, "rms": 0.0703206882000212, "peak": 0.6385914087295532, "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": "q6-warm-1.wav", "sha256": "02e965ea29137a6cd03404526af5f21e14ce9709c13704459eb6358c81723b9f", "float_sha256": "b848e80f098df3671e1c4f8ad92de29799aa589e27fac126be9022ed23775dd5", "generation_seconds": 5.635674917008146, "audio_seconds": 8.16, "rtf": 0.690646435907861, "mlx_peak_bytes": 2357940084, "rms": 0.07032068818993092, "peak": 0.6385913491249084, "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": "q6-warm-2.wav", "sha256": "287dc3b9c6d9e954070b60cf45b8bc8804dcfca9d747fb93eb6a01c93d08f80e", "float_sha256": "8d3fa5ccd8b2f2807832f5c967204a152325f24ee5f0137d945ca416f18a75a5", "generation_seconds": 5.678383709004265, "audio_seconds": 8.16, "rtf": 0.6958803564956207, "mlx_peak_bytes": 2357940084, "rms": 0.07032068817011418, "peak": 0.6385914087295532, "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": "q6-warm-3.wav", "sha256": "d66d2ed0461ad6e84d7c02c580bb8a91744ec89cc9ac2f01b93ab50a20f10102", "float_sha256": "b6c37f6d82ce3045f10a4391e943ba65b0ff37bb8ffd47227e8dd4408cf371bc", "generation_seconds": 5.838923875009641, "audio_seconds": 8.16, "rtf": 0.7155543964472599, "mlx_peak_bytes": 2357940084, "rms": 0.07032068819024384, "peak": 0.6385913491249084, "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": "q6-warm-4.wav", "sha256": "8c831438343a3dd297ac70af4e15d1d9bdea14593f7c938e18a79ff227be54ec", "float_sha256": "45730090417492647d6516a5065ad857ff50405e434f7f6f782a30ab0d6d3d85", "generation_seconds": 7.746913541996037, "audio_seconds": 8.16, "rtf": 0.9493766595583378, "mlx_peak_bytes": 2357940084, "rms": 0.07032068819328598, "peak": 0.638591468334198, "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": "q6-warm-5.wav", "sha256": "f79ce5c865c8d4a659ce6f1ebe6cd48255bd6b8997c9c02e0fb537e636c4cf47", "float_sha256": "1c11a174eac69e22ce0b91f561b00b9a87313194b8a5db6577d3242b0c0dc79f", "generation_seconds": 14.325836958989385, "audio_seconds": 8.16, "rtf": 1.755617274385954, "mlx_peak_bytes": 2357940084, "rms": 0.07032068822397626, "peak": 0.6385914087295532, "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": "q6-e2e-0-20261009.wav", "sha256": "7db3a3c56c0bd074717bde8bb232e87386c9bf3855453eb27ae5ac7be2de2896", "float_sha256": "8e9766097587deeefe0812998a8d8afef14d79f3fdeff98657ac584ae56fb228", "generation_seconds": 12.01914650000981, "audio_seconds": 8.16, "rtf": 1.4729346200992415, "mlx_peak_bytes": 2358105980, "rms": 0.07032068816482084, "peak": 0.6385914087295532, "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": 204, "speech_token_sha256": "7ed77514ea2554b589685c0af1db7af5ec1e08aefa1f9901aaa9dd596fd20045", "timing_includes_token_capture": true }, { "file": "q6-e2e-0-42.wav", "sha256": "d363fbfd30c5ed7a3ca99a8b3aa9b9f5ff03241750c7a4484942c9a894bcef8f", "float_sha256": "f01a354762481166ce32ba11cf0d62526c22b17672a28058589800ab7ee71402", "generation_seconds": 12.55688133300282, "audio_seconds": 7.68, "rtf": 1.6350105902347423, "mlx_peak_bytes": 2368620412, "rms": 0.0695731745891047, "peak": 0.6953259110450745, "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": 192, "speech_token_sha256": "6d96bde1bacfad079f0627bcfbf55b7d602b5f5f85f4277010a1d6fe3ad38969", "timing_includes_token_capture": true }, { "file": "q6-e2e-1-20261009.wav", "sha256": "9a1aa87bb1e82a5f1c1c83bc9f6d29669db41b1d90891507d0aa8daaf19a3dbb", "float_sha256": "554db80c4bdfc0f178282c2dcc2efb4490602e15d3e47e336317ece7a3b57a20", "generation_seconds": 12.766279749994283, "audio_seconds": 7.56, "rtf": 1.6886613425918364, "mlx_peak_bytes": 