Chatterbox fine-tuned model + logs

training.log

@@ -5,199 +5,513 @@ Resolved paths:
 
 Starting training...
 
-02/10/2026 03:45:28 - INFO - __main__ - Training/evaluation parameters CustomTrainingArguments(
-accelerator_config={'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None, 'use_configured_state': False},
-adam_beta1=0.9,
-adam_beta2=0.999,
-adam_epsilon=1e-08,
-auto_find_batch_size=False,
-average_tokens_across_devices=True,
-batch_eval_metrics=False,
-bf16=False,
-bf16_full_eval=False,
-data_seed=None,
-dataloader_drop_last=False,
-dataloader_num_workers=8,
-dataloader_persistent_workers=False,
-dataloader_pin_memory=False,
-dataloader_prefetch_factor=None,
-ddp_backend=None,
-ddp_broadcast_buffers=None,
-ddp_bucket_cap_mb=None,
-ddp_find_unused_parameters=None,
-ddp_timeout=1800,
-debug=[],
-deepspeed=None,
-disable_tqdm=False,
-do_eval=True,
-do_predict=False,
-do_train=True,
-early_stopping_patience=None,
-enable_jit_checkpoint=False,
-eval_accumulation_steps=None,
-eval_delay=0,
-eval_do_concat_batches=True,
-eval_on_start=True,
-eval_steps=2,
-eval_strategy=IntervalStrategy.STEPS,
-eval_use_gather_object=False,
-fp16=True,
-fp16_full_eval=False,
-fsdp=[],
-fsdp_config={'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False},
-full_determinism=False,
-gradient_accumulation_steps=1,
-gradient_checkpointing=False,
-gradient_checkpointing_kwargs=None,
-greater_is_better=None,
-group_by_length=False,
-hub_always_push=False,
-hub_model_id=None,
-hub_private_repo=None,
-hub_revision=None,
-hub_strategy=HubStrategy.EVERY_SAVE,
-hub_token=<HUB_TOKEN>,
-ignore_data_skip=False,
-include_for_metrics=[],
-include_num_input_tokens_seen=no,
-label_names=['labels_speech'],
-label_smoothing_factor=0.0,
-learning_rate=5e-05,
-length_column_name=length,
-liger_kernel_config=None,
-load_best_model_at_end=False,
-local_rank=-1,
-log_level=passive,
-log_level_replica=warning,
-log_on_each_node=True,
-logging_dir=None,
-logging_first_step=False,
-logging_nan_inf_filter=True,
-logging_steps=1.0,
-logging_strategy=IntervalStrategy.STEPS,
-lr_scheduler_kwargs=None,
-lr_scheduler_type=SchedulerType.LINEAR,
-max_grad_norm=1.0,
-max_steps=-1,
-metric_for_best_model=None,
-neftune_noise_alpha=None,
-num_train_epochs=1.0,
-optim=OptimizerNames.ADAMW_TORCH_FUSED,
-optim_args=None,
-optim_target_modules=None,
-output_dir=/app/checkpoints/v1,
-parallelism_config=None,
-per_device_eval_batch_size=8,
-per_device_train_batch_size=1,
-prediction_loss_only=False,
-project=huggingface,
-push_to_hub=False,
-remove_unused_columns=True,
-report_to=['tensorboard'],
-restore_callback_states_from_checkpoint=False,
-resume_from_checkpoint=None,
-run_name=None,
-save_on_each_node=False,
-save_only_model=False,
-save_steps=1.0,
-save_strategy=SaveStrategy.STEPS,
-save_total_limit=1,
-seed=42,
-skip_memory_metrics=True,
-tf32=None,
-torch_compile=False,
-torch_compile_backend=None,
-torch_compile_mode=None,
-torch_empty_cache_steps=None,
-trackio_space_id=trackio,
-use_cache=False,
-use_cpu=False,
-use_liger_kernel=False,
-warmup_ratio=None,
-warmup_steps=1.0,
-weight_decay=0.0,
-)
-02/10/2026 03:45:28 - INFO - __main__ - Model parameters ModelArguments(model_name_or_path='/app/hf_cache/models--ResembleAI--chatterbox/snapshots/05e904af2b5c7f8e482687a9d7336c5c824467d9', local_model_dir=None, cache_dir=None, freeze_voice_encoder=True, freeze_s3gen=True)
-02/10/2026 03:45:28 - INFO - __main__ - Data parameters DataArguments(language='hi', dataset_dir=None, metadata_file=None, dataset_name='rahul7star/hindi-speech-dataset', dataset_config_name=None, train_split_name='train', eval_split_name='validation', text_column_name='transcript', audio_column_name='audio', max_text_len=256, max_speech_len=800, audio_prompt_duration_s=3.0, eval_split_size=0.0002, preprocessing_num_workers=None, ignore_verifications=False)
-02/10/2026 03:45:28 - INFO - __main__ - Loading ChatterboxTTS model...
-02/10/2026 03:45:28 - INFO - __main__ - Loading model from Hugging Face Hub: /app/hf_cache/models--ResembleAI--chatterbox/snapshots/05e904af2b5c7f8e482687a9d7336c5c824467d9
-02/10/2026 03:45:28 - WARNING - __main__ - Could not download ve.safetensors from /app/hf_cache/models--ResembleAI--chatterbox/snapshots/05e904af2b5c7f8e482687a9d7336c5c824467d9: Repo id must be in the form 'repo_name' or 'namespace/repo_name': '/app/hf_cache/models--ResembleAI--chatterbox/snapshots/05e904af2b5c7f8e482687a9d7336c5c824467d9'. Use `repo_type` argument if needed..
-02/10/2026 03:45:28 - WARNING - __main__ - Could not download t3_mtl23ls_v2.safetensors from /app/hf_cache/models--ResembleAI--chatterbox/snapshots/05e904af2b5c7f8e482687a9d7336c5c824467d9: Repo id must be in the form 'repo_name' or 'namespace/repo_name': '/app/hf_cache/models--ResembleAI--chatterbox/snapshots/05e904af2b5c7f8e482687a9d7336c5c824467d9'. Use `repo_type` argument if needed..
-02/10/2026 03:45:28 - WARNING - __main__ - Could not download s3gen.safetensors from /app/hf_cache/models--ResembleAI--chatterbox/snapshots/05e904af2b5c7f8e482687a9d7336c5c824467d9: Repo id must be in the form 'repo_name' or 'namespace/repo_name': '/app/hf_cache/models--ResembleAI--chatterbox/snapshots/05e904af2b5c7f8e482687a9d7336c5c824467d9'. Use `repo_type` argument if needed..
-02/10/2026 03:45:28 - WARNING - __main__ - Could not download mtl_tokenizer.json from /app/hf_cache/models--ResembleAI--chatterbox/snapshots/05e904af2b5c7f8e482687a9d7336c5c824467d9: Repo id must be in the form 'repo_name' or 'namespace/repo_name': '/app/hf_cache/models--ResembleAI--chatterbox/snapshots/05e904af2b5c7f8e482687a9d7336c5c824467d9'. Use `repo_type` argument if needed..
-02/10/2026 03:45:28 - INFO - __main__ - conds.pt not found on Hub or failed to download for this model.
-02/10/2026 03:45:28 - INFO - httpx - HTTP Request: GET https://huggingface.co/api/models/ResembleAI/chatterbox/revision/main "HTTP/1.1 200 OK"
 
