Open sayakpaul opened 5 months ago
Hey @sayakpaul for text-to-image, SD
log-validation
is present : https://github.com/huggingface/diffusers/blob/33d2b5b08764d2579dd4ea555821b8986cd2dcf9/examples/text_to_image/train_text_to_image.py#L141
haven't gone through the code, will check more if we need to modify it.
Ideally, this should be delegated to log_validation()
too:
https://github.com/huggingface/diffusers/blob/33d2b5b08764d2579dd4ea555821b8986cd2dcf9/examples/text_to_image/train_text_to_image.py#L1047
As done in train_lcm_distill_lora_sdxl.py
.
Got it. @sayakpaul I missed that part. Thanks!
Hey @sayakpaul, @charchit7 I'm interested in incorporating the change to InstructPix2Pix SD
Hey @sayakpaul, @charchit7 I'm interested in incorporating the change to InstructPix2Pix SD
Hey @sang-k, I have been looking into InstructPix2Pix SD too. Would be happy to collaborate if necessary.
@coolyashas Oh, I see. I will look into InstructPix2Pix SDXL then!
Working on InstructPix2Pix SD
We provide support for running validation inference during and after training in our officially maintained training examples. This is very helpful to keep track of the training progress.
We could modularize some bits in the example to reduce the LoC.
The
train_lcm_distill_lora_sdxl.py
script already does this:https://github.com/huggingface/diffusers/blob/33d2b5b08764d2579dd4ea555821b8986cd2dcf9/examples/consistency_distillation/train_lcm_distill_lora_sdxl.py#L100
It would be nice to follow something similar for the rest of the scripts too. Here's a handy list of the scripts where we'd like to incorporate this change:
Feel free to comment here if you're interested.
Also, when opening PRs, please target one example at a time and please tag me in the PRs.