ERROR LOAD MODEL

[cosmind-rusu]

2026-04-17 10:24:25,709 - main - INFO - ============================================================
2026-04-17 10:24:25,709 - main - INFO - voicebox-server starting up...
2026-04-17 10:24:25,710 - main - INFO - Python version: 3.12.10 (tags/v3.12.10:0cc8128, Apr 8 2025, 12:21:36) [MSC v.1943 64 bit (AMD64)]
2026-04-17 10:24:25,710 - main - INFO - Executable: C:\Program Files\Voicebox\voicebox-server.exe
2026-04-17 10:24:25,710 - main - INFO - Arguments: ['\\?\C:\Program Files\Voicebox\voicebox-server.exe', '--data-dir', 'C:\Users\cosmi\AppData\Roaming\sh.voicebox.app', '--port', '17493', '--parent-pid', '8908']
2026-04-17 10:24:25,710 - main - INFO - ============================================================
2026-04-17 10:24:25,710 - main - INFO - Importing argparse...
2026-04-17 10:24:25,711 - main - INFO - Importing uvicorn...
2026-04-17 10:24:25,790 - main - INFO - Standard library imports successful
2026-04-17 10:24:25,790 - main - INFO - Importing backend.config...
2026-04-17 10:24:25,791 - main - INFO - Importing backend.database...
2026-04-17 10:24:25,977 - main - INFO - Importing backend.main (this may take a while due to torch/transformers)...
2026-04-17 10:24:28,213 - main - INFO - Backend imports successful
2026-04-17 10:24:28,214 - main - INFO - Backend variant: CPU
server.py:278: DeprecationWarning:
on_event is deprecated, use lifespan event handlers instead.
Read more about it in the
FastAPI docs for Lifespan Events.
2026-04-17 10:24:28,214 - main - INFO - Parsed arguments: host=127.0.0.1, port=17493, data_dir=C:\Users\cosmi\AppData\Roaming\sh.voicebox.app
2026-04-17 10:24:28,214 - main - INFO - Setting data directory to: C:\Users\cosmi\AppData\Roaming\sh.voicebox.app
2026-04-17 10:24:28,215 - backend.config - INFO - Data directory set to: C:\Users\cosmi\AppData\Roaming\sh.voicebox.app
2026-04-17 10:24:28,215 - main - INFO - Initializing database...
2026-04-17 10:24:28,252 - backend.database.migrations - INFO - Added voice_type column to profiles
2026-04-17 10:24:28,259 - backend.database.migrations - INFO - Added preset_engine column to profiles
2026-04-17 10:24:28,266 - backend.database.migrations - INFO - Added preset_voice_id column to profiles
2026-04-17 10:24:28,274 - backend.database.migrations - INFO - Added design_prompt column to profiles
2026-04-17 10:24:28,282 - backend.database.migrations - INFO - Added default_engine column to profiles
2026-04-17 10:24:28,294 - backend.database.migrations - INFO - Normalized 1 stored file paths
2026-04-17 10:24:28,325 - main - INFO - Database initialized successfully
2026-04-17 10:24:28,325 - main - INFO - Starting uvicorn server on 127.0.0.1:17493...
INFO: Started server process [44612]
INFO: Waiting for application startup.
2026-04-17 10:24:28,351 - backend.app - INFO - Voicebox v0.4.0 starting up
2026-04-17 10:24:28,351 - backend.app - INFO - Python 3.12.10 on Windows 11 (AMD64)
2026-04-17 10:24:28,362 - backend.app - INFO - Database: C:\Users\cosmi\AppData\Roaming\sh.voicebox.app\voicebox.db
2026-04-17 10:24:28,363 - backend.app - INFO - Data directory: C:\Users\cosmi\AppData\Roaming\sh.voicebox.app
2026-04-17 10:24:28,366 - backend.app - INFO - Profiles: 1, Generations: 2
2026-04-17 10:24:28,368 - backend.app - INFO - Backend: PYTORCH
2026-04-17 10:24:28,368 - backend.app - INFO - GPU: None (CPU only)
2026-04-17 10:24:28,372 - backend.app - INFO - Model cache: C:\Users\cosmi.cache\huggingface\hub
2026-04-17 10:24:28,372 - backend.app - INFO - Ready
2026-04-17 10:24:28,373 - watchdog - INFO - Parent watchdog started, monitoring PID 8908, server PID 44612
2026-04-17 10:24:28,373 - watchdog - INFO - Parent PID 8908 initial check: alive=True
INFO: Application startup complete.
INFO: Uvicorn running on http://127.0.0.1:17493 (Press CTRL+C to quit)
INFO: 127.0.0.1:51894 - "OPTIONS /profiles HTTP/1.1" 200 OK
INFO: 127.0.0.1:58406 - "OPTIONS /history?limit=20 HTTP/1.1" 200 OK
INFO: 127.0.0.1:56233 - "OPTIONS /effects/presets HTTP/1.1" 200 OK
INFO: 127.0.0.1:65402 - "OPTIONS /tasks/active HTTP/1.1" 200 OK
INFO: 127.0.0.1:51894 - "GET /tasks/active HTTP/1.1" 200 OK
INFO: 127.0.0.1:65402 - "GET /profiles HTTP/1.1" 200 OK
INFO: 127.0.0.1:56233 - "GET /history?limit=20 HTTP/1.1" 200 OK
INFO: 127.0.0.1:58406 - "GET /effects/presets HTTP/1.1" 200 OK
INFO: 127.0.0.1:58406 - "OPTIONS /profiles/c66eb3ab-b68e-44c3-90a2-b5dbcda6c44e HTTP/1.1" 200 OK
INFO: 127.0.0.1:58406 - "GET /profiles/c66eb3ab-b68e-44c3-90a2-b5dbcda6c44e HTTP/1.1" 200 OK
INFO: 127.0.0.1:58406 - "OPTIONS /health HTTP/1.1" 200 OK
INFO: 127.0.0.1:58406 - "GET /health HTTP/1.1" 200 OK
INFO: 127.0.0.1:58406 - "OPTIONS /models/status HTTP/1.1" 200 OK
INFO: 127.0.0.1:52655 - "OPTIONS /models/cache-dir HTTP/1.1" 200 OK
INFO: 127.0.0.1:51968 - "GET /tasks/active HTTP/1.1" 200 OK
INFO: 127.0.0.1:58406 - "GET /models/status HTTP/1.1" 200 OK
INFO: 127.0.0.1:52655 - "GET /models/cache-dir HTTP/1.1" 200 OK
INFO: 127.0.0.1:52655 - "GET /tasks/active HTTP/1.1" 200 OK
INFO: 127.0.0.1:52655 - "GET /models/status HTTP/1.1" 200 OK
INFO: 127.0.0.1:51874 - "OPTIONS /generate/ef6ab366-29db-4f51-868a-5a1b33db9c83/retry HTTP/1.1" 200 OK
INFO: 127.0.0.1:51874 - "POST /generate/ef6ab366-29db-4f51-868a-5a1b33db9c83/retry HTTP/1.1" 200 OK
INFO: 127.0.0.1:51874 - "GET /history?limit=20 HTTP/1.1" 200 OK
INFO: 127.0.0.1:53015 - "GET /generate/ef6ab366-29db-4f51-868a-5a1b33db9c83/status HTTP/1.1" 200 OK
"sox" no se reconoce como un comando interno o externo,
programa o archivo por lotes ejecutable.
2026-04-17 10:24:53,916 - sox - WARNING - SoX could not be found!
If you do not have SoX, proceed here:
- - - http://sox.sourceforge.net/ - - -
If you do (or think that you should) have SoX, double-check your
path variables.
2026-04-17 10:24:54,498 - backend.backends.pytorch_backend - INFO - Loading TTS model 1.7B on cpu...
2026-04-17 10:24:54,498 - backend.utils.hf_offline_patch - INFO - [offline-guard] qwen-tts-1.7B is cached � forcing HF_HUB_OFFLINE=1
2026-04-17 10:24:55,173 - backend.utils.progress - ERROR - Marked qwen-tts-1.7B as error: Unrecognized model in Qwen/Qwen3-TTS-12Hz-1.7B-Base. Should have a model_type key in its config.json, or contain one of the following strings in its name: aimv2, aimv2_vision_model, albert, align, altclip, apertus, arcee, aria, aria_text, audio-spectrogram-transformer, autoformer, aya_vision, bamba, bark, bart, beit, bert, bert-generation, big_bird, bigbird_pegasus, biogpt, bit, bitnet, blenderbot, blenderbot-small, blip, blip-2, blip_2_qformer, bloom, blt, bridgetower, bros, camembert, canine, chameleon, chinese_clip, chinese_clip_vision_model, clap, clip, clip_text_model, clip_vision_model, clipseg, clvp, code_llama, codegen, cohere, cohere2, cohere2_vision, colpali, colqwen2, conditional_detr, convbert, convnext, convnextv2, cpmant, csm, ctrl, cvt, d_fine, dab-detr, dac, data2vec-audio, data2vec-text, data2vec-vision, dbrx, deberta, deberta-v2, decision_transformer, deepseek_v2, deepseek_v3, deepseek_vl, deepseek_vl_hybrid, deformable_detr, deit, depth_anything, depth_pro, deta, detr, dia, diffllama, dinat, dinov2, dinov2_with_registers, dinov3_convnext, dinov3_vit, distilbert, doge, donut-swin, dots1, dpr, dpt, edgetam, edgetam_video, edgetam_vision_model, efficientformer, efficientloftr, efficientnet, electra, emu3, encodec, encoder-decoder, eomt, ernie, ernie4_5, ernie4_5_moe, ernie_m, esm, evolla, exaone4, falcon, falcon_h1, falcon_mamba, fastspeech2_conformer, fastspeech2_conformer_with_hifigan, flaubert, flava, flex_olmo, florence2, fnet, focalnet, fsmt, funnel, fuyu, gemma, gemma2, gemma3, gemma3_text, gemma3n, gemma3n_audio, gemma3n_text, gemma3n_vision, git, glm, glm4, glm4_moe, glm4v, glm4v_moe, glm4v_moe_text, glm4v_text, glpn, got_ocr2, gpt-sw3, gpt2, gpt_bigcode, gpt_neo, gpt_neox, gpt_neox_japanese, gpt_oss, gptj, gptsan-japanese, granite, granite_speech, granitemoe, granitemoehybrid, granitemoeshared, granitevision, graphormer, grounding-dino, groupvit, helium, hgnet_v2, hiera, hubert, hunyuan_v1_dense, hunyuan_v1_moe, ibert, idefics, idefics2, idefics3, idefics3_vision, ijepa, imagegpt, informer, instructblip, instructblipvideo, internvl, internvl_vision, jamba, janus, jetmoe, jukebox, kosmos-2, kosmos-2.5, kyutai_speech_to_text, layoutlm, layoutlmv2, layoutlmv3, led, levit, lfm2, lfm2_vl, lightglue, lilt, llama, llama4, llama4_text, llava, llava_next, llava_next_video, llava_onevision, longcat_flash, longformer, longt5, luke, lxmert, m2m_100, mamba, mamba2, marian, markuplm, mask2former, maskformer, maskformer-swin, mbart, mctct, mega, megatron-bert, metaclip_2, mgp-str, mimi, minimax, ministral, mistral, mistral3, mixtral, mlcd, mllama, mm-grounding-dino, mobilebert, mobilenet_v1, mobilenet_v2, mobilevit, mobilevitv2, modernbert, modernbert-decoder, moonshine, moshi, mpnet, mpt, mra, mt5, musicgen, musicgen_melody, mvp, nat, nemotron, nezha, nllb-moe, nougat, nystromformer, olmo, olmo2, olmo3, olmoe, omdet-turbo, oneformer, open-llama, openai-gpt, opt, ovis2, owlv2, owlvit, paligemma, parakeet_ctc, parakeet_encoder, patchtsmixer, patchtst, pegasus, pegasus_x, perceiver, perception_encoder, perception_lm, persimmon, phi, phi3, phi4_multimodal, phimoe, pix2struct, pixtral, plbart, poolformer, pop2piano, prompt_depth_anything, prophetnet, pvt, pvt_v2, qdqbert, qwen2, qwen2_5_omni, qwen2_5_vl, qwen2_5_vl_text, qwen2_audio, qwen2_audio_encoder, qwen2_moe, qwen2_vl, qwen2_vl_text, qwen3, qwen3_moe, qwen3_next, qwen3_omni_moe, qwen3_vl, qwen3_vl_moe, qwen3_vl_moe_text, qwen3_vl_text, rag, realm, recurrent_gemma, reformer, regnet, rembert, resnet, retribert, roberta, roberta-prelayernorm, roc_bert, roformer, rt_detr, rt_detr_resnet, rt_detr_v2, rwkv, sam, sam2, sam2_hiera_det_model, sam2_video, sam2_vision_model, sam_hq, sam_hq_vision_model, sam_vision_model, seamless_m4t, seamless_m4t_v2, seed_oss, segformer, seggpt, sew, sew-d, shieldgemma2, siglip, siglip2, siglip2_vision_model, siglip_vision_model, smollm3, smolvlm, smolvlm_vision, speech-encoder-decoder, speech_to_text, speech_to_text_2, speecht5, splinter, squeezebert, stablelm, starcoder2, superglue, superpoint, swiftformer, swin, swin2sr, swinv2, switch_transformers, t5, t5gemma, table-transformer, tapas, textnet, time_series_transformer, timesfm, timesformer, timm_backbone, timm_wrapper, trajectory_transformer, transfo-xl, trocr, tvlt, tvp, udop, umt5, unispeech, unispeech-sat, univnet, upernet, van, vaultgemma, video_llava, videomae, vilt, vipllava, vision-encoder-decoder, vision-text-dual-encoder, visual_bert, vit, vit_hybrid, vit_mae, vit_msn, vitdet, vitmatte, vitpose, vitpose_backbone, vits, vivit, vjepa2, voxtral, voxtral_encoder, wav2vec2, wav2vec2-bert, wav2vec2-conformer, wavlm, whisper, xclip, xcodec, xglm, xlm, xlm-prophetnet, xlm-roberta, xlm-roberta-xl, xlnet, xlstm, xmod, yolos, yoso, zamba, zamba2, zoedepth, qwen3_tts
Traceback (most recent call last):
File "backend\services\generation.py", line 63, in run_generation
File "backend\backends_init_.py", line 400, in load_engine_model
File "backend\backends\pytorch_backend.py", line 86, in load_model_async
File "asyncio\threads.py", line 25, in to_thread
File "concurrent\futures\thread.py", line 59, in run
File "backend\backends\pytorch_backend.py", line 111, in _load_model_sync
File "qwen_tts\inference\qwen3_tts_model.py", line 112, in from_pretrained
model = AutoModel.from_pretrained(pretrained_model_name_or_path, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\models\auto\auto_factory.py", line 549, in from_pretrained
config, kwargs = AutoConfig.from_pretrained(
^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\models\auto\configuration_auto.py", line 1380, in from_pretrained
raise ValueError(
ValueError: Unrecognized model in Qwen/Qwen3-TTS-12Hz-1.7B-Base. Should have a model_type key in its config.json, or contain one of the following strings in its name: aimv2, aimv2_vision_model, albert, align, altclip, apertus, arcee, aria, aria_text, audio-spectrogram-transformer, autoformer, aya_vision, bamba, bark, bart, beit, bert, bert-generation, big_bird, bigbird_pegasus, biogpt, bit, bitnet, blenderbot, blenderbot-small, blip, blip-2, blip_2_qformer, bloom, blt, bridgetower, bros, camembert, canine, chameleon, chinese_clip, chinese_clip_vision_model, clap, clip, clip_text_model, clip_vision_model, clipseg, clvp, code_llama, codegen, cohere, cohere2, cohere2_vision, colpali, colqwen2, conditional_detr, convbert, convnext, convnextv2, cpmant, csm, ctrl, cvt, d_fine, dab-detr, dac, data2vec-audio, data2vec-text, data2vec-vision, dbrx, deberta, deberta-v2, decision_transformer, deepseek_v2, deepseek_v3, deepseek_vl, deepseek_vl_hybrid, deformable_detr, deit, depth_anything, depth_pro, deta, detr, dia, diffllama, dinat, dinov2, dinov2_with_registers, dinov3_convnext, dinov3_vit, distilbert, doge, donut-swin, dots1, dpr, dpt, edgetam, edgetam_video, edgetam_vision_model, efficientformer, efficientloftr, efficientnet, electra, emu3, encodec, encoder-decoder, eomt, ernie, ernie4_5, ernie4_5_moe, ernie_m, esm, evolla, exaone4, falcon, falcon_h1, falcon_mamba, fastspeech2_conformer, fastspeech2_conformer_with_hifigan, flaubert, flava, flex_olmo, florence2, fnet, focalnet, fsmt, funnel, fuyu, gemma, gemma2, gemma3, gemma3_text, gemma3n, gemma3n_audio, gemma3n_text, gemma3n_vision, git, glm, glm4, glm4_moe, glm4v, glm4v_moe, glm4v_moe_text, glm4v_text, glpn, got_ocr2, gpt-sw3, gpt2, gpt_bigcode, gpt_neo, gpt_neox, gpt_neox_japanese, gpt_oss, gptj, gptsan-japanese, granite, granite_speech, granitemoe, granitemoehybrid, granitemoeshared, granitevision, graphormer, grounding-dino, groupvit, helium, hgnet_v2, hiera, hubert, hunyuan_v1_dense, hunyuan_v1_moe, ibert, idefics, idefics2, idefics3, idefics3_vision, ijepa, imagegpt, informer, instructblip, instructblipvideo, internvl, internvl_vision, jamba, janus, jetmoe, jukebox, kosmos-2, kosmos-2.5, kyutai_speech_to_text, layoutlm, layoutlmv2, layoutlmv3, led, levit, lfm2, lfm2_vl, lightglue, lilt, llama, llama4, llama4_text, llava, llava_next, llava_next_video, llava_onevision, longcat_flash, longformer, longt5, luke, lxmert, m2m_100, mamba, mamba2, marian, markuplm, mask2former, maskformer, maskformer-swin, mbart, mctct, mega, megatron-bert, metaclip_2, mgp-str, mimi, minimax, ministral, mistral, mistral3, mixtral, mlcd, mllama, mm-grounding-dino, mobilebert, mobilenet_v1, mobilenet_v2, mobilevit, mobilevitv2, modernbert, modernbert-decoder, moonshine, moshi, mpnet, mpt, mra, mt5, musicgen, musicgen_melody, mvp, nat, nemotron, nezha, nllb-moe, nougat, nystromformer, olmo, olmo2, olmo3, olmoe, omdet-turbo, oneformer, open-llama, openai-gpt, opt, ovis2, owlv2, owlvit, paligemma, parakeet_ctc, parakeet_encoder, patchtsmixer, patchtst, pegasus, pegasus_x, perceiver, perception_encoder, perception_lm, persimmon, phi, phi3, phi4_multimodal, phimoe, pix2struct, pixtral, plbart, poolformer, pop2piano, prompt_depth_anything, prophetnet, pvt, pvt_v2, qdqbert, qwen2, qwen2_5_omni, qwen2_5_vl, qwen2_5_vl_text, qwen2_audio, qwen2_audio_encoder, qwen2_moe, qwen2_vl, qwen2_vl_text, qwen3, qwen3_moe, qwen3_next, qwen3_omni_moe, qwen3_vl, qwen3_vl_moe, qwen3_vl_moe_text, qwen3_vl_text, rag, realm, recurrent_gemma, reformer, regnet, rembert, resnet, retribert, roberta, roberta-prelayernorm, roc_bert, roformer, rt_detr, rt_detr_resnet, rt_detr_v2, rwkv, sam, sam2, sam2_hiera_det_model, sam2_video, sam2_vision_model, sam_hq, sam_hq_vision_model, sam_vision_model, seamless_m4t, seamless_m4t_v2, seed_oss, segformer, seggpt, sew, sew-d, shieldgemma2, siglip, siglip2, siglip2_vision_model, siglip_vision_model, smollm3, smolvlm, smolvlm_vision, speech-encoder-decoder, speech_to_text, speech_to_text_2, speecht5, splinter, squeezebert, stablelm, starcoder2, superglue, superpoint, swiftformer, swin, swin2sr, swinv2, switch_transformers, t5, t5gemma, table-transformer, tapas, textnet, time_series_transformer, timesfm, timesformer, timm_backbone, timm_wrapper, trajectory_transformer, transfo-xl, trocr, tvlt, tvp, udop, umt5, unispeech, unispeech-sat, univnet, upernet, van, vaultgemma, video_llava, videomae, vilt, vipllava, vision-encoder-decoder, vision-text-dual-encoder, visual_bert, vit, vit_hybrid, vit_mae, vit_msn, vitdet, vitmatte, vitpose, vitpose_backbone, vits, vivit, vjepa2, voxtral, voxtral_encoder, wav2vec2, wav2vec2-bert, wav2vec2-conformer, wavlm, whisper, xclip, xcodec, xglm, xlm, xlm-prophetnet, xlm-roberta, xlm-roberta-xl, xlnet, xlstm, xmod, yolos, yoso, zamba, zamba2, zoedepth, qwen3_tts

