patrickvonplaten
Follow
Highlights
- Pro
Pinned
3,009 contributions in the last year
Contribution activity
June 2023
Created 35 commits in 7 repositories
Created 1 repository
- patrickvonplaten/audiocraft Python
Created a pull request in huggingface/diffusers that received 13 comments
[Kandinsky] Improve kandinsky API a bit
Currently the way negative prompts are used can be pretty confusing (e.g. having to run the prior twice). This PR changes the behavior so that the …
+182
−119
•
13
comments
Opened 17 other pull requests in 8 repositories
huggingface/diffusers
5
merged
1
closed
vladmandic/automatic
3
merged
YuxinWenRick/tree-ring-watermark
2
merged
huggingface/transformers
2
merged
facebookresearch/audiocraft
1
merged
AUTOMATIC1111/stable-diffusion-webui
1
closed
invoke-ai/InvokeAI
1
open
facebookresearch/fairseq
1
merged
Reviewed 37 pull requests in 7 repositories
huggingface/diffusers
25 pull requests
- manual check for checkpoints_total_limit instead of using accelerate
- remove seed
- Ldm3d first commit
-
Fixes eval generator init in
train_text_to_image_lora.py - Google drive backup support for checkpoints
- Fix schedulers zero SNR and rescale classifier free guidance
- [Community] Support StableDiffusionCanvasPipeline
- Add draft for lora text encoder scale
- training scripts accelerate checkpoint naming refactor
- Update Compel documentation for textual inversions
- [WIP]Vae preprocessor refactor (PR1)
- [docs] Model API
- move activation dispatches into helper function
- dreambooth upscaling fix added latents
- set config from original module but set compiled module on class
- Add function to remove monkey-patch for text encoder LoRA
- [docs] Load A1111 LoRA
- [docs] More API fixes
- fix inpainting pipeline when providing initial latents
- [Pipeline] Port Tune-A-Video pipeline to diffusers
- [WIP] Allow DDIMInverseScheduler to use same number of noising and denoising steps
- fixed typo in example train_text_to_image.py
- Allow setting num_cycles for cosine_with_restarts lr scheduler
- Set step_rules correctly for piecewise_constant scheduler
- Support Kohya-ss style LoRA file format (in a limited capacity)
- Some pull request reviews not shown.
patrickvonplaten/controlnet_aux
3 pull requests
huggingface/transformers
2 pull requests
facebookresearch/audiocraft
1 pull request
vladmandic/automatic
1 pull request
huggingface/api-inference-community
1 pull request
facebookresearch/fairseq
1 pull request
Created an issue in huggingface/diffusers that received 7 comments
New Discrete Diffusion Model: Wuerstchen
Model/Pipeline/Scheduler description The amazing @dome272 published a new diffusion model, called Wuerstchen: https://github.com/dome272/Wuerstchen I…
•
7
comments