diffusers

diffusers copied to clipboard

Add `disable_mmap` to option to Pipeline

1

# What does this PR do? Addressed some of the issues brought up in #12599. Using mmaped safetensors on networked volumes can lead to very slow load/device transfer times. More...

DN6

Flux2: Tensor tuples can cause issues for checkpointing

9

### Describe the bug The modulations calculated here... https://github.com/huggingface/diffusers/blob/edf36f5128abf3e6ecf92b5145115514363c58e6/src/diffusers/models/transformers/transformer_flux2.py#L716 ...return tuples of Tensors: https://github.com/huggingface/diffusers/blob/edf36f5128abf3e6ecf92b5145115514363c58e6/src/diffusers/models/transformers/transformer_flux2.py#L628 These tuples are passed from outside the transformer blocks into the checkpointed transformer blocks. If the...

dxqb

## What does this PR do? Fixes #12783 This PR temporarily disables the Sage Attention sm90 backend which is causing confetti/noisy output on SM 9.0+ (Hopper) GPUs. ## The Problem...

arrdel

## What does this PR do? Fixes #12779 This PR fixes an error in the Flux2 dreambooth training script. ## The Problem The argument parser defined the argument as (line...

arrdel

Chroma inpainting pipeline

15

I'd like someone to add an inpaiting pipeline for Chroma I created a working one here: [see the gist file](https://gist.github.com/rodjjo/1195ea079f395f8e6a81abcc2f13a141) Working example: ```python """ ChromaInpaintPipeline implements a text-guided image inpainting...

rodjjo

Flux1-Dev inference with single file ComfyUI/SD-Forge Safetensors

12

Is it possible to run inference with diffusers using a single-file safetensors created for ComfyUI/SD-Forge? It looks like FluxPipeline.from_single_file() might be intended for this purpose, but I'm getting the following...

ddpasa

### Describe the bug When using _sage_qk_int8_pv_fp8_cuda_sm90 as the attention backend on WAN2.2 I2V I notice that the output is broken: It works fine with _flash_3_hub and _sage_qk_int8_pv_fp16_cuda Does it...

dylanprins

AttributeError: type object 'Flux2ImageProcessor' has no attribute 'image_processor'

1

### Describe the bug I'm trying to run train_dreambooth_flux2_img2img.py from the DreamBooth examples, but I encountered this error. I inspected the Flux2ImageProcessor code in diffusers.pipelines.flux2.image_processor, and it turns out that...

hannalaguilar

# What does this PR do? The incorrect training logic of Dreambooth Flux2 img2img has been modified, and the issue of abnormal loss during training has been fixed. The Dreambooth...

tcaimm

### Describe the bug In train_dreambooth_lora_flux2_img2img.py on line 1422, there is a variable named validation_image_path, but shouldn’t it actually be validation_image? ` validation_image = load_image(args.validation_image_path).convert("RGB") AttributeError: 'Namespace' object has no...

hannalaguilar