repo stringclasses 1
value | github_id int64 1.27B 4.95B | github_node_id stringlengths 18 24 | number int64 8 14.3k | html_url stringlengths 49 53 | api_url stringlengths 59 63 | title stringlengths 1 402 | body stringlengths 1 62.9k ⌀ | state stringclasses 2
values | state_reason stringclasses 4
values | locked bool 2
classes | comments_count int64 0 99 | labels listlengths 0 5 | assignees listlengths 0 5 | created_at stringdate 2022-06-09 16:28:35 2026-07-22 10:32:58 | updated_at stringdate 2022-06-12 22:18:01 2026-07-22 10:32:58 | closed_at stringdate 2022-06-12 22:18:01 2026-07-21 13:17:28 ⌀ | author_association stringclasses 3
values | milestone_title stringclasses 0
values | snapshot_id stringclasses 149
values | extracted_at stringdate 2026-04-07 13:34:13 2026-07-22 10:35:16 | author_login stringlengths 3 28 | author_id int64 1.54k 296M | author_node_id stringlengths 12 20 | author_type stringclasses 3
values | author_site_admin bool 1
class |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
huggingface/diffusers | 1,999,393,950 | I_kwDOHa8MBc53LFSe | 5,845 | https://github.com/huggingface/diffusers/issues/5845 | https://api.github.com/repos/huggingface/diffusers/issues/5845 | Training DreamBooth with Pixart Alpha doesn't work. | ### Describe the bug
The model didn't generate the object I want. I trained with instance prompt "a photo of sks dog" and it generated with a random dog. Maybe there are problems with my training code scripts.
All I did is to change the code train_dreambooth.py from diffusers to adapt it with Pixart-Alpha-Pipelin... | closed | completed | false | 3 | [
"bug",
"stale"
] | [] | 2023-11-17T15:31:30Z | 2024-01-03T15:06:35Z | 2024-01-03T15:06:35Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | apluka34 | 79,373,948 | MDQ6VXNlcjc5MzczOTQ4 | User | false |
huggingface/diffusers | 1,999,557,043 | I_kwDOHa8MBc53LtGz | 5,846 | https://github.com/huggingface/diffusers/issues/5846 | https://api.github.com/repos/huggingface/diffusers/issues/5846 | diffusers/examples/text_to_image/train_text_to_image_lora_sdxl.py --resume_from_checkpoint error | ### Describe the bug
I have tried to continue training in **google colab** for diffusers/examples/text_to_image/train_text_to_image_lora_sdxl.py script with flag resume_from_checkpoint 'latest' but I have got error: No inf checks were recorded for this optimizer. I tried to solve it with different accelerate, vae mode... | closed | completed | false | 5 | [
"bug",
"stale"
] | [
"sayakpaul"
] | 2023-11-17T16:54:20Z | 2024-02-04T15:06:20Z | 2024-02-04T15:06:20Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | stpic270 | 58,371,161 | MDQ6VXNlcjU4MzcxMTYx | User | false |
huggingface/diffusers | 1,999,610,312 | I_kwDOHa8MBc53L6HI | 5,847 | https://github.com/huggingface/diffusers/issues/5847 | https://api.github.com/repos/huggingface/diffusers/issues/5847 | ControlNet + T2IAdapter Pipeline | I noticed that ControlNet is very reactive to the condition strength, that is, from a relative low strength it maintains the shape.
The next figure shows the depth map controlnet of [this](https://images.pexels.com/photos/6518723/pexels-photo-6518723.jpeg) scene, ranging from 0.1 strength to 1.0 strength, naturally ke... | closed | completed | false | 2 | [] | [] | 2023-11-17T17:26:17Z | 2023-11-21T18:59:31Z | 2023-11-21T18:59:31Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | affromero | 9,055,917 | MDQ6VXNlcjkwNTU5MTc= | User | false |
huggingface/diffusers | 1,999,637,808 | I_kwDOHa8MBc53MA0w | 5,849 | https://github.com/huggingface/diffusers/issues/5849 | https://api.github.com/repos/huggingface/diffusers/issues/5849 | Traing lora: accelerate launch train_text_to_image_lora_sdxl.py gives TypeError: main() missing 1 required positional argument: 'args' | ### Describe the bug
Hello 🤗.
I'm trying to train a LoRA SDXL on Google Cloud TPU v3-8 machine with this script: [train_text_to_image_lora_sdxl.py](https://github.com/huggingface/diffusers/blob/c896b841e48b65e561800f829c546f4cf047e634/examples/text_to_image/train_text_to_image_lora_sdxl.py).
I followed setup gui... | closed | completed | false | 13 | [
"bug",
"stale"
] | [] | 2023-11-17T17:43:56Z | 2024-02-08T15:06:23Z | 2024-02-08T15:06:22Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | PawKanarek | 15,096,514 | MDQ6VXNlcjE1MDk2NTE0 | User | false |
huggingface/diffusers | 1,999,942,747 | I_kwDOHa8MBc53NLRb | 5,851 | https://github.com/huggingface/diffusers/issues/5851 | https://api.github.com/repos/huggingface/diffusers/issues/5851 | Convert training checkpoint to a model for a piple | I need to be able to convert a training checkpoint into a model with a data structure like this: https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0/tree/main
Meaning it should be possible to load it with a pipeline (specifically `StableDiffusionXLPipeline`)
currently the checkpoint directory structur... | closed | completed | false | 3 | [
"stale"
] | [] | 2023-11-17T21:18:53Z | 2024-01-03T15:06:29Z | 2024-01-03T15:06:29Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | kopyl | 17,604,849 | MDQ6VXNlcjE3NjA0ODQ5 | User | false |
huggingface/diffusers | 2,000,264,310 | I_kwDOHa8MBc53OZx2 | 5,857 | https://github.com/huggingface/diffusers/issues/5857 | https://api.github.com/repos/huggingface/diffusers/issues/5857 | LCM lora not work well on AnimateDiff sd xl module | ### Describe the bug
I use similiar code based on https://github.com/guoyww/AnimateDiff/blob/sdxl/scripts/animate.py
I just add code to load lcm lora. And it seems that the quality of the video is pool, compared with the model without lcm lora.
I wonder how to debug and fix the problem. Any advice would be gre... | closed | completed | false | 7 | [
"bug",
"stale"
] | [] | 2023-11-18T06:06:28Z | 2024-02-07T15:05:52Z | 2024-02-07T15:05:52Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | chuck-ma | 74,402,255 | MDQ6VXNlcjc0NDAyMjU1 | User | false |
huggingface/diffusers | 1,379,555,200 | I_kwDOHa8MBc5SOleA | 586 | https://github.com/huggingface/diffusers/issues/586 | https://api.github.com/repos/huggingface/diffusers/issues/586 | Update Pillow library | ### Describe the bug
Compatibility error when installing Pillow library > 9.1
|Deprecated | Use Instead|
|----------------|----------|
|Image.LANCZOS | Image.Resampling.LANCZOS|
|Image.NEAREST | Image.Resampling.NEAREST|
|Image.BILINEAR| Image.Resampling.BILINEAR|
Here the [link](https://pillow.readthedocs... | closed | completed | false | 1 | [
"bug"
] | [
"anton-l"
] | 2022-09-20T14:51:15Z | 2022-09-21T22:02:15Z | 2022-09-21T22:02:15Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | CristoJV | 44,086,220 | MDQ6VXNlcjQ0MDg2MjIw | User | false |
huggingface/diffusers | 2,000,548,894 | I_kwDOHa8MBc53PfQe | 5,860 | https://github.com/huggingface/diffusers/issues/5860 | https://api.github.com/repos/huggingface/diffusers/issues/5860 | Tiny Autoencoder + ControlNet Pipelines doesn't work | ### Describe the bug
It seems like the slight architectural differences between autoencoder_tiny and the regular autoencoderKL are preventing it from being used with ControlNet pipelines. I'm finding that if I make a slight modification to the ControlNet pipelines, everything works fine. The only thing preventing me f... | closed | completed | false | 3 | [
"bug"
] | [] | 2023-11-18T19:28:24Z | 2024-01-15T07:25:55Z | 2023-11-27T13:17:48Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | aaronsantiago | 47,049,060 | MDQ6VXNlcjQ3MDQ5MDYw | User | false |
huggingface/diffusers | 2,000,766,852 | I_kwDOHa8MBc53QUeE | 5,861 | https://github.com/huggingface/diffusers/issues/5861 | https://api.github.com/repos/huggingface/diffusers/issues/5861 | Using ControlNet only with `UNet2DModel` | **What API design would you like to have changed or added to the library? Why?**
I want to use ControlNet only with a normal DDPM process rather than StableDiffusion, which means I need to build `ControlNetModel` with `UNet2DModel` instead of `UNet2DConditionModel`. I want only to use conditions from *ControlNet* to c... | closed | completed | false | 3 | [
"stale"
] | [] | 2023-11-19T09:16:47Z | 2024-06-18T17:32:43Z | 2024-01-03T15:06:23Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | ChengAoShen | 68,673,559 | MDQ6VXNlcjY4NjczNTU5 | User | false |
huggingface/diffusers | 2,000,866,712 | I_kwDOHa8MBc53Qs2Y | 5,862 | https://github.com/huggingface/diffusers/issues/5862 | https://api.github.com/repos/huggingface/diffusers/issues/5862 | Transformer2DModel does not support gradient checkpointing. | ### Describe the bug
Transformer2DModel does not support gradient checkpointing.
### Reproduction
`transformer = Transformer2DModel.from_pretrained(args.pretrained_teacher_model, subfolder="transformer")
transformer.enable_gradient_checkpointing()`
### Logs
```shell
Traceback (most recent call last):
... | closed | completed | false | 3 | [
"bug"
] | [] | 2023-11-19T14:09:08Z | 2024-06-28T12:12:04Z | 2023-11-27T10:51:13Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | lawrence-cj | 55,283,454 | MDQ6VXNlcjU1MjgzNDU0 | User | false |
huggingface/diffusers | 2,000,936,270 | I_kwDOHa8MBc53Q91O | 5,863 | https://github.com/huggingface/diffusers/issues/5863 | https://api.github.com/repos/huggingface/diffusers/issues/5863 | stable-diffusion-v1-4 does not have required filenames for variant=fp16 | Running a [stable_diffusion notebook ](https://github.com/fastai/diffusion-nbs/blob/master/stable_diffusion.ipynb ) from the fast.ai course on colab, I get an error asking me to open an issue here. Apparently you would even be grateful !??
