[LCM] add: locm docs. #5723
[LCM] add: locm docs. #5723
Conversation
|
The documentation is not available anymore as the PR was closed or merged. |
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>
Co-authored-by: Pedro Cuenca <pedro@huggingface.co>
|
|
||
|  | ||
|
|
||
| ## ControlNet |
There was a problem hiding this comment.
Actually can we move this section to LoRA? ControlNet only works with the lora distilled yet (maybe show an example of v15 here): https://huggingface.co/latent-consistency/lcm-lora-sdv1-5
There was a problem hiding this comment.
But the result shows it works:
| image = pipe( | ||
| prompt=prompt, | ||
| num_inference_steps=4, | ||
| guidance_scale=1, |
There was a problem hiding this comment.
| guidance_scale=1, | |
| guidance_scale=0, |
Actually guidance_scale=0 also disables it and is cleaner.
There was a problem hiding this comment.
@pcuenca tested it with 1.0.
There was a problem hiding this comment.
0 and 1 is the same
|
|
||
| * By default, _no guidance is performed_ which enables faster inference. It works for most of the prompts. This means `guidance_scale` is set to 1 (as in the example above). | ||
| * When `guidance_scale` > 1, regular classifier-free guidance is performed leading to slightly slower inference as this doubles the batch size. | ||
| * The recommended range for `guidance_scale` is [1., 2.]. |
There was a problem hiding this comment.
Can we also show img2img and Controlnet - see: https://github.com/huggingface/diffusers/pull/5723/files#r1388130749
There was a problem hiding this comment.
Very nice work! Added some updates regarding the naming :-)
It would be great if we could add some more examples to the LoRA section, e.g.:
- Img2Img
- ControlNet (note that ControlNet doesn't work fully-fine-tuned lcm, but only for LoRA)
- Adding multiple LoRAs
- Img2Img
- Inpainting
Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
* add: locm docs. * correct path * Apply suggestions from code review Co-authored-by: Pedro Cuenca <pedro@huggingface.co> * Apply suggestions from code review Co-authored-by: Pedro Cuenca <pedro@huggingface.co> * Apply suggestions from code review Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * up * add --------- Co-authored-by: Pedro Cuenca <pedro@huggingface.co> Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
* add: locm docs. * correct path * Apply suggestions from code review Co-authored-by: Pedro Cuenca <pedro@huggingface.co> * Apply suggestions from code review Co-authored-by: Pedro Cuenca <pedro@huggingface.co> * Apply suggestions from code review Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com> * up * add --------- Co-authored-by: Pedro Cuenca <pedro@huggingface.co> Co-authored-by: Patrick von Platen <patrick.v.platen@gmail.com>
No description provided.