Slices model input with output channels so the caching tracks only the noise channels, resolves channel mismatch with models like WanVideo I2V
Also fix for slicing deprecation in pytorch 2.9
Clean up a bunch of stacked and no-longer-needed tensors on the QWEN
VRAM peak (currently FFN).
With this I go from OOMing at B=37x1328x1328 to being able to
succesfully run B=47 (RTX5090).
The partial unloader path in model re-use flow skips straight to the
actual unload without any check of the patching UUID. This means that
if you do an upscale flow with a model patch on an existing model, it
will not apply your patchings.
Fix by delaying the partial_unload until after the uuid checks. This
is done by making partial_unload a model of partial_load where extra_mem
is -ve.
- Remove non-upstream packages from requirements.txt
- Make dependency check advisory only (no auto-install)
- Remove CUDA PyTorch auto-install/update flows
- Trim banner and keep minimal preflight checks
- Drop non-portable create_shortcut.ps1
- Updated PR_SUBMISSION_CHECKLIST.md with Feature Request issue creation step
- Added FINAL_SUMMARY.md with complete status and next steps
- All documentation and preparation complete
- Created FEATURE_REQUEST_ISSUE.md with complete issue content
- Created CREATE_ISSUE_INSTRUCTIONS.md with step-by-step instructions
- Created PR_COMPLIANCE_ANALYSIS.md analyzing compliance with wiki requirements
- Created SEARCH_RESULTS_SUMMARY.md documenting search for existing issues/PRs
- Updated PR_DESCRIPTION.md with:
- Issue Addressed section (explicitly states problems solved)
- Potential Concerns and Side Effects section (required by wiki)
- PR Size Note (acknowledges large PR size)
- Placeholder for Related Issue number
- Added UTF-8 encoding (chcp 65001) to fix Unicode character display in ASCII art header
- Enabled progress bars for all pip installations (--progress-bar on)
- Fixed CUDA PyTorch auto-installation logic to properly continue to ComfyUI launch
- Updated CUDA availability variables after successful installation
- Fixed misleading Restart message to accurately reflect Continue to launch
- Improved error handling and user feedback throughout the installation process
Enhanced run_comfyui.bat with:
- Automatic detection of missing critical dependencies
- Virtual environment detection and warnings
- Optional user-prompted installation with clear warnings
- Comprehensive dependency checking for all essential packages
NOTE: Author is not a professional coder and relied heavily on Cursor AI for implementation. Please review code thoroughly before merging.
