Overview

2 Releases published by 1 person

• v0.35.0 Diffusers 0.35.0: Qwen Image pipelines, Flux Kontext, Wan 2.2, and more

  published Aug 19, 2025

• v0.35.1 v0.35.1 for improvements in Qwen-Image Edit

  published Aug 20, 2025

107 Pull requests merged by 40 people

• [docs] Models

  #12248 merged Sep 5, 2025

• [docs] Inference section cleanup

  #12281 merged Sep 5, 2025

• [docs] components_to_quantize

  #12287 merged Sep 5, 2025

• [docs] Sharing pipelines/models

  #12280 merged Sep 4, 2025

• fix some typos

  #12265 merged Sep 3, 2025

• Add AttentionMixin to WanVACETransformer3DModel

  #12268 merged Sep 3, 2025

• [tests] feat: add AoT compilation tests

  #12203 merged Sep 3, 2025

• [Quantization] Add TRT-ModelOpt as a Backend

  #11173 merged Sep 3, 2025

• [docs] AutoPipeline

  #12160 merged Sep 3, 2025

• [core] use kernels to support _flash_3_hub attention backend

  #12236 merged Sep 3, 2025

• [CI] Remove big accelerator requirements from Quanto Tests

  #12266 merged Sep 3, 2025

• [Flux LoRA] fix for prior preservation and mixed precision sampling, follow up on #11873

  #12264 merged Sep 2, 2025

• [docs] Fix VAE scale factor calculation in distributed inference docs

  #12259 merged Sep 2, 2025

• [fix] fix for prior preservation and mixed precision sampling

  #11873 merged Sep 1, 2025

• Fix lora conversion function for ai-toolkit Qwen Image LoRAs

  #12261 merged Sep 1, 2025

• Add Qwen-Image-Edit Inpainting pipeline

  #12225 merged Aug 31, 2025

• Improve performance of NPU FA

  #12260 merged Aug 30, 2025

• [QwenImageEditPipeline] Add image entry in __call__ function

  #12254 merged Aug 30, 2025

• [Modular] Consolidate load_default_components into load_components

  #12217 merged Aug 28, 2025

• [Modular] Provide option to disable custom code loading globally via env variable

  #12177 merged Aug 28, 2025

• [Refactor] Move testing utils out of src

  #12238 merged Aug 28, 2025

• [modular diffusers] Fix AutoGuidance validation

  #12247 merged Aug 28, 2025

• [Modular] support standard repo

  #11944 merged Aug 28, 2025

• [docs] Reproducibility

  #12237 merged Aug 27, 2025

• [docs] Remove Flax

  #12244 merged Aug 27, 2025

• Fix Table Rendering in ReadME

  #12245 merged Aug 27, 2025

• [Qwen-Image] adding validation for guidance_scale, true_cfg_scale and negative_prompt

  #12223 merged Aug 27, 2025

• docs: correct typos in using-diffusers/other-formats

  #12243 merged Aug 26, 2025

• Fix typos and inconsistencies

  #12204 merged Aug 26, 2025

• Deprecate Flax support

  #12151 merged Aug 26, 2025

• Propose to update & upgrade SkyReels-V2

  #12167 merged Aug 26, 2025

• NPU attention refactor for FLUX

  #12209 merged Aug 26, 2025

• Fix typo: 'will ge generated' -> 'will be generated'

  #12231 merged Aug 25, 2025

• Update README.md

  #12193 merged Aug 25, 2025

• [docs] typo : corrected 'compile regions' to 'compile_regions'

  #12199 merged Aug 25, 2025

• Fix typo in LoRA

  #12228 merged Aug 25, 2025

• [docs] Diffusion pipeline

  #12148 merged Aug 25, 2025

• fix title for compile + offload quantized models

  #12233 merged Aug 25, 2025

• [docs] change wan2.1 -> wan

  #12230 merged Aug 25, 2025

• added a fast test for Qwen-Image Controlnet Pipeline

  #12226 merged Aug 25, 2025

• [refactor] Make guiders return their inputs

  #12213 merged Aug 23, 2025

• Fix PyTorch 2.3.1 compatibility: add version guard for torch.library....

