# šŸ§Ø Diffusers **URL:** https://discuss.huggingface.co/c/discussion-related-to-httpsgithubcomhuggingfacediffusers/63.md [Latest](https://discuss.huggingface.co/latest.md) Ā· [Categories](https://discuss.huggingface.co/categories.md) --- ## [About the šŸ§Ø Diffusers category](https://discuss.huggingface.co/t/about-the-diffusers-category/20582)
**Author:** [@patrickvonplaten](https://discuss.huggingface.co/u/patrickvonplaten)\ **Replies:** 0\ **Last updated:** [July 19, 2022, 4:17pm UTC](https://discuss.huggingface.co/t/about-the-diffusers-category/20582 "2022-07-19T16:17:17Z")
This category should be used for all comments / questions related to the Diffusers library: GitHub - huggingface/diffusers --- ## [Qwen Image Edit inpainting works on smooth surfaces but fails on textured floors, light and shadows. How to get seamless results?](https://discuss.huggingface.co/t/qwen-image-edit-inpainting-works-on-smooth-surfaces-but-fails-on-textured-floors-light-and-shadows-how-to-get-seamless-results/183247)
**Author:** [@s-shivansh](https://discuss.huggingface.co/u/s-shivansh)\ **Replies:** 0\ **Last updated:** [October 6, 2026, 5:34am UTC](https://discuss.huggingface.co/t/qwen-image-edit-inpainting-works-on-smooth-surfaces-but-fails-on-textured-floors-light-and-shadows-how-to-get-seamless-results/183247 "2026-10-06T05:34:34Z")
Hi everyone, Iā€™m using Qwen Image Edit with mask-based inpainting through diffusers (Python script on RunPod, not ComfyUI). My goal is object removal, for example removing a rug from a room photo and filling the area wiā€¦ --- ## [Genuine question regarding learning Deep Learning for Video](https://discuss.huggingface.co/t/genuine-question-regarding-learning-deep-learning-for-video/182751)
**Author:** [@pop123ux](https://discuss.huggingface.co/u/pop123ux)\ **Replies:** 0\ **Last updated:** [October 1, 2026, 12:47pm UTC](https://discuss.huggingface.co/t/genuine-question-regarding-learning-deep-learning-for-video/182751 "2026-10-01T12:47:33Z")
Is the diffusers course still relevant in 2026, and worth going through? --- ## [MiniMax-H3 Quantisation RAM/VRAM](https://discuss.huggingface.co/t/minimax-h3-quantisation-ram-vram/178461)
**Author:** [@artificial-illusion](https://discuss.huggingface.co/u/artificial-illusion)\ **Replies:** 1\ **Last updated:** [August 4, 2026, 10:47pm UTC](https://discuss.huggingface.co/t/minimax-h3-quantisation-ram-vram/178461 "2026-08-04T22:47:53Z")
Following the information in the pull request docs MiniMax-H3 Ā· Hugging Face based on the 24-32GB consumer card section, Iā€™m trying to load the t2va model on a machine with 96GB RAM and an R9700 32GB. Even with the 8biā€¦ --- ## [How many images to parameters?](https://discuss.huggingface.co/t/how-many-images-to-parameters/177672)
**Author:** [@8BitStudio](https://discuss.huggingface.co/u/8BitStudio)\ **Replies:** 3\ **Last updated:** [July 12, 2026, 12:45am UTC](https://discuss.huggingface.co/t/how-many-images-to-parameters/177672 "2026-07-12T00:45:54Z")
Iā€™m training a latent diffusion model and wondering what the scaling laws were, I couldnt find much info online. My current is around 430M parameters training on around 2 Million images. Iā€™m thinking of possibly scaling ā€¦ --- ## [How to convert a single safetensors file to PEFT format](https://discuss.huggingface.co/t/how-to-convert-a-single-safetensors-file-to-peft-format/173103)
**Author:** [@MMing](https://discuss.huggingface.co/u/MMing)\ **Replies:** 12\ **Last updated:** [June 8, 2026, 12:34pm UTC](https://discuss.huggingface.co/t/how-to-convert-a-single-safetensors-file-to-peft-format/173103 "2026-06-08T12:34:26Z")
Iā€™m struggle in getting adapter\_model.safetensors and adapter\_config.json from this weighted LoRA file --- ## [UNet1DModel not converging on single batch](https://discuss.huggingface.co/t/unet1dmodel-not-converging-on-single-batch/37656)
