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 | 4,212,516,302 | I_kwDOHa8MBc77FeXO | 13,425 | https://github.com/huggingface/diffusers/issues/13425 | https://api.github.com/repos/huggingface/diffusers/issues/13425 | Division by zero in rescale_noise_cfg can produce NaNs during inference | ### Describe the bug
## Bug: Division by zero in `rescale_noise_cfg` can produce NaNs
### Description
The function `rescale_noise_cfg` performs a division by `std_cfg` without any numerical stability guard:
```python
noise_pred_rescaled = noise_cfg * (std_text / std_cfg)
```
If std_cfg becomes zero, this leads to N... | open | null | false | 0 | [
"bug"
] | [] | 2026-04-06T15:32:50Z | 2026-04-06T15:32:50Z | null | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Akash504-ai | 199,394,642 | U_kgDOC-KFUg | User | false |
huggingface/diffusers | 4,216,296,333 | I_kwDOHa8MBc77T5ON | 13,429 | https://github.com/huggingface/diffusers/issues/13429 | https://api.github.com/repos/huggingface/diffusers/issues/13429 | add JoyAI-Image-Edit | ### Model/Pipeline/Scheduler description
JoyAI-Image is a unified multimodal foundation model for image understanding, text-to-image generation, and instruction-guided image editing. It combines an 8B Multimodal Large Language Model (MLLM) with a 16B Multimodal Diffusion Transformer (MMDiT). A central principle of Joy... | closed | completed | false | 0 | [] | [] | 2026-04-07T07:58:41Z | 2026-04-07T07:59:45Z | 2026-04-07T07:59:44Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Moran232 | 9,348,990 | MDQ6VXNlcjkzNDg5OTA= | User | false |
huggingface/diffusers | 1,456,556,096 | I_kwDOHa8MBc5W0UhA | 1,343 | https://github.com/huggingface/diffusers/issues/1343 | https://api.github.com/repos/huggingface/diffusers/issues/1343 | Enable xformers by default | **Is your feature request related to a problem? Please describe.**
Diffusers are too slow to train in the default case.
**Describe the solution you'd like**
If xformers is installed in the environment, xformers is enabled by default.
**Describe alternatives you've considered**
A clear and concise description o... | closed | completed | false | 6 | [] | [
"patil-suraj"
] | 2022-11-19T16:02:11Z | 2023-01-17T09:00:46Z | 2022-11-22T18:36:13Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | chinoll | 22,575,232 | MDQ6VXNlcjIyNTc1MjMy | User | false |
huggingface/diffusers | 4,216,329,419 | I_kwDOHa8MBc77UBTL | 13,430 | https://github.com/huggingface/diffusers/issues/13430 | https://api.github.com/repos/huggingface/diffusers/issues/13430 | Support for JoyAI-Image-Edit | ### Model/Pipeline/Scheduler description
JoyAI-Image is a unified multimodal foundation model for image understanding, text-to-image generation, and instruction-guided image editing. It combines an 8B Multimodal Large Language Model (MLLM) with a 16B Multimodal Diffusion Transformer (MMDiT). A central principle of Joy... | closed | completed | false | 0 | [] | [] | 2026-04-07T08:05:37Z | 2026-05-07T20:57:58Z | 2026-05-07T20:57:57Z | CONTRIBUTOR | null | 20260507T213518Z | 2026-05-07T21:35:18Z | Moran232 | 9,348,990 | MDQ6VXNlcjkzNDg5OTA= | User | false |
huggingface/diffusers | 4,238,301,874 | I_kwDOHa8MBc78n1qy | 13,439 | https://github.com/huggingface/diffusers/issues/13439 | https://api.github.com/repos/huggingface/diffusers/issues/13439 | HunyuanVideo 1.5 I2V image conditioning preprocessed at latent resolution instead of pixel resolution | ### Describe the bug
While working on the modular pipeline for HunyuanVideo 1.5 (#13389), I found a bug in prepare_cond_latents_and_mask in pipeline_hunyuan_video1_5_image2video.py.
Line 614 shadows the pixel height/width parameters with latent dims from latents.shape:
```python
def prepare_cond_latents_and_mask(sel... | closed | completed | false | 0 | [
"bug",
"pipelines"
] | [] | 2026-04-10T09:58:27Z | 2026-04-10T19:54:38Z | 2026-04-10T19:54:38Z | CONTRIBUTOR | null | 20260410T215946Z | 2026-04-10T21:59:46Z | akshan-main | 97,239,696 | U_kgDOBcvCkA | User | false |
huggingface/diffusers | 1,456,583,761 | I_kwDOHa8MBc5W0bRR | 1,344 | https://github.com/huggingface/diffusers/issues/1344 | https://api.github.com/repos/huggingface/diffusers/issues/1344 | attention.py only supports one GPU | ### Model/Pipeline/Scheduler description
Hello!
In attention.py the lines
```
_ = xformers.ops.memory_efficient_attention(
torch.randn((1, 2, 40), device="cuda"),
torch.randn((1, 2, 40), device="cuda"),
torch.randn((1, 2, 40), device... | closed | completed | false | 11 | [
"stale"
] | [] | 2022-11-19T17:12:57Z | 2023-01-13T15:04:34Z | 2023-01-13T15:04:34Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Marcophono2 | 22,599,855 | MDQ6VXNlcjIyNTk5ODU1 | User | false |
huggingface/diffusers | 4,245,912,131 | I_kwDOHa8MBc79E3pD | 13,445 | https://github.com/huggingface/diffusers/issues/13445 | https://api.github.com/repos/huggingface/diffusers/issues/13445 | Wrong default layers for Flux 2 Klein | ### Describe the bug
Thanks a lot for taking the time to review this issue 🤗
There seems to be a mismatch between the default values exposed via CLI arguments and the layers that are actually used internally in the script:
Script:
[pipeline_flux2_klein.py](https://github.com/huggingface/diffusers/blob/main/src/diff... | open | null | false | 0 | [
"bug",
"needs-code-example",
"lora",
"pipelines"
] | [] | 2026-04-11T22:24:31Z | 2026-04-11T22:26:00Z | null | NONE | null | 20260413T135838Z | 2026-04-13T13:58:38Z | forhaxed | 9,382,999 | MDQ6VXNlcjkzODI5OTk= | User | false |
huggingface/diffusers | 4,246,102,084 | I_kwDOHa8MBc79FmBE | 13,446 | https://github.com/huggingface/diffusers/issues/13446 | https://api.github.com/repos/huggingface/diffusers/issues/13446 | Insufficient trust check for custom_pipeline parameter of DiffusionPipeline.from_pretrained method | ### Describe the bug
The `DiffusionPipeline.from_pretrained` API in the Diffusers library allows users to specify a `custom_pipeline`, which can point to an external Hugging Face repository containing a `pipeline.py`. This feature enables flexible extension of pipelines but may raise critical security concerns.
The c... | closed | completed | false | 33 | [
"bug",
"pipelines"
] | [] | 2026-04-12T00:38:10Z | 2026-05-07T09:36:33Z | 2026-04-23T05:37:44Z | NONE | null | 20260507T103516Z | 2026-05-07T10:35:16Z | Vancir | 19,147,918 | MDQ6VXNlcjE5MTQ3OTE4 | User | false |
huggingface/diffusers | 4,246,948,118 | I_kwDOHa8MBc79I0kW | 13,447 | https://github.com/huggingface/diffusers/issues/13447 | https://api.github.com/repos/huggingface/diffusers/issues/13447 | Support for Exclusive Self Attention (XSA) in Diffusion Models (DiT, U-Net, etc.) | **Is your feature request related to a problem? Please describe.**
Self-attention in diffusion models often produces outputs highly similar to the token’s own value, creating redundancy and under-utilizing context from other patches.
**Describe the solution you’d like.**
Add optional support for **Exclusive Self A... | open | null | false | 0 | [
"models",
"feature-request"
] | [] | 2026-04-12T07:01:18Z | 2026-04-13T13:35:30Z | null | NONE | null | 20260413T135838Z | 2026-04-13T13:58:38Z | Bili-Sakura | 108,928,239 | U_kgDOBn4c7w | User | false |
huggingface/diffusers | 1,456,707,098 | I_kwDOHa8MBc5W05Ya | 1,345 | https://github.com/huggingface/diffusers/issues/1345 | https://api.github.com/repos/huggingface/diffusers/issues/1345 | train dreambooth with runway inpainting model | Does anyone know a way to train dream booth with runway inpainting? | closed | completed | false | 7 | [
"stale"
] | [] | 2022-11-19T23:57:48Z | 2023-02-05T19:09:09Z | 2022-12-21T13:22:50Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | loboere | 15,265,895 | MDQ6VXNlcjE1MjY1ODk1 | User | false |
huggingface/diffusers | 4,251,259,682 | I_kwDOHa8MBc79ZRMi | 13,452 | https://github.com/huggingface/diffusers/issues/13452 | https://api.github.com/repos/huggingface/diffusers/issues/13452 | [CI] Trial Sourcery PR reviews | This is a followup to #13343 that continues to consider ways in which PRs can be managed more efficiently.
You may wish to consider trialing Sourcery PR reviews, which are [free](https://www.sourcery.ai/pricing) for open-source repositories. This would facilitate the automation of at least some first-stage PR reviews,... | open | null | false | 0 | [] | [] | 2026-04-13T04:28:06Z | 2026-04-14T21:28:52Z | null | NONE | null | 20260417T173542Z | 2026-04-17T17:35:42Z | iwr-redmond | 142,086,261 | U_kgDOCHgQdQ | User | false |
huggingface/diffusers | 1,457,031,223 | I_kwDOHa8MBc5W2Ig3 | 1,348 | https://github.com/huggingface/diffusers/issues/1348 | https://api.github.com/repos/huggingface/diffusers/issues/1348 | Help replicating imagen | Hello, I'm trying to replicate imagen using the diffusers library, so far I've trained the 64x64 unet, next I have to train the 256x256 model.
My question is, what is the best parameters for the second unet?