  #12206 merged Aug 23, 2025

• add attentionmixin to qwen image

  #12219 merged Aug 22, 2025

• Support ControlNet for Qwen-Image

  #12215 merged Aug 22, 2025

• [docs] Pipeline callbacks

  #12212 merged Aug 22, 2025

• [docs] Community pipelines

  #12201 merged Aug 22, 2025

• [Docs] Add documentation for KontextInpaintingPipeline

  #12197 merged Aug 22, 2025

• enable bria integration test on xpu, passed

  #12214 merged Aug 22, 2025

• [docs] Optimized code snippets

  #12200 merged Aug 21, 2025

• continue translate document to zh

  #12194 merged Aug 20, 2025

• [chore] remove extra validation check in determine_device_map

  #12176 merged Aug 20, 2025

• Bria 3 2 pipeline

  #12010 merged Aug 20, 2025

• post release v0.35.0

  #12184 merged Aug 19, 2025

• translate document to zh

  #12179 merged Aug 19, 2025

• Performance Improve for Qwen Image Edit

  #12190 merged Aug 19, 2025

• Emergency fix for Qwen-Image-Edit

  #12188 merged Aug 19, 2025

• [Wan 2.2 LoRA] add support for 2nd transformer lora loading + wan 2.2 lightx2v lora

  #12074 merged Aug 19, 2025

• [chore] add lora button to qwenimage docs

  #12183 merged Aug 19, 2025

• Update README.md

  #12182 merged Aug 18, 2025

• [LoRA] feat: support more Qwen LoRAs from the community.

  #12170 merged Aug 18, 2025

• [docs] Clarify guidance scale in Qwen pipelines

  #12181 merged Aug 18, 2025

• fix(training_utils): wrap device in list for DiffusionPipeline

  #12178 merged Aug 18, 2025

• fix: caching allocator behaviour for quantization.

  #12172 merged Aug 18, 2025

• [Docs] typo error in qwen image

  #12144 merged Aug 18, 2025

• Minor modification to support DC-AE-turbo

  #12169 merged Aug 18, 2025

• [qwen] Qwen image edit followups

  #12166 merged Aug 18, 2025

• remove silu for CogView4

  #12150 merged Aug 18, 2025

• Qwen Image Edit Support

  #12164 merged Aug 17, 2025

• [docs] Quickstart

  #12128 merged Aug 15, 2025

• make test_gguf all pass on xpu

  #12158 merged Aug 15, 2025

• support hf_quantizer in cache warmup.

  #12043 merged Aug 14, 2025

• [core] respect local_files_only=True when using sharded checkpoints

  #12005 merged Aug 14, 2025

• feat: cuda device_map for pipelines.

  #12122 merged Aug 14, 2025

• [docs] Parallel loading of shards

  #12135 merged Aug 14, 2025

• make parallel loading flag a part of constants.

  #12137 merged Aug 14, 2025

• Fix bf15/fp16 for pipeline_wan_vace.py

  #12143 merged Aug 13, 2025

• [docs] Refresh effective and efficient doc

  #12134 merged Aug 13, 2025

• try to use deepseek with an agent to auto i18n to zh

  #12032 merged Aug 13, 2025

• [core] parallel loading of shards

  #12028 merged Aug 13, 2025

• Add QwenImage Inpainting and Img2Img pipeline

  #12117 merged Aug 13, 2025

• [Bugfix] typo fix in NPU FA

  #12129 merged Aug 12, 2025

• [docs] Modular diffusers

  #11931 merged Aug 12, 2025

• Add low_cpu_mem_usage option to from_single_file to align with from_pretrained

  #12114 merged Aug 12, 2025

• [Modular] Add experimental feature warning for Modular Diffusers

  #12127 merged Aug 12, 2025

• [docs] Cache link

  #12105 merged Aug 11, 2025

• [chore] complete the licensing statement.

  #12001 merged Aug 11, 2025

• [tests] Add inference test slices for SD3 and remove unnecessary tests

  #12106 merged Aug 11, 2025

• enable compilation in qwen image.