---

Warning: flash-attn is not installed. Will only run the manual PyTorch version. Please install flash-attn for faster inference.

---

INFO: 127.0.0.1:53015 - "GET /history?limit=20 HTTP/1.1" 200 OK
INFO: 127.0.0.1:53015 - "GET /tasks/active HTTP/1.1" 200 OK
INFO: 127.0.0.1:56957 - "GET /tasks/active HTTP/1.1" 200 OK

[cosmind-rusu]

2026-04-17 10:24:25,709 - main - INFO - ============================================================
2026-04-17 10:24:25,709 - main - INFO - voicebox-server starting up...
2026-04-17 10:24:25,710 - main - INFO - Python version: 3.12.10 (tags/v3.12.10:0cc8128, Apr 8 2025, 12:21:36) [MSC v.1943 64 bit (AMD64)]
2026-04-17 10:24:25,710 - main - INFO - Executable: C:\Program Files\Voicebox\voicebox-server.exe
2026-04-17 10:24:25,710 - main - INFO - Arguments: ['\\?\C:\Program Files\Voicebox\voicebox-server.exe', '--data-dir', 'C:\Users\cosmi\AppData\Roaming\sh.voicebox.app', '--port', '17493', '--parent-pid', '8908']
2026-04-17 10:24:25,710 - main - INFO - ============================================================
2026-04-17 10:24:25,710 - main - INFO - Importing argparse...
2026-04-17 10:24:25,711 - main - INFO - Importing uvicorn...
2026-04-17 10:24:25,790 - main - INFO - Standard library imports successful
2026-04-17 10:24:25,790 - main - INFO - Importing backend.config...
2026-04-17 10:24:25,791 - main - INFO - Importing backend.database...
2026-04-17 10:24:25,977 - main - INFO - Importing backend.main (this may take a while due to torch/transformers)...
2026-04-17 10:24:28,213 - main - INFO - Backend imports successful
2026-04-17 10:24:28,214 - main - INFO - Backend variant: CPU
server.py:278: DeprecationWarning:
on_event is deprecated, use lifespan event handlers instead.
Read more about it in the
FastAPI docs for Lifespan Events.
2026-04-17 10:24:28,214 - main - INFO - Parsed arguments: host=127.0.0.1, port=17493, data_dir=C:\Users\cosmi\AppData\Roaming\sh.voicebox.app
2026-04-17 10:24:28,214 - main - INFO - Setting data directory to: C:\Users\cosmi\AppData\Roaming\sh.voicebox.app
2026-04-17 10:24:28,215 - backend.config - INFO - Data directory set to: C:\Users\cosmi\AppData\Roaming\sh.voicebox.app
2026-04-17 10:24:28,215 - main - INFO - Initializing database...
2026-04-17 10:24:28,252 - backend.database.migrations - INFO - Added voice_type column to profiles
2026-04-17 10:24:28,259 - backend.database.migrations - INFO - Added preset_engine column to profiles
2026-04-17 10:24:28,266 - backend.database.migrations - INFO - Added preset_voice_id column to profiles
2026-04-17 10:24:28,274 - backend.database.migrations - INFO - Added design_prompt column to profiles
2026-04-17 10:24:28,282 - backend.database.migrations - INFO - Added default_engine column to profiles
2026-04-17 10:24:28,294 - backend.database.migrations - INFO - Normalized 1 stored file paths
2026-04-17 10:24:28,325 - main - INFO - Database initialized successfully
2026-04-17 10:24:28,325 - main - INFO - Starting uvicorn server on 127.0.0.1:17493...
INFO: Started server process [44612]
INFO: Waiting for application startup.
2026-04-17 10:24:28,351 - backend.app - INFO - Voicebox v0.4.0 starting up
2026-04-17 10:24:28,351 - backend.app - INFO - Python 3.12.10 on Windows 11 (AMD64)
2026-04-17 10:24:28,362 - backend.app - INFO - Database: C:\Users\cosmi\AppData\Roaming\sh.voicebox.app\voicebox.db
2026-04-17 10:24:28,363 - backend.app - INFO - Data directory: C:\Users\cosmi\AppData\Roaming\sh.voicebox.app
2026-04-17 10:24:28,366 - backend.app - INFO - Profiles: 1, Generations: 2
2026-04-17 10:24:28,368 - backend.app - INFO - Backend: PYTORCH
2026-04-17 10:24:28,368 - backend.app - INFO - GPU: None (CPU only)
2026-04-17 10:24:28,372 - backend.app - INFO - Model cache: C:\Users\cosmi.cache\huggingface\hub
2026-04-17 10:24:28,372 - backend.app - INFO - Ready
2026-04-17 10:24:28,373 - watchdog - INFO - Parent watchdog started, monitoring PID 8908, server PID 44612
2026-04-17 10:24:28,373 - watchdog - INFO - Parent PID 8908 initial check: alive=True
INFO: Application startup complete.
INFO: Uvicorn running on http://127.0.0.1:17493 (Press CTRL+C to quit)
INFO: 127.0.0.1:51894 - "OPTIONS /profiles HTTP/1.1" 200 OK
INFO: 127.0.0.1:58406 - "OPTIONS /history?limit=20 HTTP/1.1" 200 OK
INFO: 127.0.0.1:56233 - "OPTIONS /effects/presets HTTP/1.1" 200 OK
INFO: 127.0.0.1:65402 - "OPTIONS /tasks/active HTTP/1.1" 200 OK
INFO: 127.0.0.1:51894 - "GET /tasks/active HTTP/1.1" 200 OK
INFO: 127.0.0.1:65402 - "GET /profiles HTTP/1.1" 200 OK
INFO: 127.0.0.1:56233 - "GET /history?limit=20 HTTP/1.1" 200 OK
INFO: 127.0.0.1:58406 - "GET /effects/presets HTTP/1.1" 200 OK
INFO: 127.0.0.1:58406 - "OPTIONS /profiles/c66eb3ab-b68e-44c3-90a2-b5dbcda6c44e HTTP/1.1" 200 OK
INFO: 127.0.0.1:58406 - "GET /profiles/c66eb3ab-b68e-44c3-90a2-b5dbcda6c44e HTTP/1.1" 200 OK
INFO: 127.0.0.1:58406 - "OPTIONS /health HTTP/1.1" 200 OK
INFO: 127.0.0.1:58406 - "GET /health HTTP/1.1" 200 OK
INFO: 127.0.0.1:58406 - "OPTIONS /models/status HTTP/1.1" 200 OK
INFO: 127.0.0.1:52655 - "OPTIONS /models/cache-dir HTTP/1.1" 200 OK
INFO: 127.0.0.1:51968 - "GET /tasks/active HTTP/1.1" 200 OK
INFO: 127.0.0.1:58406 - "GET /models/status HTTP/1.1" 200 OK
INFO: 127.0.0.1:52655 - "GET /models/cache-dir HTTP/1.1" 200 OK
INFO: 127.0.0.1:52655 - "GET /tasks/active HTTP/1.1" 200 OK
INFO: 127.0.0.1:52655 - "GET /models/status HTTP/1.1" 200 OK
INFO: 127.0.0.1:51874 - "OPTIONS /generate/ef6ab366-29db-4f51-868a-5a1b33db9c83/retry HTTP/1.1" 200 OK
INFO: 127.0.0.1:51874 - "POST /generate/ef6ab366-29db-4f51-868a-5a1b33db9c83/retry HTTP/1.1" 200 OK
INFO: 127.0.0.1:51874 - "GET /history?limit=20 HTTP/1.1" 200 OK
INFO: 127.0.0.1:53015 - "GET /generate/ef6ab366-29db-4f51-868a-5a1b33db9c83/status HTTP/1.1" 200 OK
"sox" no se reconoce como un comando interno o externo,
programa o archivo por lotes ejecutable.
2026-04-17 10:24:53,916 - sox - WARNING - SoX could not be found!
If you do not have SoX, proceed here:
- - - http://sox.sourceforge.net/ - - -
If you do (or think that you should) have SoX, double-check your
path variables.
2026-04-17 10:24:54,498 - backend.backends.pytorch_backend - INFO - Loading TTS model 1.7B on cpu...
2026-04-17 10:24:54,498 - backend.utils.hf_offline_patch - INFO - [offline-guard] qwen-tts-1.7B is cached � forcing HF_HUB_OFFLINE=1
2026-04-17 10:24:55,173 - backend.utils.progress - ERROR - Marked qwen-tts-1.7B as error: Unrecognized model in Qwen/Qwen3-TTS-12Hz-1.7B-Base. Should have a model_type key in its config.json, or contain one of the following strings in its name: aimv2, aimv2_vision_model, albert, align, altclip, apertus, arcee, aria, aria_text, audio-spectrogram-transformer, autoformer, aya_vision, bamba, bark, bart, beit, bert, bert-generation, big_bird, bigbird_pegasus, biogpt, bit, bitnet, blenderbot, blenderbot-small, blip, blip-2, blip_2_qformer, bloom, blt, bridgetower, bros, camembert, canine, chameleon, chinese_clip, chinese_clip_vision_model, clap, clip, clip_text_model, clip_vision_model, clipseg, clvp, code_llama, codegen, cohere, cohere2, cohere2_vision, colpali, colqwen2, conditional_detr, convbert, convnext, convnextv2, cpmant, csm, ctrl, cvt, d_fine, dab-detr, dac, data2vec-audio, data2vec-text, data2vec-vision, dbrx, deberta, deberta-v2, decision_transformer, deepseek_v2, deepseek_v3, deepseek_vl, deepseek_vl_hybrid, deformable_detr, deit, depth_anything, depth_pro, deta, detr, dia, diffllama, dinat, dinov2, dinov2_with_registers, dinov3_convnext, dinov3_vit, distilbert, doge, donut-swin, dots1, dpr, dpt, edgetam, edgetam_video, edgetam_vision_model, efficientformer, efficientloftr, efficientnet, electra, emu3, encodec, encoder-decoder, eomt, ernie, ernie4_5, ernie4_5_moe, ernie_m, esm, evolla, exaone4, falcon, falcon_h1, falcon_mamba, fastspeech2_conformer, fastspeech2_conformer_with_hifigan, flaubert, flava, flex_olmo, florence2, fnet, focalnet, fsmt, funnel, fuyu, gemma, gemma2, gemma3, gemma3_text, gemma3n, gemma3n_audio, gemma3n_text, gemma3n_vision, git, glm, glm4, glm4_moe, glm4v, glm4v_moe, glm4v_moe_text, glm4v_text, glpn, got_ocr2, gpt-sw3, gpt2, gpt_bigcode, gpt_neo, gpt_neox, gpt_neox_japanese, gpt_oss, gptj, gptsan-japanese, granite, granite_speech, granitemoe, granitemoehybrid, granitemoeshared, granitevision, graphormer, grounding-dino, groupvit, helium, hgnet_v2, hiera, hubert, hunyuan_v1_dense, hunyuan_v1_moe, ibert, idefics, idefics2, idefics3, idefics3_vision, ijepa, imagegpt, informer, instructblip, instructblipvideo, internvl, internvl_vision, jamba, janus, jetmoe, jukebox, kosmos-2, kosmos-2.5, kyutai_speech_to_text, layoutlm, layoutlmv2, layoutlmv3, led, levit, lfm2, lfm2_vl, lightglue, lilt, llama, llama4, llama4_text, llava, llava_next, llava_next_video, llava_onevision, longcat_flash, longformer, longt5, luke, lxmert, m2m_100, mamba, mamba2, marian, markuplm, mask2former, maskformer, maskformer-swin, mbart, mctct, mega, megatron-bert, metaclip_2, mgp-str, mimi, minimax, ministral, mistral, mistral3, mixtral, mlcd, mllama, mm-grounding-dino, mobilebert, mobilenet_v1, mobilenet_v2, mobilevit, mobilevitv2, modernbert, modernbert-decoder, moonshine, moshi, mpnet, mpt, mra, mt5, musicgen, musicgen_melody, mvp, nat, nemotron, nezha, nllb-moe, nougat, nystromformer, olmo, olmo2, olmo3, olmoe, omdet-turbo, oneformer, open-llama, openai-gpt, opt, ovis2, owlv2, owlvit, paligemma, parakeet_ctc, parakeet_encoder, patchtsmixer, patchtst, pegasus, pegasus_x, perceiver, perception_encoder, perception_lm, persimmon, phi, phi3, phi4_multimodal, phimoe, pix2struct, pixtral, plbart, poolformer, pop2piano, prompt_depth_anything, prophetnet, pvt, pvt_v2, qdqbert, qwen2, qwen2_5_omni, qwen2_5_vl, qwen2_5_vl_text, qwen2_audio, qwen2_audio_encoder, qwen2_moe, qwen2_vl, qwen2_vl_text, qwen3, qwen3_moe, qwen3_next, qwen3_omni_moe, qwen3_vl, qwen3_vl_moe, qwen3_vl_moe_text, qwen3_vl_text, rag, realm, recurrent_gemma, reformer, regnet, rembert, resnet, retribert, roberta, roberta-prelayernorm, roc_bert, roformer, rt_detr, rt_detr_resnet, rt_detr_v2, rwkv, sam, sam2, sam2_hiera_det_model, sam2_video, sam2_vision_model, sam_hq, sam_hq_vision_model, sam_vision_model, seamless_m4t, seamless_m4t_v2, seed_oss, segformer, seggpt, sew, sew-d, shieldgemma2, siglip, siglip2, siglip2_vision_model, siglip_vision_model, smollm3, smolvlm, smolvlm_vision, speech-encoder-decoder, speech_to_text, speech_to_text_2, speecht5, splinter, squeezebert, stablelm, starcoder2, superglue, superpoint, swiftformer, swin, swin2sr, swinv2, switch_transformers, t5, t5gemma, table-transformer, tapas, textnet, time_series_transformer, timesfm, timesformer, timm_backbone, timm_wrapper, trajectory_transformer, transfo-xl, trocr, tvlt, tvp, udop, umt5, unispeech, unispeech-sat, univnet, upernet, van, vaultgemma, video_llava, videomae, vilt, vipllava, vision-encoder-decoder, vision-text-dual-encoder, visual_bert, vit, vit_hybrid, vit_mae, vit_msn, vitdet, vitmatte, vitpose, vitpose_backbone, vits, vivit, vjepa2, voxtral, voxtral_encoder, wav2vec2, wav2vec2-bert, wav2vec2-conformer, wavlm, whisper, xclip, xcodec, xglm, xlm, xlm-prophetnet, xlm-roberta, xlm-roberta-xl, xlnet, xlstm, xmod, yolos, yoso, zamba, zamba2, zoedepth, qwen3_tts
Traceback (most recent call last):
File "backend\services\generation.py", line 63, in run_generation
File "backend\backends_init_.py", line 400, in load_engine_model
File "backend\backends\pytorch_backend.py", line 86, in load_model_async
File "asyncio\threads.py", line 25, in to_thread
File "concurrent\futures\thread.py", line 59, in run
File "backend\backends\pytorch_backend.py", line 111, in _load_model_sync
File "qwen_tts\inference\qwen3_tts_model.py", line 112, in from_pretrained
model = AutoModel.from_pretrained(pretrained_model_name_or_path, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\models\auto\auto_factory.py", line 549, in from_pretrained
config, kwargs = AutoConfig.from_pretrained(
^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\models\auto\configuration_auto.py", line 1380, in from_pretrained
raise ValueError(
ValueError: Unrecognized model in Qwen/Qwen3-TTS-12Hz-1.7B-Base. Should have a model_type key in its config.json, or contain one of the following strings in its name: aimv2, aimv2_vision_model, albert, align, altclip, apertus, arcee, aria, aria_text, audio-spectrogram-transformer, autoformer, aya_vision, bamba, bark, bart, beit, bert, bert-generation, big_bird, bigbird_pegasus, biogpt, bit, bitnet, blenderbot, blenderbot-small, blip, blip-2, blip_2_qformer, bloom, blt, bridgetower, bros, camembert, canine, chameleon, chinese_clip, chinese_clip_vision_model, clap, clip, clip_text_model, clip_vision_model, clipseg, clvp, code_llama, codegen, cohere, cohere2, cohere2_vision, colpali, colqwen2, conditional_detr, convbert, convnext, convnextv2, cpmant, csm, ctrl, cvt, d_fine, dab-detr, dac, data2vec-audio, data2vec-text, data2vec-vision, dbrx, deberta, deberta-v2, decision_transformer, deepseek_v2, deepseek_v3, deepseek_vl, deepseek_vl_hybrid, deformable_detr, deit, depth_anything, depth_pro, deta, detr, dia, diffllama, dinat, dinov2, dinov2_with_registers, dinov3_convnext, dinov3_vit, distilbert, doge, donut-swin, dots1, dpr, dpt, edgetam, edgetam_video, edgetam_vision_model, efficientformer, efficientloftr, efficientnet, electra, emu3, encodec, encoder-decoder, eomt, ernie, ernie4_5, ernie4_5_moe, ernie_m, esm, evolla, exaone4, falcon, falcon_h1, falcon_mamba, fastspeech2_conformer, fastspeech2_conformer_with_hifigan, flaubert, flava, flex_olmo, florence2, fnet, focalnet, fsmt, funnel, fuyu, gemma, gemma2, gemma3, gemma3_text, gemma3n, gemma3n_audio, gemma3n_text, gemma3n_vision, git, glm, glm4, glm4_moe, glm4v, glm4v_moe, glm4v_moe_text, glm4v_text, glpn, got_ocr2, gpt-sw3, gpt2, gpt_bigcode, gpt_neo, gpt_neox, gpt_neox_japanese, gpt_oss, gptj, gptsan-japanese, granite, granite_speech, granitemoe, granitemoehybrid, granitemoeshared, granitevision, graphormer, grounding-dino, groupvit, helium, hgnet_v2, hiera, hubert, hunyuan_v1_dense, hunyuan_v1_moe, ibert, idefics, idefics2, idefics3, idefics3_vision, ijepa, imagegpt, informer, instructblip, instructblipvideo, internvl, internvl_vision, jamba, janus, jetmoe, jukebox, kosmos-2, kosmos-2.5, kyutai_speech_to_text, layoutlm, layoutlmv2, layoutlmv3, led, levit, lfm2, lfm2_vl, lightglue, lilt, llama, llama4, llama4_text, llava, llava_next, llava_next_video, llava_onevision, longcat_flash, longformer, longt5, luke, lxmert, m2m_100, mamba, mamba2, marian, markuplm, mask2former, maskformer, maskformer-swin, mbart, mctct, mega, megatron-bert, metaclip_2, mgp-str, mimi, minimax, ministral, mistral, mistral3, mixtral, mlcd, mllama, mm-grounding-dino, mobilebert, mobilenet_v1, mobilenet_v2, mobilevit, mobilevitv2, modernbert, modernbert-decoder, moonshine, moshi, mpnet, mpt, mra, mt5, musicgen, musicgen_melody, mvp, nat, nemotron, nezha, nllb-moe, nougat, nystromformer, olmo, olmo2, olmo3, olmoe, omdet-turbo, oneformer, open-llama, openai-gpt, opt, ovis2, owlv2, owlvit, paligemma, parakeet_ctc, parakeet_encoder, patchtsmixer, patchtst, pegasus, pegasus_x, perceiver, perception_encoder, perception_lm, persimmon, phi, phi3, phi4_multimodal, phimoe, pix2struct, pixtral, plbart, poolformer, pop2piano, prompt_depth_anything, prophetnet, pvt, pvt_v2, qdqbert, qwen2, qwen2_5_omni, qwen2_5_vl, qwen2_5_vl_text, qwen2_audio, qwen2_audio_encoder, qwen2_moe, qwen2_vl, qwen2_vl_text, qwen3, qwen3_moe, qwen3_next, qwen3_omni_moe, qwen3_vl, qwen3_vl_moe, qwen3_vl_moe_text, qwen3_vl_text, rag, realm, recurrent_gemma, reformer, regnet, rembert, resnet, retribert, roberta, roberta-prelayernorm, roc_bert, roformer, rt_detr, rt_detr_resnet, rt_detr_v2, rwkv, sam, sam2, sam2_hiera_det_model, sam2_video, sam2_vision_model, sam_hq, sam_hq_vision_model, sam_vision_model, seamless_m4t, seamless_m4t_v2, seed_oss, segformer, seggpt, sew, sew-d, shieldgemma2, siglip, siglip2, siglip2_vision_model, siglip_vision_model, smollm3, smolvlm, smolvlm_vision, speech-encoder-decoder, speech_to_text, speech_to_text_2, speecht5, splinter, squeezebert, stablelm, starcoder2, superglue, superpoint, swiftformer, swin, swin2sr, swinv2, switch_transformers, t5, t5gemma, table-transformer, tapas, textnet, time_series_transformer, timesfm, timesformer, timm_backbone, timm_wrapper, trajectory_transformer, transfo-xl, trocr, tvlt, tvp, udop, umt5, unispeech, unispeech-sat, univnet, upernet, van, vaultgemma, video_llava, videomae, vilt, vipllava, vision-encoder-decoder, vision-text-dual-encoder, visual_bert, vit, vit_hybrid, vit_mae, vit_msn, vitdet, vitmatte, vitpose, vitpose_backbone, vits, vivit, vjepa2, voxtral, voxtral_encoder, wav2vec2, wav2vec2-bert, wav2vec2-conformer, wavlm, whisper, xclip, xcodec, xglm, xlm, xlm-prophetnet, xlm-roberta, xlm-roberta-xl, xlnet, xlstm, xmod, yolos, yoso, zamba, zamba2, zoedepth, qwen3_tts