I am currently using the following configuration, but not yet getting good results
_class_name="UNet2... | closed | completed | false | 7 | [
"stale"
] | [] | 2022-11-20T20:15:01Z | 2022-12-29T15:03:39Z | 2022-12-29T15:03:39Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | thedarkzeno | 45,200,346 | MDQ6VXNlcjQ1MjAwMzQ2 | User | false |
huggingface/diffusers | 1,315,915,875 | I_kwDOHa8MBc5Ob0hj | 135 | https://github.com/huggingface/diffusers/issues/135 | https://api.github.com/repos/huggingface/diffusers/issues/135 | The model_card_template is not included in the library | Running the training colab fails when pushing to the Hub with
```
FileNotFoundError: [Errno 2] No such file or directory: '/usr/local/lib/python3.7/dist-packages/diffusers/utils/model_card_template.md'
```
Doing
```
!ls /usr/local/lib/python3.7/dist-packages/diffusers/utils/
```
gives
```
dummy_transfor... | closed | completed | false | 8 | [] | [] | 2022-07-24T14:02:15Z | 2022-09-13T15:35:28Z | 2022-09-13T15:35:27Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | osanseviero | 7,246,357 | MDQ6VXNlcjcyNDYzNTc= | User | false |
huggingface/diffusers | 1,457,228,183 | I_kwDOHa8MBc5W24mX | 1,350 | https://github.com/huggingface/diffusers/issues/1350 | https://api.github.com/repos/huggingface/diffusers/issues/1350 | DiffusionDet: Diffusion models for object detection | ### Model/Pipeline/Scheduler description
Recent work which leverages diffusion models for object detection task. https://arxiv.org/abs/2211.09788
Add capability to run it through HF diffusers pipeline and if possible also create benchmarks or comparison on datasets like nuScenes.
### Open source status
- [X] The ... | open | null | false | 23 | [
"New pipeline/model",
"contributions-welcome"
] | [] | 2022-11-21T03:14:44Z | 2024-11-21T11:28:29Z | null | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | 345ishaan | 7,318,028 | MDQ6VXNlcjczMTgwMjg= | User | false |
huggingface/diffusers | 1,457,585,945 | I_kwDOHa8MBc5W4P8Z | 1,351 | https://github.com/huggingface/diffusers/issues/1351 | https://api.github.com/repos/huggingface/diffusers/issues/1351 | Where to save the model locally? | I cannot believe that I really ask this but after two hours of research I didn't find a working way. I did it very often a few months ago but with my new setup - I am lost.
The only things which are different now are that I am using model 1.5 (before 1.4) and that I installed SD now with git clone. (before I downloade... | closed | completed | false | 3 | [
"stale"
] | [] | 2022-11-21T09:16:33Z | 2023-06-20T09:09:37Z | 2022-12-31T15:03:18Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Marcophono2 | 22,599,855 | MDQ6VXNlcjIyNTk5ODU1 | User | false |
huggingface/diffusers | 1,457,825,856 | I_kwDOHa8MBc5W5KhA | 1,353 | https://github.com/huggingface/diffusers/issues/1353 | https://api.github.com/repos/huggingface/diffusers/issues/1353 | how to combine train_unconditional.py with LDMPipline | ### Describe the bug
how to combine train_unconditional.py with LDMPipline
### Reproduction
_No response_
### Logs
_No response_
### System Info
diffuser=0.6.0
python=3.8.12
pycharm | closed | completed | false | 2 | [
"bug",
"stale"
] | [] | 2022-11-21T12:03:17Z | 2022-12-31T15:03:17Z | 2022-12-31T15:03:17Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | panenyaun | 67,307,930 | MDQ6VXNlcjY3MzA3OTMw | User | false |
huggingface/diffusers | 1,458,251,452 | I_kwDOHa8MBc5W6ya8 | 1,358 | https://github.com/huggingface/diffusers/issues/1358 | https://api.github.com/repos/huggingface/diffusers/issues/1358 | Safetensors support | **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 [...]
Provide a way to share and distribute files which are not unsafe (in the pickle sense with pytorch https://twitter.com/amli_art/status/1593312723898552320 )... | closed | completed | false | 3 | [
"stale"
] | [] | 2022-11-21T16:43:08Z | 2022-12-31T15:03:16Z | 2022-12-31T15:03:16Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Narsil | 204,321 | MDQ6VXNlcjIwNDMyMQ== | User | false |
huggingface/diffusers | 1,458,766,128 | I_kwDOHa8MBc5W8wEw | 1,363 | https://github.com/huggingface/diffusers/issues/1363 | https://api.github.com/repos/huggingface/diffusers/issues/1363 | [ONNX] Long prompt weight community pipeline rasies error when using LMS discrete scheduler | ### Describe the bug
When using ONNX with the custom pipeline `lpw_stable_diffusion_onnx` it raises a "unsupported operand" error. It's the same error as this other bug #967
### Reproduction
```python
from diffusers import OnnxStableDiffusionPipeline, DDIMScheduler, PNDMScheduler, LMSDiscreteScheduler
mode... | closed | completed | false | 1 | [
"bug"
] | [
"anton-l"
] | 2022-11-21T23:51:30Z | 2022-12-12T15:06:44Z | 2022-12-12T15:06:44Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | azuritecoin | 115,603,437 | U_kgDOBuP37Q | User | false |
huggingface/diffusers | 1,458,939,034 | I_kwDOHa8MBc5W9aSa | 1,364 | https://github.com/huggingface/diffusers/issues/1364 | https://api.github.com/repos/huggingface/diffusers/issues/1364 | [Community] Is there a code for pre-training stable diffusion? | This project works great. Is there a code for pre-training stable diffusion? | closed | completed | false | 5 | [
"good first issue"
] | [] | 2022-11-22T02:54:03Z | 2022-12-26T11:04:46Z | 2022-12-01T02:09:31Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | 980202006 | 24,452,502 | MDQ6VXNlcjI0NDUyNTAy | User | false |
huggingface/diffusers | 1,460,494,067 | I_kwDOHa8MBc5XDV7z | 1,367 | https://github.com/huggingface/diffusers/issues/1367 | https://api.github.com/repos/huggingface/diffusers/issues/1367 | Can't save configuration of scheduler initialized with custom `trained_betas` | ### Describe the bug
`trained_betas` is a numpy array, and it cannot be serialized to JSON. This makes it currently impossible to save the configuration of a custom scheduler that does not follow one of the predefined scheduler.
### Reproduction
```Python
import numpy as np
from diffusers import EulerDiscreteSched... | closed | completed | false | 1 | [
"bug"
] | [] | 2022-11-22T20:14:53Z | 2022-11-30T09:05:53Z | 2022-11-30T09:05:53Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | pcuenca | 1,177,582 | MDQ6VXNlcjExNzc1ODI= | User | false |
huggingface/diffusers | 1,460,616,167 | I_kwDOHa8MBc5XDzvn | 1,368 | https://github.com/huggingface/diffusers/issues/1368 | https://api.github.com/repos/huggingface/diffusers/issues/1368 | img2img results in noisy image with low num_inference_steps | I'm trying to make a video with the `StableDiffusionImg2ImgPipeline` but am encountering some unexpected behavior.
I generate the first frame by using the `StableDiffusionPipeline` with relatively many steps (e.g. 100). This frame is then slightly warped (e.g. rotation, translation, zoom). Next I feed the frame to `St... | closed | completed | false | 8 | [
"stale"
] | [] | 2022-11-22T22:04:25Z | 2023-02-05T15:03:25Z | 2023-02-05T15:03:25Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | jonasdoevenspeck | 28,825,134 | MDQ6VXNlcjI4ODI1MTM0 | User | false |
huggingface/diffusers | 1,461,373,146 | I_kwDOHa8MBc5XGsja | 1,371 | https://github.com/huggingface/diffusers/issues/1371 | https://api.github.com/repos/huggingface/diffusers/issues/1371 | Adding DreamBooth paper super resolution fine tuning | I'm currently experimenting with the dreambooth training script, and while the results are very impressive, I think it is missing the super resolution fine tuning phase?
In the dreambooth paper they fine tune a low resolution diffusion model to link a rare token with some instance images (with or without class loss)... | closed | completed | false | 2 | [
"stale"
] | [] | 2022-11-23T09:30:51Z | 2022-12-31T15:03:13Z | 2022-12-31T15:03:13Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | alicranck | 20,641,504 | MDQ6VXNlcjIwNjQxNTA0 | User | false |
huggingface/diffusers | 1,461,596,078 | I_kwDOHa8MBc5XHi-u | 1,372 | https://github.com/huggingface/diffusers/issues/1372 | https://api.github.com/repos/huggingface/diffusers/issues/1372 | Training Unconditional Model on Images with different number of channels (!=3) | Hey!
I would like to ask if you have any hints on training an unconditional diffusion model on a different number of channels. For example training it on grayscale images or even on 5-channel images. I tried it out on the [repo](https://github.com/openai/improved-diffusion) connected to the paper by OpenAi with my o... | closed | completed | false | 1 | [] | [] | 2022-11-23T11:38:05Z | 2022-11-23T13:18:26Z | 2022-11-23T13:18:25Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | PatrickSVM | 73,708,538 | MDQ6VXNlcjczNzA4NTM4 | User | false |
huggingface/diffusers | 1,461,969,389 | I_kwDOHa8MBc5XI-Ht | 1,375 | https://github.com/huggingface/diffusers/issues/1375 | https://api.github.com/repos/huggingface/diffusers/issues/1375 | PR #1365 broke importing diffusers | ### Describe the bug
When I try to import diffusers, I get the error below. The file causing the error is from PR #1365, which was merged today.
### Reproduction
```
%cd /content
!git clone -b main https://github.com/huggingface/diffusers.git
import sys
sys.path.insert(0,'/conten... | closed | completed | false | 5 | [
"bug"
] | [] | 2022-11-23T15:31:42Z | 2022-11-24T00:13:16Z | 2022-11-24T00:12:23Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | OrionFive | 13,980,562 | MDQ6VXNlcjEzOTgwNTYy | User | false |
huggingface/diffusers | 1,462,257,796 | I_kwDOHa8MBc5XKEiE | 1,377 | https://github.com/huggingface/diffusers/issues/1377 | https://api.github.com/repos/huggingface/diffusers/issues/1377 | ImportError: cannot import name 'CLIPVisionModelWithProjection' from 'transformers' (/usr/local/lib/python3.9/dist-packages/transformers/__init__.py) | ### Describe the bug
ImportError: cannot import name 'CLIPVisionModelWithProjection' from 'transformers' (/usr/local/lib/python3.9/dist-packages/transformers/__init__.py)
1 hour ago it worked nice.