 
-Downloading (incomplete total...): 0.00B [00:00, ?B/s]
+README.md:   0%|          | 0.00/591 [00:00<?, ?B/s]
+README.md: 100%|██████████| 591/591 [00:00<00:00, 3.34MB/s]
 
-Fetching 5 files:   0%|          | 0/5 [00:00<?, ?it/s]02/10/2026 03:45:28 - INFO - httpx - HTTP Request: HEAD https://huggingface.co/ResembleAI/chatterbox/resolve/05e904af2b5c7f8e482687a9d7336c5c824467d9/conds.pt "HTTP/1.1 302 Found"
-02/10/2026 03:45:28 - INFO - httpx - HTTP Request: HEAD https://huggingface.co/ResembleAI/chatterbox/resolve/05e904af2b5c7f8e482687a9d7336c5c824467d9/grapheme_mtl_merged_expanded_v1.json "HTTP/1.1 307 Temporary Redirect"
-02/10/2026 03:45:28 - INFO - httpx - HTTP Request: HEAD https://huggingface.co/ResembleAI/chatterbox/resolve/05e904af2b5c7f8e482687a9d7336c5c824467d9/s3gen.pt "HTTP/1.1 302 Found"
-02/10/2026 03:45:28 - INFO - httpx - HTTP Request: HEAD https://huggingface.co/ResembleAI/chatterbox/resolve/05e904af2b5c7f8e482687a9d7336c5c824467d9/ve.pt "HTTP/1.1 302 Found"
-02/10/2026 03:45:28 - INFO - httpx - HTTP Request: HEAD https://huggingface.co/ResembleAI/chatterbox/resolve/05e904af2b5c7f8e482687a9d7336c5c824467d9/t3_mtl23ls_v2.safetensors "HTTP/1.1 302 Found"
-02/10/2026 03:45:28 - INFO - httpx - HTTP Request: GET https://huggingface.co/api/models/ResembleAI/chatterbox/xet-read-token/05e904af2b5c7f8e482687a9d7336c5c824467d9 "HTTP/1.1 200 OK"
+Downloading data:   0%|          | 0/36 [00:00<?, ?files/s]
 
+data/train-00000-of-00036.parquet:   0%|          | 0.00/1.77G [00:00<?, ?B/s]
 
-Downloading (incomplete total...):   0%|          | 0.00/107k [00:00<?, ?B/s]02/10/2026 03:45:28 - INFO - httpx - HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/ResembleAI/chatterbox/05e904af2b5c7f8e482687a9d7336c5c824467d9/grapheme_mtl_merged_expanded_v1.json "HTTP/1.1 200 OK"
-02/10/2026 03:45:28 - INFO - httpx - HTTP Request: GET https://huggingface.co/api/models/ResembleAI/chatterbox/xet-read-token/05e904af2b5c7f8e482687a9d7336c5c824467d9 "HTTP/1.1 200 OK"
+data/train-00000-of-00036.parquet:   0%|          | 11.6k/1.77G [00:01<44:03:44, 11.1kB/s]
 
+data/train-00000-of-00036.parquet:  11%|█▏        | 203M/1.77G [00:02<00:13, 117MB/s]     
 
-Downloading (incomplete total...):   0%|          | 0.00/5.81M [00:00<?, ?B/s]02/10/2026 03:45:28 - INFO - httpx - HTTP Request: GET https://huggingface.co/api/models/ResembleAI/chatterbox/xet-read-token/05e904af2b5c7f8e482687a9d7336c5c824467d9 "HTTP/1.1 200 OK"
+data/train-00000-of-00036.parquet:  34%|███▍      | 598M/1.77G [00:03<00:04, 243MB/s]
 
+data/train-00000-of-00036.parquet:  59%|█████▉    | 1.04G/1.77G [00:04<00:02, 319MB/s]
 
-Downloading (incomplete total...):   0%|          | 0.00/1.06G [00:00<?, ?B/s]02/10/2026 03:45:28 - INFO - httpx - HTTP Request: GET https://huggingface.co/api/models/ResembleAI/chatterbox/xet-read-token/05e904af2b5c7f8e482687a9d7336c5c824467d9 "HTTP/1.1 200 OK"
+data/train-00000-of-00036.parquet:  86%|████████▌ | 1.52G/1.77G [00:05<00:00, 376MB/s]
+data/train-00000-of-00036.parquet: 100%|██████████| 1.77G/1.77G [00:05<00:00, 336MB/s]
 
