Commit History

Fix GenerationResult attribute: use audios dict, not audio_paths
003cf5e
Running
unverified

pedroapfilho commited on

Add custom tag input and fix model download (zip directory)
6ca081e
unverified

pedroapfilho commited on

Add lora_training/ and checkpoints/ to gitignore
ed3930a
unverified

pedroapfilho commited on

Use HF dataset repo as source of truth for dataset.json
6c32e21
unverified

pedroapfilho commited on

Add torchcodec dependency for torchaudio audio loading
5ae1ea5
unverified

pedroapfilho commited on

Fix meta tensor device mismatch by also patching torch.arange
f266123
unverified

pedroapfilho commited on

Patch both residual_fsq and finite_scalar_quantization tensor refs
011d30d
unverified

pedroapfilho commited on

Fix meta tensor: patch residual_fsq module's tensor reference, not torch.tensor
0c47532
unverified

pedroapfilho commited on

Remove .python-version: let HF Space use container default Python
773a0e0
unverified

pedroapfilho commited on

Fix meta tensor: monkey-patch ResidualFSQ to force CPU tensor creation
bf4c82d
unverified

pedroapfilho commited on

Fix meta tensor crash: force CPU device context during model init
b0a0560
unverified

pedroapfilho commited on

Add boot diagnostics to trace silent startup failure
0db63ea
unverified

pedroapfilho commited on

Fix: replace gr.Number with gr.Slider for max files input
0199d58
unverified

pedroapfilho commited on

Add max files limit to HF dataset download
eb72117
unverified

pedroapfilho commited on

Use HF cache for dataset downloads instead of custom local_dir
ad1969b
unverified

pedroapfilho commited on

Fix src/__init__.py: export download_hf_dataset instead of removed LoRATrainer
f7d4206
unverified

pedroapfilho commited on

Rewrite Tab 3 to wire real training pipeline with HF dataset support
3bc6e37
unverified

pedroapfilho commited on

Try _fast_init=False to prevent meta device initialization
5b76ce1

ACE-Step Custom commited on

Fix dtype parameter: use dtype instead of torch_dtype, remove device_map
140e1e2

ACE-Step Custom commited on

Fix meta tensor: add device_map=None and use torch_dtype parameter
a11c404

ACE-Step Custom commited on

Fix meta tensor error: disable low_cpu_mem_usage for model loading
23dc982

ACE-Step Custom commited on

Add vector-quantize-pytorch dependency for ACE-Step model
5e842f6

ACE-Step Custom commited on

Fix DiT initialization: pass model name only, not full path
6a590ee

ACE-Step Custom commited on

Fix TypeError: pass Path object to ensure_main_model, not string
fa7f63d

ACE-Step Custom commited on

Fix syntax error: close unterminated docstring in _load_models
78910e3

ACE-Step Custom commited on

Use official ACE-Step model downloader for automatic model downloads
ee19acb

ACE-Step Custom commited on

Monkey patch get_api_info to prevent JSON schema errors
55a1124

ACE-Step Custom commited on

Move endpoint override before launch
ccb3542

ACE-Step Custom commited on

Fix fastapi_app attribute name and move endpoint override after launch
2217362

ACE-Step Custom commited on

Add /info endpoint override and pin Gradio version
d6ad90b

ACE-Step Custom commited on

Add loguru dependency and fix Gradio 4.0.0 compatibility
9259263

ACE-Step Custom commited on

Fix engine initialization and audio I/O components
4709141

ACE-Step Custom commited on

Deploy ACE-Step Custom Edition with bug fixes
a602628

ACE-Step Custom commited on