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 | 1,392,720,890 | I_kwDOHa8MBc5TAzv6 | 692 | https://github.com/huggingface/diffusers/issues/692 | https://api.github.com/repos/huggingface/diffusers/issues/692 | Is there a example of fine-tuning diffusion model? | **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 an... | closed | completed | false | 8 | [] | [
"patil-suraj"
] | 2022-09-30T17:05:04Z | 2022-10-11T17:18:51Z | 2022-10-11T17:18:51Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | RiverDong | 54,600,128 | MDQ6VXNlcjU0NjAwMTI4 | User | false |
huggingface/diffusers | 2,126,670,541 | I_kwDOHa8MBc5-wmrN | 6,920 | https://github.com/huggingface/diffusers/issues/6920 | https://api.github.com/repos/huggingface/diffusers/issues/6920 | How to merge a lot of embedding into a single file | I create a lot of embedding through textual inversion, but I couldn't found a file to merge this ckpt
| open | null | false | 3 | [
"stale"
] | [] | 2024-02-09T08:18:42Z | 2026-05-22T07:11:06Z | null | NONE | null | 20260522T073512Z | 2026-05-22T07:35:12Z | Eggwardhan | 32,223,217 | MDQ6VXNlcjMyMjIzMjE3 | User | false |
huggingface/diffusers | 2,126,902,492 | I_kwDOHa8MBc5-xfTc | 6,921 | https://github.com/huggingface/diffusers/issues/6921 | https://api.github.com/repos/huggingface/diffusers/issues/6921 | SDXL weights shipped with diffusers deviate from the official weights | ### Describe the bug
I observed statistical differences in Stable Diffusion XL images created using the diffusers library and using [Stability AI's implementation](https://github.com/Stability-AI/generative-models).
One cause of these differences is that the weights are not identical.
For the official GitHub imp... | closed | completed | false | 6 | [
"bug"
] | [] | 2024-02-09T10:58:15Z | 2024-02-12T09:01:48Z | 2024-02-12T09:01:48Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | btlorch | 11,959,715 | MDQ6VXNlcjExOTU5NzE1 | User | false |
huggingface/diffusers | 2,127,623,703 | I_kwDOHa8MBc5-0PYX | 6,923 | https://github.com/huggingface/diffusers/issues/6923 | https://api.github.com/repos/huggingface/diffusers/issues/6923 | regression - vae load sets inccorect scaling factor | ### Describe the bug
code added via this [pr](https://github.com/huggingface/diffusers/pull/6428) (refactor update from single file)
forces scaling factor to fixed value regardless of what is being loaded
code reference:
newly added code in: <https://github.com/huggingface/diffusers/blame/7c8cab313e4c66a813d146... | closed | completed | false | 9 | [
"bug"
] | [] | 2024-02-09T18:05:39Z | 2024-02-13T15:32:42Z | 2024-02-13T15:32:42Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | vladmandic | 57,876,960 | MDQ6VXNlcjU3ODc2OTYw | User | false |
huggingface/diffusers | 2,127,747,443 | I_kwDOHa8MBc5-0tlz | 6,925 | https://github.com/huggingface/diffusers/issues/6925 | https://api.github.com/repos/huggingface/diffusers/issues/6925 | IP Adapter Image Embeds - Compatibility with ComfyUI and other libraries/apps | Since we now have this PR #6868 merged and as suggested in this comment https://github.com/huggingface/diffusers/pull/6868#issuecomment-1932797580 I'm opening this issue to discuss the option to make the embeds for IP Adapters compatible with ComfyUI and other systems.
I've prepared a collab to demonstrate the issue... | closed | completed | false | 11 | [] | [
"yiyixuxu"
] | 2024-02-09T19:39:36Z | 2024-03-01T01:38:50Z | 2024-03-01T01:38:50Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | asomoza | 5,442,875 | MDQ6VXNlcjU0NDI4NzU= | User | false |
huggingface/diffusers | 2,127,814,390 | I_kwDOHa8MBc5-0972 | 6,926 | https://github.com/huggingface/diffusers/issues/6926 | https://api.github.com/repos/huggingface/diffusers/issues/6926 | DDPMScheduler claims to have set_alpha_to_one in combination with steps_offset | When inspecting the configuration for SD/SDXL we noticed that `set_alpha_to_one` is mentioned in the scheduler's init docstring but it is nowhere to be found. It seems that DDPM either lacks this functionality, or does not need it. | open | null | false | 8 | [
"stale",
"scheduler"
] | [] | 2024-02-09T20:36:55Z | 2026-05-22T04:05:02Z | null | CONTRIBUTOR | null | 20260522T043511Z | 2026-05-22T04:35:11Z | bghira | 59,658,056 | MDQ6VXNlcjU5NjU4MDU2 | User | false |
huggingface/diffusers | 2,127,981,599 | I_kwDOHa8MBc5-1mwf | 6,928 | https://github.com/huggingface/diffusers/issues/6928 | https://api.github.com/repos/huggingface/diffusers/issues/6928 | Add option to disable colorization of depthmap for community Marigold pipeline | **Is your feature request related to a problem? Please describe.**
The Marigold pipeline has no option to disable generation of the colormap. For my use case I do not need the colormap.
**Describe the solution you'd like.**
An option to disable generation of the colormap, as I've done in [my custom branch](https:/... | closed | completed | false | 1 | [] | [] | 2024-02-09T23:38:53Z | 2024-03-09T03:51:12Z | 2024-03-09T03:51:12Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | qqii | 6,891,353 | MDQ6VXNlcjY4OTEzNTM= | User | false |
huggingface/diffusers | 2,128,861,414 | I_kwDOHa8MBc5-49jm | 6,934 | https://github.com/huggingface/diffusers/issues/6934 | https://api.github.com/repos/huggingface/diffusers/issues/6934 | Conversion to original SDXL format from `diffusers` is broken | Setup:
1. Serialize https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0 locally:
```python
from diffusers import DiffusionPipeline
pipeline = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0")
pipeline.save_pretrained("stable-diffusion-xl-base-1.0")
```
2. Then run ... | closed | completed | false | 0 | [] | [] | 2024-02-11T06:29:19Z | 2024-02-12T12:00:24Z | 2024-02-12T12:00:24Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | sayakpaul | 22,957,388 | MDQ6VXNlcjIyOTU3Mzg4 | User | false |
huggingface/diffusers | 2,128,885,005 | I_kwDOHa8MBc5-5DUN | 6,935 | https://github.com/huggingface/diffusers/issues/6935 | https://api.github.com/repos/huggingface/diffusers/issues/6935 | ip adapter face id | **Is your feature request related to a problem? Please describe.**
Can't use ip adapter face id with ip adapter face if I use the community pipeline
**Describe the solution you'd like.**
Add ip adapter face id to the official support
**Describe alternatives you've considered.**
Spending a couple days and do so... | closed | completed | false | 4 | [] | [] | 2024-02-11T08:04:53Z | 2024-11-19T09:25:51Z | 2024-11-19T09:25:47Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | alexblattner | 15,870,094 | MDQ6VXNlcjE1ODcwMDk0 | User | false |
huggingface/diffusers | 2,130,029,724 | I_kwDOHa8MBc5-9ayc | 6,949 | https://github.com/huggingface/diffusers/issues/6949 | https://api.github.com/repos/huggingface/diffusers/issues/6949 | DPMSolverSinglestepScheduler, SD generates blank images at EVEN step numbers | ### Describe the bug
If DPMSolverSinglestepScheduler is used as a schedule with a stable diffusion pipeline the image generated by stable diffusion is a blank, black image if the number of inference steps is even.