+Downloading data:   3%|▎         | 1/36 [00:05<03:05,  5.31s/files]
 
-Downloading (incomplete total...):   0%|          | 0.00/3.21G [00:00<?, ?B/s]02/10/2026 03:45:28 - INFO - httpx - HTTP Request: GET https://huggingface.co/api/resolve-cache/models/ResembleAI/chatterbox/05e904af2b5c7f8e482687a9d7336c5c824467d9/grapheme_mtl_merged_expanded_v1.json "HTTP/1.1 200 OK"
+data/train-00001-of-00036.parquet:   0%|          | 0.00/496M [00:00<?, ?B/s]
 
+data/train-00001-of-00036.parquet:   5%|▍         | 23.6M/496M [00:01<00:20, 23.6MB/s]
+data/train-00001-of-00036.parquet: 100%|██████████| 496M/496M [00:01<00:00, 263MB/s]  
 
-Downloading (incomplete total...):   0%|          | 0.00/3.21G [00:00<?, ?B/s]
+Downloading data:   6%|▌         | 2/36 [00:07<01:52,  3.31s/files]
+
+data/train-00002-of-00036.parquet:   0%|          | 0.00/461M [00:00<?, ?B/s]
+
+data/train-00002-of-00036.parquet:  15%|█▍        | 68.8M/461M [00:01<00:05, 68.6MB/s]
+data/train-00002-of-00036.parquet: 100%|█████████���| 461M/461M [00:01<00:00, 251MB/s]  
+
+Downloading data:   8%|▊         | 3/36 [00:09<01:27,  2.65s/files]
+
+data/train-00003-of-00036.parquet:   0%|          | 0.00/372M [00:00<?, ?B/s]
+
+data/train-00003-of-00036.parquet:  10%|█         | 37.5M/372M [00:01<00:08, 37.5MB/s]
+data/train-00003-of-00036.parquet: 100%|██████████| 372M/372M [00:01<00:00, 214MB/s]  
+
+Downloading data:  11%|█         | 4/36 [00:10<01:13,  2.30s/files]
+
+data/train-00004-of-00036.parquet:   0%|          | 0.00/373M [00:00<?, ?B/s]
+
+data/train-00004-of-00036.parquet:  10%|█         | 39.0M/373M [00:01<00:08, 38.8MB/s]
+data/train-00004-of-00036.parquet: 100%|██████████| 373M/373M [00:01<00:00, 231MB/s]  
+
+Downloading data:  14%|█▍        | 5/36 [00:12<01:03,  2.06s/files]
+
+data/train-00005-of-00036.parquet:   0%|          | 0.00/841M [00:00<?, ?B/s]
+
+data/train-00005-of-00036.parquet:   5%|▍         | 39.4M/841M [00:01<00:20, 39.4MB/s]
+
+data/train-00005-of-00036.parquet:  39%|███▉      | 330M/841M [00:02<00:02, 187MB/s]  
+
+data/train-00005-of-00036.parquet: 100%|██████████| 841M/841M [00:03<00:00, 286MB/s]
+data/train-00005-of-00036.parquet: 100%|██████████| 841M/841M [00:03<00:00, 249MB/s]
+
+Downloading data:  17%|█▋        | 6/36 [00:15<01:15,  2.52s/files]
+
+data/train-00006-of-00036.parquet:   0%|          | 0.00/837M [00:00<?, ?B/s]
+
+data/train-00006-of-00036.parquet:   3%|▎         | 29.3M/837M [00:01<00:27, 29.2MB/s]
+
+data/train-00006-of-00036.parquet:  39%|███▉      | 325M/837M [00:02<00:02, 186MB/s]  
+
+data/train-00006-of-00036.parquet: 100%|██████████| 837M/837M [00:03<00:00, 310MB/s]
+data/train-00006-of-00036.parquet: 100%|██████████| 837M/837M [00:03<00:00, 264MB/s]
+
+Downloading data:  19%|█▉        | 7/36 [00:19<01:19,  2.74s/files]
+
+data/train-00007-of-00036.parquet:   0%|          | 0.00/618M [00:00<?, ?B/s]
+
+data/train-00007-of-00036.parquet:   5%|▍         | 28.5M/618M [00:01<00:20, 28.4MB/s]
+
+data/train-00007-of-00036.parquet:  53%|█████▎    | 328M/618M [00:02<00:01, 187MB/s]  
+
+data/train-00007-of-00036.parquet: 100%|██████████| 618M/618M [00:03<00:00, 203MB/s]
+data/train-00007-of-00036.parquet: 100%|██████████| 618M/618M [00:03<00:00, 185MB/s]
+
+Downloading data:  22%|██▏       | 8/36 [00:22<01:22,  2.94s/files]
+
+data/train-00008-of-00036.parquet:   0%|          | 0.00/415M [00:00<?, ?B/s]
+
+data/train-00008-of-00036.parquet:   7%|▋         | 30.4M/415M [00:01<00:12, 30.4MB/s]
+data/train-00008-of-00036.parquet: 100%|██████████| 415M/415M [00:01<00:00, 247MB/s]  
+
+Downloading data:  25%|██▌       | 9/36 [00:24<01:08,  2.55s/files]
+
+data/train-00009-of-00036.parquet:   0%|          | 0.00/303M [00:00<?, ?B/s]
+
+data/train-00009-of-00036.parquet:   4%|▍         | 12.7M/303M [00:01<00:23, 12.6MB/s]
+data/train-00009-of-00036.parquet: 100%|██████████| 303M/303M [00:01<00:00, 196MB/s]  