```
pipe = StableDiffusionPipeline.from_pretrained("CompVis/stable-diffusi... | closed | completed | false | 6 | [
"stale"
] | [] | 2023-11-19T16:47:00Z | 2023-12-26T21:30:49Z | 2023-12-26T21:30:49Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | andreaskundig | 899,160 | MDQ6VXNlcjg5OTE2MA== | User | false |
huggingface/diffusers | 2,001,222,569 | I_kwDOHa8MBc53SDup | 5,865 | https://github.com/huggingface/diffusers/issues/5865 | https://api.github.com/repos/huggingface/diffusers/issues/5865 | Separate text_encoder learning rate for LoRA scripts | This is one I notice with LoRA you can train with a very high unet learning rate of 1e-4, but the text encoder being a separate entity I feel this lr is way too high, would be cool to have a bit more control over both. | closed | completed | false | 3 | [
"stale"
] | [] | 2023-11-20T02:01:52Z | 2024-01-03T15:06:21Z | 2024-01-03T15:06:21Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | jmaccall316 | 120,006,038 | U_kgDOBycllg | User | false |
huggingface/diffusers | 2,001,686,461 | I_kwDOHa8MBc53T0-9 | 5,866 | https://github.com/huggingface/diffusers/issues/5866 | https://api.github.com/repos/huggingface/diffusers/issues/5866 | AnimateDiff + ControlNet | ### Model/Pipeline/Scheduler description
Anyone interested in adding a `AnimateDiffControlNetPipeline`?
The expected behavior is to allow user to pass a list of conditions (e.g. pose) and use them to condition the generation for each frame.
see comfy UI example workflow here
https://github.com/Kosinkadink/Com... | closed | completed | false | 12 | [
"good first issue",
"New pipeline/model",
"contributions-welcome"
] | [] | 2023-11-20T09:04:06Z | 2024-10-25T22:07:48Z | 2024-10-25T22:07:47Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | yiyixuxu | 12,631,849 | MDQ6VXNlcjEyNjMxODQ5 | User | false |
huggingface/diffusers | 2,002,092,980 | I_kwDOHa8MBc53VYO0 | 5,868 | https://github.com/huggingface/diffusers/issues/5868 | https://api.github.com/repos/huggingface/diffusers/issues/5868 | Dataset access of LCM training | Hi, thanks for your good job in LCM-LoRA training ! I want to try the new feature, but it seems that i do not have access to the training data since i do not have aws account, could you please provide a new data address such as google drive or huggingface for better access? Thanks ! | closed | completed | false | 3 | [] | [
"patil-suraj"
] | 2023-11-20T12:26:30Z | 2023-12-06T09:35:38Z | 2023-12-06T09:35:37Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | littletomatodonkey | 14,270,174 | MDQ6VXNlcjE0MjcwMTc0 | User | false |
huggingface/diffusers | 2,003,411,460 | I_kwDOHa8MBc53aaIE | 5,873 | https://github.com/huggingface/diffusers/issues/5873 | https://api.github.com/repos/huggingface/diffusers/issues/5873 | Conversion to onnx does not work: Exporting the operator 'aten::scaled_dot_product_attention' to ONNX opset version 14 is not supported. Please feel free to request support or submit a pull request on PyTorch GitHub: https://github.com/pytorch/pytorch/issues. | ### Describe the bug
The script scripts/convert_stable_diffusion_checkpoint_to_onnx.py
### Reproduction
Run the script scripts/convert_stable_diffusion_checkpoint_to_onnx.py with SD 1.5 model
### Logs
_No response_
### System Info
Linux
### Who can help?
_No response_ | closed | completed | false | 4 | [
"bug",
"stale"
] | [] | 2023-11-21T04:00:34Z | 2024-01-03T15:06:18Z | 2024-01-03T15:06:18Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | kopyl | 17,604,849 | MDQ6VXNlcjE3NjA0ODQ5 | User | false |
huggingface/diffusers | 2,003,485,967 | I_kwDOHa8MBc53asUP | 5,875 | https://github.com/huggingface/diffusers/issues/5875 | https://api.github.com/repos/huggingface/diffusers/issues/5875 | conver sd_xl_base_1.0 to ONNXT model prompt error | ### Describe the bug
i want to conver sd_xl_base_1.0 to ONNXT model use conv_sd_to_onnx.py but prompt error :
global_step key not found in model
Traceback (most recent call last):
File "J:\download\StableDiffusionONNX\Stable-Diffusion-ONNX-FP16ver1.3\conv_sd_to_onnx.py", line 602, in <module>
pl = load_pipel... | closed | completed | false | 3 | [
"bug",
"stale"
] | [] | 2023-11-21T05:22:44Z | 2024-01-03T15:06:16Z | 2024-01-03T15:06:16Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | mikeduoxin | 20,881,817 | MDQ6VXNlcjIwODgxODE3 | User | false |
huggingface/diffusers | 2,003,518,965 | I_kwDOHa8MBc53a0X1 | 5,876 | https://github.com/huggingface/diffusers/issues/5876 | https://api.github.com/repos/huggingface/diffusers/issues/5876 | Add multi aspect ratio, and multi crop SCM training script | None of the diffusers training scripts support multiple crops or aspect ratios in the data loader. This is mentioned in the tech report for SDXL as a part of their training process.
In my tests diffusion model struggle to do high resolution details. I did this for base SD 1.5 and was planning to port the code to the... | open | null | false | 11 | [
"stale"
] | [
"patil-suraj"
] | 2023-11-21T05:55:58Z | 2024-04-11T15:05:01Z | null | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | nbardy | 1,278,972 | MDQ6VXNlcjEyNzg5NzI= | User | false |
huggingface/diffusers | 2,003,621,391 | I_kwDOHa8MBc53bNYP | 5,877 | https://github.com/huggingface/diffusers/issues/5877 | https://api.github.com/repos/huggingface/diffusers/issues/5877 | subprocess.CalledProcessError occurs in textual_inversion and DreamBooth | ### Describe the bug
subprocess.CalledProcessError occurs in textual_inversion and DreamBooth
### Reproduction
[This notebook](https://colab.research.google.com/github/suzukimain/diffusers_in_Colab/blob/test1/bug.ipynb) reproduces the problem.
### Logs
_No response_
### System Info
- `diffusers` version: 0.24.... | closed | completed | false | 4 | [
"bug"
] | [
"sayakpaul"
] | 2023-11-21T07:24:49Z | 2023-12-05T07:11:01Z | 2023-12-05T07:11:00Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | suzukimain | 131,413,573 | U_kgDOB9U2RQ | User | false |
huggingface/diffusers | 2,004,090,622 | I_kwDOHa8MBc53c_7- | 5,881 | https://github.com/huggingface/diffusers/issues/5881 | https://api.github.com/repos/huggingface/diffusers/issues/5881 | animatediff result is worse than origin Repo | ### Describe the bug
It's seems that animatediff result is bad, generate image is blur, however when I use the sdwebui or origin animatediff, it can get good result
diffuser result
https://github.com/huggingface/diffusers/assets/98469560/994c50d7-0568-4798-8e33-61251c77cd36
sdwebui or origin repo result
h... | closed | completed | false | 11 | [
"bug"
] | [
"DN6"
] | 2023-11-21T11:42:20Z | 2023-12-25T06:06:23Z | 2023-12-07T11:52:05Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | akk-123 | 98,469,560 | U_kgDOBd6GuA | User | false |
huggingface/diffusers | 2,004,399,720 | I_kwDOHa8MBc53eLZo | 5,882 | https://github.com/huggingface/diffusers/issues/5882 | https://api.github.com/repos/huggingface/diffusers/issues/5882 | Classifier free guidance implementation does not match the description in the paper? | The [paper](https://arxiv.org/pdf/2207.12598.pdf) describes classifier free guidance as this:
<img width="687" alt="image" src="https://github.com/huggingface/diffusers/assets/2013726/74c94bb4-09dd-4091-a3f1-239e70c76a25">
But in all of the pipelines in diffusers is implemented like this:
https://github.com/hu... | closed | completed | false | 6 | [
"stale"
] | [
"DN6",
"yiyixuxu"
] | 2023-11-21T14:23:13Z | 2024-01-26T17:28:52Z | 2024-01-26T17:28:52Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | juancamilog | 2,013,726 | MDQ6VXNlcjIwMTM3MjY= | User | false |
huggingface/diffusers | 2,004,829,491 | I_kwDOHa8MBc53f0Uz | 5,884 | https://github.com/huggingface/diffusers/issues/5884 | https://api.github.com/repos/huggingface/diffusers/issues/5884 | [IP-Adapter] adding IP adapter support to all ControlNet and T2I pipelines | we recently added IP-adapter support to many of our pipelines including all the text2img, img2img and inpaint pipelines, as well as the text2img ControlNet pipeline.
anyone interested in adding it to rest of the ControlNet pipelines and T2I pipelines? You can check out this PR to see how it should be added https://g... | closed | completed | false | 17 | [
"good first issue",
"help wanted",
"contributions-welcome"
] | [
"juancopi81",
"a-r-r-o-w"
] | 2023-11-21T17:46:47Z | 2024-11-19T09:17:50Z | 2024-11-19T09:17:49Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | yiyixuxu | 12,631,849 | MDQ6VXNlcjEyNjMxODQ5 | User | false |
huggingface/diffusers | 2,005,072,970 | I_kwDOHa8MBc53gvxK | 5,886 | https://github.com/huggingface/diffusers/issues/5886 | https://api.github.com/repos/huggingface/diffusers/issues/5886 | [IP-Adapter] support face model | We recently added IP-adapter support to many of our pipelines in diffusers! You can now very easily load your IP-Adapter into a diffusers pipeline with `pipe.load_ip_adapter(...)`;
However right now we do not support [`IPAdapterFull`](https://github.com/tencent-ailab/IP-Adapter/blob/e317f750e1829e846c944b9fdd7a2f28... | closed | completed | false | 12 | [
"good first issue",
"help wanted",
"Good second issue",
"contributions-welcome"
] | [] | 2023-11-21T20:28:02Z | 2023-12-07T16:40:41Z | 2023-12-07T16:40:41Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | yiyixuxu | 12,631,849 | MDQ6VXNlcjEyNjMxODQ5 | User | false |
huggingface/diffusers | 2,005,379,780 | I_kwDOHa8MBc53h6rE | 5,888 | https://github.com/huggingface/diffusers/issues/5888 | https://api.github.com/repos/huggingface/diffusers/issues/5888 | Default parameters for UNet1DModel | **What API design would you like to have changed or added to the library? Why?**
I am proposing a potential modification to the default parameters for `UNet1DModel`. In version 0.23.1, the `UNet1DModel` has default parameters where `down_block_types` begins with `DownBlock1DNoSkip` and `extra_in_channels` is set to ... | closed | completed | false | 3 | [
"stale"
] | [] | 2023-11-22T00:58:05Z | 2024-01-03T15:06:11Z | 2024-01-03T15:06:11Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | yonetaniryo | 6,667,994 | MDQ6VXNlcjY2Njc5OTQ= | User | false |
huggingface/diffusers | 2,005,514,581 | I_kwDOHa8MBc53iblV | 5,889 | https://github.com/huggingface/diffusers/issues/5889 | https://api.github.com/repos/huggingface/diffusers/issues/5889 | Stable video diffusion | ### Model/Pipeline/Scheduler description
Hello, yesterday Stable Diffusion open-sourced their image-to-video model. When it will be merged into Diffusers, and if possible, can Diffusers also provide the merged training code?
