OPENGIT/stable-diffusion-webui
1
0
Fork 0
mirror of https://github.com/AUTOMATIC1111/stable-diffusion-webui.git synced 2025-04-25 14:28:59 +08:00
Code Issues Packages Projects Releases Wiki Activity

make PYTORCH_CUDA_ALLOC_CONF enviornment variable placement easier to understand

ClashSAN 2023-03-01 00:55:26 +00:00
parent 328bdab7a9
commit f21ef0d5de
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
Optimizations.md

@ -43,7 +43,7 @@ Notes:
- Performance with `lowvram` preset is very low below batch size 8 and by then memory savings are not that big - Performance with `lowvram` preset is very low below batch size 8 and by then memory savings are not that big
Other possible optimizations: Other possible optimizations:
- `PYTORCH_CUDA_ALLOC_CONF=garbage_collection_threshold:0.9,max_split_size_mb:512` - adding `set PYTORCH_CUDA_ALLOC_CONF=garbage_collection_threshold:0.9,max_split_size_mb:512` in `webui-user.bat`
No performance impact and increases initial memory footprint a bit but reduces memory fragmentation in long runs No performance impact and increases initial memory footprint a bit but reduces memory fragmentation in long runs
- `opt-channelslast` - `opt-channelslast`
Hit-and-miss: seems like additional slight performance increase with higher batch sizes and slower with small sizes, but differences are within margin-of-error Hit-and-miss: seems like additional slight performance increase with higher batch sizes and slower with small sizes, but differences are within margin-of-error