### Reproduction
_No response_
### Logs
_No response_
### System Info
last, all, all | closed | completed | false | 8 | [
"bug"
] | [] | 2022-11-23T18:58:18Z | 2022-11-24T00:11:47Z | 2022-11-24T00:11:47Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | AlekseiSDev | 37,876,629 | MDQ6VXNlcjM3ODc2NjI5 | User | false |
huggingface/diffusers | 1,462,513,957 | I_kwDOHa8MBc5XLDEl | 1,379 | https://github.com/huggingface/diffusers/issues/1379 | https://api.github.com/repos/huggingface/diffusers/issues/1379 | Repeated operation in Transformer2DModel | It looks like Norm operation is repeated in the Transformer2DModel.forward method in lines:
https://github.com/huggingface/diffusers/blob/f07a16e09bb5b1cf4fa2306bfa4ea791f24fa968/src/diffusers/models/attention.py#L203
and
https://github.com/huggingface/diffusers/blob/f07a16e09bb5b1cf4fa2306bfa4ea791f24fa968/sr... | closed | completed | false | 2 | [] | [] | 2022-11-23T22:57:01Z | 2022-11-29T12:47:03Z | 2022-11-23T23:19:01Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Warvito | 4,493,399 | MDQ6VXNlcjQ0OTMzOTk= | User | false |
huggingface/diffusers | 1,318,080,087 | I_kwDOHa8MBc5OkE5X | 138 | https://github.com/huggingface/diffusers/issues/138 | https://api.github.com/repos/huggingface/diffusers/issues/138 | ImportError: LDMTextToImagePipeline requires the transformers library but it was not found in your environment. | Hi, Thanks for a fantastic model but when I am installing it on Colab, It throws the below error.
`ImportError:
LDMTextToImagePipeline requires the transformers library but it was not found in your environment. You can install it with pip.`
I have installed it via `!pip install diffusers transformers` and I am ... | closed | completed | false | 1 | [] | [] | 2022-07-26T11:07:56Z | 2022-09-13T15:35:34Z | 2022-09-13T15:35:34Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | skuma307 | 86,348,268 | MDQ6VXNlcjg2MzQ4MjY4 | User | false |
huggingface/diffusers | 1,462,595,537 | I_kwDOHa8MBc5XLW_R | 1,384 | https://github.com/huggingface/diffusers/issues/1384 | https://api.github.com/repos/huggingface/diffusers/issues/1384 | NameError: name 'pipe' is not defined | I'm trying to run the script bellow:
```
`from diffusers import LMSDiscreteScheduler
pipe.scheduler = LMSDiscreteScheduler.from_config(pipe.scheduler.config)
prompt = "a photo of an astronaut riding a horse on mars"
image = pipe(prompt).images[0]
image.save("astronaut_rides_horse.png")`
```
Ho... | closed | completed | false | 3 | [] | [] | 2022-11-24T00:32:48Z | 2022-11-24T05:54:21Z | 2022-11-24T05:54:21Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Limbicnation | 6,284,675 | MDQ6VXNlcjYyODQ2NzU= | User | false |
huggingface/diffusers | 1,462,719,226 | I_kwDOHa8MBc5XL1L6 | 1,387 | https://github.com/huggingface/diffusers/issues/1387 | https://api.github.com/repos/huggingface/diffusers/issues/1387 | Issue with tokenizer | ### Describe the bug
ValueError: Unable to create tensor, you should probably activate truncation and/or padding with 'padding=True' 'truncation=True' to have batched tensors with the same length. Perhaps your features (`input_ids` in this case) have excessive nesting (inputs type `list` where type `int` is expected... | closed | completed | false | 4 | [
"bug",
"stale"
] | [
"patil-suraj"
] | 2022-11-24T03:58:21Z | 2023-01-02T15:03:25Z | 2023-01-02T15:03:25Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | koutilya-pnvr | 24,998,956 | MDQ6VXNlcjI0OTk4OTU2 | User | false |
huggingface/diffusers | 1,462,720,077 | I_kwDOHa8MBc5XL1ZN | 1,388 | https://github.com/huggingface/diffusers/issues/1388 | https://api.github.com/repos/huggingface/diffusers/issues/1388 | Come on, come on, let's adapt the conversion script to SD 2.0 | **Is your feature request related to a problem? Please describe.**
It would be great if we could run SD 2 with cpu_offload, attention slicing, xformers, etc...
**Describe the solution you'd like**
Adapt the conversion script to SD 2.0
**Describe alternatives you've considered**
Stability AI's repo is not as fl... | closed | completed | false | 56 | [
"stale"
] | [] | 2022-11-24T04:00:00Z | 2023-01-15T15:03:25Z | 2023-01-15T15:03:25Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | piEsposito | 47,679,710 | MDQ6VXNlcjQ3Njc5NzEw | User | false |
huggingface/diffusers | 1,462,842,213 | I_kwDOHa8MBc5XMTNl | 1,389 | https://github.com/huggingface/diffusers/issues/1389 | https://api.github.com/repos/huggingface/diffusers/issues/1389 | Refactor StableDiffusion example scripts to use Optimum / BetterTransformer optimisations | **Describe the solution you'd like**
StableDiffusion example scripts are updated to make use of the Optimum / BetterTransformer optimisations.
**Describe alternatives you've considered**
That the examples can't be integrated, or that it won't provide a speedup/benefits.
**Additional context**
- https://hug... | closed | completed | false | 2 | [
"stale"
] | [] | 2022-11-24T06:43:54Z | 2023-01-02T15:03:23Z | 2023-01-02T15:03:23Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | 0xdevalias | 753,891 | MDQ6VXNlcjc1Mzg5MQ== | User | false |
huggingface/diffusers | 1,318,638,450 | I_kwDOHa8MBc5OmNNy | 139 | https://github.com/huggingface/diffusers/issues/139 | https://api.github.com/repos/huggingface/diffusers/issues/139 | Shouldn't PNDMScheduler have an add_noise() method, too? | Maybe a naive question, but I was following the tutorial notebooks and was testing to exchange schedulers and noticed that it was missing there but is present in both DDIM and DDPM. | closed | completed | false | 1 | [] | [] | 2022-07-26T18:59:00Z | 2022-07-27T15:38:53Z | 2022-07-27T15:38:52Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Quasimondo | 182,098 | MDQ6VXNlcjE4MjA5OA== | User | false |
huggingface/diffusers | 1,462,934,699 | I_kwDOHa8MBc5XMpyr | 1,391 | https://github.com/huggingface/diffusers/issues/1391 | https://api.github.com/repos/huggingface/diffusers/issues/1391 | Multiple prompts | ### Model/Pipeline/Scheduler description
Hi.
How to use:
from diffusers import FlaxStableDiffusionPipeline
pipeline, params = FlaxStableDiffusionPipeline.from_pretrained(
"runwayml/stable-diffusion-v1-5", revision="bf16", dtype=jax.numpy.bfloat16
)
to generate multiple prompts instead of single one (pr... | closed | completed | false | 8 | [] | [] | 2022-11-24T08:14:57Z | 2022-12-03T11:15:24Z | 2022-12-02T16:39:25Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | tralala87 | 104,907,171 | U_kgDOBkDBow | User | false |
huggingface/diffusers | 1,462,980,170 | I_kwDOHa8MBc5XM05K | 1,392 | https://github.com/huggingface/diffusers/issues/1392 | https://api.github.com/repos/huggingface/diffusers/issues/1392 | Stable Diffusion 2 | ### Model/Pipeline/Scheduler description
Perhaps you already have something in the works for this, but I couldn't see an existing issue for this.
Stable Diffusion 2 [just got released](https://stability.ai/blog/stable-diffusion-v2-release) on a separate [repo](https://github.com/Stability-AI/stablediffusion). The... | closed | completed | false | 14 | [] | [] | 2022-11-24T08:52:05Z | 2022-11-30T10:19:48Z | 2022-11-30T10:19:48Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | vvvm23 | 44,398,246 | MDQ6VXNlcjQ0Mzk4MjQ2 | User | false |
huggingface/diffusers | 1,463,535,101 | I_kwDOHa8MBc5XO8X9 | 1,398 | https://github.com/huggingface/diffusers/issues/1398 | https://api.github.com/repos/huggingface/diffusers/issues/1398 | Stable Diffusion 2 and Apple Silicon/MPS | ### Describe the bug
The diffusers code will run under Apple Silicon/MPS for Stable Diffusion 2.0 but the result is always noise. No images are generated.
Additionally, if you specify `pipe.enable_attention_slicing()`, then you get the following error — this is with the latest, dev source from the site and is not u... | closed | completed | false | 11 | [
"bug"
] | [] | 2022-11-24T15:49:06Z | 2022-11-30T23:26:53Z | 2022-11-25T13:18:16Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | FahimF | 181,110 | MDQ6VXNlcjE4MTExMA== | User | false |
huggingface/diffusers | 1,463,539,524 | I_kwDOHa8MBc5XO9dE | 1,399 | https://github.com/huggingface/diffusers/issues/1399 | https://api.github.com/repos/huggingface/diffusers/issues/1399 | VersatileDiffusion doesnnt work get_down_block() got an unexpected keyword argument 'use_linear_projection' | ### Describe the bug
VersatileDiffusion raise error on init
### Reproduction
Run this
https://colab.research.google.com/drive/1AnW_Sn_BNo6FpIbmiyqY3jPQib__V9bo?usp=sharing
### Logs
```shell
/usr/local/lib/python3.7/dist-packages/diffusers/pipeline_utils.py in from_pretrained(cls, pretrained_model_name_or_path, **... | closed | completed | false | 2 | [
"bug"
] | [] | 2022-11-24T15:53:01Z | 2022-11-24T22:41:33Z | 2022-11-24T22:41:33Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | hadaev8 | 20,247,085 | MDQ6VXNlcjIwMjQ3MDg1 | User | false |
huggingface/diffusers | 1,272,607,973 | I_kwDOHa8MBc5L2nTl | 14 | https://github.com/huggingface/diffusers/issues/14 | https://api.github.com/repos/huggingface/diffusers/issues/14 | Create Colab Notebook for training script | I have seen in the TO-DO list that you think to create a Colab to train the DM in examples folder.
I can do it! | closed | completed | false | 4 | [] | [] | 2022-06-15T18:47:37Z | 2022-07-21T19:18:37Z | 2022-07-21T19:02:58Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | mrm8488 | 3,653,789 | MDQ6VXNlcjM2NTM3ODk= | User | false |
huggingface/diffusers | 1,463,686,091 | I_kwDOHa8MBc5XPhPL | 1,403 | https://github.com/huggingface/diffusers/issues/1403 | https://api.github.com/repos/huggingface/diffusers/issues/1403 | StableDiffusionSafetyChecker's non-voided images are not the same as without safety checker | ### Describe the bug
For a given pipeline, the non-voided images returned by the safety checker are different from the images we get with the same pipeline without safety checker.