+
+Downloading data:  28%|██▊       | 10/36 [00:25<00:58,  2.25s/files]
+
+data/train-00010-of-00036.parquet:   0%|          | 0.00/341M [00:00<?, ?B/s]
+
+data/train-00010-of-00036.parquet:  20%|█▉        | 67.3M/341M [00:01<00:04, 66.8MB/s]
+
+data/train-00010-of-00036.parquet: 100%|██████████| 341M/341M [00:02<00:00, 180MB/s]  
+data/train-00010-of-00036.parquet: 100%|██████████| 341M/341M [00:02<00:00, 163MB/s]
+
+Downloading data:  31%|███       | 11/36 [00:27<00:55,  2.21s/files]
+
+data/train-00011-of-00036.parquet:   0%|          | 0.00/440M [00:00<?, ?B/s]
+
+data/train-00011-of-00036.parquet:   5%|▍         | 19.9M/440M [00:01<00:21, 19.9MB/s]
+
+data/train-00011-of-00036.parquet: 100%|██████████| 440M/440M [00:02<00:00, 253MB/s]  
+data/train-00011-of-00036.parquet: 100%|██████████| 440M/440M [00:02<00:00, 218MB/s]
+
+Downloading data:  33%|███▎      | 12/36 [00:29<00:52,  2.17s/files]
+
+data/train-00012-of-00036.parquet:   0%|          | 0.00/412M [00:00<?, ?B/s]
+
+data/train-00012-of-00036.parquet:  18%|█▊        | 74.7M/412M [00:01<00:04, 74.5MB/s]
+data/train-00012-of-00036.parquet: 100%|██████████| 412M/412M [00:01<00:00, 283MB/s]  
+
+Downloading data:  36%|███▌      | 13/36 [00:31<00:45,  1.96s/files]
+
+data/train-00013-of-00036.parquet:   0%|          | 0.00/397M [00:00<?, ?B/s]
+
+data/train-00013-of-00036.parquet:  14%|█▍        | 55.0M/397M [00:01<00:06, 54.3MB/s]
+data/train-00013-of-00036.parquet: 100%|██████████| 397M/397M [00:01<00:00, 261MB/s]  
+
+Downloading data:  39%|███▉      | 14/36 [00:32<00:40,  1.84s/files]
+
+data/train-00014-of-00036.parquet:   0%|          | 0.00/397M [00:00<?, ?B/s]
+
+data/train-00014-of-00036.parquet:   7%|▋         | 28.0M/397M [00:01<00:13, 27.9MB/s]
+
+data/train-00014-of-00036.parquet: 100%|██████████| 397M/397M [00:02<00:00, 185MB/s]  
+data/train-00014-of-00036.parquet: 100%|██████████| 397M/397M [00:02<00:00, 165MB/s]
+
+Downloading data:  42%|████▏     | 15/36 [00:35<00:42,  2.02s/files]
+
+data/train-00015-of-00036.parquet:   0%|          | 0.00/389M [00:00<?, ?B/s]
+
+data/train-00015-of-00036.parquet:   7%|▋         | 26.6M/389M [00:01<00:13, 26.6MB/s]
+
+data/train-00015-of-00036.parquet:  83%|████████▎ | 322M/389M [00:02<00:00, 177MB/s]  
+data/train-00015-of-00036.parquet: 100%|██████████| 389M/389M [00:02<00:00, 177MB/s]
+
+Downloading data:  44%|████▍     | 16/36 [00:37<00:41,  2.08s/files]
+
+data/train-00016-of-00036.parquet:   0%|          | 0.00/384M [00:00<?, ?B/s]
+
+data/train-00016-of-00036.parquet:  11%|█▏        | 43.8M/384M [00:01<00:07, 43.7MB/s]
+
+data/train-00016-of-00036.parquet: 100%|██████████| 384M/384M [00:02<00:00, 218MB/s]  
+data/train-00016-of-00036.parquet: 100%|██████████| 384M/384M [00:02<00:00, 192MB/s]
+
+Downloading data:  47%|████▋     | 17/36 [00:39<00:39,  2.06s/files]
+
+data/train-00017-of-00036.parquet:   0%|          | 0.00/441M [00:00<?, ?B/s]
+
+data/train-00017-of-00036.parquet:   3%|▎         | 14.5M/441M [00:01<00:29, 14.5MB/s]
+
+data/train-00017-of-00036.parquet: 100%|██████████| 441M/441M [00:02<00:00, 229MB/s]  
+data/train-00017-of-00036.parquet: 100%|██████████| 441M/441M [00:02<00:00, 200MB/s]
+
+Downloading data:  50%|█████     | 18/36 [00:41<00:38,  2.11s/files]
+
+data/train-00018-of-00036.parquet:   0%|          | 0.00/571M [00:00<?, ?B/s]
+
+data/train-00018-of-00036.parquet:   4%|▍         | 24.9M/571M [00:01<00:22, 24.8MB/s]
+
+data/train-00018-of-00036.parquet:  88%|████████▊ | 504M/571M [00:02<00:00, 290MB/s]  
+data/train-00018-of-00036.parquet: 100%|██████████| 571M/571M [00:03<00:00, 188MB/s]
+
+Downloading data:  53%|█████▎    | 19/36 [00:44<00:40,  2.40s/files]
+
+data/train-00019-of-00036.parquet:   0%|          | 0.00/413M [00:00<?, ?B/s]
+
+data/train-00019-of-00036.parquet:  10%|█         | 41.5M/413M [00:01<00:08, 41.5MB/s]