  #12061 merged Aug 11, 2025

• [Modular] More Updates for Custom Code Loading

  #11969 merged Aug 11, 2025

• [lora] support loading loras from lightx2v/Qwen-Image-Lightning

  #12119 merged Aug 11, 2025

• [core] add modular support for Flux I2I

  #12086 merged Aug 11, 2025

• Fix MultiControlNet import

  #12118 merged Aug 10, 2025

• [docs] diffusers gguf checkpoints

  #12092 merged Aug 9, 2025

• [GGUF] feat: support loading diffusers format gguf checkpoints.

  #11684 merged Aug 8, 2025

• [Modular] Fast Tests

  #11937 merged Aug 8, 2025

• [tests] device placement for non-denoiser components in group offloading LoRA tests

  #12103 merged Aug 8, 2025

• lora_conversion_utils: replace lora up/down with a/b even if transformer. in key

  #12101 merged Aug 8, 2025

• [lora] adapt new LoRA config injection method

  #11999 merged Aug 8, 2025

• [qwen] device typo

  #12099 merged Aug 7, 2025

• fix flux type hint

  #12089 merged Aug 7, 2025

• Implement Frequency-Decoupled Guidance (FDG) as a Guider

  #11976 merged Aug 7, 2025

• [tests] tighten compilation tests for quantization

  #12002 merged Aug 7, 2025

• Make prompt_2 optional in Flux Pipelines

  #12073 merged Aug 7, 2025

• Helper functions to return skip-layer compatible layers

  #12048 merged Aug 6, 2025

• Fix group offloading synchronization bug for parameter-only GroupModule's

  #12077 merged Aug 6, 2025

• [refactor] condense group offloading

  #11990 merged Aug 6, 2025

• fix input shape for WanGGUFTexttoVideoSingleFileTests

  #12081 merged Aug 6, 2025

45 Pull requests opened by 29 people

• Fix Qwen-Image long prompt dimension mismatch error (issue #12083)

  #12087 opened Aug 6, 2025

• fix(fluX image validation): convert RGB to BGR for OpenPose images

  #12090 opened Aug 7, 2025

• Add Dream 7B Diffusion Large Language Model Pipeline

  #12091 opened Aug 7, 2025

• [model] LTX Video 0.9.8

  #12095 opened Aug 7, 2025

• Add support for Multiple ControlNetXSAdapters in SDXL pipeline

  #12100 opened Aug 8, 2025

• [scheduler] Support custom sigmas/timesteps in UniPCMultistepScheduler

  #12109 opened Aug 9, 2025

• fixed bug in defining embed dim for UNet1D

  #12111 opened Aug 10, 2025

• Verified issue with test

  #12115 opened Aug 10, 2025

• Fix AttributeError of `VisualClozeProcessor`

  #12121 opened Aug 11, 2025

• [chore] remove controlnet implementations outside controlnet module.

  #12152 opened Aug 15, 2025

• `no_init_weights` in `from_pretrained` should be optional

  #12153 opened Aug 15, 2025

• make models and pipeline docs device agnostic, to cover accelerators other than CUDA

  #12161 opened Aug 15, 2025

• update document preview parts

  #12171 opened Aug 18, 2025

• feat(autoencoders): add patchify support for AutoencoderKLWan

  #12180 opened Aug 18, 2025

• Add device_map="auto" support for the diffusers pipelines with higher VRAM occupancy