It seems to come from the initialization of the `CLIPVisionModel` inside the `__init__` of `StableDiffusionSafetyChecker`. More specif... | closed | completed | false | 3 | [
"bug"
] | [] | 2022-11-24T18:12:45Z | 2022-11-25T13:27:28Z | 2022-11-25T13:27:28Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | regisss | 15,324,346 | MDQ6VXNlcjE1MzI0MzQ2 | User | false |
huggingface/diffusers | 1,463,714,944 | I_kwDOHa8MBc5XPoSA | 1,404 | https://github.com/huggingface/diffusers/issues/1404 | https://api.github.com/repos/huggingface/diffusers/issues/1404 | CLIP Guided StableDiffusion V2 | ### Describe the bug
When trying to use the CLIP guided pipeline with Stable Diffusion V2, it errors out
### Reproduction
See above
### Logs
```shell
---------------------------------------------------------------------------
AttributeError Traceback (most recent call last)
<ipython-in... | closed | completed | false | 2 | [
"bug"
] | [] | 2022-11-24T18:54:13Z | 2022-11-25T14:12:06Z | 2022-11-25T14:12:05Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | dblunk88 | 39,381,389 | MDQ6VXNlcjM5MzgxMzg5 | User | false |
huggingface/diffusers | 1,463,725,717 | I_kwDOHa8MBc5XPq6V | 1,405 | https://github.com/huggingface/diffusers/issues/1405 | https://api.github.com/repos/huggingface/diffusers/issues/1405 | pipe.enable_attention_slicing() broken on main | ### Describe the bug
pipe.enable_attention_slicing() appears to be broken on main. Discovered when attempting to run SD 2.0 example code.
### Reproduction
Setup:
```
pip install --upgrade git+https://github.com/huggingface/diffusers.git transformers accelerate scipy ftfy
pip install -U --force-reinstall torch tor... | closed | completed | false | 4 | [
"bug"
] | [] | 2022-11-24T19:10:04Z | 2022-11-24T23:12:20Z | 2022-11-24T22:54:39Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | rabidgremlin | 444,906 | MDQ6VXNlcjQ0NDkwNg== | User | false |
huggingface/diffusers | 1,463,820,872 | I_kwDOHa8MBc5XQCJI | 1,407 | https://github.com/huggingface/diffusers/issues/1407 | https://api.github.com/repos/huggingface/diffusers/issues/1407 | 😱 PR #1395 causes ValueError expected 'feature_extractor' in Onnx Pipeline with SD 2.0 Models | ### Describe the bug
@patrickvonplaten
#1395 optional "feature_extractor" is not so optional. Issue when loading SD 2.0 Models as they contain:
```
"feature_extractor": [
null,
null
],
```
### Reproduction
Install [Diffusers ](https://github.com/huggingface/diffusers/tree/cbfed0c25606... | closed | completed | false | 2 | [
"bug"
] | [] | 2022-11-24T21:18:17Z | 2022-11-25T13:51:19Z | 2022-11-25T13:51:19Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | averad | 640,619 | MDQ6VXNlcjY0MDYxOQ== | User | false |
huggingface/diffusers | 1,463,861,089 | I_kwDOHa8MBc5XQL9h | 1,409 | https://github.com/huggingface/diffusers/issues/1409 | https://api.github.com/repos/huggingface/diffusers/issues/1409 | IndexError: list index out of range | Hello, I get the error:
`IndexError: list index out of range`
when I run inference.py
```
from diffusers import StableDiffusionPipeline
import torch
model_id = "/mnt/data-ml/ML-Images2/models/diffusers/pixel-art-characters"
pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float1... | closed | completed | false | 2 | [] | [] | 2022-11-24T22:28:14Z | 2022-11-25T00:49:40Z | 2022-11-24T22:54:18Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Limbicnation | 6,284,675 | MDQ6VXNlcjYyODQ2NzU= | User | false |
huggingface/diffusers | 1,319,867,084 | I_kwDOHa8MBc5Oq5LM | 141 | https://github.com/huggingface/diffusers/issues/141 | https://api.github.com/repos/huggingface/diffusers/issues/141 | Wishlist: example/documentation for fine-tuning or continuing training of a model | This is more of a newby issue, but I would find it helpful if there was an example showing how to continue training from an existing checkpoint and/or start training a new model from an existing pre-trained model. I guess it is straightforward by loading the state dict of the old model, but maybe there are certain tric... | closed | completed | false | 11 | [
"enhancement"
] | [
"patil-suraj"
] | 2022-07-27T17:24:15Z | 2024-11-21T11:21:25Z | 2024-11-21T11:21:25Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Quasimondo | 182,098 | MDQ6VXNlcjE4MjA5OA== | User | false |
huggingface/diffusers | 1,464,339,841 | I_kwDOHa8MBc5XSA2B | 1,413 | https://github.com/huggingface/diffusers/issues/1413 | https://api.github.com/repos/huggingface/diffusers/issues/1413 | Implement v-prediction in all schedulers | This would be nice to guarantee scheduler interoperability in Stable Diffusion 2. DDIM, DDPM and EulerDiscrete already support v-prediction, the following ones remain (from the list of compatible stable diffusion schedulers):
- [ ] EulerAncestralDiscreteScheduler
- [ ] PNDMScheduler
- [ ] LMSDiscreteScheduler
- [... | closed | completed | false | 4 | [] | [] | 2022-11-25T10:28:42Z | 2022-12-01T17:16:16Z | 2022-12-01T17:16:16Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | pcuenca | 1,177,582 | MDQ6VXNlcjExNzc1ODI= | User | false |
huggingface/diffusers | 1,464,468,724 | I_kwDOHa8MBc5XSgT0 | 1,417 | https://github.com/huggingface/diffusers/issues/1417 | https://api.github.com/repos/huggingface/diffusers/issues/1417 | CPU offloading currently done in a hacky way. | ### Describe the bug
It seems like `accelerate` doesn't yet support cpu offloading of modules that have single parameters such as:
https://github.com/huggingface/diffusers/blob/29021090614641d2509155ca0021497896228999/src/diffusers/pipelines/stable_diffusion/safety_checker.py#L44
which forces one to do the followi... | closed | completed | false | 3 | [
"bug",
"stale"
] | [] | 2022-11-25T12:16:42Z | 2023-01-22T21:08:29Z | 2023-01-22T15:03:10Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | patrickvonplaten | 23,423,619 | MDQ6VXNlcjIzNDIzNjE5 | User | false |
huggingface/diffusers | 1,464,654,135 | I_kwDOHa8MBc5XTNk3 | 1,422 | https://github.com/huggingface/diffusers/issues/1422 | https://api.github.com/repos/huggingface/diffusers/issues/1422 | When I use CLIPGuidedStableDiffusion based on stable-diffusion-2, the generated result is incredibly bad, all mosaic. | ### Describe the bug
When I use CLIPGuidedStableDiffusion based on stable-diffusion-2, the generated result is incredibly bad, all mosaic.
### Reproduction
from diffusers import DiffusionPipeline
fro... | closed | completed | false | 4 | [
"bug",
"stale"
] | [] | 2022-11-25T14:55:59Z | 2023-01-03T15:04:01Z | 2023-01-03T15:04:01Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | ScottishFold007 | 36,957,508 | MDQ6VXNlcjM2OTU3NTA4 | User | false |
huggingface/diffusers | 1,464,914,992 | I_kwDOHa8MBc5XUNQw | 1,427 | https://github.com/huggingface/diffusers/issues/1427 | https://api.github.com/repos/huggingface/diffusers/issues/1427 | FlaxStableDiffusionPipeline.from_pretrained() fails to load SD-2 | ### Describe the bug
I am unable to load SD-2 with FlaxStableDiffusionPipeline.from_pretrained(). FlaxStableDiffusionPipeline is unable to load SD-2 because FlaxStableDiffusionPipeline.from_pretrained() looking for an integer value for 'attention_head_dim', while the new SD-2 model config has a list of values for 'att... | closed | completed | false | 1 | [
"bug"
] | [] | 2022-11-25T20:38:05Z | 2022-11-29T19:58:13Z | 2022-11-29T19:58:13Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | lk-wq | 55,562,097 | MDQ6VXNlcjU1NTYyMDk3 | User | false |
huggingface/diffusers | 1,464,942,484 | I_kwDOHa8MBc5XUT-U | 1,429 | https://github.com/huggingface/diffusers/issues/1429 | https://api.github.com/repos/huggingface/diffusers/issues/1429 | Dreambooth example on SD2-768 model is producing weird results | ### Describe the bug
Hi,
This only happens on the 768 model, the base and 1.5/1.4 are not affected as far as I can tell.
Here are the samples produced while training, it does generate properly before training takes place.
(keep in mind the weird prompt is on purpose, it's unique tokens to see the changes quickl... | closed | completed | false | 19 | [
"bug"
] | [] | 2022-11-25T21:41:27Z | 2022-12-13T15:09:46Z | 2022-12-13T15:09:46Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | devilismyfriend | 87,043,616 | MDQ6VXNlcjg3MDQzNjE2 | User | false |
huggingface/diffusers | 1,320,585,914 | I_kwDOHa8MBc5Otoq6 | 143 | https://github.com/huggingface/diffusers/issues/143 | https://api.github.com/repos/huggingface/diffusers/issues/143 | Running difussers with GPU | Running the example codes i see that the CPU and not the GPU is used, is there a way to use GPU instead | closed | completed | false | 4 | [
"question"
] | [] | 2022-07-28T08:34:12Z | 2022-08-15T17:27:31Z | 2022-08-15T17:27:31Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | jfdelgad | 36,968,968 | MDQ6VXNlcjM2OTY4OTY4 | User | false |
huggingface/diffusers | 1,465,070,872 | I_kwDOHa8MBc5XUzUY | 1,433 | https://github.com/huggingface/diffusers/issues/1433 | https://api.github.com/repos/huggingface/diffusers/issues/1433 | AttributeError: 'OnnxRuntimeModel' object has no attribute 'config' | when i am trying to run stable-diffusion with onnx,there is a problem.it is like this:
Traceback (most recent call last):
File "\txt2img.py", line 8, in <module>
pipe = OnnxStableDiffusionPipeline.from_pretrained("./libs/stable_diffusion_onnx", provider="CPUExecutionProvider", safety_checker=None)
File "\ve... | closed | completed | false | 7 | [] | [] | 2022-11-26T06:59:48Z | 2023-06-08T07:35:56Z | 2022-11-28T00:43:08Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | john1089 | 76,211,484 | MDQ6VXNlcjc2MjExNDg0 | User | false |
huggingface/diffusers | 1,465,115,868 | I_kwDOHa8MBc5XU-Tc | 1,434 | https://github.com/huggingface/diffusers/issues/1434 | https://api.github.com/repos/huggingface/diffusers/issues/1434 | [OOM] Memory blows out when trying to upscale images larger than 128x128 using StableDiffusionUpscalePipeline | ### Describe the bug
When trying to upscale images larger than 128x128 the progress goes to 100% and then crashes with CUDA OOM.
