Dhruv Nair
Dhruv Nair
# What does this PR do? Set max parallel T4 jobs used in slow pipeline tests so that we don't exhaust the pool. Fixes # (issue) ## Before submitting -...
# What does this PR do? SD Cascade UNet is now tested in the the single file test module. We don't need to run these tests anymore. Fixes # (issue)...
# What does this PR do? SVD input images are not being properly resized in the video processor, and is leading to a test failure since the image latents shape...
# What does this PR do? Clean up some miscellaneous stuff in the docs that was missed in the original PR Fixes # (issue) ## Before submitting - [ ]...
# What does this PR do? When loading the T5 encoder in SD3 with `from_pretrained` certain layers are are kept in fp32. Currently this is not the case with single...
# What does this PR do? There seems to be an issue with 1. Saving Kohya style weights in the dreambooth advanced lora script. The save path can output a...
### Model/Pipeline/Scheduler description [SUPIR](https://supir.xpixel.group/) is a super-resolution model that looks like it produces excellent results Github Repo: https://github.com/Fanghua-Yu/SUPIR The model is quite memory intensive, so the optimisation features available in...
# What does this PR do? Big GPU test marker isn't registered in conftest, so tests aren't running. https://github.com/huggingface/diffusers/actions/runs/15838149705/job/44645818322 Fold it into `require_big_accelerator` so that we need to use fewer...
# What does this PR do? Since we now have multiple offload options available in group offload (standard, low_cpu_mem_usage, offload_to_disk), it would make sense to create separate utility functions for...
# What does this PR do? Draft PR to address some issues with the attention modules. 1. Moves a number of transformer related blocks/objects in the attention module that would...