Here's a shared Colab ready to go
https://colab.research.google.com/drive/1AUOvWu380FDJE0J0XOWKPBqX4q... | closed | completed | false | 7 | [
"bug",
"stale",
"scheduler"
] | [] | 2024-02-12T12:23:51Z | 2024-03-13T15:54:48Z | 2024-03-13T15:54:48Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Vargol | 62,868 | MDQ6VXNlcjYyODY4 | User | false |
huggingface/diffusers | 2,130,822,704 | I_kwDOHa8MBc5_AcYw | 6,955 | https://github.com/huggingface/diffusers/issues/6955 | https://api.github.com/repos/huggingface/diffusers/issues/6955 | When trying to do inference using StableDiffusionPipeline, I get Transformers error : 'tokenizers.AddedToken' object has no attribute 'special' | Hi! I get an error when trying to load the pretrained weights for Stable Diffusion (same issue with ControlNet). Below is my code:
```
from diffusers import StableDiffusionPipeline
import os
pipe = StableDiffusionPipeline.from_pretrained("CompVis/stable-diffusion-v1-4")
pipe.to("cuda")
```
And I get an err... | closed | completed | false | 1 | [] | [] | 2024-02-12T19:27:41Z | 2024-02-12T20:05:33Z | 2024-02-12T20:05:33Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Saumya-Gupta-26 | 16,615,692 | MDQ6VXNlcjE2NjE1Njky | User | false |
huggingface/diffusers | 2,130,911,576 | I_kwDOHa8MBc5_AyFY | 6,956 | https://github.com/huggingface/diffusers/issues/6956 | https://api.github.com/repos/huggingface/diffusers/issues/6956 | StableVideoDiffusionPipeline : CUDA out of memory on SageMaker | ### Describe the bug
Hi team !
I am using Diffusers and specifically looking at [this documentation](https://huggingface.co/docs/diffusers/using-diffusers/svd), however the link to open a notebook in Studio Lab doesn't work (the notebook is not present in the GitHub repo).
I am trying to deploy https://huggingface... | closed | completed | false | 4 | [
"bug",
"stale"
] | [] | 2024-02-12T20:14:04Z | 2024-03-28T06:19:11Z | 2024-03-28T06:19:11Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | krokoko | 6,213,264 | MDQ6VXNlcjYyMTMyNjQ= | User | false |
huggingface/diffusers | 2,131,075,425 | I_kwDOHa8MBc5_BaFh | 6,957 | https://github.com/huggingface/diffusers/issues/6957 | https://api.github.com/repos/huggingface/diffusers/issues/6957 | StableDiffusionXLImg2ImgPipeline strength not working as expected | ### Describe the bug
According to the documentation:
> Conceptually, indicates how much to transform the reference image. Must be between 0 and 1. image will be used as a starting point, adding more noise to it the larger the strength. The number of denoising steps depends on the amount of noise initially added. W... | closed | completed | false | 6 | [
"bug",
"stale"
] | [] | 2024-02-12T22:11:55Z | 2024-03-19T02:42:05Z | 2024-03-14T15:18:18Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | zstiggz | 704,532 | MDQ6VXNlcjcwNDUzMg== | User | false |
huggingface/diffusers | 2,131,397,180 | I_kwDOHa8MBc5_Coo8 | 6,958 | https://github.com/huggingface/diffusers/issues/6958 | https://api.github.com/repos/huggingface/diffusers/issues/6958 | use_linear_projection not in mid_block in unet_motion_model | ### Describe the bug
https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/unets/unet_motion_model.py#L299
use_linear_projection not passed
### Reproduction
https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/unets/unet_motion_model.py#L299
use_linear_projection not passed... | closed | completed | false | 2 | [
"bug",
"contributions-welcome"
] | [] | 2024-02-13T03:30:25Z | 2024-02-26T05:19:15Z | 2024-02-26T05:19:15Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | jiarenyf | 12,405,578 | MDQ6VXNlcjEyNDA1NTc4 | User | false |
huggingface/diffusers | 1,393,825,295 | I_kwDOHa8MBc5TFBYP | 696 | https://github.com/huggingface/diffusers/issues/696 | https://api.github.com/repos/huggingface/diffusers/issues/696 | CUDA out of memory when i want to train dreambooth | ### Describe the bug
I'm using T4 with colab free , when I start training it tells me cuda error, it happens when I activate prior_preservation.
Run training
```
Launching training on one GPU.
S... | closed | completed | false | 12 | [
"bug"
] | [
"patil-suraj"
] | 2022-10-02T16:41:26Z | 2024-04-19T06:06:27Z | 2022-10-20T17:24:13Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | loboere | 15,265,895 | MDQ6VXNlcjE1MjY1ODk1 | User | false |
huggingface/diffusers | 2,131,583,850 | I_kwDOHa8MBc5_DWNq | 6,964 | https://github.com/huggingface/diffusers/issues/6964 | https://api.github.com/repos/huggingface/diffusers/issues/6964 | StableDiffusionControlNetInpaintPipeline - does not work with 4 channel Unets | ### Describe the bug
Hi,
I have tried to use an example from here `https://huggingface.co/docs/diffusers/main/en/api/pipelines/controlnet` (posted below). However, when a model is passed that has a Unet that expects 4 channels the pipeline breaks with the following error
```
RuntimeError: Given groups=1, weigh... | closed | completed | false | 10 | [
"bug"
] | [] | 2024-02-13T06:53:59Z | 2024-02-14T13:35:27Z | 2024-02-14T13:34:58Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | cryptexis | 1,158,677 | MDQ6VXNlcjExNTg2Nzc= | User | false |
huggingface/diffusers | 2,132,401,648 | I_kwDOHa8MBc5_Gd3w | 6,966 | https://github.com/huggingface/diffusers/issues/6966 | https://api.github.com/repos/huggingface/diffusers/issues/6966 | 'UNet2DConditionModel' object has no attribute 'attn_processors' when using unet.set_attn_processor() after forward pass | ### Describe the bug
Hi together,
I'm working on a usecase applying LoRA to a standard UNet2DConditionModel. I add LoRA as is done in the [example script](https://github.com/huggingface/diffusers/blob/main/examples/research_projects/lora/train_text_to_image_lora.py):
```python
lora_attn_procs = {}
for name in ... | closed | completed | false | 3 | [
"bug"
] | [] | 2024-02-13T14:04:37Z | 2024-02-18T09:13:07Z | 2024-02-18T09:13:07Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | LeggereZidrone | 114,178,846 | U_kgDOBs47Hg | User | false |
huggingface/diffusers | 2,132,710,100 | I_kwDOHa8MBc5_HpLU | 6,967 | https://github.com/huggingface/diffusers/issues/6967 | https://api.github.com/repos/huggingface/diffusers/issues/6967 | [advanced dreambooth lora sdxl script]: cannot train --with_prior_preservation, shape mismatch | ### Describe the bug
I came across this while testing new features from https://github.com/huggingface/diffusers/pull/6691 (many thanks for supporting micro-conditioning!)