With 512x512 images it's trying to allocate **256.00 GiB**!
### Reproduction
```
import requests
from PIL import Image
from io import BytesIO
from diffusers import StableDiffusionUp... | closed | completed | false | 17 | [
"bug",
"stale"
] | [] | 2022-11-26T10:46:05Z | 2023-02-08T23:38:35Z | 2023-02-05T15:03:22Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | qunash | 3,750,161 | MDQ6VXNlcjM3NTAxNjE= | User | false |
huggingface/diffusers | 1,465,309,297 | I_kwDOHa8MBc5XVthx | 1,438 | https://github.com/huggingface/diffusers/issues/1438 | https://api.github.com/repos/huggingface/diffusers/issues/1438 | training fails with torch.float16 dtype - weights are deleted in train loop | ### Describe the bug
i'm using the "textual inversion" training notebooks/scripts. could never get the TPU setup in Flax working, but i blame google for that one - they need to keep up their evil "yeah we got paying beta testers buahahaha" game
in the pytorch one, if using torch.float16 data type, on the second it... | closed | completed | false | 8 | [
"bug"
] | [
"pcuenca",
"patil-suraj"
] | 2022-11-27T02:21:58Z | 2022-12-30T14:19:48Z | 2022-12-30T14:19:48Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | krahnikblis | 84,637,076 | MDQ6VXNlcjg0NjM3MDc2 | User | false |
huggingface/diffusers | 1,465,330,398 | I_kwDOHa8MBc5XVyre | 1,439 | https://github.com/huggingface/diffusers/issues/1439 | https://api.github.com/repos/huggingface/diffusers/issues/1439 | when using diffusers and transformer: module transformers has no attribute FeatureExtractionMixin | ### Describe the bug
I'm trying to use diffusers alongside another package, CLIP Interrogator. When I install the dependencies for CLIP Interrogator my diffuser code goes kerplooie! and throws this error: module transformers has no attribute FeatureExtractionMixin
### Reproduction
a notebook with the code to rep... | closed | completed | false | 7 | [
"bug"
] | [] | 2022-11-27T04:43:14Z | 2022-12-03T01:57:11Z | 2022-11-30T16:07:42Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | TheOneTrueGuy | 3,240,071 | MDQ6VXNlcjMyNDAwNzE= | User | false |
huggingface/diffusers | 1,321,045,168 | I_kwDOHa8MBc5OvYyw | 144 | https://github.com/huggingface/diffusers/issues/144 | https://api.github.com/repos/huggingface/diffusers/issues/144 | TensorFlow and Flax/Jax | Will this library be similar to transformers and include TensorFlow and Flax implementations at some point? | closed | completed | false | 2 | [
"enhancement"
] | [] | 2022-07-28T14:43:29Z | 2022-09-13T15:36:24Z | 2022-09-13T15:36:24Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | will-rice | 25,072,137 | MDQ6VXNlcjI1MDcyMTM3 | User | false |
huggingface/diffusers | 1,465,387,644 | I_kwDOHa8MBc5XWAp8 | 1,440 | https://github.com/huggingface/diffusers/issues/1440 | https://api.github.com/repos/huggingface/diffusers/issues/1440 | Delete Cached Model | ### Describe the bug
Hi have been using diffusers, it works fine. after 10-20 model loading using
`from_pretrained("CompVis/stable-diffusion-v1-4"`
`from_pretrained("hakurei/waifu-diffusion"`
.
and tens of more,
what i found is saved model cache is building up, how can i remove a specific pre-trained mode... | closed | completed | false | 1 | [
"bug"
] | [] | 2022-11-27T09:53:50Z | 2022-11-30T01:42:56Z | 2022-11-30T01:42:56Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | adhikjoshi | 11,740,719 | MDQ6VXNlcjExNzQwNzE5 | User | false |
huggingface/diffusers | 1,465,565,218 | I_kwDOHa8MBc5XWsAi | 1,443 | https://github.com/huggingface/diffusers/issues/1443 | https://api.github.com/repos/huggingface/diffusers/issues/1443 | Let prediction type be a property on unet instead of the sampler | **What API design would you like to have changed or added to the library? Why?**
I think the prediction type is a property of the unet since it describes what it is outputting. It would be nice if it was part of the unet config rather than the sampler. I am loading the different parts of stable diffusion v2 separate... | closed | completed | false | 5 | [
"stale"
] | [
"pcuenca",
"anton-l",
"patil-suraj"
] | 2022-11-27T20:21:34Z | 2023-01-05T15:03:31Z | 2023-01-05T15:03:31Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | samedii | 1,872,050 | MDQ6VXNlcjE4NzIwNTA= | User | false |
huggingface/diffusers | 1,465,706,875 | I_kwDOHa8MBc5XXOl7 | 1,444 | https://github.com/huggingface/diffusers/issues/1444 | https://api.github.com/repos/huggingface/diffusers/issues/1444 | ImportError: cannot import name 'StableDiffusionSafetyChecker' | ### Describe the bug
`ImportError: cannot import name 'StableDiffusionSafetyChecker' from 'diffusers.pipelines.stable_diffusion' (/home/tiger/anaconda3/lib/python3.9/site-packages/diffusers/pipelines/stable_diffusion/__init__.py)`
### Reproduction
_No response_
### Logs
_No response_
### System Info
diffusers=... | closed | completed | false | 2 | [
"bug",
"stale"
] | [] | 2022-11-28T02:31:47Z | 2023-01-05T15:03:30Z | 2023-01-05T15:03:30Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | liu6381810 | 24,307,321 | MDQ6VXNlcjI0MzA3MzIx | User | false |
huggingface/diffusers | 1,465,901,717 | I_kwDOHa8MBc5XX-KV | 1,445 | https://github.com/huggingface/diffusers/issues/1445 | https://api.github.com/repos/huggingface/diffusers/issues/1445 | 🐛 AttributeError in OnnxStableDiffusionInpaintPipelineLegacy in 🤗 Diffusers 0.9.0 | ### Describe the bug 🐛
ℹ️ AttributeError in OnnxStableDiffusionInpaintPipelineLegacy in 🤗 Diffusers 0.9.0
⚠️ Error: 'OnnxRuntimeModel' object has no attribute 'config'
```
is_unet_version_less_0_9_0 = hasattr(unet.config, "_diffusers_version") and version.parse(
AttributeError: 'OnnxRuntimeModel' object has n... | closed | completed | false | 2 | [
"bug"
] | [
"anton-l"
] | 2022-11-28T06:47:03Z | 2022-11-28T13:18:15Z | 2022-11-28T13:18:15Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | averad | 640,619 | MDQ6VXNlcjY0MDYxOQ== | User | false |
huggingface/diffusers | 1,466,225,637 | I_kwDOHa8MBc5XZNPl | 1,447 | https://github.com/huggingface/diffusers/issues/1447 | https://api.github.com/repos/huggingface/diffusers/issues/1447 | [Announcement] Stable Diffusion v1-4/v1-5 have lifted gated access! => No need to be logged into hf.co anymore! | https://huggingface.co/CompVis/stable-diffusion-v1-4
https://huggingface.co/CompVis/stable-diffusion-v1-3
https://huggingface.co/CompVis/stable-diffusion-v1-2
https://huggingface.co/CompVis/stable-diffusion-v1-1
are all no longer gated!
This means you don't have to be logged in to download the model. Just doin... | closed | completed | false | 3 | [
"stale"
] | [] | 2022-11-28T11:05:59Z | 2023-03-10T13:09:26Z | 2023-01-15T15:03:22Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | patrickvonplaten | 23,423,619 | MDQ6VXNlcjIzNDIzNjE5 | User | false |
huggingface/diffusers | 1,321,803,745 | I_kwDOHa8MBc5OyR_h | 145 | https://github.com/huggingface/diffusers/issues/145 | https://api.github.com/repos/huggingface/diffusers/issues/145 | Kindly simplify training our own dataset | Hello,
Please try to simplify training dataset of our own. It is taking lot of Time and it is difficult to train using current setup.
| closed | completed | false | 4 | [] | [
"anton-l"
] | 2022-07-29T06:09:18Z | 2022-09-13T15:37:55Z | 2022-09-13T15:37:55Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | NeethanShetty | 104,005,827 | U_kgDOBjMAww | User | false |
huggingface/diffusers | 1,466,334,770 | I_kwDOHa8MBc5XZn4y | 1,451 | https://github.com/huggingface/diffusers/issues/1451 | https://api.github.com/repos/huggingface/diffusers/issues/1451 | Not able to run FlaxStableDiffusionPipeline example of CompVis/stable-diffusion-v1-4 | ### Describe the bug
I'm not able to run the Flax pipeline example of [CompVis/stable-diffusion-v1-4](https://huggingface.co/CompVis/stable-diffusion-v1-4) on Colab with GPU enabled. I get this error
```
ValueError: pmap got inconsistent sizes for array axes to be mapped:
* most axes (1530 of them) had size 1, e.... | closed | completed | false | 2 | [
"bug"
] | [
"pcuenca"
] | 2022-11-28T12:23:13Z | 2022-11-29T19:47:19Z | 2022-11-29T19:47:19Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | dzlab | 1,645,304 | MDQ6VXNlcjE2NDUzMDQ= | User | false |
huggingface/diffusers | 1,466,335,122 | I_kwDOHa8MBc5XZn-S | 1,452 | https://github.com/huggingface/diffusers/issues/1452 | https://api.github.com/repos/huggingface/diffusers/issues/1452 | [ONNX] PyTorch IO binding for faster GPU inference | **Context**
Currently [`OnnxStableDiffusionPipeline`](https://github.com/huggingface/diffusers/blob/main/src/diffusers/pipelines/stable_diffusion/pipeline_onnx_stable_diffusion.py#LL34C7-L34C34) perform unnecessary tensor casting between torch and numpy. The downsides of that are:
* The pipeline code is harder to m... | closed | completed | false | 3 | [
"enhancement",
"help wanted"
] | [] | 2022-11-28T12:23:27Z | 2024-11-21T15:27:18Z | 2024-11-21T15:27:18Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | anton-l | 26,864,830 | MDQ6VXNlcjI2ODY0ODMw | User | false |
huggingface/diffusers | 1,466,751,597 | I_kwDOHa8MBc5XbNpt | 1,456 | https://github.com/huggingface/diffusers/issues/1456 | https://api.github.com/repos/huggingface/diffusers/issues/1456 | Demo test -- ignore | Create a demo notebook | closed | not_planned | false | 0 | [] | [] | 2022-11-28T16:53:37Z | 2022-11-28T16:56:01Z | 2022-11-28T16:54:16Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | brosand | 15,368,605 | MDQ6VXNlcjE1MzY4NjA1 | User | false |
huggingface/diffusers | 1,321,836,472 | I_kwDOHa8MBc5OyZ-4 | 146 | https://github.com/huggingface/diffusers/issues/146 | https://api.github.com/repos/huggingface/diffusers/issues/146 | Wishlist: example for training a superresolution pipeline | null | closed | completed | false | 13 | [
"enhancement"
] | [] | 2022-07-29T06:52:59Z | 2022-11-09T16:47:15Z | 2022-11-09T12:42:18Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Quasimondo | 182,098 | MDQ6VXNlcjE4MjA5OA== | User | false |
huggingface/diffusers | 1,466,994,172 | I_kwDOHa8MBc5XcI38 | 1,460 | https://github.com/huggingface/diffusers/issues/1460 | https://api.github.com/repos/huggingface/diffusers/issues/1460 | Autoencoder scale values in config | **What API design would you like to have changed or added to the library? Why?**
Would be nice to have some way of getting the appropriate scale value for the VAE like saving it in the config. It is now hardcoded in the pipelines.