### Open source status
- [X] The model implementation is available.
- [X] The model weight... | closed | completed | false | 17 | [
"stale"
] | [] | 2023-11-22T03:48:39Z | 2024-01-26T15:06:51Z | 2024-01-26T15:06:51Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | zhw-zhang | 135,094,935 | U_kgDOCA1ilw | User | false |
huggingface/diffusers | 2,005,535,453 | I_kwDOHa8MBc53igrd | 5,890 | https://github.com/huggingface/diffusers/issues/5890 | https://api.github.com/repos/huggingface/diffusers/issues/5890 | Inference on train_text_to_image.py produces warning Token indices sequence length is longer than the specified maximum sequence length for this model | ### Describe the bug
How to resolve the below error,
Token indices sequence length is longer than the specified maximum sequence length for this model (82 > 77). Running this sequence through the model will result in indexing errors
The following part of your input was truncated because CLIP can only handle sequen... | closed | completed | false | 5 | [
"bug",
"stale"
] | [
"sayakpaul"
] | 2023-11-22T04:07:40Z | 2024-01-03T15:06:08Z | 2024-01-03T15:06:08Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | bhosalems | 10,846,405 | MDQ6VXNlcjEwODQ2NDA1 | User | false |
huggingface/diffusers | 2,005,819,878 | I_kwDOHa8MBc53jmHm | 5,892 | https://github.com/huggingface/diffusers/issues/5892 | https://api.github.com/repos/huggingface/diffusers/issues/5892 | Import error | ### Describe the bug
Getting import error while adding activation function
### Reproduction
ImportError Traceback (most recent call last)
[( in <cell line: 33>()
31 from torch import nn
32 get_ipython().system('pip install python-utils')
---> 33 from ..utils import USE_PEFT_BACKEND
34 from .lora import LoRAComp... | closed | completed | false | 3 | [
"bug",
"stale"
] | [] | 2023-11-22T08:29:11Z | 2024-01-09T15:06:40Z | 2024-01-09T15:06:40Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | pow9791 | 58,098,952 | MDQ6VXNlcjU4MDk4OTUy | User | false |
huggingface/diffusers | 2,006,151,336 | I_kwDOHa8MBc53k3Co | 5,894 | https://github.com/huggingface/diffusers/issues/5894 | https://api.github.com/repos/huggingface/diffusers/issues/5894 | Error RuntimeError: Invalid buffer size | ### Describe the bug
I got the error `RuntimeError: Invalid buffer size: 11.25 GB` while create simple gif with `AnimateDiffPipeline`, code is bellow.
Please help me. I use Mac M1 Pro, 16GB
### Reproduction
```
import torch
from diffusers import MotionAdapter, AnimateDiffPipeline, EulerAncestralDiscreteSche... | open | reopened | false | 36 | [
"bug",
"stale"
] | [
"DN6"
] | 2023-11-22T11:36:01Z | 2025-10-19T03:33:41Z | null | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | misteral | 716,630 | MDQ6VXNlcjcxNjYzMA== | User | false |
huggingface/diffusers | 2,006,761,793 | I_kwDOHa8MBc53nMFB | 5,897 | https://github.com/huggingface/diffusers/issues/5897 | https://api.github.com/repos/huggingface/diffusers/issues/5897 | LORA error when running train_text_to_image_lora.py, error Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu! | ### Describe the bug
I tried to experiment with LoRA training following [examples/text_to_image/README.md#training-with-lora](https://github.com/huggingface/diffusers/blob/main/examples/text_to_image/README.md#training-with-lora).
However, I got the error ```RuntimeError: Expected all tensors to be on the same dev... | closed | completed | false | 17 | [
"bug",
"stale"
] | [
"sayakpaul"
] | 2023-11-22T17:01:37Z | 2024-01-15T15:07:07Z | 2024-01-15T15:07:07Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | MhDang | 26,376,101 | MDQ6VXNlcjI2Mzc2MTAx | User | false |
huggingface/diffusers | 1,379,759,824 | I_kwDOHa8MBc5SPXbQ | 590 | https://github.com/huggingface/diffusers/issues/590 | https://api.github.com/repos/huggingface/diffusers/issues/590 | Typo in noise level formula in Karras scheduler | ### Describe the bug
Comparing the formula for noise levels in `KarrasVeScheduler`
https://github.com/huggingface/diffusers/blob/a45dca077c636cd60199bab888da4ecfcc206d18/src/diffusers/schedulers/scheduling_karras_ve.py#L107-L113
with [Karras, Tero, et al.](https://arxiv.org/abs/2206.00364) (VE column of Table... | closed | completed | false | 2 | [
"bug"
] | [
"anton-l"
] | 2022-09-20T17:27:38Z | 2022-09-24T16:24:09Z | 2022-09-24T16:24:09Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | sgrigory | 23,200,558 | MDQ6VXNlcjIzMjAwNTU4 | User | false |
huggingface/diffusers | 2,007,286,404 | I_kwDOHa8MBc53pMKE | 5,902 | https://github.com/huggingface/diffusers/issues/5902 | https://api.github.com/repos/huggingface/diffusers/issues/5902 | The GPU utilization is very low when fine-tuning SDXL in a multi-machine setup | When I fine-tune SDXL using multiple machines, the GPU utilization significantly decreases as the number of machines increases. The GPU utilization is around 95% on a single machine, but it drops to only 63% when training on two machines. However, when training LLM on two machines, there isn't a noticeable decrease in ... | closed | completed | false | 0 | [] | [] | 2023-11-23T00:18:45Z | 2023-11-24T00:25:08Z | 2023-11-24T00:25:08Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Hzj199 | 49,015,556 | MDQ6VXNlcjQ5MDE1NTU2 | User | false |
huggingface/diffusers | 2,007,401,420 | I_kwDOHa8MBc53poPM | 5,903 | https://github.com/huggingface/diffusers/issues/5903 | https://api.github.com/repos/huggingface/diffusers/issues/5903 | Generations with resolutions not divisible by 32 incur loss of quality at bottom & right edges. | ### Description
When generating images at resolutions not divisible by 32, image quality at the bottom and right edge is negatively impacted. It looks like this issue has persisted since #505 as shown in the second example image in #1571.
I haven't had the chance to do a deep dive on this issue so it's not clear ... | closed | completed | false | 3 | [
"bug",
"stale"
] | [] | 2023-11-23T02:51:37Z | 2023-12-26T21:29:24Z | 2023-12-26T21:29:23Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | lachlan-nicholson | 57,090,563 | MDQ6VXNlcjU3MDkwNTYz | User | false |
huggingface/diffusers | 2,007,526,248 | I_kwDOHa8MBc53qGto | 5,904 | https://github.com/huggingface/diffusers/issues/5904 | https://api.github.com/repos/huggingface/diffusers/issues/5904 | problem with StableDiffusionXLControlNetInpaintPipeline | the example of StableDiffusionControlNetInpaintPipeline set the mask area to -1 when make control_image, but XL version set 0 in the example. Would that be more reasonable? Why change? | closed | completed | false | 9 | [
"stale",
"inpainting"
] | [
"yiyixuxu"
] | 2023-11-23T05:49:12Z | 2024-01-15T15:07:05Z | 2024-01-15T15:07:05Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | D222097 | 50,061,868 | MDQ6VXNlcjUwMDYxODY4 | User | false |
huggingface/diffusers | 2,007,571,123 | I_kwDOHa8MBc53qRqz | 5,905 | https://github.com/huggingface/diffusers/issues/5905 | https://api.github.com/repos/huggingface/diffusers/issues/5905 | Request to add diffusion-GAN model | **Is your feature request related to a problem? Please describe.**
I don't see any models related to diffusion-GAN in diffusers library.
**Describe the solution you'd like.**
Is there a plan to support diffusion-GAN models in diffusers library?
Especially, I would like support for the latest diffusion-GAN model,... | closed | completed | false | 8 | [
"stale"
] | [] | 2023-11-23T06:39:08Z | 2024-04-13T06:57:57Z | 2024-01-18T15:05:50Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | JunbongJang | 25,048,253 | MDQ6VXNlcjI1MDQ4MjUz | User | false |
huggingface/diffusers | 2,007,819,245 | I_kwDOHa8MBc53rOPt | 5,907 | https://github.com/huggingface/diffusers/issues/5907 | https://api.github.com/repos/huggingface/diffusers/issues/5907 | DDPMScheduler does not accept `prediction_type` argument despite docs saying so | ### Describe the bug
Docs (https://huggingface.co/docs/diffusers/api/schedulers/ddpm#diffusers.DDPMScheduler) mention argument `prediction_type`, but this does not exist.
### Reproduction
```
from diffusers import DDPMScheduler
DDPMScheduler(prediction_type='sample')
```
### Logs
```shell
╭────────... | closed | completed | false | 0 | [
"bug"
] | [] | 2023-11-23T09:40:46Z | 2023-11-23T09:45:48Z | 2023-11-23T09:45:48Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | juliusbierk | 1,892,174 | MDQ6VXNlcjE4OTIxNzQ= | User | false |
huggingface/diffusers | 2,008,178,161 | I_kwDOHa8MBc53sl3x | 5,912 | https://github.com/huggingface/diffusers/issues/5912 | https://api.github.com/repos/huggingface/diffusers/issues/5912 | How to set config in VaeImageProcessor? | I created a `StableDiffusionControlNetImg2ImgPipeline` and I want to manually set the config `do_normalize` in `VaeImageProcessor`. I wonder how can I set? I look for it in the pipe.vae.config and see nothing about it. | closed | completed | false | 2 | [
"stale"
] | [] | 2023-11-23T12:54:22Z | 2023-12-26T21:29:17Z | 2023-12-26T21:29:17Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | youyuge34 | 15,784,256 | MDQ6VXNlcjE1Nzg0MjU2 | User | false |
huggingface/diffusers | 2,009,043,841 | I_kwDOHa8MBc53v5OB | 5,917 | https://github.com/huggingface/diffusers/issues/5917 | https://api.github.com/repos/huggingface/diffusers/issues/5917 | convert_ldm_unet_checkpoint fails after upgrading diffusers from 0.21.0 to 0.22.0 | ### Describe the bug
unet model: https://civitai.com/models/50696/qteamix-q
unet config: https://huggingface.co/runwayml/stable-diffusion-v1-5/blob/main/unet/config.json
Following code run correctly in diffusers 0.21.0
```
import torch
import safetensors
from diffusers import UNet2DConditionModel
from diffuse... | closed | completed | false | 5 | [
"bug"
] | [
"DN6"
] | 2023-11-24T03:36:17Z | 2023-12-08T12:37:18Z | 2023-12-08T12:37:18Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | xiaoyong-z | 31,178,013 | MDQ6VXNlcjMxMTc4MDEz | User | false |
huggingface/diffusers | 2,009,122,575 | I_kwDOHa8MBc53wMcP | 5,918 | https://github.com/huggingface/diffusers/issues/5918 | https://api.github.com/repos/huggingface/diffusers/issues/5918 | stabilityai/stable-diffusion-2-1-base is missing diffusion_pytorch_model.fp16.safetensors | Hello! Filing an issue here as requested :)
```
/home/paperspace/.local/lib/python3.9/site-packages/diffusers/utils/hub_utils.py:314: FutureWarning: You are loading the variant fp16 from stabilityai/stable-diffusion-2-1-base via `revision='fp16'`. This behavior is deprecated and will be removed in diffusers v1. One... | closed | completed | false | 2 | [
"stale"
] | [] | 2023-11-24T05:27:25Z | 2023-12-26T21:29:12Z | 2023-12-26T21:29:12Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | thedch | 13,357,734 | MDQ6VXNlcjEzMzU3NzM0 | User | false |
huggingface/diffusers | 2,009,315,219 | I_kwDOHa8MBc53w7eT | 5,919 | https://github.com/huggingface/diffusers/issues/5919 | https://api.github.com/repos/huggingface/diffusers/issues/5919 | AttributeError: 'Linear' object has no attribute 'set_lora_layer' | ### Describe the bug
I use (ttps://github.com/huggingface/diffusers/blob/main/examples/text_to_image/train_text_to_image_lora.py to train lora and raise an error ```AttributeError: Linear' object has no attribute 'set_lora_layer```.