  #12187 opened Aug 19, 2025

• [Wan 2.2 VAE] fix VAE tiling encode/decode

  #12191 opened Aug 19, 2025

• [scheduler] Add SDE Samplers: Curved Euler and Overshoot

  #12192 opened Aug 19, 2025

• [Modular] implement requirements validation for custom blocks

  #12196 opened Aug 20, 2025

• [wip][quantization] incorporate `nunchaku`

  #12207 opened Aug 21, 2025

• [feature] Support Flux TensorRT Pipeline

  #12218 opened Aug 22, 2025

• [Modular] Qwen

  #12220 opened Aug 22, 2025

• [Refactor] Move transformer blocks out of attention module

  #12229 opened Aug 25, 2025

• [quant] allow `components_to_quantize` to be a non-list for single components

  #12234 opened Aug 25, 2025

• DeepSpeed adaption for flux-kontext

  #12240 opened Aug 26, 2025

• [packaging] make `torch` a required dependency

  #12242 opened Aug 26, 2025

• [docs] Schedulers

  #12246 opened Aug 27, 2025

• Add Step1X-Edit Pipeline

  #12249 opened Aug 28, 2025

• Fix WanVACEPipeline to allow prompt to be None and skip encoding step

  #12251 opened Aug 28, 2025

• [docs] Model formats

  #12256 opened Aug 28, 2025

• Add Wan2.2-S2V

  #12258 opened Aug 29, 2025

• fix incorrect sigmas of DPMSolverMultistepInverseScheduler for flow_prediction

  #12262 opened Aug 30, 2025

• Fix #12116: preserve boolean dtype for attention masks in ChromaPipeline

  #12263 opened Aug 30, 2025

• Support Flux Kontext in modular (T2I and I2I)

  #12269 opened Sep 2, 2025

• Deprecate slicing and tiling methods from `DiffusionPipeline`

  #12271 opened Sep 3, 2025

• [modular] fix flux modular pipelines for t2i and i2i

  #12272 opened Sep 3, 2025

• [quantization] feat: support aobaseconfig classes in `TorchAOConfig`

  #12275 opened Sep 3, 2025

• [tests] disable xformer tests for pipelines it isn't popular.

  #12277 opened Sep 3, 2025

• Bump torch from 2.2.0 to 2.8.0 in /examples/research_projects/realfill

  #12279 opened Sep 3, 2025

• [core] feat: support group offloading at the pipeline level

  #12283 opened Sep 4, 2025

• [docs] Distributed inference

  #12285 opened Sep 4, 2025

• [docs] Pipeline group offloading

  #12286 opened Sep 4, 2025

• [WIP] Add InfiniteTalk Audio-Driven Video Generation Pipeline

  #12288 opened Sep 5, 2025

• Fix many type hint errors

  #12289 opened Sep 5, 2025

• adjust criteria for marigold-intrinsics example on XPU

  #12290 opened Sep 5, 2025

• refactor: improve handling when only masked_image_latents are provided in Flux.1 Fill dev

  #12293 opened Sep 5, 2025

40 Issues closed by 24 people

• Failed to finetune the pre-trained model of 'stable-diffusion-v1-4' on image inpainting task

  #12165 closed Sep 4, 2025

• Questions about train_text_to_image_lora_sdxl.py

  #12282 closed Sep 3, 2025

• [BUG] Misleading ValueError when subclassing StableDiffusionImg2ImgPipeline with a mismatched __init__ signature

  #12255 closed Aug 30, 2025

• Create modular pipeline from existing pipeline

  #11915 closed Aug 28, 2025

• Add Bria 3.2 pipeline - next-generation commercial-ready text-to-image model

  #11978 closed Aug 26, 2025

• The diffusers0.34.0 installed by pip is not aligned with the source code

  #12149 closed Aug 26, 2025

• Error creating pre-trained FluxPipeline with diffusers "Cannot instantiate this tokenizer from a slow version"

  #9063 closed Aug 26, 2025

• wan2.2 inference error Bug Report: `NoneType` object has no attribute `cache_context` when switching from `transformer` to `transformer2`

  #12126 closed Aug 25, 2025

• [Contribution welcome] adding a fast test for Qwen-Image Controlnet Pipeline

  #12222 closed Aug 25, 2025

• Version 0.35.0 AutoencoderKL failed with PyTorch 2.3.1: AutoAttributeError: module 'torch.library' has no attribute 'custom_op'

  #12195 closed Aug 23, 2025

• Bad Image Quality generated by Qwen Image

  #12175 closed Aug 21, 2025

• [Feature Request] Settability of Multiple Attention Types in Attention Dispatcher

  #12210 closed Aug 21, 2025

• AttributeError: 'bool' object has no attribute '__module__'. Did you mean: '__mod__'?