+data/train-00019-of-00036.parquet: 100%|██████████| 413M/413M [00:01<00:00, 211MB/s]  
+
+Downloading data:  56%|█████▌    | 20/36 [00:46<00:36,  2.27s/files]
+
+data/train-00020-of-00036.parquet:   0%|          | 0.00/417M [00:00<?, ?B/s]
+
+data/train-00020-of-00036.parquet:   8%|▊         | 34.8M/417M [00:01<00:11, 34.6MB/s]
+data/train-00020-of-00036.parquet: 100%|██████████| 417M/417M [00:01<00:00, 234MB/s]  
+
+Downloading data:  58%|█████▊    | 21/36 [00:48<00:32,  2.14s/files]
+
+data/train-00021-of-00036.parquet:   0%|          | 0.00/401M [00:00<?, ?B/s]
+
+data/train-00021-of-00036.parquet:   0%|          | 26.6k/401M [00:01<4:29:15, 24.8kB/s]
+
+data/train-00021-of-00036.parquet: 100%|██████████| 401M/401M [00:02<00:00, 189MB/s]    
+data/train-00021-of-00036.parquet: 100%|██████████| 401M/401M [00:02<00:00, 164MB/s]
+
+Downloading data:  61%|██████    | 22/36 [00:51<00:31,  2.24s/files]
+
+data/train-00022-of-00036.parquet:   0%|          | 0.00/373M [00:00<?, ?B/s]
+
+data/train-00022-of-00036.parquet:   4%|▍         | 14.0M/373M [00:01<00:25, 14.0MB/s]
+data/train-00022-of-00036.parquet: 100%|██████████| 373M/373M [00:01<00:00, 214MB/s]  
+
+Downloading data:  64%|██████▍   | 23/36 [00:52<00:27,  2.09s/files]
+
+data/train-00023-of-00036.parquet:   0%|          | 0.00/374M [00:00<?, ?B/s]
+
+data/train-00023-of-00036.parquet:   0%|          | 30.4k/374M [00:01<3:31:02, 29.5kB/s]
+
+data/train-00023-of-00036.parquet: 100%|██████████| 374M/374M [00:02<00:00, 210MB/s]    
+data/train-00023-of-00036.parquet: 100%|██████████| 374M/374M [00:02<00:00, 179MB/s]
+
+Downloading data:  67%|██████▋   | 24/36 [00:55<00:25,  2.11s/files]
+
+data/train-00024-of-00036.parquet:   0%|          | 0.00/531M [00:00<?, ?B/s]
+
+data/train-00024-of-00036.parquet:   4%|▍         | 21.0M/531M [00:01<00:24, 21.0MB/s]
+
+data/train-00024-of-00036.parquet:  73%|███████▎  | 386M/531M [00:02<00:00, 223MB/s]  
+data/train-00024-of-00036.parquet: 100%|██████████| 531M/531M [00:02<00:00, 190MB/s]
+
+Downloading data:  69%|██████▉   | 25/36 [00:57<00:25,  2.32s/files]
+
+data/train-00025-of-00036.parquet:   0%|          | 0.00/542M [00:00<?, ?B/s]
+
+data/train-00025-of-00036.parquet:   6%|▌         | 33.2M/542M [00:01<00:15, 33.2MB/s]
+
+data/train-00025-of-00036.parquet: 100%|██████████| 542M/542M [00:02<00:00, 212MB/s]  
+data/train-00025-of-00036.parquet: 100%|██████████| 542M/542M [00:02<00:00, 193MB/s]
+
+Downloading data:  72%|███████▏  | 26/36 [01:00<00:24,  2.48s/files]
+
+data/train-00026-of-00036.parquet:   0%|          | 0.00/434M [00:00<?, ?B/s]
+
+data/train-00026-of-00036.parquet:   8%|▊         | 36.0M/434M [00:01<00:11, 35.5MB/s]
+
+data/train-00026-of-00036.parquet: 100%|██████████| 434M/434M [00:02<00:00, 200MB/s]  
+data/train-00026-of-00036.parquet: 100%|██████████| 434M/434M [00:02<00:00, 179MB/s]
+
+Downloading data:  75%|███████▌  | 27/36 [01:03<00:22,  2.47s/files]
+
+data/train-00027-of-00036.parquet:   0%|          | 0.00/366M [00:00<?, ?B/s]
+
+data/train-00027-of-00036.parquet:   8%|▊         | 29.6M/366M [00:01<00:11, 29.6MB/s]
+data/train-00027-of-00036.parquet: 100%|██████████| 366M/366M [00:01<00:00, 202MB/s]  
+
+Downloading data:  78%|███████▊  | 28/36 [01:05<00:18,  2.28s/files]
+
+data/train-00028-of-00036.parquet:   0%|          | 0.00/374M [00:00<?, ?B/s]
+
+data/train-00028-of-00036.parquet:   8%|▊         | 29.5M/374M [00:01<00:11, 29.4MB/s]
+data/train-00028-of-00036.parquet: 100%|██████████| 374M/374M [00:01<00:00, 225MB/s]  
+
+Downloading data:  81%|████████  | 29/36 [01:06<00:14,  2.10s/files]
+
+data/train-00029-of-00036.parquet:   0%|          | 0.00/402M [00:00<?, ?B/s]
+
+data/train-00029-of-00036.parquet:   5%|▌         | 21.8M/402M [00:01<00:17, 21.8MB/s]
+data/train-00029-of-00036.parquet: 100%|██████████| 402M/402M [00:01<00:00, 231MB/s]  
+
+Downloading data:  83%|████████▎ | 30/36 [01:08<00:12,  2.00s/files]
 
