lllyasviel / lllyasviel/stable-diffusion-webui-forge
[Feature Request]: use potential of medvram, or do not force lowvram for 4GB or less GPUs
- Dominant language
- Python
- Stars
- 13k
- Forks
- 1.7k
- PR merge metrics
- No merged PRs in 30d
Description
### Is there an existing issue for this?
- [X] I have searched the existing issues and checked the recent builds/commits
### What would your feature do ?
I have 2GB gpu and found sd-webui-forge forces lowvram mode. But medvram has significantly performance boost, and 2 gb is enought for medvram today, thanks to fp8
```
stock --medvram:
[01:22<00:00, 4.13s/it]
forge:
[01:36<00:00, 4.86s/it]
forge --always-normal-vram:
[01:34<00:00, 4.83s/it]
```
`--always-normal-vram` from the tip is not about this
```
Total VRAM 2001 MB, total RAM 19904 MB
Trying to enable lowvram mode because your GPU seems to have 4GB or less. If you don't want this use: --always-normal-vram
xformers version: 0.0.23.post1
Set vram state to: LOW_VRAM
Device: cuda:0 NVIDIA GeForce MX150 : native
```
It would be nice to enable lowvram only for OOM in different case
### Proposed workflow
1. Go to ....
2. Press ....
3. ...
### Additional information
_No response_
Contributor guide
No contributing guide indexed for this repository
Research direction
Start by locating the VRAM-mode selection path that emits the “Trying to enable lowvram mode” message and handles --always-normal-vram. Reproduce the reported 2GB GPU comparison, then verify that lowvram is not forced solely by the GPU size and that the intended medvram behavior remains available with OOM fallback.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- machine-learning, performance
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100