Using [`train_dreambooth_lora_sdxl_advanced.py`](https://github.com/huggingface/diffusers/blob/main/examples/advanced_diffusion_training/train_d... | closed | completed | false | 1 | [
"bug"
] | [
"linoytsaban"
] | 2024-02-13T16:24:26Z | 2024-02-15T06:50:07Z | 2024-02-15T06:50:06Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | To-jak | 33,063,955 | MDQ6VXNlcjMzMDYzOTU1 | User | false |
huggingface/diffusers | 2,133,039,205 | I_kwDOHa8MBc5_I5hl | 6,969 | https://github.com/huggingface/diffusers/issues/6969 | https://api.github.com/repos/huggingface/diffusers/issues/6969 | Expanded init fields in StableDiffusionPipeline cause incompatibilities with many/most inherited pipelines | ### Describe the bug
class `StableDiffusionPipeline` has its init section as:
```py
def __init__(
self,
vae: AutoencoderKL,
text_encoder: CLIPTextModel,
tokenizer: CLIPTokenizer,
unet: UNet2DConditionModel,
scheduler: KarrasDiffusionSchedulers,
saf... | open | null | false | 22 | [
"bug",
"good first issue",
"help wanted",
"wip",
"contributions-welcome"
] | [
"asomoza"
] | 2024-02-13T20:00:36Z | 2025-10-31T16:15:36Z | null | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | vladmandic | 57,876,960 | MDQ6VXNlcjU3ODc2OTYw | User | false |
huggingface/diffusers | 2,133,339,427 | I_kwDOHa8MBc5_KC0j | 6,970 | https://github.com/huggingface/diffusers/issues/6970 | https://api.github.com/repos/huggingface/diffusers/issues/6970 | Stable Cascade Diffusers not working with FP16 | You know that there is a demo released here : https://huggingface.co/spaces/multimodalart/stable-cascade
I am using it locally. Works perfectly fine
But I have to change float type to float16 from bfloat16 on Kaggle
That is where error starts
```
\venv\lib\site-packages\diffusers\utils\pil_utils.py:43: ... | closed | completed | false | 17 | [
"bug",
"stale"
] | [
"kashif",
"DN6",
"yiyixuxu",
"sayakpaul"
] | 2024-02-14T00:18:32Z | 2024-11-23T00:15:37Z | 2024-11-23T00:15:37Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | FurkanGozukara | 19,240,467 | MDQ6VXNlcjE5MjQwNDY3 | User | false |
huggingface/diffusers | 2,133,806,972 | I_kwDOHa8MBc5_L098 | 6,974 | https://github.com/huggingface/diffusers/issues/6974 | https://api.github.com/repos/huggingface/diffusers/issues/6974 | unload_textual_inversion doesn't unload tokenizer_2 & text_encoder_2 | ### Describe the bug
unload_textual_inversion unloads only the pipeline tokenizer and text_encoder.
I can submit a PR if you like
### Reproduction
Following [Advanced Lora training inference examples ](https://huggingface.co/blog/sdxl_lora_advanced_script#inference)
```
# load embeddings to the text encoder... | closed | completed | false | 4 | [
"bug"
] | [] | 2024-02-14T08:29:48Z | 2024-02-22T02:53:26Z | 2024-02-22T02:53:26Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | H3zi | 12,728,968 | MDQ6VXNlcjEyNzI4OTY4 | User | false |
huggingface/diffusers | 1,393,852,542 | I_kwDOHa8MBc5TFIB- | 698 | https://github.com/huggingface/diffusers/issues/698 | https://api.github.com/repos/huggingface/diffusers/issues/698 | Any Webui for local Dreambooth yet? I mean, to use DB to train. | Although I'm trying my best to dive into the who PY scene I'm at a loss. Has anyone made a Webui for DB training on your local yet? If so, where is it at? If not, is anyone working on one? | closed | completed | false | 5 | [
"good first issue",
"help wanted"
] | [] | 2022-10-02T18:14:47Z | 2023-03-18T18:41:55Z | 2023-03-18T18:41:55Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Duemellon | 113,479,041 | U_kgDOBsONgQ | User | false |
huggingface/diffusers | 2,135,473,639 | I_kwDOHa8MBc5_SL3n | 6,981 | https://github.com/huggingface/diffusers/issues/6981 | https://api.github.com/repos/huggingface/diffusers/issues/6981 | fp16 overflow when combine scheduler that supports`rescale_betas_zero_snr=True` with InstructPix2PixPipeline | ### Describe the bug
InstructPix2Pix pipeline has a flag ```scheduler_is_in_sigma_space = hasattr(self.scheduler, "sigmas")``` which result latent value scale back and forth from sigma space for calculation of classifer free guidance, which seems not necessary, and especially when setting `rescale_betas_zero_snr=Tru... | closed | completed | false | 16 | [
"bug",
"stale"
] | [] | 2024-02-15T01:33:46Z | 2024-03-29T15:24:59Z | 2024-03-29T15:24:59Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | erliding | 42,523,282 | MDQ6VXNlcjQyNTIzMjgy | User | false |
huggingface/diffusers | 2,135,871,814 | I_kwDOHa8MBc5_TtFG | 6,984 | https://github.com/huggingface/diffusers/issues/6984 | https://api.github.com/repos/huggingface/diffusers/issues/6984 | refactors pipelines inheriting from `StableDiffusionPipeline` | We have a couple of community pipelines that used `StableDiffusionPipeline` as the base class instead of `DiffusionPipeline`. Most of them are broken after we changed the signature of `StableDiffusionPipeline`. see more details on the issue here https://github.com/huggingface/diffusers/issues/6969#issuecomment-19441628... | closed | completed | false | 17 | [
"good first issue",
"help wanted",
"Good second issue",
"contributions-welcome"
] | [] | 2024-02-15T08:01:17Z | 2025-07-15T16:37:12Z | 2025-07-15T16:36:41Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | yiyixuxu | 12,631,849 | MDQ6VXNlcjEyNjMxODQ5 | User | false |
huggingface/diffusers | 1,393,869,873 | I_kwDOHa8MBc5TFMQx | 699 | https://github.com/huggingface/diffusers/issues/699 | https://api.github.com/repos/huggingface/diffusers/issues/699 | Add support for negative prompts | **Is your feature request related to a problem? Please describe.**
Some forks stable diffusion added [support for negative prompts](https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Negative-prompt).