**Author:** [@StableDiffuser317](https://discuss.huggingface.co/u/StableDiffuser317)\ **Replies:** 4\ **Last updated:** [May 10, 2026, 11:13am UTC](https://discuss.huggingface.co/t/unet1dmodel-not-converging-on-single-batch/37656 "2026-05-10T11:13:46Z")
Hello, Iā€™m currently trying to implement a latent generative diffuser using a custom-trained VAE which maps 3x256x256 to 1x256 and a UNet1DModel with the following parameters : UNet1DModel(in\_channels = 1, out\_channelsā€¦ --- ## [Cosmos 2.5 Diffusers Issue](https://discuss.huggingface.co/t/cosmos-2-5-diffusers-issue/172998)
**Author:** [@advpatel](https://discuss.huggingface.co/u/advpatel)\ **Replies:** 1\ **Last updated:** [February 2, 2026, 10:28pm UTC](https://discuss.huggingface.co/t/cosmos-2-5-diffusers-issue/172998 "2026-02-02T22:28:40Z")
Error when running: \`\`\`from diffusers import DiffusionPipeline pipe = DiffusionPipeline.from\_pretrained(ā€œnvidia/Cosmos-Predict2.5-2Bā€, dtype=torch.bfloat16, device\_map=ā€œcudaā€)\`\`\` \[rank0\]: Traceback (most recent call lā€¦ --- ## [\[Need Advice\] Maintaining Product Fidelity & Texture in Generative AI Mockup Automation (Stable Diffusion/Gemini)](https://discuss.huggingface.co/t/need-advice-maintaining-product-fidelity-texture-in-generative-ai-mockup-automation-stable-diffusion-gemini/172102)
**Author:** [@sangwopark19](https://discuss.huggingface.co/u/sangwopark19)\ **Replies:** 1\ **Last updated:** [December 30, 2025, 9:54am UTC](https://discuss.huggingface.co/t/need-advice-maintaining-product-fidelity-texture-in-generative-ai-mockup-automation-stable-diffusion-gemini/172102 "2025-12-30T09:54:45Z")
Our team is building an automated pipeline for e-commerce merchandise. We merge character IP assets onto product blanks (mugs, t-shirts, phone cases, etc.) to create realistic mockups. Currently, weā€™re using a Gemini-basā€¦ --- ## [Diffusers-workflow - A JSON-based workflow engine for Diffusers](https://discuss.huggingface.co/t/diffusers-workflow-a-json-based-workflow-engine-for-diffusers/171960)
**Author:** [@dkackman](https://discuss.huggingface.co/u/dkackman)\ **Replies:** 0\ **Last updated:** [December 25, 2025, 7:46pm UTC](https://discuss.huggingface.co/t/diffusers-workflow-a-json-based-workflow-engine-for-diffusers/171960 "2025-12-25T19:46:11Z")
Hi all, Iā€™ve been dabbling with diffusion and diffusers since the early days and never made the jump to tools like ComfyUI. Instead, I kept iterating on my own approach, and it eventually grew into something that might ā€¦ --- ## [ControlNetModel.from\_single\_file problem](https://discuss.huggingface.co/t/controlnetmodel-from-single-file-problem/171434)
**Author:** [@carmatic](https://discuss.huggingface.co/u/carmatic)\ **Replies:** 1\ **Last updated:** [December 12, 2025, 5:53am UTC](https://discuss.huggingface.co/t/controlnetmodel-from-single-file-problem/171434 "2025-12-12T05:53:44Z")
Hello I am trying to convert alfredplpl/ControlNetForSD2 Ā· Hugging Face to PyTorch Using this script: from diffusers import ControlNetModel controlnet = ControlNetModel.from\_single\_file( "./models/picasso11OpenPoā€¦ --- ## [FLUX-Schnell: How to generate 4 single object outputs using num\_outputs=4 without object mixing?](https://discuss.huggingface.co/t/flux-schnell-how-to-generate-4-single-object-outputs-using-num-outputs-4-without-object-mixing/171163)
