8c4eb8c3f1
fix: pin transformers to 4.49.0 for DiffRhythm compatibility
...
Build and Push RunPod Docker Image / build-and-push (push) Successful in 13s
Resolves tensor dimension mismatch error in rotary position embeddings.
DiffRhythm requires transformers 4.49.0 - newer versions (4.50+) cause
"The size of tensor a (32) must match the size of tensor b (64)" error
due to transformer block initialization changes.
Updated pivoine_diffrhythm.py documentation to reflect actual root cause
and link to upstream GitHub issues #44 and #48 .
References:
- https://github.com/billwuhao/ComfyUI_DiffRhythm/issues/44
- https://github.com/billwuhao/ComfyUI_DiffRhythm/issues/48
🤖 Generated with [Claude Code](https://claude.com/claude-code )
Co-Authored-By: Claude <noreply@anthropic.com >
2025-11-24 18:14:40 +01:00
67d41c3923
fix: patch infer_utils.decode_audio instead of DiffRhythmNode.infer
...
Build and Push RunPod Docker Image / build-and-push (push) Successful in 13s
The correct function to patch is decode_audio from infer_utils module,
which is where chunked VAE decoding actually happens. This intercepts
the call at the right level to force chunked=False.
🤖 Generated with [Claude Code](https://claude.com/claude-code )
Co-Authored-By: Claude <noreply@anthropic.com >
2025-11-24 17:28:30 +01:00
1981b7b256
fix: monkey-patch DiffRhythm infer function to force chunked=False
...
Build and Push RunPod Docker Image / build-and-push (push) Successful in 14s
The previous approach of overriding diffrhythmgen wasn't working because
ComfyUI doesn't pass the chunked parameter when it's not in INPUT_TYPES.
This fix monkey-patches the infer() function at module level to always
force chunked=False, preventing the tensor dimension mismatch error.
🤖 Generated with [Claude Code](https://claude.com/claude-code )
Co-Authored-By: Claude <noreply@anthropic.com >
2025-11-24 17:24:22 +01:00
5096e3ffb5
feat: add Pivoine custom ComfyUI nodes for DiffRhythm
...
Build and Push RunPod Docker Image / build-and-push (push) Successful in 14s
Add custom node wrapper PivoineDiffRhythmRun that fixes tensor dimension
mismatch error by disabling chunked VAE decoding. The original DiffRhythm
node's overlap=32 parameter conflicts with the VAE's 64-channel architecture.
Changes:
- Add comfyui/nodes/pivoine_diffrhythm.py: Custom node wrapper
- Add comfyui/nodes/__init__.py: Package initialization
- Add arty.yml setup/pivoine-nodes: Deployment script for symlink
- Update all 4 DiffRhythm workflows to use PivoineDiffRhythmRun
Technical details:
- Inherits from DiffRhythmRun to avoid upstream patching
- Forces chunked=False in diffrhythmgen() override
- Requires more VRAM (~12-16GB) but RTX 4090 has 24GB
- Category: 🌸 Pivoine/Audio for easy identification
🤖 Generated with [Claude Code](https://claude.com/claude-code )
Co-Authored-By: Claude <noreply@anthropic.com >
2025-11-24 16:28:54 +01:00