I tested this in the `StableDiffusionPipeline` and it seems to work that way with diffusers as well. I cu... | closed | completed | false | 5 | [] | [] | 2022-10-02T19:15:57Z | 2022-10-04T17:33:12Z | 2022-10-04T17:33:12Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | allo- | 1,482,341 | MDQ6VXNlcjE0ODIzNDE= | User | false |
huggingface/diffusers | 2,138,725,242 | I_kwDOHa8MBc5_elt6 | 6,996 | https://github.com/huggingface/diffusers/issues/6996 | https://api.github.com/repos/huggingface/diffusers/issues/6996 | if "text_embeds" not in added_cond_kwargs: TypeError: argument of type 'NoneType' is not iterable | ### Describe the bug
when i use StableDiffusionControlNetPipeline by loading local model to generate image,error happen in controlnet.py line 775 :
if "text_embeds" not in added_cond_kwargs:
### Reproduction
import torch
import numpy as np
from controlnet_aux import MidasDetector
from diffusers.utils imp... | closed | completed | false | 9 | [
"bug",
"stale"
] | [] | 2024-02-16T14:22:19Z | 2024-11-23T00:13:38Z | 2024-11-23T00:13:38Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | oftenliu | 20,364,898 | MDQ6VXNlcjIwMzY0ODk4 | User | false |
huggingface/diffusers | 2,140,109,686 | I_kwDOHa8MBc5_j3t2 | 7,000 | https://github.com/huggingface/diffusers/issues/7000 | https://api.github.com/repos/huggingface/diffusers/issues/7000 | src/diffusers/utils /outputs.py: Deprecation Warning for torch.utils._pytree._register_pytree_node in latest PyTorch | ### Describe the bug
Hi! The latest version of PyTorch has deprecated torch.utils._pytree._register_pytree_node, recommending the use of torch.utils._pytree.register_pytree_node (without the leading underscore). Perhaps not an urgent issue but wondering might affect long-term compatibility. Happened when I run diffu... | closed | completed | false | 3 | [
"bug",
"contributions-welcome"
] | [] | 2024-02-17T13:53:27Z | 2024-04-12T11:03:29Z | 2024-02-18T15:45:46Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | zyinghua | 49,605,250 | MDQ6VXNlcjQ5NjA1MjUw | User | false |
huggingface/diffusers | 2,140,535,492 | I_kwDOHa8MBc5_lfrE | 7,001 | https://github.com/huggingface/diffusers/issues/7001 | https://api.github.com/repos/huggingface/diffusers/issues/7001 | Denoising SDXL iteration images for coherent image previews that a user could understand | Hello,
I'm currently trying to create image previews with SDXL. This works! However, the image output are very noisy. A very long time ago I found a solution to this for sd1.5 but unfortunately it has been lost to time.
How would I go about denoising these images so they are a little more coherent to a human vie... | open | null | false | 7 | [
"stale",
"should-move-to-discussion"
] | [] | 2024-02-17T19:58:58Z | 2026-05-22T07:11:13Z | null | NONE | null | 20260522T073512Z | 2026-05-22T07:35:12Z | MonkeeMan1 | 98,274,841 | U_kgDOBduOGQ | User | false |
huggingface/diffusers | 2,140,621,900 | I_kwDOHa8MBc5_l0xM | 7,002 | https://github.com/huggingface/diffusers/issues/7002 | https://api.github.com/repos/huggingface/diffusers/issues/7002 | StableDiffusionImg2ImgPipeline RuntimeError: Invalid buffer size: 361.00 GB | ### Describe the bug
StableDiffusionImg2ImgPipeline is giving the error `RuntimeError: Invalid buffer size: 361.00 GB`. Its working fine with StableDiffusionControlNetImg2ImgPipeline and I didn't change any code between the two (the image used as input is the same).
### Reproduction
```
image = generator(
im... | closed | completed | false | 2 | [
"bug",
"needs-code-example"
] | [] | 2024-02-17T21:19:10Z | 2024-02-18T03:22:20Z | 2024-02-18T03:21:57Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | ballenvironment | 111,252,136 | U_kgDOBqGSqA | User | false |
huggingface/diffusers | 2,140,632,379 | I_kwDOHa8MBc5_l3U7 | 7,003 | https://github.com/huggingface/diffusers/issues/7003 | https://api.github.com/repos/huggingface/diffusers/issues/7003 | StableDiffusionInpaintPipeline expected shape tensor(..., device='meta', size=(320, 9, 3, 3)), but got torch.Size([320, 4, 3, 3]) | ### Describe the bug
StableDiffusionInpaintPipeline and StableDiffusionControlnetInpaintPipeline both give an error that the tensor shape doesn't match.
I tried to use the provided solution but it doesn't seem to work at least not with a local file.
```py
generator = StableDiffusionInpaintPipeline.from_single_... | closed | completed | false | 7 | [
"bug",
"needs-code-example",
"single_file"
] | [] | 2024-02-17T21:57:09Z | 2024-12-19T02:29:36Z | 2024-02-20T16:27:15Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | ballenvironment | 111,252,136 | U_kgDOBqGSqA | User | false |
huggingface/diffusers | 2,140,808,971 | I_kwDOHa8MBc5_micL | 7,009 | https://github.com/huggingface/diffusers/issues/7009 | https://api.github.com/repos/huggingface/diffusers/issues/7009 | TextToVideoSDPipeline could not generate resonable results when loading ali-vilab/text-to-video-ms-1.7b | ### Describe the bug
https://github.com/huggingface/diffusers/assets/26198430/c50d5393-f027-4685-9408-021d35ae8da0
### Reproduction
```python
import torch
from diffusers import DiffusionPipeline
from diffusers.utils import export_to_video
pipe = DiffusionPipeline.from_pretrained("damo-vilab/text-... | closed | completed | false | 6 | [
"bug"
] | [] | 2024-02-18T08:13:43Z | 2025-03-29T17:14:49Z | 2024-11-19T09:28:47Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Zeqiang-Lai | 26,198,430 | MDQ6VXNlcjI2MTk4NDMw | User | false |
huggingface/diffusers | 2,140,931,578 | I_kwDOHa8MBc5_nAX6 | 7,010 | https://github.com/huggingface/diffusers/issues/7010 | https://api.github.com/repos/huggingface/diffusers/issues/7010 | How to set export HF_HOME on Kaggle? | Kaggle temporary disk is slow once again and I want models to be downloaded into working directory.
I have used the below command but it didn't work. Which command I need?
`!export HF_HOME="/kaggle/working"`
| closed | completed | false | 2 | [
"bug"
] | [] | 2024-02-18T11:15:21Z | 2024-02-18T14:39:08Z | 2024-02-18T14:39:08Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | FurkanGozukara | 19,240,467 | MDQ6VXNlcjE5MjQwNDY3 | User | false |
huggingface/diffusers | 2,141,022,239 | I_kwDOHa8MBc5_nWgf | 7,011 | https://github.com/huggingface/diffusers/issues/7011 | https://api.github.com/repos/huggingface/diffusers/issues/7011 | Regularize Latent input for SD and SDXL img2img | I am currently trying to implement seed interpolation in SDNext. This can be done smoothly for SD and SDXL pipelines but not for img2img
Currently latents can be passed to `StableDiffusionXLImg2ImgPipeline` but they are promptly ignored here https://github.com/huggingface/diffusers/blob/8974c50bff1f51c5a1208723af88d... | closed | completed | false | 10 | [
"help wanted",
"contributions-welcome"
] | [] | 2024-02-18T14:51:24Z | 2024-04-29T00:53:30Z | 2024-04-29T00:53:30Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | AI-Casanova | 54,461,896 | MDQ6VXNlcjU0NDYxODk2 | User | false |
huggingface/diffusers | 2,141,057,420 | I_kwDOHa8MBc5_nfGM | 7,012 | https://github.com/huggingface/diffusers/issues/7012 | https://api.github.com/repos/huggingface/diffusers/issues/7012 | v_prediction in text_to_image weird images | ### Describe the bug
Using `--prediction_type="v_prediction"` with the example `text_to_image_lora.py` script leads to very weird images:
![image](https://github.com/huggingface/diffusers/assets/30842617/0ed8206... | closed | completed | false | 5 | [
"bug",
"stale"
] | [] | 2024-02-18T15:58:09Z | 2024-03-29T15:25:13Z | 2024-03-29T15:25:12Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | BurgerAndreas | 30,842,617 | MDQ6VXNlcjMwODQyNjE3 | User | false |
huggingface/diffusers | 2,141,141,282 | I_kwDOHa8MBc5_nzki | 7,014 | https://github.com/huggingface/diffusers/issues/7014 | https://api.github.com/repos/huggingface/diffusers/issues/7014 | Move `ip_adapter_face_id` from community to main pipeline | Currently `examples/community/ip_adapter_face_id.py` lives in the community folder. However, its usage based on: https://huggingface.co/h94/IP-Adapter-FaceID showcase 385K downloads in the last month
<img width="514" alt="image" src="https://github.com/huggingface/diffusers/assets/788417/a63f1f5c-3996-4bda-bb42-41cae... | closed | completed | false | 7 | [] | [
"fabiorigano"
] | 2024-02-18T19:06:43Z | 2024-04-19T00:13:28Z | 2024-04-19T00:13:28Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | apolinario | 788,417 | MDQ6VXNlcjc4ODQxNw== | User | false |
huggingface/diffusers | 1,394,305,819 | I_kwDOHa8MBc5TG2sb | 702 | https://github.com/huggingface/diffusers/issues/702 | https://api.github.com/repos/huggingface/diffusers/issues/702 | Reverse sampling for schedulers | ### Model/Pipeline/Scheduler description
Some of the samplers (e.g. DDIM) are fully reversible. A simple API addition of `reverse_step()` or a mixin can be implemented. This will allow interpolating between images not generated by diffusion models.