**Author:** [@fauad123](https://discuss.huggingface.co/u/fauad123)\ **Replies:** 1\ **Last updated:** [December 5, 2025, 3:09am UTC](https://discuss.huggingface.co/t/flux-schnell-how-to-generate-4-single-object-outputs-using-num-outputs-4-without-object-mixing/171163 "2025-12-05T03:09:36Z")
Hi everyone, I am working on an icon generator using Replicateā€™s FLUX-Schnell model. Goal: Generate 4 icon style images (one centered object per image, plain background) that share consistent style. Issue: If I passā€¦ --- ## [Help with TemporalNetV2 in PyTorch](https://discuss.huggingface.co/t/help-with-temporalnetv2-in-pytorch/171131)
**Author:** [@carmatic](https://discuss.huggingface.co/u/carmatic)\ **Replies:** 3\ **Last updated:** [December 4, 2025, 6:13pm UTC](https://discuss.huggingface.co/t/help-with-temporalnetv2-in-pytorch/171131 "2025-12-04T18:13:41Z")
I am trying to use this controlnet: CiaraRowles/TemporalNet2 Ā· Hugging Face The files are provided as .safetensors, .ckpt, and .yaml However my PyTorch project SPECIFICALLY requires it as a ā€˜diffusion\_pytorch\_model.binā€¦ --- ## [Loading an image to SAG pipeline](https://discuss.huggingface.co/t/loading-an-image-to-sag-pipeline/167744)
**Author:** [@haobaiguy](https://discuss.huggingface.co/u/haobaiguy)\ **Replies:** 2\ **Last updated:** [August 26, 2025, 11:43am UTC](https://discuss.huggingface.co/t/loading-an-image-to-sag-pipeline/167744 "2025-08-26T11:43:10Z")
I realize that this is a deprecated pipeline, but I still want to give it a try in refining an existing image. Does anyone know how I can pass an image to StableDiffusionSAGPipeline? I can tell you what Iā€™ve tried, but ā€¦ --- ## [Cosmos Predict 2 Tensor Dimension issue](https://discuss.huggingface.co/t/cosmos-predict-2-tensor-dimension-issue/165678)
**Author:** [@fireflymll](https://discuss.huggingface.co/u/fireflymll)\ **Replies:** 1\ **Last updated:** [August 5, 2025, 6:39am UTC](https://discuss.huggingface.co/t/cosmos-predict-2-tensor-dimension-issue/165678 "2025-08-05T06:39:17Z")
Hi, I have Cosmos Predict 2B Video2World running on a ec2 gpu instance (Ampere architecture). However, doing inference i keep getting the issue: Processing failed: The size of tensor a (24) must match the size of tensoā€¦ --- ## [Gradient Issue after using torch.no\_grad](https://discuss.huggingface.co/t/gradient-issue-after-using-torch-no-grad/164699)
**Author:** [@sasuke-ss1](https://discuss.huggingface.co/u/sasuke-ss1)\ **Replies:** 1\ **Last updated:** [July 27, 2025, 10:25pm UTC](https://discuss.huggingface.co/t/gradient-issue-after-using-torch-no-grad/164699 "2025-07-27T22:25:22Z")
Hi, I am training a diffusion model. I am facing a weird issue where the gradient tracking is not working. If I run this code segement noise\_pred = self.unet( latent\_model\_input, t, encoder\_hidden\_states=prā€¦ --- ## [Floating point exception with nightly pytorch and cuda](https://discuss.huggingface.co/t/floating-point-exception-with-nightly-pytorch-and-cuda/163308)
**Author:** [@MxtAppz](https://discuss.huggingface.co/u/MxtAppz)\ **Replies:** 4\ **Last updated:** [July 17, 2025, 3:37am UTC](https://discuss.huggingface.co/t/floating-point-exception-with-nightly-pytorch-and-cuda/163308 "2025-07-17T03:37:02Z")
Hi, First of all excuse me if this post is off-topic but I believe this issue is caused by Diffusers, but Iā€™m not opening an issue in git because itā€™s probably a misconfiguration rather than a bug. So Iā€™m new here in trā€¦ --- ## [Finetune Stable Diffusion with depth controlnet on interior datasets](https://discuss.huggingface.co/t/finetune-stable-diffusion-with-depth-controlnet-on-interior-datasets/160192)
**Author:** [@AsmaAILab](https://discuss.huggingface.co/u/AsmaAILab)\ **Replies:** 4\ **Last updated:** [July 6, 2025, 5:27am UTC](https://discuss.huggingface.co/t/finetune-stable-diffusion-with-depth-controlnet-on-interior-datasets/160192 "2025-07-06T05:27:21Z")
I am creating an interior design app, and I am stuck in fine-tuning stable diffusion using controlnet depth condition. But I am stuck on whether I should fine-tune only ControlNet for depth condition or to train Stable Dā€¦ --- ## [Debugging Custom Stable Diffusion Pipeline for 1D Signal Generation](https://discuss.huggingface.co/t/debugging-custom-stable-diffusion-pipeline-for-1d-signal-generation/161425)
