[Docs] add: 8bit inference with pixart alpha #5814
Conversation
|
|
||
| </Tip> | ||
|
|
||
| ## Running the `PixArtAlphaPipeline` in under 8GB GPU VRAM |
There was a problem hiding this comment.
Let's put this content into a Tips section (like how it is here) so it is more aligned with the other model API docs. I think this'll improve the docs experience since all the pages will be consistent and users know what to expect if they're going from one page to the next
There was a problem hiding this comment.
I don't think this section is a good candidate to be put as a tip, though. Taking this as an example, I see all the points as advice (or tips). They very well align with what's supposed to be a tip. But I don't think execution under constrained environments can be a tip.
There was a problem hiding this comment.
Good point, maybe we can put it in a Usage section then? I think it's better to have something more generalizable to keep the pages consistent (like DeepFloyd IF or Pix2Pix).
Co-authored-by: Steven Liu <[email protected]>
|
The documentation is not available anymore as the PR was closed or merged. |
* add: 8bit inference with pixart alpha * Apply suggestions from code review Co-authored-by: Steven Liu <[email protected]> * add: note on 4bit. * Apply suggestions from code review Co-authored-by: Steven Liu <[email protected]> * address comment --------- Co-authored-by: Steven Liu <[email protected]> Co-authored-by: Patrick von Platen <[email protected]>
What does this PR do?
This PR adds a section to the PixArt Alpha docs about how to perform inference in 8bit so that the GPU memory requirements are kept under 8GB.
Cc: @lawrence-cj