### Open source status
- [X] The model implementation is available
-... | closed | completed | false | 12 | [
"stale"
] | [] | 2022-10-03T07:58:44Z | 2022-11-20T19:45:34Z | 2022-11-20T19:45:34Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | neverix | 46,641,404 | MDQ6VXNlcjQ2NjQxNDA0 | User | false |
huggingface/diffusers | 2,141,527,615 | I_kwDOHa8MBc5_pR4_ | 7,021 | https://github.com/huggingface/diffusers/issues/7021 | https://api.github.com/repos/huggingface/diffusers/issues/7021 | Import error HF_HOME | ### Describe the bug
ImportError: cannot import name 'HF_HOME' from 'huggingface_hub.constants'
### Reproduction
from diffusers import (
AutoencoderKL,
DDPMScheduler,
PNDMScheduler,
StableDiffusionPipeline,
UNet2DConditionModel,
)
### Logs
```shell
/home/wz1937/miniconda3/envs/learn/lib/p... | closed | completed | false | 2 | [
"bug"
] | [] | 2024-02-19T05:17:32Z | 2025-02-21T03:09:20Z | 2024-02-19T05:22:57Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Will-Zhao0 | 72,349,413 | MDQ6VXNlcjcyMzQ5NDEz | User | false |
huggingface/diffusers | 2,143,448,833 | I_kwDOHa8MBc5_wm8B | 7,033 | https://github.com/huggingface/diffusers/issues/7033 | https://api.github.com/repos/huggingface/diffusers/issues/7033 | IP-Adapter attention masking warning logs | https://github.com/huggingface/diffusers/blob/d8d208acdea43a274bd74ed27119a4f95e8e0946/src/diffusers/models/attention_processor.py#L517
https://github.com/huggingface/diffusers/blob/d8d208acdea43a274bd74ed27119a4f95e8e0946/src/diffusers/models/attention_processor.py#L1202
hi!
Is the call method of the AttnProces... | closed | completed | false | 8 | [
"stale"
] | [] | 2024-02-20T03:10:52Z | 2024-06-12T15:50:15Z | 2024-06-12T15:50:14Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Honey-666 | 76,274,942 | MDQ6VXNlcjc2Mjc0OTQy | User | false |
huggingface/diffusers | 2,143,791,493 | I_kwDOHa8MBc5_x6mF | 7,037 | https://github.com/huggingface/diffusers/issues/7037 | https://api.github.com/repos/huggingface/diffusers/issues/7037 | delete me | delete me | closed | completed | false | 1 | [] | [] | 2024-02-20T08:24:13Z | 2024-02-22T02:45:15Z | 2024-02-22T02:45:15Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | differential-diffusion | 134,973,080 | U_kgDOCAuGmA | User | false |
huggingface/diffusers | 2,143,796,639 | I_kwDOHa8MBc5_x72f | 7,038 | https://github.com/huggingface/diffusers/issues/7038 | https://api.github.com/repos/huggingface/diffusers/issues/7038 | Differential Diffusion: Giving Each Pixel Its Strength | ### Model/Pipeline/Scheduler description
Hello,
I would like to suggest merging my paper: Differential Diffusion: Giving Each Pixel Its Strength.
The paper allows a user to edit a picture by a change map that describes how much each region should change.
The editing process is typically guided by textual instruct... | closed | completed | false | 23 | [] | [] | 2024-02-20T08:27:30Z | 2025-04-07T08:15:50Z | 2024-04-02T04:15:31Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | exx8 | 8,540,180 | MDQ6VXNlcjg1NDAxODA= | User | false |
huggingface/diffusers | 1,394,400,244 | I_kwDOHa8MBc5THNv0 | 704 | https://github.com/huggingface/diffusers/issues/704 | https://api.github.com/repos/huggingface/diffusers/issues/704 | [Community] Accept latents as optional input in other pipelines | The Stable Diffusion pipeline accepts latents as an optional input for reproducibility purposes, see #247.
In conversation with @tmabraham, he suggested extending this feature to other pipelines too. In addition to reproducibility efforts, this could come in handy for a potential integration with the `fastai` librar... | closed | completed | false | 5 | [
"good first issue"
] | [] | 2022-10-03T09:12:47Z | 2023-03-18T18:42:01Z | 2023-03-18T18:42:01Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | pcuenca | 1,177,582 | MDQ6VXNlcjExNzc1ODI= | User | false |
huggingface/diffusers | 2,144,179,988 | I_kwDOHa8MBc5_zZcU | 7,041 | https://github.com/huggingface/diffusers/issues/7041 | https://api.github.com/repos/huggingface/diffusers/issues/7041 | Segmentation fault when num_inference_steps > 1 | ### Describe the bug
When I'm running StableDiffusionXLPipeline with either num_inference_steps > 1 or batch > 1, I'm getting segmentation fault.
### Reproduction
```
from diffusers import (
DPMSolverSinglestepScheduler,
StableDiffusionXLPipeline,
)
pipe = StableDiffusionXLPipeline.from_pretrained(
... | closed | completed | false | 8 | [
"bug"
] | [] | 2024-02-20T11:39:14Z | 2024-02-21T12:38:59Z | 2024-02-21T12:38:58Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | mdomaradzki-cosmose | 155,545,349 | U_kgDOCUVvBQ | User | false |
huggingface/diffusers | 2,144,793,463 | I_kwDOHa8MBc5_1vN3 | 7,043 | https://github.com/huggingface/diffusers/issues/7043 | https://api.github.com/repos/huggingface/diffusers/issues/7043 | Dreambooth with custom captions | Hi, there is an advanced lora training script that supports custom captions. Can the same be done with regular dreambooth training? Or how can I modify this script to train dreambooth instead of lora?