**Author:** [@ThuyVuPhuong](https://discuss.huggingface.co/u/ThuyVuPhuong)\ **Replies:** 1\ **Last updated:** [July 2, 2025, 6:59am UTC](https://discuss.huggingface.co/t/debugging-custom-stable-diffusion-pipeline-for-1d-signal-generation/161425 "2025-07-02T06:59:16Z")
Hi everyone, Iā€™m building a custom Stable Diffusion pipeline to generate 1D signal data instead of images. Hereā€™s what Iā€™ve implemented so far: :white\_check\_mark: Trained a custom autoencoder to encode raw 1D signals ā€¦ --- ## [solving mazes with diffusion models](https://discuss.huggingface.co/t/solving-mazes-with-diffusion-models/161301)
**Author:** [@omarlahia](https://discuss.huggingface.co/u/omarlahia)\ **Replies:** 0\ **Last updated:** [July 1, 2025, 10:34am UTC](https://discuss.huggingface.co/t/solving-mazes-with-diffusion-models/161301 "2025-07-01T10:34:58Z")
Hi everyone!! :waving\_hand: I recently started an experimental project to explore whether diffusion models can learn to solve mazes, and if these models are capable of learning structured patterns like paths from start ā€¦ --- ## [Cloud not import module AutoImageProcessor](https://discuss.huggingface.co/t/cloud-not-import-module-autoimageprocessor/160857)
**Author:** [@aniket2025](https://discuss.huggingface.co/u/aniket2025)\ **Replies:** 1\ **Last updated:** [June 27, 2025, 8:22am UTC](https://discuss.huggingface.co/t/cloud-not-import-module-autoimageprocessor/160857 "2025-06-27T08:22:22Z")
Currently I am just copy pasting the code for Inapinting using SDXL and IPAdapter as provided in the IPAdapter section. The code is: Importing required packages import torch from diffusers import AutoPipelineForImage2Iā€¦ --- ## [SDXL Image to Image, howto](https://discuss.huggingface.co/t/sdxl-image-to-image-howto/54624)
**Author:** [@delon](https://discuss.huggingface.co/u/delon)\ **Replies:** 8\ **Last updated:** [June 1, 2025, 9:38pm UTC](https://discuss.huggingface.co/t/sdxl-image-to-image-howto/54624 "2025-06-01T21:38:56Z")
Hi all Iā€™ve been using the SD1.5 image to image diffusers and theyā€™ve been working really well. The idea is that I take a basic drawing and make it real based on the prompt. Iā€™m trying to move over to SDXL but I can sā€¦ --- ## [Discusison of PR for StableDiffusionXLPipeline](https://discuss.huggingface.co/t/discusison-of-pr-for-stablediffusionxlpipeline/156756)
**Author:** [@ppbrown](https://discuss.huggingface.co/u/ppbrown)\ **Replies:** 2\ **Last updated:** [May 26, 2025, 2:50am UTC](https://discuss.huggingface.co/t/discusison-of-pr-for-stablediffusionxlpipeline/156756 "2025-05-26T02:50:56Z")
This is about the draft PR, I was planning to discuss it in discord, but the bot wont let me see the channels, even though it says I am already ā€œRegisteredā€ for it. Cut-n-pasting the PR description: This makes the ā€œā€¦ --- ## [Slow Loading of Diffusers](https://discuss.huggingface.co/t/slow-loading-of-diffusers/155874)
**Author:** [@bradley6597](https://discuss.huggingface.co/u/bradley6597)\ **Replies:** 1\ **Last updated:** [May 19, 2025, 11:56am UTC](https://discuss.huggingface.co/t/slow-loading-of-diffusers/155874 "2025-05-19T11:56:55Z")
I am currently running diffusers==0.33.1 and finding that the import times of modules are super slow. For context, I am running this on an A100 GPU instance on Google Colab and this occurs when I restart a session and thā€¦ --- ## [Guidance on Training Stable Diffusion Models for Image Generation with Multiple Object Categories](https://discuss.huggingface.co/t/guidance-on-training-stable-diffusion-models-for-image-generation-with-multiple-object-categories/57405)
