Dev #7
Merged
Dev #7
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
- Added CUDA development tools (nvidia-cuda-dev, nvidia-cuda-toolkit) - Added SageAttention compilation from source with optimized parallel flags - Includes cleanup of source directory after installation 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
- Removed SageAttention compilation from Dockerfile to avoid build failures - Added SageAttention installation to start.sh after PyTorch is available - Installation happens at runtime with proper dependency resolution - Uses optimized parallel compilation flags for faster build 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
- Removed nvidia-cuda-dev and nvidia-cuda-toolkit packages - These are not needed since SageAttention is now installed at runtime - Reduces Docker image size and build complexity 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
- Changed from ubuntu22.04 to ubuntu24.04 for latest packages - Maintains CUDA 12.8.0 compatibility 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
- Remove wan2.2_s2v_14B_fp8_scaled.safetensors from model_config_wan.json - Add new Pusa LoRA models (HIGH and LOW variants) to model config - Enable ComfyUI_essentials custom node installation - Add ComfyUI-WanVideoWrapper and ComfyUI_LayerStyle custom nodes - Enable SageAttention flag for ComfyUI startup 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
- Replace complex source compilation with simple pip install - Use uv pip install sageattention instead of building from source - This should resolve the SageAttention package not found error 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
- Revert to source compilation for SageAttention 2.2.0 with SageAttention2++ - Use pip install -e . instead of python setup.py install for better compatibility - Keep parallel compilation flags for faster build 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
- Use uv pip consistently for SageAttention installation - Add runtime check before ComfyUI starts to ensure SageAttention is available - Install SageAttention if missing during startup to prevent --use-sage-attention errors - This ensures SageAttention is always available regardless of installation path 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
- Add --no-build-isolation flag to SageAttention installation - This allows the build process to access already installed PyTorch - Fixes ModuleNotFoundError: No module named 'torch' during SageAttention build 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
- Check for nvcc command before attempting SageAttention installation - Skip SageAttention installation gracefully if CUDA development tools unavailable - Conditionally start ComfyUI with/without --use-sage-attention flag - Prevents build errors when NVCC is missing 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
- Remove all source compilation attempts for SageAttention - Add installation of prebuilt SageAttention wheel from HuggingFace - Use sageattention-2.1.1-cp312-cp312-linux_x86_64.whl to avoid build issues - Re-enable --use-sage-attention flag for ComfyUI startup 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
- Replace wget download with direct URL installation - Upgrade from sageattention-2.1.1 to sageattention-2.2.0 - Use Kijai's PrecompiledWheels repository for better reliability - Simplify installation process by removing separate download step 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
- Add ComfyUI-MelBandRoFormer to the list of custom nodes - This node provides audio separation and processing capabilities 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
…ention support - Added USE_SAGE_ATTENTION environment variable with default value "false" - Modified ComfyUI startup to conditionally use --use-sage-attention flag based on env var - Allows users to enable/disable SageAttention without modifying the script 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
- Remove model_config_wan.json - Add model_config_wan22.json - Simplified models_config.json by removing unused models 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
- Remove model_config_wan22.json - Restore original model_config_wan.json file 🤖 Generated with [Claude Code](https://claude.ai/code) Co-Authored-By: Claude <noreply@anthropic.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.