[examples/advanced_diffusion_training/train_dreambooth_lora_sd15_advanced.py](https://github.com/huggingface/diffuse... | closed | completed | false | 3 | [] | [] | 2024-02-20T16:32:49Z | 2024-02-21T20:10:32Z | 2024-02-21T20:10:32Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | ballenvironment | 111,252,136 | U_kgDOBqGSqA | User | false |
huggingface/diffusers | 2,146,575,951 | I_kwDOHa8MBc5_8iZP | 7,046 | https://github.com/huggingface/diffusers/issues/7046 | https://api.github.com/repos/huggingface/diffusers/issues/7046 | Implementation error of v_prediction in eulara and eular | I came across a potential inconsistency in the implementation of v_prediction within the Euler scheduler when examining the source code. Specifically, in the computation of pred_original_sample, the formula is given as:
`pred_original_sample = model_output * (-sigma / (sigma**2 + 1) ** 0.5) + (sample / (sigma**2 + 1))... | open | null | false | 7 | [
"stale"
] | [] | 2024-02-21T11:59:59Z | 2025-07-21T01:38:25Z | null | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | yelanyelan | 11,421,623 | MDQ6VXNlcjExNDIxNjIz | User | false |
huggingface/diffusers | 2,146,853,370 | I_kwDOHa8MBc5_9mH6 | 7,049 | https://github.com/huggingface/diffusers/issues/7049 | https://api.github.com/repos/huggingface/diffusers/issues/7049 | IP adapter basic example causes ValueError: too many values to unpack (expected 3) | ### Describe the bug
I tried to run the basic example from [this tutorial ](https://huggingface.co/docs/diffusers/main/en/using-diffusers/ip_adapter?tasks=Inpainting). Building the model works, but it crashes on inference.
### Reproduction
```
from diffusers import AutoPipelineForText2Image
from diffusers.utils im... | closed | completed | false | 7 | [
"bug",
"IPAdapter"
] | [
"yiyixuxu"
] | 2024-02-21T14:10:42Z | 2024-02-25T10:05:47Z | 2024-02-25T10:05:47Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | AlmogDeepKeep | 126,669,158 | U_kgDOB4zRZg | User | false |
huggingface/diffusers | 2,147,633,859 | I_kwDOHa8MBc6AAkrD | 7,053 | https://github.com/huggingface/diffusers/issues/7053 | https://api.github.com/repos/huggingface/diffusers/issues/7053 | UNet3DConditionModel does not use class conditioning | ### Describe the bug
Hello guys,
I found a small inconsistency. In `UNet3DConditionModel` forward method, `class_labels` is [defined](https://github.com/huggingface/diffusers/blob/6fedbd850a282ef21009f99a394f619ed7420137/src/diffusers/models/unets/unet_3d_condition.py#L540C13-L540C25) but is never used after that.... | open | null | false | 7 | [
"bug",
"stale"
] | [] | 2024-02-21T20:24:45Z | 2024-04-21T15:04:02Z | null | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | ilan-theodoro | 7,650,238 | MDQ6VXNlcjc2NTAyMzg= | User | false |
huggingface/diffusers | 2,147,766,664 | I_kwDOHa8MBc6ABFGI | 7,054 | https://github.com/huggingface/diffusers/issues/7054 | https://api.github.com/repos/huggingface/diffusers/issues/7054 | LoraLoaderMixin.load_lora_weights() empties state_dict passed as input param. | ### Describe the bug
The case happens when LoRA's are opened beforehand and stored as state_dict for optimization purposes (lasy-load, re-use etc) in production.
After `pipe.load_lora_weights(adapter_state_dict, adapter_name=adapter_name)` is called, `adapter_state_dict` becomes empty dict and can't be reused to ... | closed | completed | false | 1 | [
"bug",
"peft"
] | [] | 2024-02-21T21:45:05Z | 2024-02-27T01:42:25Z | 2024-02-27T01:42:25Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | slep0v | 37,597,789 | MDQ6VXNlcjM3NTk3Nzg5 | User | false |
huggingface/diffusers | 2,148,248,242 | I_kwDOHa8MBc6AC6qy | 7,056 | https://github.com/huggingface/diffusers/issues/7056 | https://api.github.com/repos/huggingface/diffusers/issues/7056 | sdxl text to image lora fine tuning with train_text_encoder parameter got me an error. | ### Describe the bug
Hi, I'm trying to fine-tune the SDXL 1.0 model using diffusers 0.26.2 with ` train_text_encoder` parameter. But it got me an error.
However, training without the parameter, it works fine.
I forked 0.26.2 [training file](https://github.com/huggingface/diffusers/blob/v0.26.2/examples/text_to_i... | closed | completed | false | 1 | [
"bug"
] | [] | 2024-02-22T05:44:38Z | 2024-02-22T06:01:41Z | 2024-02-22T05:59:21Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | haandol | 124,808 | MDQ6VXNlcjEyNDgwOA== | User | false |
huggingface/diffusers | 2,148,361,261 | I_kwDOHa8MBc6ADWQt | 7,057 | https://github.com/huggingface/diffusers/issues/7057 | https://api.github.com/repos/huggingface/diffusers/issues/7057 | Ckpt of this model doesn't work | ### Describe the bug
Hello, I tried loading this model but the ckpt version doesn't seem to work and only generates noisy/broken images when loaded into StableDiffusionPipeline.
https://huggingface.co/Mitsua/mitsua-diffusion-one/blob/main/mitsua-diffusion-one.ckpt
But it works if you load the model with diffuser... | closed | completed | false | 3 | [
"bug"
] | [] | 2024-02-22T07:02:19Z | 2024-11-20T00:31:46Z | 2024-11-20T00:31:46Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | ballenvironment | 111,252,136 | U_kgDOBqGSqA | User | false |
huggingface/diffusers | 2,148,863,472 | I_kwDOHa8MBc6AFQ3w | 7,059 | https://github.com/huggingface/diffusers/issues/7059 | https://api.github.com/repos/huggingface/diffusers/issues/7059 | PixArt-δ ControlNet with diffusers | PixArt-δ ControlNet and [Checkpoint](https://huggingface.co/PixArt-alpha/PixArt-ControlNet/tree/main) has been released and it seems that the diffusers framework does not currently support. Is it possible to add a PixArtControlNetPipeline just like controlnet pipelines of sd 1.5?
Since we already have PixArtAlphaPipel... | open | null | false | 4 | [
"community-examples",
"stale",
"contributions-welcome"
] | [] | 2024-02-22T11:35:08Z | 2024-04-17T15:03:56Z | null | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | fangxia1997 | 27,539,274 | MDQ6VXNlcjI3NTM5Mjc0 | User | false |
huggingface/diffusers | 1,394,653,121 | I_kwDOHa8MBc5TILfB | 706 | https://github.com/huggingface/diffusers/issues/706 | https://api.github.com/repos/huggingface/diffusers/issues/706 | How to make Stable Diffusion work with DDIM | It seems that using DDIM with the `StableDiffusionPIpeline` gives odd results
```python
pipe = StableDiffusionPipeline.from_pretrained(
"CompVis/stable-diffusion-v1-4",
scheduler=DDIMScheduler(beta_start=0.00085, beta_end=0.012, beta_schedule="scaled_linear", steps_offset=1, clip_sample=False),
use_au... | closed | completed | false | 3 | [] | [] | 2022-10-03T12:29:29Z | 2022-10-05T17:23:47Z | 2022-10-05T17:23:47Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | NouamaneTazi | 29,777,165 | MDQ6VXNlcjI5Nzc3MTY1 | User | false |
huggingface/diffusers | 2,149,291,077 | I_kwDOHa8MBc6AG5RF | 7,061 | https://github.com/huggingface/diffusers/issues/7061 | https://api.github.com/repos/huggingface/diffusers/issues/7061 | docstring for imagic community pipeline inference is incorrect | ### Describe the bug
The docstring for the community pipeline has incorrect parameters.
