Spaces:
Running
on
Zero
Windows Support ?
Trying to run this locally but its requirement is Flash-Attention 3 which is developed for (Requirements: H100 / H800 GPU, CUDA >= 12.3.)
Would it be possible to rollback to Flash-Attention2 for compatibilty ?
(env) PS C:\Users\thesi\OneDrive\Desktop\Qwen-Image-Edit-Angles> python app.py
WARNING:torchao.kernel.intmm:Warning: Detected no triton, on systems without Triton certain kernels will not work
W1107 11:26:15.400000 22128 env\Lib\site-packages\torch\distributed\elastic\multiprocessing\redirects.py:29] NOTE: Redirects are currently not supported in Windows or MacOs.
Fetching 0 files: 0it [00:00, ?it/s]
Fetching 4 files: 100%|██████████████████████████████████████████████████████████████████████████| 4/4 [00:00<?, ?it/s]
Loading checkpoint shards: 100%|█████████████████████████████████████████████████████████| 4/4 [00:13<00:00, 3.48s/it]
Loading pipeline components...: 0%| | 0/6 [00:00<?, ?it/s]torch_dtype is deprecated! Use dtype instead!
Loading checkpoint shards: 100%|█████████████████████████████████████████████████████████| 4/4 [00:00<00:00, 71.42it/s]
Loading pipeline components...: 100%|████████████████████████████████████████████████████| 6/6 [00:00<00:00, 6.71it/s]
Expected types for transformer: (<class 'diffusers.models.transformers.transformer_qwenimage.QwenImageTransformer2DModel'>,), got <class 'qwenimage.transformer_qwenimage.QwenImageTransformer2DModel'>.
Traceback (most recent call last):
File "C:\Users\thesi\OneDrive\Desktop\Qwen-Image-Edit-Angles\app.py", line 52, in
pipe.transformer.set_attn_processor(QwenDoubleStreamAttnProcessorFA3())
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\thesi\OneDrive\Desktop\Qwen-Image-Edit-Angles\qwenimage\qwen_fa3_processor.py", line 57, in init
_ensure_fa3_available()
File "C:\Users\thesi\OneDrive\Desktop\Qwen-Image-Edit-Angles\qwenimage\qwen_fa3_processor.py", line 20, in _ensure_fa3_available
raise ImportError(
ImportError: FlashAttention-3 via Hugging Face kernels is required. Tried get_kernel('kernels-community/vllm-flash-attn3') and failed with:
Kernel kernels-community/vllm-flash-attn3 at revision main does not have build: torch29-cu130-x86_64-windows
(env) PS C:\Users\thesi\OneDrive\Desktop\Qwen-Image-Edit-Angles>