  #12044 closed Aug 21, 2025

• Add support for guidance-interval during sampling

  #12198 closed Aug 20, 2025

• Wan2.2 Lightning LoRA Loading Fails

  #12147 closed Aug 19, 2025

• LoRA loading compatibility with some Qwen Image LoRAs

  #12159 closed Aug 18, 2025

• When training the Qwen-Image model using --bnb_quantization_config_path, the following error occurred: 'torch.device' object is not iterable.

  #12157 closed Aug 18, 2025

• "Requested to load AutoencoderKL"

  #12123 closed Aug 17, 2025

• IndexError: index 0 is out of bounds for dimension 0 with size 0

  #12104 closed Aug 17, 2025

• dje

  #12154 closed Aug 17, 2025

• cool man

  #12156 closed Aug 15, 2025

• monsieur

  #12155 closed Aug 15, 2025

• WanVACEPipeline doesn't work with bfloat16/foat16 vae

  #12141 closed Aug 13, 2025

• Add pruna integration for loading model through diffusers.from_pretrained / pipeline

  #11700 closed Aug 12, 2025

• [Wan2.2] pipeline_wan miss the 'shift' parameter which used by Wan2.2-A14B-diffusers.

  #12094 closed Aug 10, 2025

• Flux1.Dev Kohya Loras text encoder layers no more supported

  #12053 closed Aug 9, 2025

• [Performance] Issue on *SanaLinearAttnProcessor2_0 family. 1.06X speedup can be reached with a simple change.

  #11499 closed Aug 8, 2025

• Qwen image transformers doesn't currently support from_single_file (i.e. GGUFs)

  #12098 closed Aug 8, 2025

• Wan 2.2 with LightX2V offloading tries to multiply tensors from different devices and fails

  #12052 closed Aug 8, 2025

• Qwen-Image long prompt will cause error

  #12083 closed Aug 8, 2025

• [lora] `exclude_modules` won't consider the modules names in `target_modules`

  #11874 closed Aug 8, 2025

• Qwen Image incorrect device assignment during prompt encode

  #12066 closed Aug 7, 2025

• WanTransformer3DModel.from_single_file wont load Wan2.2 GGUF (NotImplementedError: Cannot copy out of meta tensor; no data)

  #12009 closed Aug 7, 2025

• CPU Offloading Inefficiency: UNet/Transformer Repeatedly Moved Between CPU/GPU During Loops

  #11872 closed Aug 7, 2025

• Frequency-Decoupled Guidance (FDG) for diffusion models

  #11956 closed Aug 7, 2025

• Qwen Image : Image Editing Inference

  #12080 closed Aug 6, 2025

• Groupoffloading introduce bad results

  #11981 closed Aug 6, 2025

• Fusing Lightx2v lora on Wan2.2 GGUF fails

  #12047 closed Aug 6, 2025

• Wan 2.2 VAE forward fails

  #12039 closed Aug 6, 2025

• WanVACETransformer3DModel load gguf error

  #12085 closed Aug 6, 2025

43 Issues opened by 40 people

• [Qwen-image] encoder_hidden_states_mask is not used

  #12294 opened Sep 6, 2025

• Why use torch.repeat instead of torch.repeat_interleave in train_dreambooth_lora_sdxl

  #12292 opened Sep 5, 2025

• Attention test seems uncompatible with latest xformers

  #12284 opened Sep 4, 2025

• The image is blurry.

  #12276 opened Sep 3, 2025

• New HunyuanVideoI2V PR changes to transformer_hunyuan_video.py

  #12273 opened Sep 3, 2025

• From 0.35.0: DLL load failed while importing libtriton: A dynamic link library (DLL) initialization routine failed.

  #12270 opened Sep 2, 2025

• [Looking for community contribution] support Wan 2.2 S2V: an audio-driven cinematic video generation model

  #12257 opened Aug 29, 2025

• stabilityai/stable-diffusion-2-base is missing fp16 files' so that the correct variant file can be added

  #12253 opened Aug 28, 2025

• WAN2.1 FLF2V: Incorrect MASK Creation????