### Reproduction
1. Clone or view the code at https://github.com/huggingface/diffusers/blob/main/examples/community/imagic_stable_diffusion.py
### Logs
_No response_
### System Info
This is a documentation fix, so no need for... | closed | completed | false | 2 | [
"bug"
] | [] | 2024-02-22T15:08:32Z | 2024-02-23T21:42:34Z | 2024-02-23T21:42:33Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | chongdashu | 1,178,121 | MDQ6VXNlcjExNzgxMjE= | User | false |
huggingface/diffusers | 2,150,184,277 | I_kwDOHa8MBc6AKTVV | 7,064 | https://github.com/huggingface/diffusers/issues/7064 | https://api.github.com/repos/huggingface/diffusers/issues/7064 | [ip-adapter]fix IP-adapter support for SAG, panorama pipeline | We have added a `IPAdapterTexterMixin` https://github.com/huggingface/diffusers/pull/6862. However, we have discovered that the IP adapter support was not implemented correctly in the following pipelines:
- StableDiffusionSAGPipeline
- StableDiffusionPanoramaPipeline
If anyone is interested in fixing the IP-ada... | closed | completed | false | 8 | [
"help wanted",
"contributions-welcome"
] | [] | 2024-02-23T00:43:25Z | 2024-03-25T02:07:03Z | 2024-03-25T02:07:03Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | yiyixuxu | 12,631,849 | MDQ6VXNlcjEyNjMxODQ5 | User | false |
huggingface/diffusers | 2,150,327,586 | I_kwDOHa8MBc6AK2Ui | 7,066 | https://github.com/huggingface/diffusers/issues/7066 | https://api.github.com/repos/huggingface/diffusers/issues/7066 | HELP, can I do img2img on multi imgs at the same time? | In my case, I need to do img2img on multiple images, I saw the parameter "image" can be a list of PIL.image.iamge but when I used a list of images as input, I got an error like below.
> File "/mnt/disk1/1/code/ip-adapter/infer-face-diffusers-lcm-upscaleM1.py", line 107, in ip
image_hires = pipes_img2img[i](ip_a... | open | null | false | 7 | [
"stale"
] | [
"yiyixuxu"
] | 2024-02-23T04:03:07Z | 2024-10-12T15:09:13Z | null | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | blx0102 | 17,761,044 | MDQ6VXNlcjE3NzYxMDQ0 | User | false |
huggingface/diffusers | 2,152,338,756 | I_kwDOHa8MBc6AShVE | 7,079 | https://github.com/huggingface/diffusers/issues/7079 | https://api.github.com/repos/huggingface/diffusers/issues/7079 | Transformer2DModel error when using patches and norm_type=layer_norm | ### Describe the bug
Using the setting norm_type=layer_norm will cause the transformer to use nn.LayerNorm (as intended). But here in the forward code it calls norm.emb(....) on a nn.LayerNorm module
https://github.com/huggingface/diffusers/blob/738c9869572c45619056556673cc3b732455fac7/src/diffusers/models/transfor... | closed | completed | false | 3 | [
"bug"
] | [] | 2024-02-24T14:25:35Z | 2024-03-08T04:11:03Z | 2024-03-08T04:11:03Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | theAdamColton | 72,479,734 | MDQ6VXNlcjcyNDc5NzM0 | User | false |
huggingface/diffusers | 2,152,353,571 | I_kwDOHa8MBc6ASk8j | 7,080 | https://github.com/huggingface/diffusers/issues/7080 | https://api.github.com/repos/huggingface/diffusers/issues/7080 | SDXL control net training inference output defer from wandb | ### Describe the bug
I am training a controlnet using the diffusers script. I have set it to save a checkpoint every 200 steps. However, when I try to use the safetensor file for inference, the output is completely different from the one reported in wandb.
The inference code is the same as the one in the [README]... | closed | completed | false | 11 | [
"bug",
"stale"
] | [] | 2024-02-24T15:05:41Z | 2024-08-15T07:11:38Z | 2024-06-30T05:41:47Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | brycegoh | 16,372,429 | MDQ6VXNlcjE2MzcyNDI5 | User | false |
huggingface/diffusers | 2,152,550,245 | I_kwDOHa8MBc6ATU9l | 7,082 | https://github.com/huggingface/diffusers/issues/7082 | https://api.github.com/repos/huggingface/diffusers/issues/7082 | [Bug:V0.26.3] Cannot import name 'CpuOffload' from 'accelerate.hooks' | ### Describe the bug
Thanks for your great work.
```
RuntimeError: Failed to import diffusers.models.unets.unet_2d_condition because of the following error (look up to see its traceback):
Failed to import diffusers.loaders.unet because of the following error (look up to see its traceback):
cannot import name '... | closed | completed | false | 3 | [
"bug"
] | [] | 2024-02-25T01:18:16Z | 2024-12-09T06:22:47Z | 2024-03-05T00:46:52Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Aisuko | 8,053,949 | MDQ6VXNlcjgwNTM5NDk= | User | false |
huggingface/diffusers | 2,152,585,186 | I_kwDOHa8MBc6ATdfi | 7,084 | https://github.com/huggingface/diffusers/issues/7084 | https://api.github.com/repos/huggingface/diffusers/issues/7084 | AttributeError in StableDiffusionGLIGENPipeline: 'load_lora_weights' Method Not Found | AttributeError: 'StableDiffusionGLIGENPipeline' object has no attribute 'load_lora_weights'
https://github.com/huggingface/diffusers/blob/66f94eaa0c68a893b2aba1ec9f79ee7890786fba/src/diffusers/pipelines/stable_diffusion_gligen/pipeline_stable_diffusion_gligen.py#L102
How to integrate the lora loaders into this pi... | closed | completed | false | 2 | [] | [] | 2024-02-25T03:48:43Z | 2024-03-01T07:06:49Z | 2024-03-01T07:06:49Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | maliozer | 1,297,458 | MDQ6VXNlcjEyOTc0NTg= | User | false |
huggingface/diffusers | 2,152,624,505 | I_kwDOHa8MBc6ATnF5 | 7,085 | https://github.com/huggingface/diffusers/issues/7085 | https://api.github.com/repos/huggingface/diffusers/issues/7085 | how to train controlnet with lora? | train full controlnet need much resource and time, so how to train controlnet with lora?
| closed | completed | false | 1 | [
"should-move-to-discussion"
] | [] | 2024-02-25T06:31:47Z | 2024-03-03T06:38:35Z | 2024-03-03T06:38:35Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | akk-123 | 98,469,560 | U_kgDOBd6GuA | User | false |
huggingface/diffusers | 2,152,804,576 | I_kwDOHa8MBc6AUTDg | 7,088 | https://github.com/huggingface/diffusers/issues/7088 | https://api.github.com/repos/huggingface/diffusers/issues/7088 | Vague error: `ValueError: With local_files_only set to False, you must first locally save the tokenizer in the following path: 'openai/clip-vit-large-patch14'.` how to fix? | Trying to convert a .safetensors stable diffusion model to whatever the format is that hugging face requires. It throws a vague nonsequitur of an error:
`pipe = diffusers.StableDiffusionPipeline.from_single_file(str(aPathlibPath/"vodkaByFollowfoxAI_v40.safetensors") )`
```...
