diffusers

diffusers copied to clipboard

Bugfix for dreambooth flux2 img2img2

2

# What does this PR do? I'm training the flux2 img2img with 8 GPUs, the running scripts is using the script from examples/dreambooth/README_flux2.md: ``` bash accelerate launch train_dreambooth_lora_flux2_img2img.py \ --pretrained_model_name_or_path=black-forest-labs/FLUX.2-dev...

leisuzz

from_pipe: Keep existing dtype by default

1

# What does this PR do? Fixes a regression introduced in #10816, causing `from_pipe` to convert pipelines to float32 by default. Fixes #12754 ## Before submitting - [ ] This...

missionfloyd

# What does this PR do? This PR adds a training script for **Latent Consistency Model (LCM) distillation** applied to **InstructPix2Pix with Stable Diffusion XL**. This enables fast, few-step image...

mzeynali

# What does this PR do? Fixes #12760 @toilaluan @DN6 I am doubtful of these remaining cases - (can be found using `TODO-context` comment) src - ```bash (calls: 2, wrapped:...

omkar-334

Can diffusers support loading and running FLUX with fp8 ?

5

This is how I use diffusers to load flux model: ``` import torch from diffusers import FluxPipeline pipe = FluxPipeline.from_pretrained( "/ckptstorage/repo/pretrained_weights/black-forest-labs/FLUX.1-dev", torch_dtype=torch.float16, ) device = torch.device(f"cuda:{device_number}" if torch.cuda.is_available() else "cpu")...

EmmaThompson123

Fix QwenImage txt_seq_lens handling

8

# What does this PR do? - Removes the redundant `txt_seq_lens` plumbing from all QwenImage pipelines and modular steps; the transformer now infers text length from encoder inputs/masks and validates...

kashif

[core] gracefully error out when attn-backend x cp combo isn't supported.

1

# What does this PR do? We should be able to error out when an attention backend isn't supported with CP. Refer to https://github.com/huggingface/diffusers/pull/12829#issuecomment-3645237672 and https://github.com/huggingface/diffusers/pull/12829#issuecomment-3645582823. Additionally, we specify `parallel_config`...

sayakpaul

# What does this PR do? Even though we will have separate unit-level testing for attention backends in https://github.com/huggingface/diffusers/pull/12822/, I think it's still nice to have integration tests. This PR:...

sayakpaul

Fix qwen encoder hidden states mask

15

# What does this PR do? Fixes the QwenImage encoder to properly apply `encoder_hidden_states_mask` when passed to the model. Previously, the mask parameter was accepted but ignored, causing padding tokens...

cdutr

[Wan] Optimize time & memory

2

# What does this PR do? This PR reduces the time and space used when running Wan. I have successfully tested the performance improvement and I have done a crash...

Fabrice-TIERCELIN