### Reproduction
training examples/text_to_image/train_text_to_image_lora.py
### Lo... | closed | completed | false | 22 | [
"bug",
"stale"
] | [] | 2023-11-24T08:33:46Z | 2024-05-20T09:55:52Z | 2024-01-18T15:05:47Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | aihao2000 | 51,043,929 | MDQ6VXNlcjUxMDQzOTI5 | User | false |
huggingface/diffusers | 1,379,815,840 | I_kwDOHa8MBc5SPlGg | 592 | https://github.com/huggingface/diffusers/issues/592 | https://api.github.com/repos/huggingface/diffusers/issues/592 | [Debugging] Stable Diffusion | To quickly debug new functionality for stable diffusion you can use this dummy model:
For PyTorch:
```python
from diffusers import DiffusionPipeline
pipe = DiffusionPipeline.from_pretrained("hf-internal-testing/tiny-stable-diffusion-pipe")
```
For Flax:
```python
from diffusers import FlaxStableDiffus... | closed | completed | false | 5 | [
"stale"
] | [] | 2022-09-20T18:16:27Z | 2024-05-20T22:04:57Z | 2022-10-24T20:43:15Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | patrickvonplaten | 23,423,619 | MDQ6VXNlcjIzNDIzNjE5 | User | false |
huggingface/diffusers | 2,009,540,698 | I_kwDOHa8MBc53xyha | 5,921 | https://github.com/huggingface/diffusers/issues/5921 | https://api.github.com/repos/huggingface/diffusers/issues/5921 | Negative Previous timestep in DDIMScheduler | ### Describe the bug
When executing a pipeline with a DDIM scheduler usually in the beginning the inference steps are set to a custom number and a sequence of corresponding timesteps is calculated. Furthermore, when doing a step with the DDIMScheduler the previous timestep is calculated. If a model (e.g. Unconditional... | closed | completed | false | 1 | [
"bug",
"stale"
] | [
"yiyixuxu"
] | 2023-11-24T11:07:51Z | 2024-01-03T15:05:55Z | 2024-01-03T15:05:55Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | sidney1505 | 16,155,252 | MDQ6VXNlcjE2MTU1MjUy | User | false |
huggingface/diffusers | 2,010,042,509 | I_kwDOHa8MBc53ztCN | 5,923 | https://github.com/huggingface/diffusers/issues/5923 | https://api.github.com/repos/huggingface/diffusers/issues/5923 | Multi-GPU training Issue: Stuck execution with text-to-image fine-tuning | When I executed the bash script below on multiple GPUs,
```bash
export MODEL_NAME="CompVis/stable-diffusion-v1-4"
export DATASET_NAME="lambdalabs/pokemon-blip-captions"
accelerate launch --mixed_precision="fp16" train_text_to_image.py \
--pretrained_model_name_or_path=$MODEL_NAME \
--dataset_name=$DATASE... | closed | completed | false | 8 | [] | [
"sayakpaul"
] | 2023-11-24T17:40:50Z | 2024-05-14T17:15:28Z | 2023-11-28T13:17:41Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | hanweikung | 643,825 | MDQ6VXNlcjY0MzgyNQ== | User | false |
huggingface/diffusers | 2,010,181,365 | I_kwDOHa8MBc530O71 | 5,924 | https://github.com/huggingface/diffusers/issues/5924 | https://api.github.com/repos/huggingface/diffusers/issues/5924 | [runtime] Spike in memory usage when running VAE ("segmind/SSD-1B", "stabilityai/stable-diffusion-2") | ### Describe the bug
When running inference, the VAE decoders for SD2 and SSD-1B are only ~1/10th the size of the unet, but cause a GPU memory spike when run (sharp increase of 2GB+), causing OOM on my device
Would be able to help investigate if advice is provided. I think there's somewhere where we can early rel... | closed | completed | false | 11 | [
"bug",
"stale"
] | [] | 2023-11-24T20:09:27Z | 2024-08-07T02:56:15Z | 2024-01-03T15:05:53Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | jon-chuang | 9,093,549 | MDQ6VXNlcjkwOTM1NDk= | User | false |
huggingface/diffusers | 2,010,258,719 | I_kwDOHa8MBc530h0f | 5,925 | https://github.com/huggingface/diffusers/issues/5925 | https://api.github.com/repos/huggingface/diffusers/issues/5925 | Add ZipLora to enhance multi-lora inference | **Is your feature request related to a problem? Please describe.**
Refer to https://ziplora.github.io/
**Describe the solution you'd like.**
When multiple LoRAs are passed to the model, you should have the ability to use ziplora to merge them.
**Additional context.**
arxiv paper: https://arxiv.org/abs/2311.136... | closed | completed | false | 4 | [
"stale"
] | [] | 2023-11-24T22:24:02Z | 2024-01-15T15:07:01Z | 2024-01-15T15:07:01Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | AmericanPresidentJimmyCarter | 110,263,573 | U_kgDOBpJ9FQ | User | false |
huggingface/diffusers | 2,010,439,617 | I_kwDOHa8MBc531N_B | 5,926 | https://github.com/huggingface/diffusers/issues/5926 | https://api.github.com/repos/huggingface/diffusers/issues/5926 | Kandinsky 3.0 | ### Model/Pipeline/Scheduler description
Kandinsky 3.0 is an open-source text-to-image diffusion model built upon the Kandinsky2-x model family.
### Open source status
- [X] The model implementation is available.
- [X] The model weights are available (Only relevant if addition is not a scheduler).
### Provide usefu... | closed | completed | false | 2 | [] | [] | 2023-11-25T02:25:16Z | 2023-12-01T12:12:11Z | 2023-12-01T12:12:11Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | trouble-maker007 | 73,164,596 | MDQ6VXNlcjczMTY0NTk2 | User | false |
huggingface/diffusers | 2,010,694,690 | I_kwDOHa8MBc532MQi | 5,927 | https://github.com/huggingface/diffusers/issues/5927 | https://api.github.com/repos/huggingface/diffusers/issues/5927 | IP-Adapter + Reference_adain+attn + Inpaint_only+lama | ### Model/Pipeline/Scheduler description
Hello! I'm looking for a pipeline with these three components. Does anyone know if it has already been implemented?
### Open source status
- [ ] The model implementation is available.
- [ ] The model weights are available (Only relevant if addition is not a scheduler).
### P... | closed | completed | false | 1 | [] | [] | 2023-11-25T15:15:50Z | 2023-12-01T12:11:47Z | 2023-12-01T12:11:46Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | igorriti | 71,894,332 | MDQ6VXNlcjcxODk0MzMy | User | false |
huggingface/diffusers | 2,010,761,882 | I_kwDOHa8MBc532cqa | 5,929 | https://github.com/huggingface/diffusers/issues/5929 | https://api.github.com/repos/huggingface/diffusers/issues/5929 | runwayml/stable-diffusion-v1-5 is missing fp16 files | Please add the correct variant file. | closed | completed | false | 2 | [
"stale"
] | [] | 2023-11-25T18:47:59Z | 2023-12-26T21:28:53Z | 2023-12-26T21:28:53Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | odeb1 | 48,140,495 | MDQ6VXNlcjQ4MTQwNDk1 | User | false |
huggingface/diffusers | 1,379,866,470 | I_kwDOHa8MBc5SPxdm | 593 | https://github.com/huggingface/diffusers/issues/593 | https://api.github.com/repos/huggingface/diffusers/issues/593 | hight res stable diffusion gives repeated images | ### Describe the bug
I have tried to run SD on 1024x1024 and 1664x1664x resolutions, but the result is that I get almost always repeated objects in the generated picture. Is there any simple way to repair it?
For example, results of:
1024x1024 generation is: https://ibb.co/4RKhS6y
1664x1664 resolution is: https... | closed | completed | false | 11 | [
"bug"
] | [] | 2022-09-20T19:02:59Z | 2023-05-01T11:51:41Z | 2022-10-19T13:27:49Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | batrlatom | 11,804,368 | MDQ6VXNlcjExODA0MzY4 | User | false |
huggingface/diffusers | 2,010,902,358 | I_kwDOHa8MBc532-9W | 5,931 | https://github.com/huggingface/diffusers/issues/5931 | https://api.github.com/repos/huggingface/diffusers/issues/5931 | Generation using StableDiffusionPipeline is not deterministic | ### Describe the bug
Generating using the same seed and generator outputs different latents. Tested on `stabilityai/stable-diffusion-2-base` and `stabilityai/stable-diffusion-2-1`
Used CPU generator from [docs](https://huggingface.co/docs/diffusers/using-diffusers/reproducibility)
### Reproduction
```python... | closed | completed | false | 5 | [
"bug",
"stale"
] | [] | 2023-11-26T05:48:08Z | 2024-01-15T15:06:59Z | 2024-01-15T15:06:59Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | nreHieW | 81,154,837 | MDQ6VXNlcjgxMTU0ODM3 | User | false |
huggingface/diffusers | 2,010,928,213 | I_kwDOHa8MBc533FRV | 5,932 | https://github.com/huggingface/diffusers/issues/5932 | https://api.github.com/repos/huggingface/diffusers/issues/5932 | RuntimeError: Given groups=1, weight of size [320, 4, 3, 3], expected input[1, 3, 512, 512] to have 4 channels, but got 3 channels instead | ### Describe the bug
I'm trying to follow the Dreambooth training example and I'm getting this error:
RuntimeError: Given groups=1, weight of size [320, 4, 3, 3], expected input[1, 3, 512, 512] to have 4 channels, but got 3 channels instead
This is the full window:
accelerate launch train_dreambooth.py --pre... | closed | completed | false | 11 | [
"bug",
"stale"
] | [] | 2023-11-26T07:15:38Z | 2025-04-11T07:15:32Z | 2024-01-26T15:06:45Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | TwinkyW | 152,054,747 | U_kgDOCRAr2w | User | false |
huggingface/diffusers | 2,010,961,838 | I_kwDOHa8MBc533Neu | 5,933 | https://github.com/huggingface/diffusers/issues/5933 | https://api.github.com/repos/huggingface/diffusers/issues/5933 | Dreambooth Dataset __len__ returns wrong dataset length in Dreambooth example in `train_dreambooth.py` | ### Describe the bug
Wrong dataset length in the`DreamBoothDataset` in the `train_dreambooth.py` script in the dreambooth example.