-Downloading (incomplete total...):   0%|          | 13.5M/3.21G [00:01<07:02, 7.56MB/s]
+data/train-00030-of-00036.parquet:   0%|          | 0.00/400M [00:00<?, ?B/s]
 
-Downloading (incomplete total...):   3%|▎         | 80.6M/3.21G [00:04<02:36, 19.9MB/s]
+data/train-00030-of-00036.parquet:   3%|▎         | 12.8M/400M [00:01<00:30, 12.8MB/s]
+
+data/train-00030-of-00036.parquet: 100%|██████████| 400M/400M [00:02<00:00, 199MB/s]  
+data/train-00030-of-00036.parquet: 100%|██████████| 400M/400M [00:02<00:00, 175MB/s]
+
+Downloading data:  86%|████████▌ | 31/36 [01:10<00:10,  2.09s/files]
+
+data/train-00031-of-00036.parquet:   0%|          | 0.00/382M [00:00<?, ?B/s]
+
+data/train-00031-of-00036.parquet:   2%|▏         | 7.72M/382M [00:01<00:48, 7.70MB/s]
+data/train-00031-of-00036.parquet: 100%|██████████| 382M/382M [00:01<00:00, 209MB/s]  
+
+Downloading data:  89%|████████▉ | 32/36 [01:12<00:08,  2.02s/files]
+
+data/train-00032-of-00036.parquet:   0%|          | 0.00/380M [00:00<?, ?B/s]
+
+data/train-00032-of-00036.parquet:   5%|▍         | 18.4M/380M [00:01<00:19, 18.4MB/s]
+data/train-00032-of-00036.parquet: 100%|██████████| 380M/380M [00:01<00:00, 196MB/s]  
+
+Downloading data:  92%|█████████▏| 33/36 [01:14<00:06,  2.00s/files]
+
+data/train-00033-of-00036.parquet:   0%|          | 0.00/376M [00:00<?, ?B/s]
+
+data/train-00033-of-00036.parquet:   1%|          | 3.43M/376M [00:01<01:48, 3.42MB/s]
+data/train-00033-of-00036.parquet: 100%|██████████| 376M/376M [00:01<00:00, 201MB/s]  
+
+Downloading data:  94%|█████████▍| 34/36 [01:16<00:03,  1.97s/files]
+
+data/train-00034-of-00036.parquet:   0%|          | 0.00/372M [00:00<?, ?B/s]
+
+data/train-00034-of-00036.parquet:   4%|▍         | 14.3M/372M [00:01<00:25, 14.3MB/s]
+data/train-00034-of-00036.parquet: 100%|██████████| 372M/372M [00:01<00:00, 220MB/s]  
+
+Downloading data:  97%|█████████▋| 35/36 [01:18<00:01,  1.90s/files]
+
+data/train-00035-of-00036.parquet:   0%|          | 0.00/373M [00:00<?, ?B/s]
+
+data/train-00035-of-00036.parquet:   5%|▌         | 19.1M/373M [00:01<00:18, 19.1MB/s]
+data/train-00035-of-00036.parquet: 100%|██████████| 373M/373M [00:01<00:00, 207MB/s]  
+
+Downloading data: 100%|██████████| 36/36 [01:20<00:00,  1.87s/files]
+Downloading data: 100%|██████████| 36/36 [01:20<00:00,  2.22s/files]
+
+
+data/validation-00000-of-00001.parquet:   0%|          | 0.00/162M [00:00<?, ?B/s]
+
+data/validation-00000-of-00001.parquet:  37%|███▋      | 59.3M/162M [00:01<00:01, 59.2MB/s]
+data/validation-00000-of-00001.parquet: 100%|██████████| 162M/162M [00:01<00:00, 137MB/s]  
+
+
+data/test-00000-of-00001.parquet:   0%|          | 0.00/306M [00:00<?, ?B/s]
+
+data/test-00000-of-00001.parquet:   3%|▎         | 10.7M/306M [00:01<00:27, 10.6MB/s]
+
+data/test-00000-of-00001.parquet: 100%|██████████| 306M/306M [00:02<00:00, 169MB/s]  
+data/test-00000-of-00001.parquet: 100%|██████████| 306M/306M [00:02<00:00, 147MB/s]
+
+
+Generating train split:   0%|          | 0/145152 [00:00<?, ? examples/s]
+
+Generating train split:   0%|          | 200/145152 [00:01<15:32, 155.47 examples/s]
+
+Generating train split:   0%|          | 500/145152 [00:02<11:15, 214.23 examples/s]
+
+Generating train split:   1%|          | 800/145152 [00:03<11:05, 216.89 examples/s]
+
+Generating train split:   1%|          | 1100/145152 [00:05<10:52, 220.70 examples/s]
+
+Generating train split:   1%|          | 1400/145152 [00:06<10:16, 233.33 examples/s]
+
+Generating train split:   1%|          | 1700/145152 [00:07<09:57, 240.15 examples/s]
+
+Generating train split:   1%|▏         | 2000/145152 [00:08<10:01, 238.10 examples/s]
+
+Generating train split:   3%|▎         | 4132/145152 [00:09<03:14, 725.45 examples/s]
+
+Generating train split:   4%|▍         | 6232/145152 [00:10<02:08, 1085.02 examples/s]
+
+Generating train split:   6%|▌         | 8264/145152 [00:11<01:42, 1339.46 examples/s]
+
+Generating train split:   7%|▋         | 10464/145152 [00:13<01:25, 1570.42 examples/s]
+
+Generating train split:   9%|▉         | 13296/145152 [00:14<01:09, 1890.88 examples/s]
+
+Generating train split:  11%|█▏        | 16428/145152 [00:15<00:58, 2186.18 examples/s]
+
+Generating train split:  13%|█▎        | 19428/145152 [00:16<00:53, 2362.66 examples/s]
+
+Generating train split:  15%|█▌        | 21860/145152 [00:17<01:02, 1973.51 examples/s]
+
+Generating train split:  17%|█▋        | 23960/145152 [00:19<01:12, 1665.91 examples/s]
+
+Generating train split:  18%|█▊        | 25792/145152 [00:21<01:18, 1518.35 examples/s]
+
+Generating train split:  19%|█▉        | 27492/145152 [00:22<01:22, 1419.11 examples/s]
+
+Generating train split:  20%|█▉        | 29024/145152 [00:24<01:24, 1366.99 examples/s]
+
+Generating train split:  21%|██        | 30824/145152 [00:25<01:17, 1466.19 examples/s]