  #12241 opened Aug 26, 2025

• Support for InfiniteTalk

  #12239 opened Aug 26, 2025

• Support fun control variants for Wan

  #12235 opened Aug 25, 2025

• [Feature] Integrate 'CHORDS: Diffusion Sampling Accelerator with Multi-core Hierarchical ODE Solvers'

  #12227 opened Aug 24, 2025

• Option to cache text embedding vectors

  #12224 opened Aug 23, 2025

• [Looking for community contribution] support DiffSynth Controlnet in diffusers

  #12221 opened Aug 22, 2025

• Qwen-Image-Edit Inferior Results Compared to ComfyUI

  #12216 opened Aug 22, 2025

• Transformer2DModel onnx convertation problems

  #12208 opened Aug 21, 2025

• [Features] Planning to support TensorRT for Flux Pipelines

  #12202 opened Aug 21, 2025

• When using the qwen-image model to generate images, diffusers takes twice as long as diffsynth-studio

  #12189 opened Aug 19, 2025

• `WanVACETransformer3DModel` does not inherit `AttentionMixin` while `WanTransformer3DModel` does

  #12186 opened Aug 19, 2025

• What's the difference between DreamBooth LoRa and traditional LoRa?

  #12185 opened Aug 19, 2025

• 'super' object has no attribute '__getattr__'

  #12174 opened Aug 18, 2025

• Wan i2i with latent upscale

  #12173 opened Aug 18, 2025

• Not compatible with Flash ATTN 2.8.2

  #12162 opened Aug 16, 2025

• Wan 2.1 / Wan2.2 Lightning & Sparse Distillation & Tiny VAE Acceleration Methods

  #12146 opened Aug 14, 2025

• resize_token_embeddings (?) (different result from new and old version)

  #12145 opened Aug 13, 2025

• The IP Adapter provides different conditional guidance for each frame generated by AnimateDiff.

  #12139 opened Aug 13, 2025

• requirest for the model use

  #12138 opened Aug 13, 2025

• How to use Diffusers to Convert Safetensors SDXL 1.0 to Onnx?

  #12136 opened Aug 13, 2025

• inference with lora suddenly significantly slower

  #12131 opened Aug 12, 2025

• CompVis/stable-diffusion-v1-4 is missing fp16 files

  #12125 opened Aug 12, 2025

• For qwen-image training file, Maybe "shuffle" of dataloader should be "False" when custom_instance_prompts is not None and cache_latents is False?

  #12124 opened Aug 11, 2025

• How to train a lora with distilled flux model, such as flux-schnell???

  #12120 opened Aug 11, 2025

• Attention masking in Chroma pipeline

  #12116 opened Aug 10, 2025

• WanImageToVideoPipeline: Given groups=1, weight of size [160, 12, 3, 3, 3], expected input[1, 3, 3, 258, 258] to have 12 channels, but got 3 channels instead

  #12113 opened Aug 10, 2025

• Various Chroma pipeline issues

  #12112 opened Aug 10, 2025

• Bug in initialization of UNet1DModel GaussianFourier time projection

  #12110 opened Aug 10, 2025

• Qwen Image and Chroma pipeline breaks using schedulers that enable flow matching by parameter.

  #12108 opened Aug 9, 2025

• accelerator.init_trackers error when try with a custom object such as list

  #12107 opened Aug 9, 2025

• Support training where width does not equal height for Qwen-Image

  #12102 opened Aug 8, 2025

• Wan2.2 TI2V-5B VRAM OOM at the end

  #12097 opened Aug 7, 2025

• WanVACEPipeline - doesn't work with apply_group_offloading

  #12096 opened Aug 7, 2025

• Will `cosmos-transfer1` be supported in diffusers in the future?