The `__len__` method should return the length of the total dataset which includes class and instance images.
Currently, the method returns either the number `num_instance_images`... | closed | completed | false | 2 | [
"bug"
] | [] | 2023-11-26T09:08:17Z | 2023-11-26T20:32:31Z | 2023-11-26T20:30:53Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | maxmynter | 32,773,644 | MDQ6VXNlcjMyNzczNjQ0 | User | false |
huggingface/diffusers | 2,010,966,814 | I_kwDOHa8MBc533Ose | 5,934 | https://github.com/huggingface/diffusers/issues/5934 | https://api.github.com/repos/huggingface/diffusers/issues/5934 | Error when using torch.compile | Hi,
With torch.compile I get below error. without torch.compile everything is OK.
```
/mnt/saeed.khanehgir/txt2image/menv/lib/python3.10/site-packages/torch/_prims/__init__.py:256: UserWarning: ComplexHalf support is experimental and many operators don't support it yet. (Triggered internally at ../aten/src/ATen... | closed | completed | false | 10 | [] | [] | 2023-11-26T09:25:30Z | 2024-04-04T18:34:36Z | 2023-12-04T05:01:41Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | saeedkhanehgir | 65,589,645 | MDQ6VXNlcjY1NTg5NjQ1 | User | false |
huggingface/diffusers | 2,011,249,756 | I_kwDOHa8MBc534Txc | 5,936 | https://github.com/huggingface/diffusers/issues/5936 | https://api.github.com/repos/huggingface/diffusers/issues/5936 | [doc] add a doc page for kandinsky 3.0! | we recently added kandinsky 3.0 to diffusers https://github.com/huggingface/diffusers/pull/5913
one of the to-do remaining is to add a doc page for kandinsky 3.0.
Anyone up to it?
current dog page: https://huggingface.co/docs/diffusers/main/en/api/pipelines/kandinsky3#kandinsky-3
info about the model:
gi... | closed | completed | false | 1 | [
"good first issue",
"help wanted",
"contributions-welcome"
] | [] | 2023-11-26T22:56:07Z | 2023-11-29T14:32:02Z | 2023-11-29T14:32:01Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | yiyixuxu | 12,631,849 | MDQ6VXNlcjEyNjMxODQ5 | User | false |
huggingface/diffusers | 2,011,412,705 | I_kwDOHa8MBc5347jh | 5,940 | https://github.com/huggingface/diffusers/issues/5940 | https://api.github.com/repos/huggingface/diffusers/issues/5940 | SDXLControlNetImg2Img refiner support | When trying the SDXLControlnetImg2Img pipeline with refiner model, its failing. Is there a plan for the support or is it already supported? | closed | completed | false | 2 | [
"stale"
] | [] | 2023-11-27T03:27:05Z | 2024-01-09T15:06:16Z | 2024-01-09T15:06:15Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | polavishnu4444 | 2,263,240 | MDQ6VXNlcjIyNjMyNDA= | User | false |
huggingface/diffusers | 2,011,504,929 | I_kwDOHa8MBc535SEh | 5,941 | https://github.com/huggingface/diffusers/issues/5941 | https://api.github.com/repos/huggingface/diffusers/issues/5941 | [SDXL LoRA DreamBooth training] follow-up from #5508 | Some thoughts:
* Reflect the updates introduced in https://github.com/huggingface/diffusers/pull/5508 in README_sdxl.md. In particular, focus on:
* The simple improvements. If a blog post is being planned, just briefly mentioning the supported stuff and referencing the blog post work for me.
* Interplay between... | closed | completed | false | 2 | [
"stale"
] | [] | 2023-11-27T05:19:52Z | 2024-01-09T15:06:14Z | 2024-01-09T15:06:14Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | sayakpaul | 22,957,388 | MDQ6VXNlcjIyOTU3Mzg4 | User | false |
huggingface/diffusers | 2,011,617,924 | I_kwDOHa8MBc535tqE | 5,942 | https://github.com/huggingface/diffusers/issues/5942 | https://api.github.com/repos/huggingface/diffusers/issues/5942 | How to prepare dataset for text-guided image to image generation | As the title suggests, I want to use stable diffusion to fine-tune my own dataset. How should I build it? I have tried:
--input_image
--xx.jpg
--xx.jpg
--output_image
--yy.jpg
--yy.jpg
metadata.csv
but it did't work ,can anybody help? | closed | completed | false | 2 | [
"stale"
] | [] | 2023-11-27T06:58:57Z | 2024-01-09T15:06:12Z | 2024-01-09T15:06:12Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | feelme0461 | 62,633,341 | MDQ6VXNlcjYyNjMzMzQx | User | false |
huggingface/diffusers | 2,011,967,721 | I_kwDOHa8MBc537DDp | 5,947 | https://github.com/huggingface/diffusers/issues/5947 | https://api.github.com/repos/huggingface/diffusers/issues/5947 | Follow The Policy Objective of Diffusion-DPO | **Is your feature request related to a problem? Please describe.**
A clear and concise description of what the problem is. Ex. I'm always frustrated when [...].
**Describe the solution you'd like.**
A clear and concise description of what you want to happen.
**Describe alternatives you've considered.**
A clear... | closed | completed | false | 1 | [] | [] | 2023-11-27T10:31:11Z | 2023-12-01T12:10:22Z | 2023-12-01T12:10:22Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | 1073521013 | 26,042,094 | MDQ6VXNlcjI2MDQyMDk0 | User | false |
huggingface/diffusers | 2,013,166,550 | I_kwDOHa8MBc53_nvW | 5,951 | https://github.com/huggingface/diffusers/issues/5951 | https://api.github.com/repos/huggingface/diffusers/issues/5951 | Pixart training | ### Describe the bug
I was trying to run PixArt training of 512x512 model following your tutorial, but got this error:
`{'clip_sample', 'clip_sample_range'} was not found in config. Values will be initialized to default values.`
(more info in the "logs" section)
What is pixart: https://github.com/PixArt-alpha/P... | closed | completed | false | 10 | [
"bug",
"stale"
] | [] | 2023-11-27T21:50:05Z | 2024-07-02T13:35:11Z | 2024-01-18T15:05:42Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | kopyl | 17,604,849 | MDQ6VXNlcjE3NjA0ODQ5 | User | false |
huggingface/diffusers | 2,013,492,786 | I_kwDOHa8MBc54A3Yy | 5,953 | https://github.com/huggingface/diffusers/issues/5953 | https://api.github.com/repos/huggingface/diffusers/issues/5953 | StableDiffusionControlNetImg2ImgPipeline will crash when image and control_image sizes are inconsistent | ### Describe the bug
you can see error message `"RuntimeError: The size of tensor a (96) must match the size of tensor b (64) at non-singleton"` after run the code as follows.
I think the reason for this is that the `control_image` and `image` sizes are inconsistent. we can see the source code as follows:
http... | closed | completed | false | 3 | [
"bug",
"stale"
] | [
"yiyixuxu"
] | 2023-11-28T02:50:03Z | 2024-01-18T15:05:40Z | 2024-01-18T15:05:40Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Moskize91 | 14,900,772 | MDQ6VXNlcjE0OTAwNzcy | User | false |
huggingface/diffusers | 2,013,732,185 | I_kwDOHa8MBc54Bx1Z | 5,954 | https://github.com/huggingface/diffusers/issues/5954 | https://api.github.com/repos/huggingface/diffusers/issues/5954 | Kandinsky22 Inference Error | ### Describe the bug
Initialization Error on multi-gpu environment
### Reproduction
```from diffusers import KandinskyV22PriorPipeline, KandinskyV22Pipeline
from diffusers.utils import load_image
import PIL
import torch
from torchvision import transforms
import time
class Kandinsky:
def __init__(s... | closed | completed | false | 1 | [
"bug"
] | [] | 2023-11-28T06:13:49Z | 2023-11-28T12:18:42Z | 2023-11-28T12:18:42Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | geret1 | 75,851,744 | MDQ6VXNlcjc1ODUxNzQ0 | User | false |
huggingface/diffusers | 2,014,006,109 | I_kwDOHa8MBc54C0td | 5,956 | https://github.com/huggingface/diffusers/issues/5956 | https://api.github.com/repos/huggingface/diffusers/issues/5956 | Degraded Image Generation Performance in SD-XL U-Net Fine-tuning | ### Describe the bug
Problem:
I am experiencing degraded image generation performance while fine-tuning the U-Net part of the SDXL model using the train_text_to_image_sdxl.py script from the diffusers library. The generated results are noticeably worse compared to training on SD1.5 with the exact same dataset.