+
+Generating train split:  22%|██▏       | 32656/145152 [00:26<01:13, 1522.14 examples/s]
+
+Generating train split:  24%|██▍       | 34956/145152 [00:27<01:04, 1712.43 examples/s]
+
+Generating train split:  26%|██▋       | 38288/145152 [00:28<00:50, 2132.95 examples/s]
+
+Generating train split:  29%|██▊       | 41520/145152 [00:29<00:42, 2427.94 examples/s]
+
+Generating train split:  31%|███       | 44452/145152 [00:30<00:39, 2559.95 examples/s]
+
+Generating train split:  33%|███▎      | 47352/145152 [00:31<00:39, 2462.75 examples/s]
+
+Generating train split:  34%|███▍      | 49884/145152 [00:32<00:39, 2422.60 examples/s]
+
+Generating train split:  36%|███▌      | 52516/145152 [00:33<00:37, 2463.02 examples/s]
+
+Generating train split:  38%|███▊      | 55116/145152 [00:34<00:36, 2479.32 examples/s]
+
+Generating train split:  40%|███▉      | 57948/145152 [00:35<00:34, 2503.63 examples/s]
+
+Generating train split:  42%|████▏     | 60580/145152 [00:36<00:33, 2535.19 examples/s]
+
+Generating train split:  44%|████▎     | 63480/145152 [00:37<00:31, 2569.53 examples/s]
+
+Generating train split:  46%|████▌     | 66312/145152 [00:38<00:30, 2618.17 examples/s]
+
+Generating train split:  48%|████▊     | 69144/145152 [00:39<00:28, 2639.65 examples/s]
+
+Generating train split:  49%|████▉     | 71844/145152 [00:40<00:28, 2603.44 examples/s]
+
+Generating train split:  51%|█████▏    | 74576/145152 [00:42<00:30, 2323.08 examples/s]
+
+Generating train split:  53%|█████▎    | 77208/145152 [00:43<00:30, 2210.35 examples/s]
+
+Generating train split:  55%|█████▌    | 79908/145152 [00:44<00:28, 2308.45 examples/s]
+
+Generating train split:  57%|█████▋    | 82440/145152 [00:45<00:26, 2367.41 examples/s]
+
+Generating train split:  59%|█████▊    | 85272/145152 [00:46<00:24, 2422.16 examples/s]
+
+Generating train split:  61%|██████    | 87972/145152 [00:47<00:23, 2469.83 examples/s]
+
+Generating train split:  63%|██████▎   | 90804/145152 [00:49<00:21, 2544.65 examples/s]
+
+Generating train split:  65%|██████▍   | 93636/145152 [00:50<00:19, 2622.46 examples/s]
+
+Generating train split:  67%|██████▋   | 96636/145152 [00:51<00:17, 2724.14 examples/s]
+
+Generating train split:  69%|██████▊   | 99468/145152 [00:52<00:18, 2430.66 examples/s]
+
+Generating train split:  70%|███████   | 102000/145152 [00:53<00:19, 2238.26 examples/s]
+
+Generating train split:  72%|███████▏  | 104400/145152 [00:55<00:19, 2133.52 examples/s]
+
+Generating train split:  74%|███████▎  | 106832/145152 [00:56<00:18, 2089.83 examples/s]
+
+Generating train split:  76%|███████▌  | 109764/145152 [00:57<00:15, 2273.20 examples/s]
+
+Generating train split:  78%|██████���▊  | 112764/145152 [00:58<00:13, 2438.22 examples/s]
+
+Generating train split:  80%|███████▉  | 115596/145152 [00:59<00:11, 2516.56 examples/s]
+
+Generating train split:  82%|████████▏ | 118428/145152 [01:00<00:10, 2533.33 examples/s]
+
+Generating train split:  84%|████████▎ | 121260/145152 [01:01<00:09, 2546.56 examples/s]
+
+Generating train split:  85%|████████▌ | 123960/145152 [01:02<00:08, 2580.24 examples/s]
+
+Generating train split:  87%|████████▋ | 126792/145152 [01:03<00:07, 2620.04 examples/s]
+
+Generating train split:  89%|████████▉ | 129624/145152 [01:04<00:05, 2663.67 examples/s]
+
+Generating train split:  91%|█████████▏| 132624/145152 [01:05<00:04, 2705.80 examples/s]
+
+Generating train split:  93%|█████████▎| 135456/145152 [01:06<00:03, 2734.67 examples/s]
+
+Generating train split:  95%|█████████▌| 138588/145152 [01:07<00:02, 2775.60 examples/s]
+
+Generating train split:  97%|█████████▋| 141420/145152 [01:08<00:01, 2784.34 examples/s]
+
+Generating train split:  99%|█████████▉| 144420/145152 [01:10<00:00, 2815.47 examples/s]
+Generating train split: 100%|██████████| 145152/145152 [01:10<00:00, 2068.64 examples/s]
+
+
+Generating validation split:   0%|          | 0/239 [00:00<?, ? examples/s]
+Generating validation split: 100%|██████████| 239/239 [00:00<00:00, 271.32 examples/s]
+
+
+Generating test split:   0%|          | 0/418 [00:00<?, ? examples/s]
+
+Generating test split:  48%|████▊     | 200/418 [00:01<00:01, 151.91 examples/s]
+Generating test split: 100%|██████████| 418/418 [00:01<00:00, 235.25 examples/s]
+
+
+Downloading (incomplete total...): 0.00B [00:00, ?B/s]
+
+Fetching 5 files:   0%|          | 0/5 [00:00<?, ?it/s]
+
+Downloading (incomplete total...):   0%|          | 0.00/5.70M [00:00<?, ?B/s]
+
+Downloading (incomplete total...):   0%|          | 0.00/5.81M [00:00<?, ?B/s]
+
+Downloading (incomplete total...):   0%|          | 0.00/2.15G [00:00<?, ?B/s]
+
+Downloading (incomplete total...):   0%|          | 0.00/3.21G [00:00<?, ?B/s]
+
+Downloading (incomplete total...):   0%|          | 0.00/3.21G [00:00<?, ?B/s]
 