---

Warning: flash-attn is not installed. Will only run the manual PyTorch version. Please install flash-attn for faster inference.

---

INFO: 127.0.0.1:53015 - "GET /history?limit=20 HTTP/1.1" 200 OK
INFO: 127.0.0.1:53015 - "GET /tasks/active HTTP/1.1" 200 OK
INFO: 127.0.0.1:56957 - "GET /tasks/active HTTP/1.1" 200 OK
INFO: 127.0.0.1:53291 - "GET /tasks/active HTTP/1.1" 200 OK
INFO: 127.0.0.1:65383 - "GET /tasks/active HTTP/1.1" 200 OK
INFO: 127.0.0.1:54082 - "OPTIONS /backend/cuda-status HTTP/1.1" 200 OK
INFO: 127.0.0.1:54082 - "GET /backend/cuda-status HTTP/1.1" 200 OK
INFO: 127.0.0.1:59902 - "OPTIONS /backend/download-cuda HTTP/1.1" 200 OK
INFO: 127.0.0.1:59902 - "POST /backend/download-cuda HTTP/1.1" 200 OK
2026-04-17 10:26:48,045 - backend.services.cuda - INFO - Starting CUDA backend download for v0.4.0 (server=yes, libs=yes)
INFO: 127.0.0.1:59902 - "GET /backend/cuda-status HTTP/1.1" 200 OK
INFO: 127.0.0.1:59902 - "GET /backend/cuda-progress HTTP/1.1" 200 OK
2026-04-17 10:26:48,335 - backend.utils.progress - INFO - SSE client subscribed to cuda-backend, total listeners: 1
2026-04-17 10:26:48,335 - backend.utils.progress - INFO - Sending initial progress for cuda-backend: downloading
2026-04-17 10:26:48,660 - httpx - INFO - HTTP Request: HEAD https://github.com/jamiepine/voicebox/releases/download/v0.4.0/voicebox-server-cuda.tar.gz "HTTP/1.1 302 Found"
2026-04-17 10:26:48,810 - httpx - INFO - HTTP Request: HEAD https://release-assets.githubusercontent.com/github-production-release-asset/1141782198/3db93613-8c63-4fd2-81f4-c3da29c50ec3?sp=r&sv=2018-11-09&sr=b&spr=https&se=2026-04-17T09%3A14%3A11Z&rscd=attachment%3B+filename%3Dvoicebox-server-cuda.tar.gz&rsct=application%2Foctet-stream&skoid=96c2d410-5711-43a1-aedd-ab1947aa7ab0&sktid=398a6654-997b-47e9-b12b-9515b896b4de&skt=2026-04-17T08%3A13%3A50Z&ske=2026-04-17T09%3A14%3A11Z&sks=b&skv=2018-11-09&sig=9GKNV69qv6RwJx2vpZBgcNElWfbhpBOZs11cQipprgE%3D&jwt=eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJpc3MiOiJnaXRodWIuY29tIiwiYXVkIjoicmVsZWFzZS1hc3NldHMuZ2l0aHVidXNlcmNvbnRlbnQuY29tIiwia2V5Ijoia2V5MSIsImV4cCI6MTc3NjQxODAwOSwibmJmIjoxNzc2NDE0NDA5LCJwYXRoIjoicmVsZWFzZWFzc2V0cHJvZHVjdGlvbi5ibG9iLmNvcmUud2luZG93cy5uZXQifQ.UoqOgAUEVuq3DQ4d3TZkkVkDCERyBf7Fn3SNLrXcc9s&response-content-disposition=attachment%3B%20filename%3Dvoicebox-server-cuda.tar.gz&response-content-type=application%2Foctet-stream "HTTP/1.1 200 OK"
2026-04-17 10:26:49,010 - httpx - INFO - HTTP Request: HEAD https://github.com/jamiepine/voicebox/releases/download/v0.4.0/cuda-libs-cu128-v1.tar.gz "HTTP/1.1 302 Found"
2026-04-17 10:26:49,119 - httpx - INFO - HTTP Request: HEAD https://release-assets.githubusercontent.com/github-production-release-asset/1141782198/9a363c7e-9f95-46dc-ba06-973f7c553d44?sp=r&sv=2018-11-09&sr=b&spr=https&se=2026-04-17T09%3A11%3A25Z&rscd=attachment%3B+filename%3Dcuda-libs-cu128-v1.tar.gz&rsct=application%2Foctet-stream&skoid=96c2d410-5711-43a1-aedd-ab1947aa7ab0&sktid=398a6654-997b-47e9-b12b-9515b896b4de&skt=2026-04-17T08%3A10%3A54Z&ske=2026-04-17T09%3A11%3A25Z&sks=b&skv=2018-11-09&sig=2wxWvXq8XCtWgt1LFW9QoIL7vaWqqM1v8H9fQE77uzM%3D&jwt=eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJpc3MiOiJnaXRodWIuY29tIiwiYXVkIjoicmVsZWFzZS1hc3NldHMuZ2l0aHVidXNlcmNvbnRlbnQuY29tIiwia2V5Ijoia2V5MSIsImV4cCI6MTc3NjQxODAwOSwibmJmIjoxNzc2NDE0NDA5LCJwYXRoIjoicmVsZWFzZWFzc2V0cHJvZHVjdGlvbi5ibG9iLmNvcmUud2luZG93cy5uZXQifQ.UoqOgAUEVuq3DQ4d3TZkkVkDCERyBf7Fn3SNLrXcc9s&response-content-disposition=attachment%3B%20filename%3Dcuda-libs-cu128-v1.tar.gz&response-content-type=application%2Foctet-stream "HTTP/1.1 200 OK"
2026-04-17 10:26:49,119 - backend.services.cuda - INFO - Total download size: 2939.9 MB
2026-04-17 10:26:49,314 - httpx - INFO - HTTP Request: GET https://github.com/jamiepine/voicebox/releases/download/v0.4.0/voicebox-server-cuda.tar.gz.sha256 "HTTP/1.1 302 Found"
2026-04-17 10:26:49,423 - httpx - INFO - HTTP Request: GET https://release-assets.githubusercontent.com/github-production-release-asset/1141782198/7e3bac82-5098-45cd-a8f7-04264711120e?sp=r&sv=2018-11-09&sr=b&spr=https&se=2026-04-17T09%3A11%3A20Z&rscd=attachment%3B+filename%3Dvoicebox-server-cuda.tar.gz.sha256&rsct=application%2Foctet-stream&skoid=96c2d410-5711-43a1-aedd-ab1947aa7ab0&sktid=398a6654-997b-47e9-b12b-9515b896b4de&skt=2026-04-17T08%3A10%3A59Z&ske=2026-04-17T09%3A11%3A20Z&sks=b&skv=2018-11-09&sig=tSR1etf8Z2YiI4%2FQILFAltKODjixvG0XYcYQ%2FiJPG%2Bk%3D&jwt=eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJpc3MiOiJnaXRodWIuY29tIiwiYXVkIjoicmVsZWFzZS1hc3NldHMuZ2l0aHVidXNlcmNvbnRlbnQuY29tIiwia2V5Ijoia2V5MSIsImV4cCI6MTc3NjQxNDcwOSwibmJmIjoxNzc2NDE0NDA5LCJwYXRoIjoicmVsZWFzZWFzc2V0cHJvZHVjdGlvbi5ibG9iLmNvcmUud2luZG93cy5uZXQifQ.4gsgsbzNpNT32XOPFRR7nKV-b65Om4aCnW3KUuA8rQw&response-content-disposition=attachment%3B%20filename%3Dvoicebox-server-cuda.tar.gz.sha256&response-content-type=application%2Foctet-stream "HTTP/1.1 200 OK"
2026-04-17 10:26:49,424 - backend.services.cuda - INFO - CUDA server: expected SHA-256: 16b617eb982872cc...
2026-04-17 10:26:49,463 - httpx - INFO - HTTP Request: GET https://github.com/jamiepine/voicebox/releases/download/v0.4.0/voicebox-server-cuda.tar.gz "HTTP/1.1 302 Found"
2026-04-17 10:26:49,517 - httpx - INFO - HTTP Request: GET https://release-assets.githubusercontent.com/github-production-release-asset/1141782198/3db93613-8c63-4fd2-81f4-c3da29c50ec3?sp=r&sv=2018-11-09&sr=b&spr=https&se=2026-04-17T09%3A14%3A11Z&rscd=attachment%3B+filename%3Dvoicebox-server-cuda.tar.gz&rsct=application%2Foctet-stream&skoid=96c2d410-5711-43a1-aedd-ab1947aa7ab0&sktid=398a6654-997b-47e9-b12b-9515b896b4de&skt=2026-04-17T08%3A13%3A50Z&ske=2026-04-17T09%3A14%3A11Z&sks=b&skv=2018-11-09&sig=9GKNV69qv6RwJx2vpZBgcNElWfbhpBOZs11cQipprgE%3D&jwt=eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJpc3MiOiJnaXRodWIuY29tIiwiYXVkIjoicmVsZWFzZS1hc3NldHMuZ2l0aHVidXNlcmNvbnRlbnQuY29tIiwia2V5Ijoia2V5MSIsImV4cCI6MTc3NjQxODAwOSwibmJmIjoxNzc2NDE0NDA5LCJwYXRoIjoicmVsZWFzZWFzc2V0cHJvZHVjdGlvbi5ibG9iLmNvcmUud2luZG93cy5uZXQifQ.UoqOgAUEVuq3DQ4d3TZkkVkDCERyBf7Fn3SNLrXcc9s&response-content-disposition=attachment%3B%20filename%3Dvoicebox-server-cuda.tar.gz&response-content-type=application%2Foctet-stream "HTTP/1.1 200 OK"
INFO: 127.0.0.1:58834 - "GET /backend/cuda-status HTTP/1.1" 200 OK
INFO: 127.0.0.1:58834 - "GET /tasks/active HTTP/1.1" 200 OK
INFO: 127.0.0.1:58834 - "GET /models/progress/cuda-backend HTTP/1.1" 200 OK
2026-04-17 10:26:58,423 - backend.utils.progress - INFO - SSE client subscribed to cuda-backend, total listeners: 2
2026-04-17 10:26:58,423 - backend.utils.progress - INFO - Sending initial progress for cuda-backend: downloading
2026-04-17 10:27:07,691 - backend.services.cuda - INFO - CUDA server: integrity verified
2026-04-17 10:27:29,922 - backend.services.cuda - INFO - CUDA server: extracted to C:\Users\cosmi\AppData\Roaming\sh.voicebox.app\backends\cuda
INFO: 127.0.0.1:60471 - "GET /backend/cuda-status HTTP/1.1" 200 OK
INFO: 127.0.0.1:56140 - "GET /health HTTP/1.1" 200 OK
INFO: 127.0.0.1:54306 - "GET /tasks/active HTTP/1.1" 200 OK
2026-04-17 10:27:30,357 - httpx - INFO - HTTP Request: GET https://github.com/jamiepine/voicebox/releases/download/v0.4.0/cuda-libs-cu128-v1.tar.gz.sha256 "HTTP/1.1 302 Found"
2026-04-17 10:27:30,493 - httpx - INFO - HTTP Request: GET https://release-assets.githubusercontent.com/github-production-release-asset/1141782198/1e295796-acf2-4f9d-84ab-c1c62895fc80?sp=r&sv=2018-11-09&sr=b&spr=https&se=2026-04-17T09%3A04%3A58Z&rscd=attachment%3B+filename%3Dcuda-libs-cu128-v1.tar.gz.sha256&rsct=application%2Foctet-stream&skoid=96c2d410-5711-43a1-aedd-ab1947aa7ab0&sktid=398a6654-997b-47e9-b12b-9515b896b4de&skt=2026-04-17T08%3A04%3A56Z&ske=2026-04-17T09%3A04%3A58Z&sks=b&skv=2018-11-09&sig=L0eNHGeopZu6fSrlKYRGo42E02jV%2FbzQDRK6Xjbst%2Bk%3D&jwt=eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJpc3MiOiJnaXRodWIuY29tIiwiYXVkIjoicmVsZWFzZS1hc3NldHMuZ2l0aHVidXNlcmNvbnRlbnQuY29tIiwia2V5Ijoia2V5MSIsImV4cCI6MTc3NjQxNDc1MSwibmJmIjoxNzc2NDE0NDUxLCJwYXRoIjoicmVsZWFzZWFzc2V0cHJvZHVjdGlvbi5ibG9iLmNvcmUud2luZG93cy5uZXQifQ.rf-AAysu1-V1OyF31WR_B-DGgCCRYqkXghC5aQp_6lA&response-content-disposition=attachment%3B%20filename%3Dcuda-libs-cu128-v1.tar.gz.sha256&response-content-type=application%2Foctet-stream "HTTP/1.1 200 OK"
2026-04-17 10:27:30,493 - backend.services.cuda - INFO - CUDA libraries: expected SHA-256: 7d52ba9d2bbb0265...
2026-04-17 10:27:30,531 - httpx - INFO - HTTP Request: GET https://github.com/jamiepine/voicebox/releases/download/v0.4.0/cuda-libs-cu128-v1.tar.gz "HTTP/1.1 302 Found"