**Author:** [@livingbox](https://discuss.huggingface.co/u/livingbox)\ **Replies:** 3\ **Last updated:** [May 5, 2025, 4:32am UTC](https://discuss.huggingface.co/t/guidance-on-training-stable-diffusion-models-for-image-generation-with-multiple-object-categories/57405 "2025-05-05T04:32:58Z")
Hii Community, I am currently working on training a Stable Diffusion model for image generation, and I have encountered a few challenges that I believe your expertise can help me overcome. Scenario: I am in the procesā€¦ --- ## [Multiple conditioning for SD](https://discuss.huggingface.co/t/multiple-conditioning-for-sd/153410)
**Author:** [@LegumMagister](https://discuss.huggingface.co/u/LegumMagister)\ **Replies:** 2\ **Last updated:** [May 5, 2025, 4:02am UTC](https://discuss.huggingface.co/t/multiple-conditioning-for-sd/153410 "2025-05-05T04:02:10Z")
Hi, I have not worked before with diffusion models. How can I use multiple conditions like 5 images and 5 corresponding descriptions and get 1 image and 1 description as an output? For the description I believe I shouldā€¦ --- ## [StableDiffusionInpaintPipeline 'NoneType' is not iterable](https://discuss.huggingface.co/t/stablediffusioninpaintpipeline-nonetype-is-not-iterable/152054)
**Author:** [@realredowl](https://discuss.huggingface.co/u/realredowl)\ **Replies:** 1\ **Last updated:** [April 24, 2025, 4:17am UTC](https://discuss.huggingface.co/t/stablediffusioninpaintpipeline-nonetype-is-not-iterable/152054 "2025-04-24T04:17:44Z")
from diffusers import StableDiffusionInpaintPipeline import torch from PIL import Image # Configurazione dispositivo device = torch.device("cuda" if torch.cuda.is\_available() else "cpu") # Configurazione modello model\_ā€¦ --- ## [Pepeline error in Colab](https://discuss.huggingface.co/t/pepeline-error-in-colab/148148)
**Author:** [@Amit71965](https://discuss.huggingface.co/u/Amit71965)\ **Replies:** 1\ **Last updated:** [March 31, 2025, 7:42am UTC](https://discuss.huggingface.co/t/pepeline-error-in-colab/148148 "2025-03-31T07:42:31Z")
For below code: Step 5: Load the base model and LoRA weights from diffusers import DiffusionPipeline import torch print(ā€œLoading base FLUX.1-dev modelā€¦ā€) Load the base model pipe = DiffusionPipeline.from\_pretrained( ā€¦ --- ## [Issue with DDPM Training on Stanford Cars: Noise-Only Samples with Small Batch Sizes](https://discuss.huggingface.co/t/issue-with-ddpm-training-on-stanford-cars-noise-only-samples-with-small-batch-sizes/147375)
**Author:** [@belgou](https://discuss.huggingface.co/u/belgou)\ **Replies:** 1\ **Last updated:** [March 26, 2025, 7:33am UTC](https://discuss.huggingface.co/t/issue-with-ddpm-training-on-stanford-cars-noise-only-samples-with-small-batch-sizes/147375 "2025-03-26T07:33:12Z")
Hi everyone, Iā€™m training a Denoising Diffusion Probabilistic Model (DDPM) on the Stanford Cars dataset, and Iā€™ve noticed a strange issue when using smaller batch sizes. Whatā€™s Happening? Iā€™m currently training at 64Ɨā€¦ --- ## [How to get intermeidate output images](https://discuss.huggingface.co/t/how-to-get-intermeidate-output-images/29144)
**Author:** [@dkackman](https://discuss.huggingface.co/u/dkackman)\ **Replies:** 4\ **Last updated:** [March 18, 2025, 6:02am UTC](https://discuss.huggingface.co/t/how-to-get-intermeidate-output-images/29144 "2025-03-18T06:02:32Z")
Is it possible to get the images at each denoising step via the Diffusers library? I am sure Iā€™ve seen it done but canā€™t find where or how. [Next page](https://discuss.huggingface.co/c/discussion-related-to-httpsgithubcomhuggingfacediffusers/63.md?page=1)