I am referring to these two:
```
1 / 0.08333
1 / 0.18215
```
**What use case... | closed | completed | false | 10 | [
"stale"
] | [
"williamberman",
"patrickvonplaten"
] | 2022-11-28T19:45:45Z | 2023-01-04T00:51:09Z | 2023-01-04T00:51:08Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | samedii | 1,872,050 | MDQ6VXNlcjE4NzIwNTA= | User | false |
huggingface/diffusers | 1,467,267,726 | I_kwDOHa8MBc5XdLqO | 1,462 | https://github.com/huggingface/diffusers/issues/1462 | https://api.github.com/repos/huggingface/diffusers/issues/1462 | StableDiffusionInpaintPipeline - unassigned variable bug when inferencing with multiple images and masks | ### Describe the bug
The `StableDiffusionInpaintPipeline` seems to support `torch.Tensor` as inputs. However, the code contains a bug that made the inference fail.
The bug is located at line 654 and line 675 in [pipeline_stable_diffusion_inpaint.py](https://github.com/huggingface/diffusers/blob/v0.9.0/src/diffus... | closed | completed | false | 4 | [
"bug"
] | [] | 2022-11-29T00:24:26Z | 2022-12-05T11:18:03Z | 2022-12-05T11:18:03Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | zxydi1992 | 9,076,986 | MDQ6VXNlcjkwNzY5ODY= | User | false |
huggingface/diffusers | 1,467,441,984 | I_kwDOHa8MBc5Xd2NA | 1,463 | https://github.com/huggingface/diffusers/issues/1463 | https://api.github.com/repos/huggingface/diffusers/issues/1463 | from datasets import load_dataset ModuleNotFoundError: No module named 'datasets' | Traceback (most recent call last):
File "train_text_to_image.py", line 17, in <module>
from datasets import load_dataset
ModuleNotFoundError: No module named 'datasets'
Traceback (most recent call last):
File "/usr/local/bin/accelerate", line 8, in <module>
sys.exit(main())
File "/usr/local/lib/pyt... | closed | completed | false | 3 | [
"stale"
] | [] | 2022-11-29T04:41:53Z | 2023-01-07T15:02:49Z | 2023-01-07T15:02:49Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | krishnaadithya | 13,948,265 | MDQ6VXNlcjEzOTQ4MjY1 | User | false |
huggingface/diffusers | 1,467,560,764 | I_kwDOHa8MBc5XeTM8 | 1,464 | https://github.com/huggingface/diffusers/issues/1464 | https://api.github.com/repos/huggingface/diffusers/issues/1464 | Is it possible to use VersatileDiffusion for Seq2Seq generation training? | Hi,
Thanks for the amazing project for diffusion models! I'm interested in Seq2Seq generation with diffusion models. Thus is it possible to use VersatileDiffusion model for Seq2Seq generation training? Or are you planning to add text-to-text generation pipeline in Diffusers?
Thanks! | closed | completed | false | 2 | [
"stale"
] | [] | 2022-11-29T07:06:36Z | 2023-01-07T15:02:48Z | 2023-01-07T15:02:48Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | HZQ950419 | 37,740,571 | MDQ6VXNlcjM3NzQwNTcx | User | false |
huggingface/diffusers | 1,467,745,408 | I_kwDOHa8MBc5XfASA | 1,466 | https://github.com/huggingface/diffusers/issues/1466 | https://api.github.com/repos/huggingface/diffusers/issues/1466 | Dreambooth example is not compatible with Windows or MacOS | ### Describe the bug
Dreambooth example uses multiprocess DataLoader (`num_workers=1`) and it causes error when pickling `collate_fn` because it is not top-level function.
#### Workaround
Training works if I disable multiprocessing by removing `num_workers` from call to `torch.utils.data.DataLoader`
### Repro... | closed | completed | false | 4 | [
"bug"
] | [
"patil-suraj"
] | 2022-11-29T09:39:16Z | 2023-01-05T05:37:43Z | 2022-12-06T13:41:55Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Dragollla | 15,144,728 | MDQ6VXNlcjE1MTQ0NzI4 | User | false |
huggingface/diffusers | 1,323,388,273 | I_kwDOHa8MBc5O4U1x | 147 | https://github.com/huggingface/diffusers/issues/147 | https://api.github.com/repos/huggingface/diffusers/issues/147 | Image inpainting | Hi,
2 quick questions around this:
- Is there any colab / guiding doc around leveraging this model for image inpainting?
- Given a source person + t-shirt image, how can i use a guided text prompt (i.e. "show person wearing this t-shirt") to generate an image of the same? | closed | completed | false | 8 | [
"enhancement"
] | [
"anton-l"
] | 2022-07-31T07:45:48Z | 2024-11-19T09:15:16Z | 2024-11-19T09:15:16Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | ghost | 10,137 | MDQ6VXNlcjEwMTM3 | User | false |
huggingface/diffusers | 1,468,259,942 | I_kwDOHa8MBc5Xg95m | 1,473 | https://github.com/huggingface/diffusers/issues/1473 | https://api.github.com/repos/huggingface/diffusers/issues/1473 | Can some of you help me look at my code? | ### Describe the bug
clip guided generates a very poor result，
![image](https://user-image... | closed | completed | false | 4 | [
"bug",
"stale"
] | [] | 2022-11-29T15:28:39Z | 2023-01-08T15:03:54Z | 2023-01-08T15:03:54Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | ScottishFold007 | 36,957,508 | MDQ6VXNlcjM2OTU3NTA4 | User | false |
huggingface/diffusers | 1,468,547,686 | I_kwDOHa8MBc5XiEJm | 1,475 | https://github.com/huggingface/diffusers/issues/1475 | https://api.github.com/repos/huggingface/diffusers/issues/1475 | training sd with dreambooth/textual inversion with two different class of images | ### Describe the bug
Hi,
I am following stable diffusion training with the dreambooth example at local machine. Although the following link gives every details about training sd, however I am unable to identify utilizing this example when I have two or more different classes of images.:
https://github.com/huggingf... | closed | completed | false | 15 | [
"bug",
"stale"
] | [
"williamberman",
"patil-suraj"
] | 2022-11-29T18:53:18Z | 2023-04-11T18:08:40Z | 2023-02-27T15:04:15Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | hamzafar | 6,842,017 | MDQ6VXNlcjY4NDIwMTc= | User | false |
huggingface/diffusers | 1,468,745,770 | I_kwDOHa8MBc5Xi0gq | 1,477 | https://github.com/huggingface/diffusers/issues/1477 | https://api.github.com/repos/huggingface/diffusers/issues/1477 | RuntimeError: CUDA error: CUBLAS_STATUS_INVALID_VALUE when calling `cublasGemmStridedBatchedExFix...` | ### Describe the bug
The error was thrown when executing the following codes on SD V2 inpainting . It works fine on SD v1. My cuda and torch version are listed below.
```
RuntimeError: CUDA error: CUBLAS_STATUS_INVALID_VALUE when calling `cublasGemmStridedBatchedExFix( handle, opa, opb, m, n, k, (void*)(&falp... | closed | completed | false | 4 | [
"bug",
"stale"
] | [] | 2022-11-29T21:46:28Z | 2023-04-25T09:03:25Z | 2023-01-24T15:03:44Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | alfredcs | 3,374,971 | MDQ6VXNlcjMzNzQ5NzE= | User | false |
huggingface/diffusers | 1,468,763,867 | I_kwDOHa8MBc5Xi47b | 1,478 | https://github.com/huggingface/diffusers/issues/1478 | https://api.github.com/repos/huggingface/diffusers/issues/1478 | StableDiffusionUpscalePipeline cannot handle more than 1 image output | ### Describe the bug
Using the `StableDiffusionUpscalePipeline` when called to output more than one image will crash. This occurs when using the `num_images_per_prompt` parameter or providing lists with multiple elements for the `prompt` and `image` inputs. Looks like some part of the logic for preparing the image and... | closed | completed | false | 4 | [
"bug"
] | [] | 2022-11-29T22:06:03Z | 2023-03-23T00:04:04Z | 2022-12-05T12:28:56Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | zookae | 1,692,386 | MDQ6VXNlcjE2OTIzODY= | User | false |
huggingface/diffusers | 1,468,972,464 | I_kwDOHa8MBc5Xjr2w | 1,479 | https://github.com/huggingface/diffusers/issues/1479 | https://api.github.com/repos/huggingface/diffusers/issues/1479 | DummyObjects are impeding PyCharm | I'm subclassing StableDiffusionPipeline (because it seems like that's the intended way to make a DiffusionPipeline that is still able to take advantage of StableDiffusionPipeline's methods to enable attention slicing, decode latents, etc.), and I've noticed a distinct lack of help from the IDE when it comes to attribut... | closed | completed | false | 8 | [] | [
"anton-l"
] | 2022-11-30T02:56:00Z | 2022-12-08T13:42:27Z | 2022-12-08T13:02:13Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | keturn | 83,819 | MDQ6VXNlcjgzODE5 | User | false |
huggingface/diffusers | 1,323,706,466 | I_kwDOHa8MBc5O5ihi | 148 | https://github.com/huggingface/diffusers/issues/148 | https://api.github.com/repos/huggingface/diffusers/issues/148 | what's the paper of conditional unet ? | Hello.