2351507764, "rms": 0.07092462632610838, "peak": 0.631426215171814, "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": 189, "speech_token_sha256": "e1e7b029b5515302998aeb9ea96bb50f012655eb011e4c2147af906b1057835f", "timing_includes_token_capture": true }, { "file": "q6-e2e-1-42.wav", "sha256": "095678cad0ef3103908070117d8777bb753d3369f2d5225cd9115cd38a5f8555", "float_sha256": "b051799abbbac9722cf540152b5c2b2bfd3c6a07b8c97c1ad94d1088cf6f683f", "generation_seconds": 8.369348958993214, "audio_seconds": 5.24, "rtf": 1.5972039998078653, "mlx_peak_bytes": 2087780112, "rms": 0.07740916808939496, "peak": 0.5414543151855469, "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": 131, "speech_token_sha256": "2cb90e9898cd8e5bf500d6d8acd2c4d35a7214754c1b41f8b402a00fdbf39ed3", "timing_includes_token_capture": true }, { "file": "q6-e2e-2-20261009.wav", "sha256": "c89d23eb0517a995e24ac7941092abc0f2fed0de696012009c91347b7e4d2c8d", "float_sha256": "814cf4a5efd5c30eeebd6649a0812293b49285f2e962dce26894702af9f953a4", "generation_seconds": 13.923164959007408, "audio_seconds": 6.24, "rtf": 2.2312764357383665, "mlx_peak_bytes": 2201189292, "rms": 0.07404004689237437, "peak": 0.5328736901283264, "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": 156, "speech_token_sha256": "c3b0e506e528ccf443ad7d59c30052d4e0d9b2405e3f4c915bc99c4f4c8c561e", "timing_includes_token_capture": true }, { "file": "q6-e2e-2-42.wav", "sha256": "e4f6a7ce83682c90be86049fd9ab9e215c3fd0db0498156682a73bb5afc25104", "float_sha256": "75dc75bce179152614f4a32abfa5577db32c28b76ac780db94d4256321acc8b3", "generation_seconds": 10.565146042004926, "audio_seconds": 5.52, "rtf": 1.9139757322472695, "mlx_peak_bytes": 2117791696, "rms": 0.0671833139561052, "peak": 0.5190531015396118, "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": 138, "speech_token_sha256": "22d74cf6e1363329233e30f3cb2bce8f0c28b4db7a7b4f60e9645dbbc166cf97", "timing_includes_token_capture": true }, { "file": "q6-e2e-3-20261009.wav", "sha256": "52ff7e1415a5e84d44e950ecb759cf5febb5477487de5fa693a7e67d34f7a85e", "float_sha256": "b4d78c48f727343a81b7fdf8cf96579d633d8be2d97fd245282025890025c904", "generation_seconds": 17.865502125001512, "audio_seconds": 8.68, "rtf": 2.0582375720047827, "mlx_peak_bytes": 2404052804, "rms": 0.06801295676034007, "peak": 0.5194692015647888, "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": 217, "speech_token_sha256": "c2b3c4049e779dc3d7186a0809dfbbbce887428f82b4d9d440c8ad0f9734ad23", "timing_includes_token_capture": true }, { "file": "q6-e2e-3-42.wav", "sha256": "17f8828f7acdc2c88d1f38229f765e185ca9b474374940afa2d195f5a1903c44", "float_sha256": "3fad7af2984b47e9dab8cb6d8182022549a561857002124bc316d179fd76b2bb", "generation_seconds": 19.795125750009902, "audio_seconds": 8.56, "rtf": 2.3125146904217173, "mlx_peak_bytes": 2395430036, "rms": 0.07739094059282195, "peak": 0.6726484298706055, "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": 214, "speech_token_sha256": "37565725c6c00dabae33788ef392bcbb3fc8e472c59b2ba8f61f21aec0d6e87f", "timing_includes_token_capture": true }, { "file": "q6-e2e-4-20261009.wav", "sha256": "3fb0186ad1a048ad45520b8fa04fd9d9a317ff05ea5606ee3408ada12ee24eba", "float_sha256": "303ceb1cdaa069ef8c9bbfb855f108524732b375451f8a70e69536d8030a3c90", "generation_seconds": 28.033478792000096, "audio_seconds": 16.0, "rtf": 1.752092424500006, "mlx_peak_bytes": 2726726812, "rms": 0.04827491137621913, "peak": 0.543483316898346, "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": 400, "speech_token_sha256": "b8cfc930fdb0bb227e7c3eb3b963984290d8515c1ba1ab5a227def72bc7299b3", "timing_includes_token_capture": true }, { "file": "q6-e2e-4-42.wav", "sha256": "9cc42fc943f17b11d0afdb679ff2be771587d2fdaf234f2342675423e38d5a7e", "float_sha256": "feec941fb9078fb5784f32e34d355dfa1ddc997ff49c5b7451e82827917b1882", "generation_seconds": 11.545202250010334, "audio_seconds": 6.0, "rtf": 1.9242003750017223, "mlx_peak_bytes": 2175808940, "rms": 0.07766717742671307, "peak": 0.5446796417236328, "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": 150, "speech_token_sha256": "15f27c28fcdbb5f5a4224caca9f13c734470774317085d104f5274841b811a62", "timing_includes_token_capture": true }, { "file": "q6-e2e-5-20261009.wav", "sha256": "157942a9e470658a090a87545a6837cfe554dbf1ed6f0179887191e8dd6ebfad", "float_sha256": "023f18f7fe04079b242cac67440a210ef8620ff7609c65266e86066cad5e5409", "generation_seconds": 13.617624666992924, "audio_seconds": 6.32, "rtf": 2.1546874473090067, "mlx_peak_bytes": 2214764236, "rms": 0.08740317694195888, "peak": 0.6582881808280945, "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": 158, "speech_token_sha256": "09db590bfd06bcb83404264752971ca8a075c6473184d8cd3414a6f31ebb4db2", "timing_includes_token_capture": true }, { "file": "q6-e2e-5-42.wav", "sha256": "ff2276b52da9ad4ffeb80289768de3f1632a501322a16124d7c4835b2f22c9fe", "float_sha256": "743ba93c0a2a699ae8671bbedf8a8f39f32c7064a2e6e2c1bcc6849b69120be2", "generation_seconds": 18.807340041996213, "audio_seconds": 11.32, "rtf": 1.6614257987629162, "mlx_peak_bytes": 2507828622, "rms": 0.039381651241290026, "peak": 0.45455706119537354, "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": 283, "speech_token_sha256": "47e45c0538a2a36ebbe2354fa03e476c63308bab84e5fd3151f26f9681ac342b", "timing_includes_token_capture": true } ], "controlled_decoder": [ { "file": "q6-decoder-0-20261009.wav", "sha256": "c9e173d2de253049d732268cd329af5b3a1a5d1a46bb86ece3e9dbd31353488a", "float_sha256": "6b42a593fd7318bb433510df1f5ad22353e8930a16aade504b7c900a02a82280", "generation_seconds": 4.7368449160130695, "audio_seconds": 8.48, "rtf": 0.5585902023600318, "mlx_peak_bytes": 2391366532, "rms": 0.06878221431280203, "peak": 0.5855569243431091, "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": "q6-decoder-0-42.wav", "sha256": "d47e1eae8f70014821d1ef7a1c073e1b01b19a9709063c132f45f8523b1e1408", "float_sha256": "36795c99fa2166cc106355cc6a187033f97e336b97635adc492025a207b5b31e", "generation_seconds": 4.2490122910094215, "audio_seconds": 7.84, "rtf": 0.5419658534450793, "mlx_peak_bytes": 2393343828, "rms": 0.06906707856186947, "peak": 0.5365857481956482, "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": "q6-decoder-1-20261009.wav", "sha256": "589cb88c8c6ae9cac52ac8a7e7255284c76bcdd4f8b09aba58a2342f84bcebdd", "float_sha256": "d5f0bd7e0ce178afca746c224a39188ac6424fddb273551b86bedb537c9db817", "generation_seconds": 3.209237082992331, "audio_seconds": 5.8, "rtf": 0.5533167384469536, "mlx_peak_bytes": 2147532144, "rms": 0.0797693398520808, "peak": 0.6153287291526794, "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": "q6-decoder-1-42.wav", "sha256": "711a045b76bcf429e4211d4913a0198f8aaf5ea623de2d621833afede43e9571", "float_sha256": "6c24a381eb86e6ee278de1827b718038d45853abc3662295f1fe3bc9c5bfb420", "generation_seconds": 3.4158001670002704, "audio_seconds": 5.6, "rtf": 0.6099643155357626, "mlx_peak_bytes": 2123217748, "rms": 0.08018098276436847, "peak": 0.5273110866546631, "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": "q6-decoder-2-20261009.wav", "sha256": "1532ac185a9c955d08f87aed848ff310f73c18fd8b54028c9d7fa973c5756f17", "float_sha256": "a1fe1584fd6aa6f32a3480818b3947a4f094ed86cff808487be6508498477642", "generation_seconds": 