2026-04-17 10:27:30,584 - httpx - INFO - HTTP Request: GET https://release-assets.githubusercontent.com/github-production-release-asset/1141782198/9a363c7e-9f95-46dc-ba06-973f7c553d44?sp=r&sv=2018-11-09&sr=b&spr=https&se=2026-04-17T09%3A11%3A25Z&rscd=attachment%3B+filename%3Dcuda-libs-cu128-v1.tar.gz&rsct=application%2Foctet-stream&skoid=96c2d410-5711-43a1-aedd-ab1947aa7ab0&sktid=398a6654-997b-47e9-b12b-9515b896b4de&skt=2026-04-17T08%3A10%3A54Z&ske=2026-04-17T09%3A11%3A25Z&sks=b&skv=2018-11-09&sig=2wxWvXq8XCtWgt1LFW9QoIL7vaWqqM1v8H9fQE77uzM%3D&jwt=eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJpc3MiOiJnaXRodWIuY29tIiwiYXVkIjoicmVsZWFzZS1hc3NldHMuZ2l0aHVidXNlcmNvbnRlbnQuY29tIiwia2V5Ijoia2V5MSIsImV4cCI6MTc3NjQxODAwOSwibmJmIjoxNzc2NDE0NDA5LCJwYXRoIjoicmVsZWFzZWFzc2V0cHJvZHVjdGlvbi5ibG9iLmNvcmUud2luZG93cy5uZXQifQ.UoqOgAUEVuq3DQ4d3TZkkVkDCERyBf7Fn3SNLrXcc9s&response-content-disposition=attachment%3B%20filename%3Dcuda-libs-cu128-v1.tar.gz&response-content-type=application%2Foctet-stream "HTTP/1.1 200 OK"
INFO: 127.0.0.1:58228 - "GET /backend/cuda-status HTTP/1.1" 200 OK
INFO: 127.0.0.1:65304 - "GET /backend/cuda-status HTTP/1.1" 200 OK
INFO: 127.0.0.1:56779 - "GET /tasks/active HTTP/1.1" 200 OK
INFO: 127.0.0.1:56779 - "GET /health HTTP/1.1" 200 OK
INFO: 127.0.0.1:56779 - "GET /backend/cuda-status HTTP/1.1" 200 OK
INFO: 127.0.0.1:50786 - "GET /backend/cuda-status HTTP/1.1" 200 OK
2026-04-17 10:28:16,069 - backend.services.cuda - INFO - CUDA libraries: integrity verified
2026-04-17 10:28:45,293 - backend.services.cuda - INFO - CUDA libraries: extracted to C:\Users\cosmi\AppData\Roaming\sh.voicebox.app\backends\cuda
2026-04-17 10:28:45,500 - backend.services.cuda - INFO - CUDA backend ready at C:\Users\cosmi\AppData\Roaming\sh.voicebox.app\backends\cuda
2026-04-17 10:28:45,501 - backend.utils.progress - INFO - Marked cuda-backend as complete
2026-04-17 10:28:45,501 - backend.utils.progress - INFO - Download complete for cuda-backend, closing SSE connection
2026-04-17 10:28:45,501 - backend.utils.progress - INFO - SSE client unsubscribed from cuda-backend, remaining listeners: 1
2026-04-17 10:28:45,501 - backend.utils.progress - INFO - Download complete for cuda-backend, closing SSE connection
2026-04-17 10:28:45,501 - backend.utils.progress - INFO - SSE client unsubscribed from cuda-backend, remaining listeners: 0
INFO: 127.0.0.1:50786 - "GET /backend/cuda-status HTTP/1.1" 200 OK
INFO: 127.0.0.1:55762 - "GET /tasks/active HTTP/1.1" 200 OK
INFO: 127.0.0.1:54688 - "GET /health HTTP/1.1" 200 OK
INFO: 127.0.0.1:54629 - "GET /backend/cuda-status HTTP/1.1" 200 OK
INFO: 127.0.0.1:54633 - "POST /shutdown HTTP/1.1" 200 OK
2026-04-17 10:28:58,483 - main - INFO - ============================================================
2026-04-17 10:28:58,483 - main - INFO - voicebox-server starting up...
2026-04-17 10:28:58,483 - main - INFO - Python version: 3.12.10 (tags/v3.12.10:0cc8128, Apr 8 2025, 12:21:36) [MSC v.1943 64 bit (AMD64)]
2026-04-17 10:28:58,483 - main - INFO - Executable: C:\Users\cosmi\AppData\Roaming\sh.voicebox.app\backends\cuda\voicebox-server-cuda.exe
2026-04-17 10:28:58,483 - main - INFO - Arguments: ['C:\Users\cosmi\AppData\Roaming\sh.voicebox.app\backends\cuda\voicebox-server-cuda.exe', '--data-dir', 'C:\Users\cosmi\AppData\Roaming\sh.voicebox.app', '--port', '17493', '--parent-pid', '8908']
2026-04-17 10:28:58,483 - main - INFO - ============================================================
2026-04-17 10:28:58,483 - main - INFO - Importing argparse...
2026-04-17 10:28:58,486 - main - INFO - Importing uvicorn...
2026-04-17 10:28:58,576 - main - INFO - Standard library imports successful
2026-04-17 10:28:58,576 - main - INFO - Importing backend.config...
2026-04-17 10:28:58,577 - main - INFO - Importing backend.database...
2026-04-17 10:28:58,758 - main - INFO - Importing backend.main (this may take a while due to torch/transformers)...
2026-04-17 10:29:01,924 - main - INFO - Backend imports successful
2026-04-17 10:29:01,925 - main - INFO - Backend variant: CUDA
server.py:278: DeprecationWarning:
on_event is deprecated, use lifespan event handlers instead.
Read more about it in the
FastAPI docs for Lifespan Events.
2026-04-17 10:29:01,925 - main - INFO - Parsed arguments: host=127.0.0.1, port=17493, data_dir=C:\Users\cosmi\AppData\Roaming\sh.voicebox.app
2026-04-17 10:29:01,925 - main - INFO - Setting data directory to: C:\Users\cosmi\AppData\Roaming\sh.voicebox.app
2026-04-17 10:29:01,926 - backend.config - INFO - Data directory set to: C:\Users\cosmi\AppData\Roaming\sh.voicebox.app
2026-04-17 10:29:01,926 - main - INFO - Initializing database...
2026-04-17 10:29:01,971 - main - INFO - Database initialized successfully
2026-04-17 10:29:01,971 - main - INFO - Starting uvicorn server on 127.0.0.1:17493...
INFO: Started server process [43744]
INFO: Waiting for application startup.
2026-04-17 10:29:01,998 - backend.app - INFO - Voicebox v0.4.0 starting up
2026-04-17 10:29:01,999 - backend.app - INFO - Python 3.12.10 on Windows 11 (AMD64)
2026-04-17 10:29:02,010 - backend.app - INFO - Database: C:\Users\cosmi\AppData\Roaming\sh.voicebox.app\voicebox.db
2026-04-17 10:29:02,010 - backend.app - INFO - Data directory: C:\Users\cosmi\AppData\Roaming\sh.voicebox.app
2026-04-17 10:29:02,014 - backend.app - INFO - Profiles: 1, Generations: 2
2026-04-17 10:29:02,014 - backend.app - INFO - Backend: PYTORCH
2026-04-17 10:29:02,034 - backend.app - INFO - GPU: CUDA (NVIDIA GeForce RTX 5080)
2026-04-17 10:29:02,036 - backend.app - INFO - Model cache: C:\Users\cosmi.cache\huggingface\hub
2026-04-17 10:29:02,036 - backend.app - INFO - Ready
2026-04-17 10:29:02,038 - watchdog - INFO - Parent watchdog started, monitoring PID 8908, server PID 43744
2026-04-17 10:29:02,038 - watchdog - INFO - Parent PID 8908 initial check: alive=True
2026-04-17 10:29:02,416 - backend.services.cuda - INFO - CUDA binary is up to date (server=v0.4.0, libs=cu128-v1)
INFO: Application startup complete.
INFO: Uvicorn running on http://127.0.0.1:17493 (Press CTRL+C to quit)
INFO: 127.0.0.1:64350 - "GET /health HTTP/1.1" 200 OK
INFO: 127.0.0.1:64350 - "GET /health HTTP/1.1" 200 OK
INFO: 127.0.0.1:59310 - "GET /backend/cuda-status HTTP/1.1" 200 OK
INFO: 127.0.0.1:55113 - "GET /models/cache-dir HTTP/1.1" 200 OK
INFO: 127.0.0.1:60044 - "GET /models/status HTTP/1.1" 200 OK
INFO: 127.0.0.1:56714 - "GET /tasks/active HTTP/1.1" 200 OK
INFO: 127.0.0.1:55113 - "GET /profiles/c66eb3ab-b68e-44c3-90a2-b5dbcda6c44e HTTP/1.1" 200 OK
INFO: 127.0.0.1:60044 - "GET /history?limit=20 HTTP/1.1" 200 OK
INFO: 127.0.0.1:56714 - "GET /profiles HTTP/1.1" 200 OK
INFO: 127.0.0.1:64873 - "GET /effects/presets HTTP/1.1" 200 OK
INFO: 127.0.0.1:64873 - "POST /generate/ef6ab366-29db-4f51-868a-5a1b33db9c83/retry HTTP/1.1" 200 OK
INFO: 127.0.0.1:64873 - "GET /history?limit=20 HTTP/1.1" 200 OK
INFO: 127.0.0.1:60044 - "GET /generate/ef6ab366-29db-4f51-868a-5a1b33db9c83/status HTTP/1.1" 200 OK
"sox" no se reconoce como un comando interno o externo,
programa o archivo por lotes ejecutable.
2026-04-17 10:29:23,386 - sox - WARNING - SoX could not be found!
If you do not have SoX, proceed here:
- - - http://sox.sourceforge.net/ - - -
If you do (or think that you should) have SoX, double-check your
path variables.
2026-04-17 10:29:23,960 - backend.backends.pytorch_backend - INFO - Loading TTS model 1.7B on cuda...
2026-04-17 10:29:23,960 - backend.utils.hf_offline_patch - INFO - [offline-guard] qwen-tts-1.7B is cached � forcing HF_HUB_OFFLINE=1
2026-04-17 10:29:24,629 - backend.utils.progress - ERROR - Marked qwen-tts-1.7B as error: Unrecognized model in Qwen/Qwen3-TTS-12Hz-1.7B-Base. Should have a model_type key in its config.json, or contain one of the following strings in its name: aimv2, aimv2_vision_model, albert, align, altclip, apertus, arcee, aria, aria_text, audio-spectrogram-transformer, autoformer, aya_vision, bamba, bark, bart, beit, bert, bert-generation, big_bird, bigbird_pegasus, biogpt, bit, bitnet, blenderbot, blenderbot-small, blip, blip-2, blip_2_qformer, bloom, blt, bridgetower, bros, camembert, canine, chameleon, chinese_clip, chinese_clip_vision_model, clap, clip, clip_text_model, clip_vision_model, clipseg, clvp, code_llama, codegen, cohere, cohere2, cohere2_vision, colpali, colqwen2, conditional_detr, convbert, convnext, convnextv2, cpmant, csm, ctrl, cvt, d_fine, dab-detr, dac, data2vec-audio, data2vec-text, data2vec-vision, dbrx, deberta, deberta-v2, decision_transformer, deepseek_v2, deepseek_v3, deepseek_vl, deepseek_vl_hybrid, deformable_detr, deit, depth_anything, depth_pro, deta, detr, dia, diffllama, dinat, dinov2, dinov2_with_registers, dinov3_convnext, dinov3_vit, distilbert, doge, donut-swin, dots1, dpr, dpt, edgetam, edgetam_video, edgetam_vision_model, efficientformer, efficientloftr, efficientnet, electra, emu3, encodec, encoder-decoder, eomt, ernie, ernie4_5, ernie4_5_moe, ernie_m, esm, evolla, exaone4, falcon, falcon_h1, falcon_mamba, fastspeech2_conformer, fastspeech2_conformer_with_hifigan, flaubert, flava, flex_olmo, florence2, fnet, focalnet, fsmt, funnel, fuyu, gemma, gemma2, gemma3, gemma3_text, gemma3n, gemma3n_audio, gemma3n_text, gemma3n_