Can you tell me the paper name in [condional_unet architecture](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/unet_2d_condition.py) ?
I wonder what paper this model structure is based on.
What paper is this model used in? | closed | completed | false | 2 | [] | [] | 2022-08-01T02:20:02Z | 2022-08-04T02:02:52Z | 2022-08-04T02:02:52Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | taki0112 | 26,713,992 | MDQ6VXNlcjI2NzEzOTky | User | false |
huggingface/diffusers | 1,468,988,068 | I_kwDOHa8MBc5Xjvqk | 1,480 | https://github.com/huggingface/diffusers/issues/1480 | https://api.github.com/repos/huggingface/diffusers/issues/1480 | `deprecate` should point to the usage of the deprecated thing | This warning message:
> /opt/hostedtoolcache/Python/3.10.8/x64/lib/python3.10/site-packages/diffusers/utils/deprecation_utils.py:35: FutureWarning: It is deprecated to pass a pretrained model name or path to `from_config`.If you were trying to load a scheduler, please use <class 'diffusers.schedulers.scheduling_lms_di... | closed | completed | false | 1 | [] | [] | 2022-11-30T03:19:49Z | 2022-12-01T15:28:12Z | 2022-12-01T15:28:12Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | keturn | 83,819 | MDQ6VXNlcjgzODE5 | User | false |
huggingface/diffusers | 1,469,219,240 | I_kwDOHa8MBc5XkoGo | 1,481 | https://github.com/huggingface/diffusers/issues/1481 | https://api.github.com/repos/huggingface/diffusers/issues/1481 | EulerDiscreteScheduler is not complete and sign of c_out is wrong | ### Describe the bug
Hi,
the implementation of EDM method here https://github.com/huggingface/diffusers/blob/main/src/diffusers/schedulers/scheduling_euler_discrete.py misses the second-order correction step implemented here in the original paper https://arxiv.org/pdf/2206.00364.pdf:
please see https://github.com/... | closed | completed | false | 5 | [
"bug",
"stale"
] | [] | 2022-11-30T08:10:55Z | 2023-01-11T15:05:13Z | 2023-01-11T15:05:13Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | cameronalonso2 | 73,615,224 | MDQ6VXNlcjczNjE1MjI0 | User | false |
huggingface/diffusers | 1,469,329,699 | I_kwDOHa8MBc5XlDEj | 1,482 | https://github.com/huggingface/diffusers/issues/1482 | https://api.github.com/repos/huggingface/diffusers/issues/1482 | Diffusers Stable Diffusion upscale pipeline | ### Describe the bug
I get this message when I try to upscale a image
(/tmp/pip-install-r2dsw42o/xformers_584b8336664f4e63bb32e09d7dcaeac6/third_party/flash-attention/csrc/flash_attn/src/fmha_fprop_fp16_kernel.sm80.cu:68): invalid argument
I can't find the culprit unfortunately, thank you!!
### Reproduction
_... | closed | completed | false | 3 | [
"bug"
] | [] | 2022-11-30T09:37:43Z | 2022-12-01T14:47:49Z | 2022-12-01T14:47:49Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | UglyStupidHonest | 68,284,461 | MDQ6VXNlcjY4Mjg0NDYx | User | false |
huggingface/diffusers | 1,469,448,229 | I_kwDOHa8MBc5XlgAl | 1,484 | https://github.com/huggingface/diffusers/issues/1484 | https://api.github.com/repos/huggingface/diffusers/issues/1484 | fp16 does not work with euler_a scheduler | ### Describe the bug
fp16 does not work with euler_a scheduler: it displays nearly black image.
using checkpoint: stable-diffusion-v1.5
### Reproduction
- Input prompt: a boy.
- euler_a scheduler: default parameters.
- load stable-diffusion-v1.5 with fp16
### Logs
_No response_
### System Info
Ubuntu 18, 2... | closed | completed | false | 5 | [
"bug",
"stale"
] | [] | 2022-11-30T11:04:58Z | 2023-01-08T15:03:52Z | 2023-01-08T15:03:52Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | KingStorm | 8,191,470 | MDQ6VXNlcjgxOTE0NzA= | User | false |
huggingface/diffusers | 1,469,498,983 | I_kwDOHa8MBc5XlsZn | 1,485 | https://github.com/huggingface/diffusers/issues/1485 | https://api.github.com/repos/huggingface/diffusers/issues/1485 | The original model and the diffusers model do not give the same results |
When the original model and the diffusers model are run separately img2img produces different results with the same seed
| closed | completed | false | 10 | [
"stale"
] | [] | 2022-11-30T11:43:40Z | 2023-01-12T15:04:35Z | 2023-01-12T15:04:35Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | dingjingzhen | 21,244,263 | MDQ6VXNlcjIxMjQ0MjYz | User | false |
huggingface/diffusers | 1,469,552,796 | I_kwDOHa8MBc5Xl5ic | 1,486 | https://github.com/huggingface/diffusers/issues/1486 | https://api.github.com/repos/huggingface/diffusers/issues/1486 | Possible Bug in DDIMScheduler when prediction_type = 'sample' | ### Describe the bug
https://github.com/huggingface/diffusers/blob/v0.9.0/src/diffusers/schedulers/scheduling_ddim.py#L303
https://github.com/huggingface/diffusers/blob/v0.9.0/src/diffusers/schedulers/scheduling_ddim.py#L278
It seems that the ``model_output'' at L303 should be the predicted epsilon. However, it ... | closed | completed | false | 4 | [
"bug",
"stale"
] | [] | 2022-11-30T12:24:08Z | 2023-02-05T15:03:20Z | 2023-02-05T15:03:20Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | JianxinMa | 7,543,016 | MDQ6VXNlcjc1NDMwMTY= | User | false |
huggingface/diffusers | 1,323,987,304 | I_kwDOHa8MBc5O6nFo | 149 | https://github.com/huggingface/diffusers/issues/149 | https://api.github.com/repos/huggingface/diffusers/issues/149 | Functionality of `use_conv` | Hello
[L312](https://github.com/huggingface/diffusers/blob/cde0ed162a127b17f1b4d4b16ff7f736cf04e690/src/diffusers/models/resnet.py#L312) in `use_conv=False` is it correct ?
if correct, then this line is same to [`fir=='sde_vp'`](https://github.com/huggingface/diffusers/blob/cde0ed162a127b17f1b4d4b16ff7f736cf04e69... | closed | completed | false | 2 | [] | [] | 2022-08-01T07:53:44Z | 2022-09-13T15:41:11Z | 2022-09-13T15:41:11Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | taki0112 | 26,713,992 | MDQ6VXNlcjI2NzEzOTky | User | false |
huggingface/diffusers | 1,469,768,960 | I_kwDOHa8MBc5XmuUA | 1,491 | https://github.com/huggingface/diffusers/issues/1491 | https://api.github.com/repos/huggingface/diffusers/issues/1491 | question on set_timesteps | Thank you for releasing a wonderful repository! Let me ask a newbie question.
**What API design would you like to have changed or added to the library? Why?**
`diffusers.DDPMScheduler.set_timesteps`
**What use case would this enable or better enable? Can you give us a code example?**
I guess `diffusers.DDPMSche... | closed | completed | false | 2 | [
"stale"
] | [] | 2022-11-30T14:53:41Z | 2023-01-08T15:03:50Z | 2023-01-08T15:03:50Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | naoto0804 | 7,067,992 | MDQ6VXNlcjcwNjc5OTI= | User | false |
huggingface/diffusers | 1,470,068,250 | I_kwDOHa8MBc5Xn3Ya | 1,495 | https://github.com/huggingface/diffusers/issues/1495 | https://api.github.com/repos/huggingface/diffusers/issues/1495 | Cannot import name 'init_git_repo' from 'diffusers.hub_utils' | ### Describe the bug
To be honest I'm not sure if this is a bug, but since it started happening all of a sudden in the last few hours I'm going to assume it is.
I'm using a Colab setup with Dreambooth to fine tune Stable Diffusion 1.4, the error happens when executing the cell that tries to import the 'init_git_repo'... | closed | completed | false | 5 | [
"bug",
"stale"
] | [] | 2022-11-30T18:40:29Z | 2023-01-10T10:08:13Z | 2023-01-08T15:03:49Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | FogoVoar | 112,202,663 | U_kgDOBrATpw | User | false |
huggingface/diffusers | 1,470,285,615 | I_kwDOHa8MBc5Xoscv | 1,497 | https://github.com/huggingface/diffusers/issues/1497 | https://api.github.com/repos/huggingface/diffusers/issues/1497 | None | None | closed | not_planned | false | 0 | [] | [] | 2022-11-30T22:12:31Z | 2022-11-30T22:14:18Z | 2022-11-30T22:14:18Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | zaqqwerty | 12,915,808 | MDQ6VXNlcjEyOTE1ODA4 | User | false |
huggingface/diffusers | 1,470,308,038 | I_kwDOHa8MBc5Xox7G | 1,498 | https://github.com/huggingface/diffusers/issues/1498 | https://api.github.com/repos/huggingface/diffusers/issues/1498 | code-level documentation | hi! while wrenching around in the internals, there are lots of places that make me scratch my head. often this kind of thing is newb ignorance, but i find (generally speaking, in my experience reading code) just as often that there are leftover artifacts from some early proof of concept, which just get wrapped up in th... | closed | completed | false | 2 | [
"stale"
] | [] | 2022-11-30T22:33:35Z | 2023-01-08T15:03:48Z | 2023-01-08T15:03:48Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | krahnikblis | 84,637,076 | MDQ6VXNlcjg0NjM3MDc2 | User | false |
huggingface/diffusers | 1,326,285,001 | I_kwDOHa8MBc5PDYDJ | 150 | https://github.com/huggingface/diffusers/issues/150 | https://api.github.com/repos/huggingface/diffusers/issues/150 | Reading model from a local file | Greetings.
