RAM not being fully released after WAN 2.2 generation
- Dominant language
- Python
- Stars
- 133k
- Forks
- 15.7k
- Avg merge
- 1d 6h
- Merged PRs (30d)
- 155
Description
### Custom Node Testing
- [ ] I have tried disabling custom nodes and the issue persists (see [how to disable custom nodes](https://docs.comfy.org/troubleshooting/custom-node-issues#step-1%3A-test-with-all-custom-nodes-disabled) if you need help)
### Expected Behavior
RAM should be fully empty besides the small amount comfyui keeps in reserve.
### Actual Behavior
I cannot disable custom nodes to see if this persists. I am using torchcompile and sagattention, however the issue persisted before that. Running pretty much any generation results (even after clearing node cache and releasing models) in circa 30gb RAM being reserved for no purpose. With each generation this memory leaked increases until ram is saturated and spills into swap.
### Steps to Reproduce
Run a wan 2.2 video generation.
### Debug Logs
```powershell
there are none
```
### Other
RTX 5090
64gb ram
100gb swap
swappiness 80
zram after swap
Contributor guide
Assessment
This issue has not been assessed yet.