Here ... | closed | completed | false | 12 | [
"bug"
] | [] | 2023-11-28T09:26:46Z | 2023-11-29T12:46:46Z | 2023-11-29T12:44:08Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | wddwzwhhxx | 37,566,045 | MDQ6VXNlcjM3NTY2MDQ1 | User | false |
huggingface/diffusers | 2,014,020,306 | I_kwDOHa8MBc54C4LS | 5,957 | https://github.com/huggingface/diffusers/issues/5957 | https://api.github.com/repos/huggingface/diffusers/issues/5957 | Make LoRACompatibleConv padding_mode work | **Is your feature request related to a problem? Please describe.**
Due to #4315 ,the forward method is implemented in https://github.com/huggingface/diffusers/blob/50a749e90990932d49556ee54a333278559722f3/src/diffusers/models/lora.py#L354-L365
If you modify the padding mode of LoRACompatibleConv to achieve some other... | closed | completed | false | 6 | [] | [
"sayakpaul"
] | 2023-11-28T09:34:40Z | 2024-02-27T00:05:14Z | 2024-02-27T00:05:14Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | jinghuan-Chen | 42,742,857 | MDQ6VXNlcjQyNzQyODU3 | User | false |
huggingface/diffusers | 1,380,048,350 | I_kwDOHa8MBc5SQd3e | 596 | https://github.com/huggingface/diffusers/issues/596 | https://api.github.com/repos/huggingface/diffusers/issues/596 | clip-guided stable diffusion correctness | ### Describe the bug
the clip-guided pipeline uses these text-embeddings:
https://github.com/huggingface/diffusers/blob/2345481c0e21f1bd84c0d85b866b57d34506d836/examples/community/clip_guided_stable_diffusion.py#L285-L288
defined earlier as:
https://github.com/huggingface/diffusers/blob/2345481c0e21f1bd84c0d8... | closed | completed | false | 6 | [
"bug"
] | [
"patil-suraj"
] | 2022-09-20T22:06:05Z | 2023-04-11T20:33:35Z | 2022-09-28T15:41:35Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | keturn | 83,819 | MDQ6VXNlcjgzODE5 | User | false |
huggingface/diffusers | 2,014,267,399 | I_kwDOHa8MBc54D0gH | 5,960 | https://github.com/huggingface/diffusers/issues/5960 | https://api.github.com/repos/huggingface/diffusers/issues/5960 | ControlNET SDXL train - RuntimeWarning: invalid value encountered in cast images = (images * 255).round().astype("uint8") | ### Describe the bug
When running the train_controlnet_sdxl.py
https://github.com/huggingface/diffusers/blob/main/examples/controlnet/README_sdxl.md
With:
- Runpod Jupyter Notrebooks on an A100 GPU
- Same settings as tutorial, and a custom dataset with exact same structure of tutorial I get this:
/workspace/... | closed | completed | false | 4 | [
"bug"
] | [] | 2023-11-28T11:52:04Z | 2023-11-28T13:04:50Z | 2023-11-28T13:04:50Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | alelordelo | 17,708,751 | MDQ6VXNlcjE3NzA4NzUx | User | false |
huggingface/diffusers | 2,014,397,555 | I_kwDOHa8MBc54EURz | 5,962 | https://github.com/huggingface/diffusers/issues/5962 | https://api.github.com/repos/huggingface/diffusers/issues/5962 | fine tuning with many objects | Hi @thedarkzeno and @patil-suraj
I used your script and it worked fine
now, i want to know that if i can use this script for many objects and give any object one prompt. | closed | completed | false | 1 | [] | [] | 2023-11-28T13:06:08Z | 2023-12-01T12:07:48Z | 2023-12-01T12:07:48Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | hosnasn1987 | 145,333,136 | U_kgDOCKmbkA | User | false |
huggingface/diffusers | 2,014,413,923 | I_kwDOHa8MBc54EYRj | 5,963 | https://github.com/huggingface/diffusers/issues/5963 | https://api.github.com/repos/huggingface/diffusers/issues/5963 | Kandinsky 3.0 fails when passing in embeds instead of prompts | ### Describe the bug
Kandinsky 3.0 fails you you attempt to add embeds rather than prompts.
Since the text model for K3.0 is so heavy this is probably needed more the usual to reduce memory usage and speed, you really don't want to encode the text prompt multiple times if you can avoid it.
```
Traceback (most r... | closed | completed | false | 1 | [
"bug"
] | [
"yiyixuxu"
] | 2023-11-28T13:15:00Z | 2023-12-01T19:15:01Z | 2023-12-01T19:15:01Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Vargol | 62,868 | MDQ6VXNlcjYyODY4 | User | false |
huggingface/diffusers | 2,014,656,535 | I_kwDOHa8MBc54FTgX | 5,965 | https://github.com/huggingface/diffusers/issues/5965 | https://api.github.com/repos/huggingface/diffusers/issues/5965 | test | https://hackerone.com/reports/2247439 | closed | completed | false | 0 | [
"bug",
"good first issue"
] | [] | 2023-11-28T15:05:37Z | 2023-11-29T15:37:14Z | 2023-11-29T15:37:14Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | github-actions[bot] | 41,898,282 | MDM6Qm90NDE4OTgyODI= | Bot | false |
huggingface/diffusers | 2,014,777,277 | I_kwDOHa8MBc54Fw-9 | 5,966 | https://github.com/huggingface/diffusers/issues/5966 | https://api.github.com/repos/huggingface/diffusers/issues/5966 | ControlNet for Stable Diffusion 2.1 | **Is your feature request related to a problem? Please describe.**
I tried to use the `scripts/convert_stable_diffusion_controlnet_to_onnx.py` script to convert [thibaud's Canny ControlNet for SD 2.1](https://huggingface.co/thibaud/controlnet-sd21-canny-diffusers) to ONNX format but it did not work, specifically a seg... | closed | completed | false | 3 | [
"stale"
] | [] | 2023-11-28T16:01:13Z | 2024-01-15T15:06:48Z | 2024-01-15T15:06:48Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | jdp8 | 68,886,123 | MDQ6VXNlcjY4ODg2MTIz | User | false |
huggingface/diffusers | 2,015,487,240 | I_kwDOHa8MBc54IeUI | 5,967 | https://github.com/huggingface/diffusers/issues/5967 | https://api.github.com/repos/huggingface/diffusers/issues/5967 | [Announcement] Stable Diffusion Turbo | Stable Diffusion Turbo works out of the box with Diffusers.
See example usage here: https://huggingface.co/stabilityai/sdxl-turbo#diffusers
The checkpoint has the exact same architecture as SDXL-base and by default generates images of size 512x512. It can be used for everything that SDXL can be used for (Img2Img,... | closed | completed | false | 4 | [
"stale"
] | [] | 2023-11-28T23:24:14Z | 2024-01-05T10:52:21Z | 2024-01-05T10:52:21Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | patrickvonplaten | 23,423,619 | MDQ6VXNlcjIzNDIzNjE5 | User | false |
huggingface/diffusers | 2,015,521,584 | I_kwDOHa8MBc54Imsw | 5,968 | https://github.com/huggingface/diffusers/issues/5968 | https://api.github.com/repos/huggingface/diffusers/issues/5968 | Text2Video and/or Img2Video Conversion Script to ONNX | **Is your feature request related to a problem? Please describe.**
Not sure if I'm getting ahead of myself since some of the models that I will mention are new but I am not able to convert certain Text2Video or Img2Video models from Diffusers format to ONNX format. Some of the models are:
1. Text2Video
- [damo... | closed | completed | false | 2 | [
"stale"
] | [] | 2023-11-28T23:50:44Z | 2024-01-15T15:06:46Z | 2024-01-15T15:06:46Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | jdp8 | 68,886,123 | MDQ6VXNlcjY4ODg2MTIz | User | false |
huggingface/diffusers | 2,015,656,327 | I_kwDOHa8MBc54JHmH | 5,969 | https://github.com/huggingface/diffusers/issues/5969 | https://api.github.com/repos/huggingface/diffusers/issues/5969 | DDIM eta=1 NaNs with ZSNR patch | ### Describe the bug
When using DDIMScheduler and eta=0, the zsnr beta rescale applies correctly and the scheduler works more or less as intended. With eta=1, the tensor NaNs from what I'm assuming is an incorrectly applied ZSNR patch.
### Reproduction
Load a ZSNR model like `ptx0/terminus-xl-gamma-v1` then `img = p... | closed | completed | false | 2 | [
"bug",
"scheduler"
] | [
"yiyixuxu"
] | 2023-11-29T02:02:57Z | 2024-02-07T15:05:43Z | 2024-02-07T15:05:43Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Beinsezii | 39,478,211 | MDQ6VXNlcjM5NDc4MjEx | User | false |
huggingface/diffusers | 1,380,260,931 | I_kwDOHa8MBc5SRRxD | 597 | https://github.com/huggingface/diffusers/issues/597 | https://api.github.com/repos/huggingface/diffusers/issues/597 | train_unconditional.py fails on call to evaluate() | ### Describe the bug
On the first attempt to call `evaluate()` in the train_loop() function, the DDPMPipeline fails due to an incompatible output form the model.
This raises an AttributeError: 'dict' object has no attribute 'shape'
![image](https://user-images.githubusercontent.com/3916001/191408604-a4335643-6fb... | closed | completed | false | 6 | [
"bug"
] | [
"anton-l"
] | 2022-09-21T03:36:01Z | 2022-10-17T11:09:57Z | 2022-10-17T11:09:57Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | jdk514 | 3,916,001 | MDQ6VXNlcjM5MTYwMDE= | User | false |
huggingface/diffusers | 2,015,941,337 | I_kwDOHa8MBc54KNLZ | 5,972 | https://github.com/huggingface/diffusers/issues/5972 | https://api.github.com/repos/huggingface/diffusers/issues/5972 | `Min-SNR weighting` in Training controlnet document (main branch) | ### Describe the bug
Hello! I'm a user who recently discovered `diffusers` and I'm enjoying learning about it.
I noticed that in the main branch documentation, there is a section on `Training > ControlNet` that introduces the [min snr weighting](https://huggingface.co/docs/diffusers/main/en/training/controlnet#m... | closed | completed | false | 4 | [
"bug",
"stale"
] | [] | 2023-11-29T06:58:18Z | 2024-01-15T15:06:41Z | 2024-01-15T15:06:41Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | GoGiants1 | 69,342,392 | MDQ6VXNlcjY5MzQyMzky | User | false |
huggingface/diffusers | 2,016,117,244 | I_kwDOHa8MBc54K4H8 | 5,973 | https://github.com/huggingface/diffusers/issues/5973 | https://api.github.com/repos/huggingface/diffusers/issues/5973 | use multi subject dreambooth for inpainting models | Hi @kopsahlong
i want to use your script for inpainting model
i can use this script(examples/research_projects/dreambooth_inpaint) for defining one subject for model, but i want to define many objects and fine tune inpainting model by them.
can you help me? | closed | completed | false | 1 | [] | [] | 2023-11-29T08:59:38Z | 2023-12-01T12:02:31Z | 2023-12-01T12:02:30Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | hosnasn1987 | 145,333,136 | U_kgDOCKmbkA | User | false |
huggingface/diffusers | 2,016,492,612 | I_kwDOHa8MBc54MTxE | 5,976 | https://github.com/huggingface/diffusers/issues/5976 | https://api.github.com/repos/huggingface/diffusers/issues/5976 | Support IP Adapter for AnimateDiffPipeline | Like title says, request is to support newly IP adapter (via PR #5713) for existing AnimateDiffPipeline
as AnimateDiff would greatly benefit from style stabilization brought by IP adapter
CC @yiyixuxu @DN6 @sayakpaul @patrickvonplaten | closed | completed | false | 7 | [
"contributions-welcome"
] | [
"DN6",
"yiyixuxu"
] | 2023-11-29T12:33:57Z | 2023-12-01T17:58:14Z | 2023-11-29T17:28:40Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | vladmandic | 57,876,960 | MDQ6VXNlcjU3ODc2OTYw | User | false |
huggingface/diffusers | 2,016,551,926 | I_kwDOHa8MBc54MiP2 | 5,977 | https://github.com/huggingface/diffusers/issues/5977 | https://api.github.com/repos/huggingface/diffusers/issues/5977 | Do not save `unet.unet` keys when training LoRAs | ### Describe the bug
When training LoRAs, sometimes they end up with a `unet.unet` key like in this case https://huggingface.co/sayakpaul/new-lora-check-v15/blob/main/pytorch_lora_weights.safetensors5/tree/main
```
unet.unet.up_blocks.1.attentions.2.transformer_blocks.1.attn2.processor.to_v_lora.up.weight
```
Wh... | closed | completed | false | 3 | [
"bug"
] | [
"sayakpaul"
] | 2023-11-29T13:04:31Z | 2023-12-02T16:05:17Z | 2023-12-02T16:05:17Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | apolinario | 788,417 | MDQ6VXNlcjc4ODQxNw== | User | false |
huggingface/diffusers | 2,016,560,821 | I_kwDOHa8MBc54Mka1 | 5,978 | https://github.com/huggingface/diffusers/issues/5978 | https://api.github.com/repos/huggingface/diffusers/issues/5978 | PDAE and Disdiff | ### Model/Pipeline/Scheduler description
There are two diffusion model with autoencoder，whose name are PDAE and Disdiff.