Consider I have a `model.ckpt` file, and it's trained and saved on my local computer. It's not hosted on huggingface and let's consider it's a little bit confidential and cannot be hosted anywhere. How can I feed it into `from_pretrained()` function? | closed | completed | false | 4 | [] | [] | 2022-08-02T19:08:13Z | 2022-09-13T15:42:07Z | 2022-09-13T15:42:06Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | prp-e | 1,078,404 | MDQ6VXNlcjEwNzg0MDQ= | User | false |
huggingface/diffusers | 1,470,570,262 | I_kwDOHa8MBc5Xpx8W | 1,500 | https://github.com/huggingface/diffusers/issues/1500 | https://api.github.com/repos/huggingface/diffusers/issues/1500 | [Community Pipelines] lpw_stable_diffusion.py incompatible with Diffusers v0.10.0 | ### Describe the bug
When trying to load Stable Diffusion v2.0 with Diffusers v0.10.0 you are greeted with error below
```
ValueError: Pipeline <class 'diffusers_modules.git.lpw_stable_diffusion.StableDiffusionLongPromptWeightingPipeline'> expected {'tokenizer', 'text_encoder', 'unet', 'vae', 'safety_checker', '... | closed | completed | false | 8 | [
"bug"
] | [] | 2022-12-01T02:40:36Z | 2022-12-12T15:07:35Z | 2022-12-12T15:07:35Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | WASasquatch | 1,151,589 | MDQ6VXNlcjExNTE1ODk= | User | false |
huggingface/diffusers | 1,470,704,363 | I_kwDOHa8MBc5XqSrr | 1,501 | https://github.com/huggingface/diffusers/issues/1501 | https://api.github.com/repos/huggingface/diffusers/issues/1501 | why 2.55 GB ckpt becomes 3.44 GB unet | is it possible to reduce size? | closed | completed | false | 19 | [] | [] | 2022-12-01T05:27:12Z | 2022-12-22T23:36:48Z | 2022-12-22T23:36:48Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | camenduru | 54,370,274 | MDQ6VXNlcjU0MzcwMjc0 | User | false |
huggingface/diffusers | 1,470,964,771 | I_kwDOHa8MBc5XrSQj | 1,502 | https://github.com/huggingface/diffusers/issues/1502 | https://api.github.com/repos/huggingface/diffusers/issues/1502 | [Discussion]Any practice for deploying model inferences provided by diffusers into a web service? | I'm trying deploying model inferences provided by diffusers into a web service.
I'm planning to use [fastapi](https://github.com/tiangolo/fastapi) or [BentoML](https://github.com/bentoml/BentoML) to build, but haven't found anything that works with diffusers directly.
I would greatly appreciate if you have good advic... | closed | completed | false | 2 | [
"stale"
] | [] | 2022-12-01T09:20:41Z | 2023-01-08T15:03:47Z | 2023-01-08T15:03:47Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | VXenomac | 21,958,044 | MDQ6VXNlcjIxOTU4MDQ0 | User | false |
huggingface/diffusers | 1,471,506,165 | I_kwDOHa8MBc5XtWb1 | 1,506 | https://github.com/huggingface/diffusers/issues/1506 | https://api.github.com/repos/huggingface/diffusers/issues/1506 | Weighted Prompts for Diffusers stable diffusion pipeline | I could not find anything for diffusers and unfortunately I'm not on the Level yet where I can implement it myself. :)
It would be amazing to be able to weight prompts like "a dog with a hat:0.5"
Thank you for this amazing library !!
| closed | completed | false | 36 | [
"stale"
] | [] | 2022-12-01T15:22:14Z | 2023-03-22T15:03:49Z | 2023-03-22T15:03:49Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | UglyStupidHonest | 68,284,461 | MDQ6VXNlcjY4Mjg0NDYx | User | false |
huggingface/diffusers | 1,326,405,143 | I_kwDOHa8MBc5PD1YX | 151 | https://github.com/huggingface/diffusers/issues/151 | https://api.github.com/repos/huggingface/diffusers/issues/151 | Asychronous multiple GPU multiple model training with shared replay buffer | Let's say we have multiple diffusion models, as in the Cascading diffusion model paper.
Is there an easy way to setup training such that each conditional model is trained simultaneously on different GPUs? What about with a shared replay buffer, that each conditional model can access? | closed | completed | false | 2 | [] | [] | 2022-08-02T21:12:12Z | 2022-09-13T15:43:08Z | 2022-09-13T15:43:08Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | richardrl | 5,181,990 | MDQ6VXNlcjUxODE5OTA= | User | false |
huggingface/diffusers | 1,471,915,754 | I_kwDOHa8MBc5Xu6bq | 1,514 | https://github.com/huggingface/diffusers/issues/1514 | https://api.github.com/repos/huggingface/diffusers/issues/1514 | Use consistent random number generation across hardware | **Is your feature request related to a problem? Please describe.**
`pytorch.randn` is not consistent across hardware devices (See https://github.com/pytorch/pytorch/issues/84234).
`diffusers` calls `torch.randn` on the device computation is run on (typically 'cuda'). As a result, results produced with the exact s... | closed | completed | false | 13 | [
"stale"
] | [
"pcuenca",
"williamberman",
"anton-l",
"patil-suraj"
] | 2022-12-01T20:26:08Z | 2023-03-05T20:51:47Z | 2023-03-05T20:51:47Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | antoche | 1,627,384 | MDQ6VXNlcjE2MjczODQ= | User | false |
huggingface/diffusers | 1,472,179,650 | I_kwDOHa8MBc5Xv63C | 1,517 | https://github.com/huggingface/diffusers/issues/1517 | https://api.github.com/repos/huggingface/diffusers/issues/1517 | Dynamic modules fail when using a cache shared across users | ### Describe the bug
If multiple users attempt to share the same models cache (e.g., on a company internal shared file system), custom pipelines fail to load. As encountered in the test suite:
```
self = <tests.test_pipelines.CustomPipelineTests testMethod=test_local_custom_pipeline_file>
def test_local_cus... | closed | completed | false | 5 | [
"bug",
"good first issue",
"help wanted"
] | [] | 2022-12-02T01:27:08Z | 2022-12-21T21:23:51Z | 2022-12-21T21:23:50Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | antoche | 1,627,384 | MDQ6VXNlcjE2MjczODQ= | User | false |
huggingface/diffusers | 1,472,524,831 | I_kwDOHa8MBc5XxPIf | 1,519 | https://github.com/huggingface/diffusers/issues/1519 | https://api.github.com/repos/huggingface/diffusers/issues/1519 | Batch input yields noisy images for StableDiffusionPipeline on Apple M2 | ### Describe the bug
I am using the [StableDiffusionPipeline](https://github.com/huggingface/diffusers/tree/main/src/diffusers/pipelines) from the Hugging Face [Diffusers](https://huggingface.co/docs/diffusers/index) library on an M2 Mac. When I try to generate 1 image from 1 prompt, the output looks fine, but when I ... | closed | completed | false | 4 | [
"bug"
] | [
"pcuenca"
] | 2022-12-02T08:54:11Z | 2024-01-10T09:32:38Z | 2023-02-04T15:03:15Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | n-dijkstra-enjins | 119,493,737 | U_kgDOBx9UaQ | User | false |
huggingface/diffusers | 1,472,946,388 | I_kwDOHa8MBc5Xy2DU | 1,522 | https://github.com/huggingface/diffusers/issues/1522 | https://api.github.com/repos/huggingface/diffusers/issues/1522 | UNet2DConditionModel: TypeError: forward() got an unexpected keyword argument 'encoder_hidden_states' | ### Describe the bug
Hi everyone,
I try to use the UNet2DConditionModel with images as conditional input (for super-resolution purpose) in a Pytorch Lightning code. However, when I call the unet.forward function, this error message appears:
TypeError: forward() got an unexpected keyword argument ‘encoder_hidden_st... | closed | completed | false | 9 | [
"bug"
] | [] | 2022-12-02T14:23:52Z | 2022-12-07T18:12:56Z | 2022-12-07T18:12:56Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | rousseau | 398,895 | MDQ6VXNlcjM5ODg5NQ== | User | false |
huggingface/diffusers | 1,472,984,148 | I_kwDOHa8MBc5Xy_RU | 1,523 | https://github.com/huggingface/diffusers/issues/1523 | https://api.github.com/repos/huggingface/diffusers/issues/1523 | 关于速度 | I tested on v100 and found that the speed increase was only 50%, is this normal? Is there any other trick that can speed up the 512*512 resolution, the current elapsed time is a single unet 50ms. thanks. | closed | completed | false | 3 | [
"stale"
] | [] | 2022-12-02T14:47:38Z | 2023-01-09T15:03:51Z | 2023-01-09T15:03:51Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | baymin0220 | 39,220,091 | MDQ6VXNlcjM5MjIwMDkx | User | false |
huggingface/diffusers | 1,473,090,975 | I_kwDOHa8MBc5XzZWf | 1,524 | https://github.com/huggingface/diffusers/issues/1524 | https://api.github.com/repos/huggingface/diffusers/issues/1524 | examples/textual_inversion.py Poor quality training outcome with SD2.0 | ### Describe the bug
examples/textual_inversion.py Latest commit [6c56f05](https://github.com/huggingface/diffusers/commit/6c56f05097f7d3c561f02dc1c27e3dd7e9f88ce1)
This commit trains against SD1.5 well.
It does train SD2.0 but NOT well. Perhaps it is not using SD2.0's OpenCLIP, still using SD1.x CLIP?
###... | closed | completed | false | 2 | [
"bug",
"stale"
] | [] | 2022-12-02T16:00:43Z | 2023-01-10T15:03:29Z | 2023-01-10T15:03:29Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | zagglez | 61,600,248 | MDQ6VXNlcjYxNjAwMjQ4 | User | false |
huggingface/diffusers | 1,473,297,698 | I_kwDOHa8MBc5X0L0i | 1,528 | https://github.com/huggingface/diffusers/issues/1528 | https://api.github.com/repos/huggingface/diffusers/issues/1528 | Diffusers->SD CKPT script doesn't work for V2 Model. | ### Describe the bug
Extracting the V2-768 model with the current script (updated today) and then compiling again does not result in a working model.
### Reproduction
Extract the V2-768 checkpoint using [convert_original_stable_diffusion_to_diffusers.py](https://github.com/huggingface/diffusers/blob/main/scripts/con... | closed | completed | false | 15 | [
"bug",
"stale"
] | [] | 2022-12-02T18:47:13Z | 2023-02-07T15:03:49Z | 2023-02-07T15:03:49Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | d8ahazard | 1,633,844 | MDQ6VXNlcjE2MzM4NDQ= | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.