Spaces:
Paused
Paused
Commit History
-
Update requirements.txt
Verified signature538ff7c -
Update README.md
Verified signature3880696
-
Update webui.py
Verified signaturea4b297a -
Update webui.py
Verified signature195665b -
Update webui.py
Verified signaturec63e2ac -
Update webui.py
Verified signature66acc7e -
Upload 3 files
Verified signatureb1ae84d -
Update webui.py
Verified signatureb72b23c
-
Update webui.py
Verified signature1c6422c -
Update webui.py
Verified signatureca8a00e -
Upload 2 files
Verified signaturecf32427 -
Update webui.py
Verified signaturedcde3bd -
Update webui.py
Verified signature6e05a61 -
Update webui.py
Verified signature1e61d28 -
Fix LlamaAttention compatibility with transformers 4.53.0
The transformers 4.53 API requires `position_embeddings` (cos/sin rotary tuple) as a positional arg to LlamaAttention.forward() and returns 2 values instead of 3. Update DiffLlama to compute rotary embeddings and pass them through the decoder layers. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>4a3004b -
Fix NeMo CUDA error 35 on ZeroGPU by disabling CUDA graphs in decoding config
The previous fix disabled CUDA graphs on the decoding_computer object, but transcribe(timestamps=True) calls change_decoding_strategy() which rebuilds the decoder from cfg, re-enabling CUDA graphs. Setting cfg.decoding.greedy.use_cuda_graph_decoder=False persists across rebuilds. Also loads the English ASR model eagerly at startup so ZeroGPU can properly hijack CUDA calls at global scope, instead of lazy-loading inside a @spaces.GPU-decorated function. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>e36805d -
Make _get_device() ZeroGPU-aware so models target CUDA
On ZeroGPU Spaces torch.cuda.is_available() returns False at module-load time because the GPU is only attached inside @spaces.GPU. Check spaces.config.Config.zero_gpu as a fallback so all models are placed on CUDA and work correctly once the GPU is attached. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>d3bb466 -
Update requirements.txt
Verified signature5907c41 -
Disable NeMo CUDA Graphs to fix CUDA failure 35 on RNNT decoding
NeMo 2.6 RNNT decoding uses CUDA Graphs which fail with error 35 on CUDA 12.8 + PyTorch 2.9 (NVIDIA-NeMo/NeMo#15145). Disable them on the parakeet-tdt model after loading. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>a263769 -
Fix ZeroGPU nested @spaces.GPU causing CUDA failure 35
Remove @spaces.GPU from internal AppState methods (run_preprocess, run_svs, run_svs_from_paths) since they are already called from Gradio-facing functions that hold the GPU context. Also enable share=True in app.py launch. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>930ae2b -
Update requirements.txt
Verified signature477ec17 -
Update requirements.txt
Verified signature5e851cc -
Update preprocess/tools/lyric_transcription.py
Verified signatureb773a39 -
Update preprocess/tools/g2p.py
Verified signature0196c78 -
Update requirements.txt
Verified signaturec25f3e5 -
Update requirements.txt
Verified signaturec6e10b4 -
Update requirements.txt
Verified signature1058f25 -
Update app.py
Verified signature9ece03f -
Update requirements.txt
Verified signature7eb880c -
Update requirements.txt
Verified signaturee8743a1 -
Update webui.py
Verified signaturebde23dc -
Fix model download reliability and silent import failures
Extract ensure_pretrained_models() into ensure_models.py and call it from webui.py before AppState init, so models are downloaded regardless of entry point. Narrow except clauses in preprocess/tools/__init__.py from Exception to ImportError so real errors propagate instead of silently setting classes to None. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>19b877f -
Update requirements.txt
Verified signatured775c8a -
Upload webui.py with huggingface_hub
Verified signature02e5782 -
Upload app.py with huggingface_hub
Verified signature3ea55ea -
Upload requirements.txt with huggingface_hub
Verified signaturef9f3297 -
Upload README.md with huggingface_hub
Verified signature9447616 -
Upload requirements.txt with huggingface_hub
Verified signaturecbd5e26 -
Upload requirements.txt with huggingface_hub
Verified signaturef132868 -
Upload folder using huggingface_hub
Verified signaturec7f3ffb -
initial commit
Verified signaturea81bc3b