vision, git, glm, glm4, glm4_moe, glm4v, glm4v_moe, glm4v_moe_text, glm4v_text, glpn, got_ocr2, gpt-sw3, gpt2, gpt_bigcode, gpt_neo, gpt_neox, gpt_neox_japanese, gpt_oss, gptj, gptsan-japanese, granite, granite_speech, granitemoe, granitemoehybrid, granitemoeshared, granitevision, graphormer, grounding-dino, groupvit, helium, hgnet_v2, hiera, hubert, hunyuan_v1_dense, hunyuan_v1_moe, ibert, idefics, idefics2, idefics3, idefics3_vision, ijepa, imagegpt, informer, instructblip, instructblipvideo, internvl, internvl_vision, jamba, janus, jetmoe, jukebox, kosmos-2, kosmos-2.5, kyutai_speech_to_text, layoutlm, layoutlmv2, layoutlmv3, led, levit, lfm2, lfm2_vl, lightglue, lilt, llama, llama4, llama4_text, llava, llava_next, llava_next_video, llava_onevision, longcat_flash, longformer, longt5, luke, lxmert, m2m_100, mamba, mamba2, marian, markuplm, mask2former, maskformer, maskformer-swin, mbart, mctct, mega, megatron-bert, metaclip_2, mgp-str, mimi, minimax, ministral, mistral, mistral3, mixtral, mlcd, mllama, mm-grounding-dino, mobilebert, mobilenet_v1, mobilenet_v2, mobilevit, mobilevitv2, modernbert, modernbert-decoder, moonshine, moshi, mpnet, mpt, mra, mt5, musicgen, musicgen_melody, mvp, nat, nemotron, nezha, nllb-moe, nougat, nystromformer, olmo, olmo2, olmo3, olmoe, omdet-turbo, oneformer, open-llama, openai-gpt, opt, ovis2, owlv2, owlvit, paligemma, parakeet_ctc, parakeet_encoder, patchtsmixer, patchtst, pegasus, pegasus_x, perceiver, perception_encoder, perception_lm, persimmon, phi, phi3, phi4_multimodal, phimoe, pix2struct, pixtral, plbart, poolformer, pop2piano, prompt_depth_anything, prophetnet, pvt, pvt_v2, qdqbert, qwen2, qwen2_5_omni, qwen2_5_vl, qwen2_5_vl_text, qwen2_audio, qwen2_audio_encoder, qwen2_moe, qwen2_vl, qwen2_vl_text, qwen3, qwen3_moe, qwen3_next, qwen3_omni_moe, qwen3_vl, qwen3_vl_moe, qwen3_vl_moe_text, qwen3_vl_text, rag, realm, recurrent_gemma, reformer, regnet, rembert, resnet, retribert, roberta, roberta-prelayernorm, roc_bert, roformer, rt_detr, rt_detr_resnet, rt_detr_v2, rwkv, sam, sam2, sam2_hiera_det_model, sam2_video, sam2_vision_model, sam_hq, sam_hq_vision_model, sam_vision_model, seamless_m4t, seamless_m4t_v2, seed_oss, segformer, seggpt, sew, sew-d, shieldgemma2, siglip, siglip2, siglip2_vision_model, siglip_vision_model, smollm3, smolvlm, smolvlm_vision, speech-encoder-decoder, speech_to_text, speech_to_text_2, speecht5, splinter, squeezebert, stablelm, starcoder2, superglue, superpoint, swiftformer, swin, swin2sr, swinv2, switch_transformers, t5, t5gemma, table-transformer, tapas, textnet, time_series_transformer, timesfm, timesformer, timm_backbone, timm_wrapper, trajectory_transformer, transfo-xl, trocr, tvlt, tvp, udop, umt5, unispeech, unispeech-sat, univnet, upernet, van, vaultgemma, video_llava, videomae, vilt, vipllava, vision-encoder-decoder, vision-text-dual-encoder, visual_bert, vit, vit_hybrid, vit_mae, vit_msn, vitdet, vitmatte, vitpose, vitpose_backbone, vits, vivit, vjepa2, voxtral, voxtral_encoder, wav2vec2, wav2vec2-bert, wav2vec2-conformer, wavlm, whisper, xclip, xcodec, xglm, xlm, xlm-prophetnet, xlm-roberta, xlm-roberta-xl, xlnet, xlstm, xmod, yolos, yoso, zamba, zamba2, zoedepth, qwen3_tts
Traceback (most recent call last):
File "backend\services\generation.py", line 63, in run_generation
File "backend\backends_init_.py", line 400, in load_engine_model
File "backend\backends\pytorch_backend.py", line 86, in load_model_async
File "asyncio\threads.py", line 25, in to_thread
File "concurrent\futures\thread.py", line 59, in run
File "backend\backends\pytorch_backend.py", line 118, in _load_model_sync
File "qwen_tts\inference\qwen3_tts_model.py", line 112, in from_pretrained
model = AutoModel.from_pretrained(pretrained_model_name_or_path, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\models\auto\auto_factory.py", line 549, in from_pretrained
config, kwargs = AutoConfig.from_pretrained(
^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\models\auto\configuration_auto.py", line 1380, in from_pretrained
raise ValueError(
ValueError: Unrecognized model in Qwen/Qwen3-TTS-12Hz-1.7B-Base. Should have a model_type key in its config.json, or contain one of the following strings in its name: aimv2, aimv2_vision_model, albert, align, altclip, apertus, arcee, aria, aria_text, audio-spectrogram-transformer, autoformer, aya_vision, bamba, bark, bart, beit, bert, bert-generation, big_bird, bigbird_pegasus, biogpt, bit, bitnet, blenderbot, blenderbot-small, blip, blip-2, blip_2_qformer, bloom, blt, bridgetower, bros, camembert, canine, chameleon, chinese_clip, chinese_clip_vision_model, clap, clip, clip_text_model, clip_vision_model, clipseg, clvp, code_llama, codegen, cohere, cohere2, cohere2_vision, colpali, colqwen2, conditional_detr, convbert, convnext, convnextv2, cpmant, csm, ctrl, cvt, d_fine, dab-detr, dac, data2vec-audio, data2vec-text, data2vec-vision, dbrx, deberta, deberta-v2, decision_transformer, deepseek_v2, deepseek_v3, deepseek_vl, deepseek_vl_hybrid, deformable_detr, deit, depth_anything, depth_pro, deta, detr, dia, diffllama, dinat, dinov2, dinov2_with_registers, dinov3_convnext, dinov3_vit, distilbert, doge, donut-swin, dots1, dpr, dpt, edgetam, edgetam_video, edgetam_vision_model, efficientformer, efficientloftr, efficientnet, electra, emu3, encodec, encoder-decoder, eomt, ernie, ernie4_5, ernie4_5_moe, ernie_m, esm, evolla, exaone4, falcon, falcon_h1, falcon_mamba, fastspeech2_conformer, fastspeech2_conformer_with_hifigan, flaubert, flava, flex_olmo, florence2, fnet, focalnet, fsmt, funnel, fuyu, gemma, gemma2, gemma3, gemma3_text, gemma3n, gemma3n_audio, gemma3n_text, gemma3n_vision, git, glm, glm4, glm4_moe, glm4v, glm4v_moe, glm4v_moe_text, glm4v_text, glpn, got_ocr2, gpt-sw3, gpt2, gpt_bigcode, gpt_neo, gpt_neox, gpt_neox_japanese, gpt_oss, gptj, gptsan-japanese, granite, granite_speech, granitemoe, granitemoehybrid, granitemoeshared, granitevision, graphormer, grounding-dino, groupvit, helium, hgnet_v2, hiera, hubert, hunyuan_v1_dense, hunyuan_v1_moe, ibert, idefics, idefics2, idefics3, idefics3_vision, ijepa, imagegpt, informer, instructblip, instructblipvideo, internvl, internvl_vision, jamba, janus, jetmoe, jukebox, kosmos-2, kosmos-2.5, kyutai_speech_to_text, layoutlm, layoutlmv2, layoutlmv3, led, levit, lfm2, lfm2_vl, lightglue, lilt, llama, llama4, llama4_text, llava, llava_next, llava_next_video, llava_onevision, longcat_flash, longformer, longt5, luke, lxmert, m2m_100, mamba, mamba2, marian, markuplm, mask2former, maskformer, maskformer-swin, mbart, mctct, mega, megatron-bert, metaclip_2, mgp-str, mimi, minimax, ministral, mistral, mistral3, mixtral, mlcd, mllama, mm-grounding-dino, mobilebert, mobilenet_v1, mobilenet_v2, mobilevit, mobilevitv2, modernbert, modernbert-decoder, moonshine, moshi, mpnet, mpt, mra, mt5, musicgen, musicgen_melody, mvp, nat, nemotron, nezha, nllb-moe, nougat, nystromformer, olmo, olmo2, olmo3, olmoe, omdet-turbo, oneformer, open-llama, openai-gpt, opt, ovis2, owlv2, owlvit, paligemma, parakeet_ctc, parakeet_encoder, patchtsmixer, patchtst, pegasus, pegasus_x, perceiver, perception_encoder, perception_lm, persimmon, phi, phi3, phi4_multimodal, phimoe, pix2struct, pixtral, plbart, poolformer, pop2piano, prompt_depth_anything, prophetnet, pvt, pvt_v2, qdqbert, qwen2, qwen2_5_omni, qwen2_5_vl, qwen2_5_vl_text, qwen2_audio, qwen2_audio_encoder, qwen2_moe, qwen2_vl, qwen2_vl_text, qwen3, qwen3_moe, qwen3_next, qwen3_omni_moe, qwen3_vl, qwen3_vl_moe, qwen3_vl_moe_text, qwen3_vl_text, rag, realm, recurrent_gemma, reformer, regnet, rembert, resnet, retribert, roberta, roberta-prelayernorm, roc_bert, roformer, rt_detr, rt_detr_resnet, rt_detr_v2, rwkv, sam, sam2, sam2_hiera_det_model, sam2_video, sam2_vision_model, sam_hq, sam_hq_vision_model, sam_vision_model, seamless_m4t, seamless_m4t_v2, seed_oss, segformer, seggpt, sew, sew-d, shieldgemma2, siglip, siglip2, siglip2_vision_model, siglip_vision_model, smollm3, smolvlm, smolvlm_vision, speech-encoder-decoder, speech_to_text, speech_to_text_2, speecht5, splinter, squeezebert, stablelm, starcoder2, superglue, superpoint, swiftformer, swin, swin2sr, swinv2, switch_transformers, t5, t5gemma, table-transformer, tapas, textnet, time_series_transformer, timesfm, timesformer, timm_backbone, timm_wrapper, trajectory_transformer, transfo-xl, trocr, tvlt, tvp, udop, umt5, unispeech, unispeech-sat, univnet, upernet, van, vaultgemma, video_llava, videomae, vilt, vipllava, vision-encoder-decoder, vision-text-dual-encoder, visual_bert, vit, vit_hybrid, vit_mae, vit_msn, vitdet, vitmatte, vitpose, vitpose_backbone, vits, vivit, vjepa2, voxtral, voxtral_encoder, wav2vec2, wav2vec2-bert, wav2vec2-conformer, wavlm, whisper, xclip, xcodec, xglm, xlm, xlm-prophetnet, xlm-roberta, xlm-roberta-xl, xlnet, xlstm, xmod, yolos, yoso, zamba, zamba2, zoedepth, qwen3_tts