3.640969957996276, "audio_seconds": 5.72, "rtf": 0.6365332094399084, "mlx_peak_bytes": 2138881164, "rms": 0.07493713008270449, "peak": 0.5467548370361328, "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": "q6-decoder-2-42.wav", "sha256": "06b47eb40b11a134e054698216629467ee7e79324d5be2d280054196ebaad7cd", "float_sha256": "110aa296372d907d72ebaa2352f14302eda99e701c06b9399cc79ab991da3df6", "generation_seconds": 2.798599000001559, "audio_seconds": 5.32, "rtf": 0.5260524436093156, "mlx_peak_bytes": 2090186876, "rms": 0.075133834226878, "peak": 0.5394142270088196, "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": "q6-decoder-3-20261009.wav", "sha256": "2df12eb007655ad79a780981f1c9e24dc3b9daeb390cd80c0040fc6e69e4e35b", "float_sha256": "1a0a5c3b5989fb9bb7a4e0acc90144850db72b09d745bab662b23bc7a2706e01", "generation_seconds": 3.2694579580129357, "audio_seconds": 6.52, "rtf": 0.5014506070571988, "mlx_peak_bytes": 2235639884, "rms": 0.08740238392193435, "peak": 0.6106457710266113, "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": "q6-decoder-3-42.wav", "sha256": "95ac4ddcff25f3fa79f7452f7bab690cf9975f9b8a59978a6865cad4eac87a80", "float_sha256": "9351b26c9a3a854dfa31e74212792514a8d4be31659d12f66a96892d06a85d5c", "generation_seconds": 3.317587666999316, "audio_seconds": 6.4, "rtf": 0.5183730729686431, "mlx_peak_bytes": 2219435556, "rms": 0.0819871067186929, "peak": 0.586404025554657, "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": "q6-decoder-4-20261009.wav", "sha256": "b488c37223ac8447c343e697a90d67d69fc145f941814b716cf71c677eba196b", "float_sha256": "d578e101caa217dbdca9bba25c27ebc4fb2d243df5a9506a2568f12246d3f9d0", "generation_seconds": 2.920131584003684, "audio_seconds": 6.08, "rtf": 0.4802848000006059, "mlx_peak_bytes": 2181629716, "rms": 0.08650933380179598, "peak": 0.7735304236412048, "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": "q6-decoder-4-42.wav", "sha256": "1f1596b372244544a024b7703f3518c2a54f85e3b2646cf0a34f46dccff9daed", "float_sha256": "b4c7146c13bafc95d25bd9ab0b320ab8e346cf27b307b73a3c75f89d6b857a63", "generation_seconds": 3.0369312920083757, "audio_seconds": 6.16, "rtf": 0.4930083266247363, "mlx_peak_bytes": 2190755780, "rms": 0.08165177386751649, "peak": 0.7709804177284241, "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": "q6-decoder-5-20261009.wav", "sha256": "47d863d485cf7743637c164b4cae1b12b4f52089be8b9fcb26b74a50f58db37b", "float_sha256": "554ddbec8ba08eb31da572c232c4daf2544faa50bd1d1d87459c1e5ec3d5e5b0", "generation_seconds": 3.371877208002843, "audio_seconds": 6.08, "rtf": 0.5545850671057307, "mlx_peak_bytes": 2181629716, "rms": 0.09677324552799009, "peak": 0.7728903889656067, "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": "q6-decoder-5-42.wav", "sha256": "4a8fc77175878a3164bbf8456a2908c23d03175ca5faf777fc03c2b7d0ade851", "float_sha256": "060b9db5d4855e9f65240e81ef9285f3feff77736b65fa73bb0794327dd4fc5e", "generation_seconds": 4.537853374989936, "audio_seconds": 8.28, "rtf": 0.5480499245156928, "mlx_peak_bytes": 2370656416, "rms": 0.08213864372641763, "peak": 0.9573540687561035, "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": 2726726812, "process_peak_rss_bytes": 1462452224, "warm_median_seconds": 5.838923875009641, "warm_min_seconds": 5.635674917008146, "warm_max_seconds": 14.325836958989385, "warm_median_rtf": 0.7155543964472599, "warm_same_audio": false, "warm_profile_peak_mlx_bytes": 2357940084, "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." }