-Downloading (incomplete total...):   7%|▋         | 215M/3.21G [00:05<01:06, 45.2MB/s] 
+Downloading (incomplete total...):   0%|          | 13.5M/3.21G [00:01<04:28, 11.9MB/s]
 
-Downloading (incomplete total...):   9%|▉         | 282M/3.21G [00:07<01:07, 43.1MB/s]
+Downloading (incomplete total...):   2%|▏         | 64.6M/3.21G [00:08<07:11, 7.29MB/s]
 
-Downloading (incomplete total...):  23%|██▎       | 743M/3.21G [00:08<00:17, 142MB/s] 
+Downloading (incomplete total...):   4%|▍         | 132M/3.21G [00:09<03:16, 15.7MB/s] 
 
-Downloading (incomplete total...):  50%|████▉     | 1.59G/3.21G [00:09<00:05, 308MB/s]
+Downloading (incomplete total...):  23%|██▎       | 748M/3.21G [00:10<00:21, 115MB/s] 
 
-Fetching 5 files:  60%|██████    | 3/5 [00:10<00:06,  3.44s/it]
+Downloading (incomplete total...):  62%|██████▏   | 2.00G/3.21G [00:11<00:03, 331MB/s]
 
-Downloading (incomplete total...):  73%|███████▎  | 2.35G/3.21G [00:10<00:02, 424MB/s]
-Fetching 5 files: 100%|██████████| 5/5 [00:11<00:00,  2.22s/it]
+Fetching 5 files:  60%|██████    | 3/5 [00:12<00:08,  4.05s/it]
+Fetching 5 files: 100%|██████████| 5/5 [00:12<00:00,  2.54s/it]
 
 
-Download complete: 100%|██████████| 3.21G/3.21G [00:11<00:00, 424MB/s]                
-Download complete: 100%|██████████| 3.21G/3.21G [00:16<00:00, 189MB/s]
+Download complete: 100%|██████████| 3.21G/3.21G [00:12<00:00, 331MB/s]                
+Download complete: 100%|██████████| 3.21G/3.21G [00:13<00:00, 244MB/s]
 /usr/local/lib/python3.13/site-packages/diffusers/models/lora.py:393: FutureWarning: `LoRACompatibleLinear` is deprecated and will be removed in version 1.0.0. Use of `LoRACompatibleLinear` is deprecated. Please switch to PEFT backend by installing PEFT: `pip install peft`.
   deprecate("LoRACompatibleLinear", "1.0.0", deprecation_message)
-02/10/2026 03:45:46 - INFO - root - input frame rate=25
-02/10/2026 03:45:47 - INFO - httpx - HTTP Request: HEAD https://huggingface.co/ResembleAI/chatterbox/resolve/main/Cangjie5_TC.json "HTTP/1.1 307 Temporary Redirect"
-02/10/2026 03:45:47 - INFO - httpx - HTTP Request: HEAD https://huggingface.co/api/resolve-cache/models/ResembleAI/chatterbox/05e904af2b5c7f8e482687a9d7336c5c824467d9/Cangjie5_TC.json "HTTP/1.1 200 OK"
-02/10/2026 03:45:47 - INFO - httpx - HTTP Request: GET https://huggingface.co/api/resolve-cache/models/ResembleAI/chatterbox/05e904af2b5c7f8e482687a9d7336c5c824467d9/Cangjie5_TC.json "HTTP/1.1 200 OK"
 
 
 Cangjie5_TC.json:   0%|          | 0.00/1.92M [00:00<?, ?B/s]
-Cangjie5_TC.json: 100%|██████████| 1.92M/1.92M [00:00<00:00, 19.7MB/s]
+Cangjie5_TC.json: 100%|██████████| 1.92M/1.92M [00:00<00:00, 42.7MB/s]
 Downloading: "https://github.com/explosion/spacy-pkuseg/releases/download/v0.0.26/spacy_ontonotes.zip" to /root/.pkuseg/spacy_ontonotes.zip
 
 
   0%|          | 0/34567143 [00:00<?, ?it/s]
-100%|██████████| 34567143/34567143 [00:00<00:00, 164494667.16it/s]
-02/10/2026 03:45:49 - INFO - __main__ - Voice Encoder frozen.
-02/10/2026 03:45:49 - INFO - __main__ - S3Gen model frozen.
-02/10/2026 03:45:49 - INFO - __main__ - T3 model set to trainable.
-02/10/2026 03:45:49 - INFO - __main__ - Loading and processing dataset...
+100%|██████████| 34567143/34567143 [00:00<00:00, 172484889.55it/s]
 Traceback (most recent call last):
-  File "/app/chatterbox-multilingual-finetuning/src/finetune_t3.py", line 860, in <module>
+  File "/app/chatterbox-multilingual-finetuning/src/finetune_t3.py", line 570, in <module>
     main()
     ~~~~^^
-  File "/app/chatterbox-multilingual-finetuning/src/finetune_t3.py", line 651, in main
-    train_hf_dataset = train_hf_dataset.map(decode_audio)
-                       ^^^^^^^^^^^^^^^^
-UnboundLocalError: cannot access local variable 'train_hf_dataset' where it is not associated with a value
+  File "/app/chatterbox-multilingual-finetuning/src/finetune_t3.py", line 549, in main
+    text_pad_token_id=t3_config.start_text_token,
+                      ^^^^^^^^^^^^^^^^^^^^^^^^^^
+  File "/usr/local/lib/python3.13/site-packages/torch/nn/modules/module.py", line 1965, in __getattr__
+    raise AttributeError(
+        f"'{type(self).__name__}' object has no attribute '{name}'"
+    )
+AttributeError: 'T3' object has no attribute 'start_text_token'