---

Warning: flash-attn is not installed. Will only run the manual PyTorch version. Please install flash-attn for faster inference.

---

INFO: 127.0.0.1:60044 - "GET /history?limit=20 HTTP/1.1" 200 OK
INFO: 127.0.0.1:60044 - "GET /tasks/active HTTP/1.1" 200 OK
INFO: 127.0.0.1:57680 - "GET /tasks/active HTTP/1.1" 200 OK
INFO: 127.0.0.1:59330 - "OPTIONS /generate/2af49cc4-bbd3-473c-952b-8c92c5b3353b/retry HTTP/1.1" 200 OK
INFO: 127.0.0.1:59330 - "POST /generate/2af49cc4-bbd3-473c-952b-8c92c5b3353b/retry HTTP/1.1" 200 OK
INFO: 127.0.0.1:59330 - "GET /history?limit=20 HTTP/1.1" 200 OK
INFO: 127.0.0.1:63594 - "GET /generate/2af49cc4-bbd3-473c-952b-8c92c5b3353b/status HTTP/1.1" 200 OK
2026-04-17 10:30:18,244 - root - WARNING - Failed import k2 with error No module named 'k2'. Swoosh functions will fallback to PyTorch implementation, leading to slower speed and higher memory consumption.
pydub\utils.py:170: RuntimeWarning: Couldn't find ffmpeg or avconv - defaulting to ffmpeg, but may not work
2026-04-17 10:30:19,980 - backend.backends.luxtts_backend - INFO - Loading LuxTTS on cuda...
Fetching 11 files: 0%| | 0/11 [00:00<?, ?it/s]
Fetching 11 files: 9%|9 | 1/11 [00:00<00:05, 1.67it/s]
Fetching 11 files: 82%|########1 | 9/11 [00:00<00:00, 15.85it/s]
Fetching 11 files: 100%|##########| 11/11 [00:00<00:00, 13.68it/s]
Xet Storage is enabled for this repo, but the 'hf_xet' package is not installed. Falling back to regular HTTP download. For better performance, install the package with: pip install huggingface_hub[hf_xet] or pip install hf_xet
2026-04-17 10:30:22,684 - huggingface_hub.file_download - WARNING - Xet Storage is enabled for this repo, but the 'hf_xet' package is not installed. Falling back to regular HTTP download. For better performance, install the package with: pip install huggingface_hub[hf_xet] or pip install hf_xet
pytorch_model.bin: 0%| | 0.00/290M [00:00<?, ?B/s]
pytorch_model.bin: 4%|3 | 10.5M/290M [00:00<00:04, 66.4MB/s]
pytorch_model.bin: 7%|7 | 21.0M/290M [00:00<00:03, 74.8MB/s]
pytorch_model.bin: 11%|# | 31.5M/290M [00:00<00:03, 77.9MB/s]
pytorch_model.bin: 14%|#4 | 41.9M/290M [00:00<00:03, 79.8MB/s]
pytorch_model.bin: 18%|#8 | 52.4M/290M [00:00<00:02, 80.4MB/s]
pytorch_model.bin: 22%|##1 | 62.9M/290M [00:00<00:02, 81.0MB/s]
pytorch_model.bin: 25%|##5 | 73.4M/290M [00:00<00:02, 81.4MB/s]
pytorch_model.bin: 29%|##8 | 83.9M/290M [00:01<00:02, 81.7MB/s]
pytorch_model.bin: 32%|###2 | 94.4M/290M [00:01<00:02, 81.8MB/s]
pytorch_model.bin: 36%|###6 | 105M/290M [00:01<00:02, 82.0MB/s]
pytorch_model.bin: 40%|###9 | 115M/290M [00:01<00:02, 82.0MB/s]
pytorch_model.bin: 43%|####3 | 126M/290M [00:01<00:02, 82.0MB/s]
pytorch_model.bin: 47%|####6 | 136M/290M [00:01<00:01, 82.1MB/s]
pytorch_model.bin: 51%|##### | 147M/290M [00:01<00:01, 82.1MB/s]
pytorch_model.bin: 54%|#####4 | 157M/290M [00:01<00:01, 82.1MB/s]
pytorch_model.bin: 58%|#####7 | 168M/290M [00:02<00:01, 82.1MB/s]
pytorch_model.bin: 61%|######1 | 178M/290M [00:02<00:01, 82.2MB/s]
pytorch_model.bin: 65%|######4 | 189M/290M [00:02<00:01, 82.1MB/s]
pytorch_model.bin: 69%|######8 | 199M/290M [00:02<00:01, 82.2MB/s]
pytorch_model.bin: 72%|#######2 | 210M/290M [00:02<00:00, 82.1MB/s]
pytorch_model.bin: 76%|#######5 | 220M/290M [00:02<00:00, 82.1MB/s]
pytorch_model.bin: 79%|#######9 | 231M/290M [00:02<00:00, 82.2MB/s]
pytorch_model.bin: 83%|########3 | 241M/290M [00:02<00:00, 82.2MB/s]
pytorch_model.bin: 87%|########6 | 252M/290M [00:03<00:00, 82.2MB/s]
pytorch_model.bin: 90%|######### | 262M/290M [00:03<00:00, 82.2MB/s]
pytorch_model.bin: 94%|#########3| 273M/290M [00:03<00:00, 82.1MB/s]
pytorch_model.bin: 97%|#########7| 283M/290M [00:03<00:00, 82.1MB/s]
pytorch_model.bin: 100%|##########| 290M/290M [00:03<00:00, 81.4MB/s]
2026-04-17 10:30:27,957 - backend.utils.progress - ERROR - Marked luxtts as error: Could not load model openai/whisper-base with any of the following classes: (<class 'transformers.models.auto.modeling_auto.AutoModelForCTC'>, <class 'transformers.models.auto.modeling_auto.AutoModelForSpeechSeq2Seq'>). See the original errors:
while loading with AutoModelForCTC, an error is thrown:
Traceback (most recent call last):
File "transformers\pipelines\base.py", line 293, in infer_framework_load_model
model = model_class.from_pretrained(model, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\models\auto\auto_factory.py", line 607, in from_pretrained
raise ValueError(
ValueError: Unrecognized configuration class <class 'transformers.models.whisper.configuration_whisper.WhisperConfig'> for this kind of AutoModel: AutoModelForCTC.
Model type should be one of Data2VecAudioConfig, HubertConfig, MCTCTConfig, ParakeetCTCConfig, SEWConfig, SEWDConfig, UniSpeechConfig, UniSpeechSatConfig, Wav2Vec2Config, Wav2Vec2BertConfig, Wav2Vec2ConformerConfig, WavLMConfig.
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "transformers\pipelines\base.py", line 311, in infer_framework_load_model
model = model_class.from_pretrained(model, **fp32_kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\models\auto\auto_factory.py", line 607, in from_pretrained
raise ValueError(
ValueError: Unrecognized configuration class <class 'transformers.models.whisper.configuration_whisper.WhisperConfig'> for this kind of AutoModel: AutoModelForCTC.
Model type should be one of Data2VecAudioConfig, HubertConfig, MCTCTConfig, ParakeetCTCConfig, SEWConfig, SEWDConfig, UniSpeechConfig, UniSpeechSatConfig, Wav2Vec2Config, Wav2Vec2BertConfig, Wav2Vec2ConformerConfig, WavLMConfig.
while loading with AutoModelForSpeechSeq2Seq, an error is thrown:
Traceback (most recent call last):
File "transformers\pipelines\base.py", line 293, in infer_framework_load_model
model = model_class.from_pretrained(model, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\models\auto\auto_factory.py", line 604, in from_pretrained
return model_class.from_pretrained(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\modeling_utils.py", line 277, in _wrapper
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "transformers\modeling_utils.py", line 4900, in from_pretrained
checkpoint_files, sharded_metadata = _get_resolved_checkpoint_files(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\modeling_utils.py", line 1128, in _get_resolved_checkpoint_files
raise OSError(
OSError: openai/whisper-base does not appear to have a file named pytorch_model.bin but there is a file for TensorFlow weights. Use from_tf=True to load this model from those weights.
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "transformers\pipelines\base.py", line 311, in infer_framework_load_model
model = model_class.from_pretrained(model, **fp32_kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\models\auto\auto_factory.py", line 604, in from_pretrained
return model_class.from_pretrained(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\modeling_utils.py", line 277, in _wrapper
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "transformers\modeling_utils.py", line 4900, in from_pretrained
checkpoint_files, sharded_metadata = _get_resolved_checkpoint_files(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\modeling_utils.py", line 1128, in get_resolved_checkpoint_files
raise OSError(
OSError: openai/whisper-base does not appear to have a file named pytorch_model.bin but there is a file for TensorFlow weights. Use from_tf=True to load this model from those weights.
Traceback (most recent call last):
File "backend\services\generation.py", line 63, in run_generation
File "backend\backends_init.py", line 404, in load_engine_model
File "backend\backends\luxtts_backend.py", line 65, in load_model
File "asyncio\threads.py", line 25, in to_thread
File "concurrent\futures\thread.py", line 59, in run
File "backend\backends\luxtts_backend.py", line 87, in load_model_sync
File "zipvoice\luxvoice.py", line 27, in init
model, feature_extractor, vocos, tokenizer, transcriber = load_models_gpu(model_path, device=device)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "zipvoice\modeling_utils.py", line 102, in load_models_gpu
transcriber = pipeline("automatic-speech-recognition", model="openai/whisper-base", device=device)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\pipelines_init.py", line 1027, in pipeline
framework, model = infer_framework_load_model(
^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\pipelines\base.py", line 333, in infer_framework_load_model
raise ValueError(
ValueError: Could not load model openai/whisper-base with any of the following classes: (<class 'transformers.models.auto.modeling_auto.AutoModelForCTC'>, <class 'transformers.models.auto.modeling_auto.AutoModelForSpeechSeq2Seq'>). See the original errors:
while loading with AutoModelForCTC, an error is thrown:
Traceback (most recent call last):
File "transformers\pipelines\base.py", line 293, in infer_framework_load_model
model = model_class.from_pretrained(model, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\models\auto\auto_factory.py", line 607, in from_pretrained
raise ValueError(
ValueError: Unrecognized configuration class <class 'transformers.models.whisper.configuration_whisper.WhisperConfig'> for this kind of AutoModel: AutoModelForCTC.
Model type should be one of Data2VecAudioConfig, HubertConfig, MCTCTConfig, ParakeetCTCConfig, SEWConfig, SEWDConfig, UniSpeechConfig, UniSpeechSatConfig, Wav2Vec2Config, Wav2Vec2BertConfig, Wav2Vec2ConformerConfig, WavLMConfig.
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "transformers\pipelines\base.py", line 311, in infer_framework_load_model
model = model_class.from_pretrained(model, **fp32_kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\models\auto\auto_factory.py", line 607, in from_pretrained
raise ValueError(
ValueError: Unrecognized configuration class <class 'transformers.models.whisper.configuration_whisper.WhisperConfig'> for this kind of AutoModel: AutoModelForCTC.
Model type should be one of Data2VecAudioConfig, HubertConfig, MCTCTConfig, ParakeetCTCConfig, SEWConfig, SEWDConfig, UniSpeechConfig, UniSpeechSatConfig, Wav2Vec2Config, Wav2Vec2BertConfig, Wav2Vec2ConformerConfig, WavLMConfig.
while loading with AutoModelForSpeechSeq2Seq, an error is thrown:
Traceback (most recent call last):
File "transformers\pipelines\base.py", line 293, in infer_framework_load_model
model = model_class.from_pretrained(model, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\models\auto\auto_factory.py", line 604, in from_pretrained
return model_class.from_pretrained(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\modeling_utils.py", line 277, in _wrapper
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "transformers\modeling_utils.py", line 4900, in from_pretrained
checkpoint_files, sharded_metadata = _get_resolved_checkpoint_files(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\modeling_utils.py", line 1128, in _get_resolved_checkpoint_files
raise OSError(
OSError: openai/whisper-base does not appear to have a file named pytorch_model.bin but there is a file for TensorFlow weights. Use from_tf=True to load this model from those weights.
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "transformers\pipelines\base.py", line 311, in infer_framework_load_model
model = model_class.from_pretrained(model, **fp32_kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\models\auto\auto_factory.py", line 604, in from_pretrained
return model_class.from_pretrained(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\modeling_utils.py", line 277, in _wrapper
return func(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^
File "transformers\modeling_utils.py", line 4900, in from_pretrained
checkpoint_files, sharded_metadata = _get_resolved_checkpoint_files(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "transformers\modeling_utils.py", line 1128, in _get_resolved_checkpoint_files
raise OSError(
OSError: openai/whisper-base does not appear to have a file named pytorch_model.bin but there is a file for TensorFlow weights. Use from_tf=True to load this model from those weights.
INFO: 127.0.0.1:63594 - "GET /history?limit=20 HTTP/1.1" 200 OK
INFO: 127.0.0.1:63594 - "GET /tasks/active HTTP/1.1" 200 OK

[cosmind-rusu]

[cosmind-rusu]

[cosmind-rusu]

nvidia-smi
Fri Apr 17 10:31:53 2026
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 595.97 Driver Version: 595.97 CUDA Version: 13.2 |
+-----------------------------------------+------------------------+----------------------+
| GPU Name Driver-Model | Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|=========================================+========================+======================|
| 0 NVIDIA GeForce RTX 5080 WDDM | 00000000:06:00.0 On | N/A |
| 0% 45C P3 26W / 360W | 2362MiB / 16303MiB | 0% Default |
| | | N/A |
+-----------------------------------------+------------------------+----------------------+

[xxx-xxx-xxx-xxx-xxx-xxx]

Yeah, I can confirm the new update seems to be even worse than before.

[aidenszolosi]

I attest to experiencing this issue.

[jgharding]

I can neither download nor use downloaded models now, as i get the error

please unset the HF_HUB_OFFLINE environment variable

tried uninstalling everything and reinstalling, but it still happens.

[shaoyuyang-tal]

环境信息

• 系统版本: macOS tahoe 新品是m2
• Voicebox 版本: v0.4.5
• 网络环境: 中国大陆（无法直接访问 HuggingFace）

问题描述

软件始终卡在 Connecting to HuggingFace... 转圈，无法识别已手动放置的本地 Qwen3-TTS 模型文件，无法进行文本转语音。

复现步骤

1. 手动下载了 mlx-community/Qwen3-TTS-12Hz-1.7B-Base-bf16 模型文件。
2. 将模型文件夹重命名为 mlx-community--Qwen3-TTS-12Hz-1.7B-Base-bf16，并完整放入以下路径：
   /Users/syyang/.cache/huggingface/hub/
3. 文件夹内包含 model.safetensors, config.json 等所有核心文件，结构完整。
4. 重启 Voicebox，进入「模型」页面，点击 Qwen TTS 1.7B。
5. 软件始终显示 Connecting to HuggingFace...，无法识别本地文件，也无法离线加载模型。

预期行为

软件应优先读取本地 Hugging Face 缓存目录中已存在的模型文件，无需联网校验即可加载。

附加信息

• 已尝试创建 .no_connect 文件并清理缓存，问题依旧存在。
• 由于网络限制，无法通过软件内置的下载功能从 HuggingFace 拉取模型。
• 软件的强制联网校验逻辑，导致国内用户即使拥有本地模型也无法使用。