  #12084 opened Aug 6, 2025

• wan2.1 vae take more gpu memory after compile

  #12082 opened Aug 6, 2025

35 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• Context Parallel w/ Ring & Ulysses & Unified Attention

  #11941 commented on Sep 4, 2025 • 22 new comments

• [modular diffusers] Wan I2V/FLF2V

  #11997 commented on Aug 6, 2025 • 1 new comment

• Qwen Image prompt encoding is not padding to max seq len

  #12075 commented on Aug 6, 2025 • 0 new comments

• Support for control-lora

  #10686 commented on Aug 21, 2025 • 0 new comments

• Add FluxPAGPipeline with support for PAG

  #11510 commented on Aug 10, 2025 • 0 new comments

• [bugfix] Fix the issue that will save the model card if push to hub is not applied in HiDream

  #11629 commented on Aug 14, 2025 • 0 new comments

• Add MAGI-1: Autoregressive Video Generation at Scale

  #11713 commented on Aug 29, 2025 • 0 new comments

• [wip] Higgs quantization

  #11797 commented on Aug 7, 2025 • 0 new comments

• Propose to refactor `norm_out` in `SkyReelsV2Transformer3DModel` for device handling in multi-GPU case

  #11850 commented on Sep 3, 2025 • 0 new comments

• add kontext into auto-pipeline

  #11875 commented on Aug 28, 2025 • 0 new comments

• Add F5 TTS pipeline

  #11958 commented on Sep 6, 2025 • 0 new comments

• torch.compile compatibility with varlen APIs

  #11970 commented on Aug 25, 2025 • 0 new comments

• Fixed "NameError: name 'ftfy' is not defined" in Wan pipeline by directly importing ftfy.

  #12023 commented on Aug 16, 2025 • 0 new comments

• [lora] factor out the overlaps in `save_lora_weights()`.

  #12027 commented on Sep 2, 2025 • 0 new comments

• Add Flix Fill ControlNet

  #12035 commented on Aug 13, 2025 • 0 new comments

• `UNet2DConditionModel`: add support for QK Normalization by propagating `qk_norm` value from config through to child attention modules

  #12051 commented on Aug 11, 2025 • 0 new comments

• [WIP][modular] sdxl methods refactor

  #12067 commented on Aug 10, 2025 • 0 new comments

• gguf new quant type support (with demo)

  #12076 commented on Aug 18, 2025 • 0 new comments

• Problem with provided example validation input in the Flux Control finetuning example

  #12078 commented on Aug 7, 2025 • 0 new comments

• Dream 7B

  #12017 commented on Aug 7, 2025 • 0 new comments

• Support multiple control nets in the `StableDiffusionControlNetXSPipeline`/`StableDiffusionXLControlNetXSPipeline`

  #8434 commented on Aug 8, 2025 • 0 new comments

• Try load gguf for FluxKontextPipeline but invalid input size

  #11839 commented on Aug 9, 2025 • 0 new comments

• AnimateDiff SparseCtrl RGB does not work as expected

  #9508 commented on Aug 10, 2025 • 0 new comments

• CogvideoX VAE decoder consumes significantly more memory in the latest version

  #10091 commented on Aug 10, 2025 • 0 new comments

• deepspeed train flux1 dreambooth lora can not save model

  #9393 commented on Aug 13, 2025 • 0 new comments

• train_dreambooth_lora_flux_kontext.py: error: the following arguments are required: --instance_prompt

  #11973 commented on Aug 20, 2025 • 0 new comments

• How to finetune Flux-dev full params, 80G OOM ...

  #10196 commented on Aug 20, 2025 • 0 new comments

• [New Feature] Flux PuLID

  #10141 commented on Aug 26, 2025 • 0 new comments

• [Flux ControlNet] Add support for de-distilled models with CFG

  #9635 commented on Aug 28, 2025 • 0 new comments

• Error RuntimeError: Invalid buffer size

  #5894 commented on Aug 30, 2025 • 0 new comments

• Run FLUX-controlnet zero3 training failed: 'weight' must be 2-D

  #10969 commented on Sep 1, 2025 • 0 new comments

• Dreambooth finetune FLUX dev CLIPTextModel

  #10925 commented on Sep 3, 2025 • 0 new comments

• FluxFillPipeline quality deteriorates over multiple iterations of inpainting

  #10204 commented on Sep 3, 2025 • 0 new comments

• AttributeError: _hf_hook caused by delattr in hooks.remove_hook_from_module()

  #10729 commented on Sep 5, 2025 • 0 new comments

• Adds Img2Img2 Variant for HunyuanDiT.

  #9041 commented on Aug 28, 2025 • 0 new comments