If diffuser support PDAE quickly, it would be a good job.
### Open source status
- [X] The model implementation is available.
- [x] The model weights are available (Only relevant if addition i... | closed | completed | false | 4 | [
"community-examples",
"stale",
"contributions-welcome"
] | [] | 2023-11-29T13:09:31Z | 2024-01-18T15:05:35Z | 2024-01-18T15:05:35Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | YChienHung | 41,673,864 | MDQ6VXNlcjQxNjczODY0 | User | false |
huggingface/diffusers | 1,380,272,866 | I_kwDOHa8MBc5SRUri | 598 | https://github.com/huggingface/diffusers/issues/598 | https://api.github.com/repos/huggingface/diffusers/issues/598 | INVALID_ARGUMENT : Got invalid dimensions for input: latent_model_input // Whenever I modify the prompt height and weight from 512x512 | Hi
I read the instruction mentioning changing the "create_onnx.py" file, but that one is nowhere to be found?
So I went ahead and modified save_onnyx.py aswell as dml_onny.py, is that correct?
Anyway, the error message I am facing is:
`File "C:\Users\...\Desktop\StableDif2\amd_venv\lib\site-packages\onnxruntime\cap... | closed | completed | false | 4 | [] | [] | 2022-09-21T03:55:54Z | 2022-09-25T20:16:15Z | 2022-09-21T12:11:34Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Stable777 | 114,045,211 | U_kgDOBswxGw | User | false |
huggingface/diffusers | 2,016,708,939 | I_kwDOHa8MBc54NIlL | 5,980 | https://github.com/huggingface/diffusers/issues/5980 | https://api.github.com/repos/huggingface/diffusers/issues/5980 | Controlnet SDXL - generated models as ".pth" or ".safetensors"? | I am following the tutorial on:
https://github.com/huggingface/diffusers/blob/main/examples/controlnet/README_sdxl.md
After training, I located this under the checkpoints folder:
diffusion_pytorch_model.safetensors
config.json
I was expecting a ".pth" model like (as its common on WebUI ControlNET folder):
con... | closed | completed | false | 1 | [] | [] | 2023-11-29T14:23:18Z | 2023-11-29T14:34:42Z | 2023-11-29T14:34:41Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | alelordelo | 17,708,751 | MDQ6VXNlcjE3NzA4NzUx | User | false |
huggingface/diffusers | 2,016,776,593 | I_kwDOHa8MBc54NZGR | 5,981 | https://github.com/huggingface/diffusers/issues/5981 | https://api.github.com/repos/huggingface/diffusers/issues/5981 | Bug for DPMSolverMultistepScheduler | ### Describe the bug
AttributeError: 'DPMSolverMultistepScheduler' object has no attribute 'sigmas'.
```python
from diffusers import DPMSolverMultistepScheduler
import torch
scheduler = DPMSolverMultistepScheduler()
x = torch.ones(10, 2)
t = torch.tensor([0, 1, 2, 3, 4, 5, 6, 7, 8, 9]).long()
noise = torch.... | closed | completed | false | 5 | [
"bug"
] | [
"yiyixuxu"
] | 2023-11-29T14:56:34Z | 2023-12-01T17:16:18Z | 2023-12-01T17:16:18Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | JinGuang-cuhksz | 97,371,992 | U_kgDOBc3HWA | User | false |
huggingface/diffusers | 2,017,230,726 | I_kwDOHa8MBc54PH-G | 5,984 | https://github.com/huggingface/diffusers/issues/5984 | https://api.github.com/repos/huggingface/diffusers/issues/5984 | fp16 variant causing system RAM issue on colab with train_dreambooth_lora_sdxl | ### Describe the bug
This is kind of a weird one. When I try to specify the variant arg it downloads the right version but before the GPU can start running the System RAM craps out at 12.7 GB (max) and you get the dreaded kill signal (not even cuda out of memory error)
Tried to diagnose what the problem was by cha... | closed | completed | false | 11 | [
"bug",
"stale"
] | [] | 2023-11-29T19:03:22Z | 2024-02-07T15:05:41Z | 2024-02-07T15:05:41Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | jmaccall316 | 120,006,038 | U_kgDOBycllg | User | false |
huggingface/diffusers | 2,017,809,157 | I_kwDOHa8MBc54RVMF | 5,989 | https://github.com/huggingface/diffusers/issues/5989 | https://api.github.com/repos/huggingface/diffusers/issues/5989 | safetensors.torch bug | ### Describe the bug
When using
`import safetensors`: https://github.com/huggingface/diffusers/blob/f72b28c75b2b4b720a5d8de78556694cf4b893fd/src/diffusers/models/modeling_utils.py#L25
then `safetensors.torch`: https://github.com/huggingface/diffusers/blob/f72b28c75b2b4b720a5d8de78556694cf4b893fd/src/diffusers/model... | closed | completed | false | 4 | [
"bug"
] | [] | 2023-11-30T03:34:13Z | 2023-12-04T08:40:29Z | 2023-12-04T08:40:28Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | lawrence-cj | 55,283,454 | MDQ6VXNlcjU1MjgzNDU0 | User | false |
huggingface/diffusers | 1,380,307,554 | I_kwDOHa8MBc5SRdJi | 599 | https://github.com/huggingface/diffusers/issues/599 | https://api.github.com/repos/huggingface/diffusers/issues/599 | Implement Deepspeed with the DiffusionPipeline | **Is your feature request related to a problem? Please describe.**
Trying to train Stable Diffusion on your local box using the Dreambooth method causes out-of-memory issues on an RTX 3090. I'd like to be able to train such a model without having to have an A100 80GB
**Describe the solution you'd like**
Implement... | closed | completed | false | 9 | [
"stale"
] | [] | 2022-09-21T04:42:15Z | 2023-07-23T05:42:34Z | 2022-12-24T15:03:55Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | vanakema | 1,727,697 | MDQ6VXNlcjE3Mjc2OTc= | User | false |
huggingface/diffusers | 2,017,825,044 | I_kwDOHa8MBc54RZEU | 5,990 | https://github.com/huggingface/diffusers/issues/5990 | https://api.github.com/repos/huggingface/diffusers/issues/5990 | Stable video diffusion inpainting pipeline implementation issues | ### Model/Pipeline/Scheduler description
Hi, thank you for integrating stable video diffusion pipeline. I tried to implement a simple inpainting pipeline, inspired by the legacy inpainting pipeline, but encountered issue in adding noise to the inpainting latents.
In particular, the `scheduler.add_noise` function s... | closed | completed | false | 11 | [
"stale"
] | [] | 2023-11-30T03:55:02Z | 2024-01-18T15:05:31Z | 2024-01-18T15:05:31Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | caoandong | 16,655,650 | MDQ6VXNlcjE2NjU1NjUw | User | false |
huggingface/diffusers | 2,017,851,731 | I_kwDOHa8MBc54RflT | 5,992 | https://github.com/huggingface/diffusers/issues/5992 | https://api.github.com/repos/huggingface/diffusers/issues/5992 | `llm_guided_diffusion` example code gets `AttributeError: 'bool' object has no attribute '__module__'` | ### Describe the bug
I simply followed the instructions [here](https://github.com/huggingface/diffusers/blob/main/examples/community/README.md#llm-grounded-diffusion) to try out the `llm_grounded_diffusion` pipeline and it doesn't work out of the box.
### Reproduction
Started with making sure I'm using main
```
... | closed | completed | false | 2 | [
"bug"
] | [] | 2023-11-30T04:27:52Z | 2023-11-30T20:15:15Z | 2023-11-30T20:15:15Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | KristianMischke | 4,390,379 | MDQ6VXNlcjQzOTAzNzk= | User | false |
huggingface/diffusers | 2,018,222,429 | I_kwDOHa8MBc54S6Fd | 5,994 | https://github.com/huggingface/diffusers/issues/5994 | https://api.github.com/repos/huggingface/diffusers/issues/5994 | runtime error when using T4 on Google Colab with Cuda 11.8 and 12 | ### Describe the bug
When running the script
train_text_to_image_lora.py of the examples folder of the diffusers library,
on Google Colab with a standard T4 GPU, trying with both Cuda 11.8 and 12,
this error happens:
RuntimeError: Expected is_sm80 || is_sm90 to be true, but got false.
Such error doesn't happe... | closed | completed | false | 2 | [
"bug"
] | [] | 2023-11-30T09:25:19Z | 2023-12-13T20:34:34Z | 2023-11-30T09:36:24Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | javismiles | 1,000,452 | MDQ6VXNlcjEwMDA0NTI= | User | false |
huggingface/diffusers | 2,018,269,981 | I_kwDOHa8MBc54TFsd | 5,995 | https://github.com/huggingface/diffusers/issues/5995 | https://api.github.com/repos/huggingface/diffusers/issues/5995 | cannot recurrent the lora effect on the webui | trained my lora in diffusers, It works in local inference,but I cant recurrent it on the webui.
![image](ht... | closed | not_planned | false | 9 | [] | [] | 2023-11-30T09:52:36Z | 2023-12-27T07:19:34Z | 2023-12-27T07:19:34Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | DoctorTar | 147,801,017 | U_kgDOCM9DuQ | User | false |
huggingface/diffusers | 2,018,348,260 | I_kwDOHa8MBc54TYzk | 5,997 | https://github.com/huggingface/diffusers/issues/5997 | https://api.github.com/repos/huggingface/diffusers/issues/5997 | ControlNET SDXL tutorial - "You are using a model of type clip_text_model to instantiate a model of type" | ### Describe the bug
I am following this tutorial:
https://github.com/huggingface/diffusers/blob/main/examples/controlnet/README_sdxl.md
When running the model with same settings I get this warning:
`You are using a model of type clip_text_model to instantiate a model of type . This is not supported for all con... | closed | completed | false | 2 | [
"bug"
] | [] | 2023-11-30T10:35:10Z | 2023-12-01T11:55:24Z | 2023-12-01T11:55:24Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | alelordelo | 17,708,751 | MDQ6VXNlcjE3NzA4NzUx | User | false |
huggingface/diffusers | 2,018,583,495 | I_kwDOHa8MBc54USPH | 5,998 | https://github.com/huggingface/diffusers/issues/5998 | https://api.github.com/repos/huggingface/diffusers/issues/5998 | TyperError in VersatileDiffusionPipeline: get_down_block() got an unexpected keyword argument 'transformer_layers_per_block' | ### Describe the bug
I wan to use the VersatileDiffusionPipeline but I get an error when loading the model with from_pretrained().
```
TypeError Traceback (most recent call last)
[<ipython-input-7-2673c81923ce>](https://localhost:8080/#) in <cell line: 1>()
----> 1 vdm = diffuse... | closed | completed | false | 4 | [
"bug",
"good first issue",
"contributions-welcome"
] | [] | 2023-11-30T12:38:01Z | 2023-12-01T16:26:52Z | 2023-12-01T15:27:00Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | fotinidelig | 44,481,663 | MDQ6VXNlcjQ0NDgxNjYz | User | false |
huggingface/diffusers | 2,018,728,357 | I_kwDOHa8MBc54U1ml | 5,999 | https://github.com/huggingface/diffusers/issues/5999 | https://api.github.com/repos/huggingface/diffusers/issues/5999 | Dtype error in training controlnet with additional lora layers | ### Describe the bug
I add lora layers to unet model according to `train_text_to_image_lora.py` in `train_text_to_image_controlnet.py`, while the training looks smooth, but when validation, the errors of different dtype data occurred in unet.forward() function. I checked before the pipeline constructed, all paramete... | closed | completed | false | 12 | [
"bug"
] | [] | 2023-11-30T13:58:39Z | 2023-12-06T07:03:21Z | 2023-12-04T09:04:22Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | OrangeSodahub | 54,439,582 | MDQ6VXNlcjU0NDM5NTgy | User | false |
huggingface/diffusers | 2,018,759,071 | I_kwDOHa8MBc54U9Gf | 6,000 | https://github.com/huggingface/diffusers/issues/6000 | https://api.github.com/repos/huggingface/diffusers/issues/6000 | nm | ... | closed | completed | false | 0 | [] | [] | 2023-11-30T14:15:06Z | 2023-11-30T16:25:10Z | 2023-11-30T16:24:58Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | rmac85 | 90,567,763 | MDQ6VXNlcjkwNTY3NzYz | User | false |
huggingface/diffusers | 2,018,765,512 | I_kwDOHa8MBc54U-rI | 6,001 | https://github.com/huggingface/diffusers/issues/6001 | https://api.github.com/repos/huggingface/diffusers/issues/6001 | StableDiffusionXLInpaintPipeline.from_single_file for refiner returns back Img2ImgPipeline | ### Describe the bug
When using this call:
StableDiffusionXLInpaintPipeline.from_single_file(<path>/sd_xl_refiner_1.0.safetensors), we get an instance of Img2ImgPipeline rather than Inpainting pipeline.
This seems related to [4186](https://github.com/huggingface/diffusers/issues/4186), but in this case Img2ImgPipeli... | closed | completed | false | 6 | [
"bug"
] | [] | 2023-11-30T14:18:35Z | 2024-01-04T06:56:28Z | 2024-01-04T06:56:28Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | darshats | 56,063,876 | MDQ6VXNlcjU2MDYzODc2 | User | false |
huggingface/diffusers | 2,019,704,235 | I_kwDOHa8MBc54Yj2r | 6,008 | https://github.com/huggingface/diffusers/issues/6008 | https://api.github.com/repos/huggingface/diffusers/issues/6008 | IP adapter plus face is failing with LCM Lora - new diffusers (v0.24.0) release. | ### Describe the bug
I am getting key error, while trying to change the ip adapter to "ip-adapter-plus-face_sd15.bin" model of h94/IP-Adapter
### Reproduction
Just take sample code of IP adapter - LCM L... | closed | completed | false | 4 | [
"bug",
"stale"
] | [] | 2023-11-30T23:44:39Z | 2024-01-04T06:40:06Z | 2024-01-04T06:40:06Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | AyushUnleashed | 77,686,324 | MDQ6VXNlcjc3Njg2MzI0 | User | false |
huggingface/diffusers | 2,020,285,236 | I_kwDOHa8MBc54axs0 | 6,013 | https://github.com/huggingface/diffusers/issues/6013 | https://api.github.com/repos/huggingface/diffusers/issues/6013 | Add `unload_textual_inversion()` method | **Is your feature request related to a problem? Please describe.**
Currently you can add textual inversion with `.load_textual_inversion()`. But once the tokenizer and text encoder have the new embeddings, it's complex to remove the embeddings and get it back to the original state.
**Describe the solution you'd lik... | closed | completed | false | 8 | [
"Good second issue",
"contributions-welcome"
] | [] | 2023-12-01T07:30:58Z | 2024-01-29T16:26:23Z | 2024-01-29T16:26:23Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | apolinario | 788,417 | MDQ6VXNlcjc4ODQxNw== | User | false |
huggingface/diffusers | 2,020,788,137 | I_kwDOHa8MBc54csep | 6,016 | https://github.com/huggingface/diffusers/issues/6016 | https://api.github.com/repos/huggingface/diffusers/issues/6016 | torch.Size() issue with convert_original_stable_diffusion_to_diffusers | ### Describe the bug
Hello,
I attempted to convert a model from Civitai (https://civitai.com/models/132632?modelVersionId=168582) to Diffusers, but encountered an error:
```
ValueError: Trying to set a tensor of shape torch.Size([320, 9, 3, 3]) in "weight" (which has shape torch.Size([320, 4, 3, 3])), this lo... | closed | completed | false | 7 | [
"bug"
] | [] | 2023-12-01T12:28:17Z | 2024-07-05T17:08:36Z | 2023-12-01T16:49:02Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | zhuraromdev | 78,348,856 | MDQ6VXNlcjc4MzQ4ODU2 | User | false |
huggingface/diffusers | 2,021,874,307 | I_kwDOHa8MBc54g1qD | 6,026 | https://github.com/huggingface/diffusers/issues/6026 | https://api.github.com/repos/huggingface/diffusers/issues/6026 | MACos stable-diffusion-webui图像生成 终端报错 | ### Describe the bug
本人使用的是MAC M1芯片。安装部署都没有问题，输入描述词点击生成后，无反应，查看终端，python报错 ANE Scheduling Error 报错内容填写在下方，请问怎么处理这个问题
### Reproduction
Python 3.10.13 (main, Aug 24 2023, 12:59:26) [Clang 15.0.0 (clang-1500.0.40.1)]
Version: v1.6.0-2-g4afaaf8a
Commit hash: 4afaaf8a020c1df457bcf7250cb1c7f609699fa7
Launching Web UI... | closed | completed | false | 3 | [
"bug",
"stale"
] | [] | 2023-12-02T05:58:26Z | 2024-01-16T15:05:42Z | 2024-01-16T15:05:42Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | YuuLinz | 152,678,217 | U_kgDOCRmvSQ | User | false |
huggingface/diffusers | 2,022,124,775 | I_kwDOHa8MBc54hyzn | 6,028 | https://github.com/huggingface/diffusers/issues/6028 | https://api.github.com/repos/huggingface/diffusers/issues/6028 | Kandinsky 3.0 "CUDA Out of memory" error | ### Describe the bug
Kandinsky 3.0 fails with "Out of memory" error when the pipeline starts to work.
When I try other models, like SDXL, there are no problems with it and code lines like "pipe.to('cuda')" work without problems, but when I try Kandinsky3 there are.
GPU: 1x T4 GPU (Google colab)
### Reproducti... | closed | completed | false | 15 | [
"bug"
] | [] | 2023-12-02T17:01:23Z | 2023-12-14T07:45:39Z | 2023-12-13T20:37:01Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Atm4x | 96,025,052 | U_kgDOBbk53A | User | false |
huggingface/diffusers | 2,022,540,806 | I_kwDOHa8MBc54jYYG | 6,032 | https://github.com/huggingface/diffusers/issues/6032 | https://api.github.com/repos/huggingface/diffusers/issues/6032 | improve image quality | Hi,
I use Stable diffusion XL. One problem is that when an object is associated with another object، there is no continuity and the image produced is not of quality. For example, when a person holds an umbrella, there is no continuity between the umbrella and the person's hand, as in the image below.
| closed | not_planned | false | 1 | [] | [] | 2023-12-03T20:01:58Z | 2023-12-04T06:15:54Z | 2023-12-04T06:15:54Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Adeel-Intizar | 56,041,154 | MDQ6VXNlcjU2MDQxMTU0 | User | false |
huggingface/diffusers | 2,022,693,716 | I_kwDOHa8MBc54j9tU | 6,034 | https://github.com/huggingface/diffusers/issues/6034 | https://api.github.com/repos/huggingface/diffusers/issues/6034 | Error converting VAE safetensor to diffusers / UnicodeDecodeError: 'utf-8' codec can't decode byte 0xc9 in position 25944: invalid continuation byte | 
I am facing the above error while converting a sd model from civitai which has config file and vae safetensor separately, i only get this error while trying to convert vae safetensor to diffusers format using the ... | closed | completed | false | 8 | [
"stale"
] | [] | 2023-12-03T20:05:10Z | 2024-01-15T15:06:26Z | 2024-01-15T15:06:26Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Adeel-Intizar | 56,041,154 | MDQ6VXNlcjU2MDQxMTU0 | User | false |
huggingface/diffusers | 2,022,845,716 | I_kwDOHa8MBc54ki0U | 6,036 | https://github.com/huggingface/diffusers/issues/6036 | https://api.github.com/repos/huggingface/diffusers/issues/6036 | Cross attention map utils 🚀 | Hi,
We have implemented cross attention map visualization code for diffusers. We haven't been able to find attention map code for diffusers yet. So, we're sharing it in case anyone need it. 😊
@taemin6697 @rlaehghks5 @Hyunmin-H
![t2i_image](https://github.com/huggingface/diffusers/assets/75835117/0656533c-4059-4eef-... | closed | completed | false | 14 | [] | [] | 2023-12-04T01:32:10Z | 2024-04-05T04:08:06Z | 2023-12-06T15:20:15Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | wooyeolbaek | 75,835,117 | MDQ6VXNlcjc1ODM1MTE3 | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.