[1241](file:///C:/Users/openSourc... | closed | completed | false | 13 | [
"stale",
"single_file"
] | [
"DN6"
] | 2024-02-25T15:03:07Z | 2024-09-17T21:56:26Z | 2024-09-17T21:56:26Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | openSourcerer9000 | 61,931,826 | MDQ6VXNlcjYxOTMxODI2 | User | false |
huggingface/diffusers | 1,395,097,161 | I_kwDOHa8MBc5TJ35J | 709 | https://github.com/huggingface/diffusers/issues/709 | https://api.github.com/repos/huggingface/diffusers/issues/709 | [Documentation] "r" as method docstring for forward | ### Describe the bug
As seen [here](https://github.com/huggingface/diffusers/blame/688031c592a08832387761971f1e6ca504a5900b/src/diffusers/models/unet_2d.py#L173), the forward docstrings of the UNet classes have "r" as the method description. Why is this?
https://github.com/huggingface/diffusers/blob/688031c592a0883... | closed | completed | false | 1 | [
"bug"
] | [] | 2022-10-03T17:44:23Z | 2022-10-04T13:06:12Z | 2022-10-04T13:06:12Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | natolambert | 10,695,622 | MDQ6VXNlcjEwNjk1NjIy | User | false |
huggingface/diffusers | 2,152,837,966 | I_kwDOHa8MBc6AUbNO | 7,090 | https://github.com/huggingface/diffusers/issues/7090 | https://api.github.com/repos/huggingface/diffusers/issues/7090 | "KeyError: 'embedding'" in "InstantID Pipeline" | ### Describe the bug
I use the code of "InstantID Pipeline":
```# !pip install opencv-python transformers accelerate insightface
import diffusers
from diffusers.utils import load_image
from diffusers.models import ControlNetModel
import cv2
import torch
import numpy as np
from PIL import Image
from insi... | closed | completed | false | 8 | [
"bug"
] | [] | 2024-02-25T16:22:58Z | 2025-07-17T10:08:37Z | 2024-03-04T03:42:11Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | ScottishFold007 | 36,957,508 | MDQ6VXNlcjM2OTU3NTA4 | User | false |
huggingface/diffusers | 2,153,377,432 | I_kwDOHa8MBc6AWe6Y | 7,098 | https://github.com/huggingface/diffusers/issues/7098 | https://api.github.com/repos/huggingface/diffusers/issues/7098 | ControlneXS doesn't work right | ### Describe the bug
@sayakpaul
https://github.com/huggingface/diffusers/blob/main/examples/research_projects/controlnetxs/infer_sd_controlnetxs.py
![cnxs_sd canny](https://github.com/huggingface/diffusers/as... | closed | completed | false | 7 | [
"bug"
] | [] | 2024-02-26T06:43:19Z | 2024-02-29T20:12:18Z | 2024-02-27T08:13:06Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | GodOfSmallThings | 20,166,060 | MDQ6VXNlcjIwMTY2MDYw | User | false |
huggingface/diffusers | 2,154,137,723 | I_kwDOHa8MBc6AZYh7 | 7,100 | https://github.com/huggingface/diffusers/issues/7100 | https://api.github.com/repos/huggingface/diffusers/issues/7100 | Invalid pipeline links | null | closed | completed | false | 0 | [] | [] | 2024-02-26T13:06:53Z | 2024-04-14T14:45:12Z | 2024-02-26T13:10:10Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | zyinghua | 49,605,250 | MDQ6VXNlcjQ5NjA1MjUw | User | false |
huggingface/diffusers | 2,154,152,419 | I_kwDOHa8MBc6AZcHj | 7,101 | https://github.com/huggingface/diffusers/issues/7101 | https://api.github.com/repos/huggingface/diffusers/issues/7101 | can't use animevae.pt | ### Describe the bug
tried to use https://huggingface.co/akibanzu/animevae/tree/main
but it fails to be loaded
### Reproduction
download this: https://huggingface.co/akibanzu/animevae/resolve/main/animevae.pt?download=true
then use
```
import argparse
import io
import requests
import torch
import yaml... | closed | completed | false | 6 | [
"bug"
] | [] | 2024-02-26T13:14:11Z | 2024-02-27T15:25:21Z | 2024-02-27T15:25:21Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | alexblattner | 15,870,094 | MDQ6VXNlcjE1ODcwMDk0 | User | false |
huggingface/diffusers | 2,154,341,939 | I_kwDOHa8MBc6AaKYz | 7,102 | https://github.com/huggingface/diffusers/issues/7102 | https://api.github.com/repos/huggingface/diffusers/issues/7102 | StableDiffusionControlNetPipeline - MulticontrolNet & Multi Prompt Will Cause Bug | ### Describe the bug
When I use StableDiffusionControlNetPipeline, I have 2 controlnets, and i want to generate 4 prompt in one inference, i found a bug when the pipeline are checking params.
As the doc said, the params for image and prompt are as following:
```
pipeline(
prompt = [prompt,prompt,prompt,pro... | closed | completed | false | 2 | [
"bug",
"stale"
] | [] | 2024-02-26T14:37:41Z | 2024-03-29T09:53:24Z | 2024-03-29T09:53:24Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Fantast616 | 45,848,260 | MDQ6VXNlcjQ1ODQ4MjYw | User | false |
huggingface/diffusers | 2,156,028,591 | I_kwDOHa8MBc6AgmKv | 7,112 | https://github.com/huggingface/diffusers/issues/7112 | https://api.github.com/repos/huggingface/diffusers/issues/7112 | IP adapter image embeds dimension | ### Describe the bug
https://github.com/huggingface/diffusers/blob/main/src/diffusers/pipelines/stable_diffusion/pipeline_stable_diffusion.py#L537C1-L538C1
In this line, ```single_image_embeds = torch.stack([single_image_embeds] * num_images_per_prompt, dim=0)``` will add a new dimension to single_image_embeds,maki... | closed | completed | false | 4 | [
"bug",
"stale"
] | [] | 2024-02-27T08:55:16Z | 2024-03-29T02:49:22Z | 2024-03-29T02:49:22Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | caiqi | 8,349,599 | MDQ6VXNlcjgzNDk1OTk= | User | false |
huggingface/diffusers | 2,156,237,467 | I_kwDOHa8MBc6AhZKb | 7,114 | https://github.com/huggingface/diffusers/issues/7114 | https://api.github.com/repos/huggingface/diffusers/issues/7114 | train_text_to_image_sdxl.py error:CUDA out of memory | ### Describe the bug
I tried to run the following script to train the sdxl_model, but encountered a CUDA out of memory error. After debugging, I found that during training: unet.dtype=torch.float32. I tried adding [unet.to](http://unet.to/)(accelerator.device, dtype=weight_dtype), but encountered another error: Atte... | closed | completed | false | 11 | [
"bug",
"stale",
"training"
] | [] | 2024-02-27T10:31:42Z | 2024-06-06T03:25:51Z | 2024-03-29T02:14:43Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Siannodel | 22,951,233 | MDQ6VXNlcjIyOTUxMjMz | User | false |
huggingface/diffusers | 2,156,469,155 | I_kwDOHa8MBc6AiRuj | 7,119 | https://github.com/huggingface/diffusers/issues/7119 | https://api.github.com/repos/huggingface/diffusers/issues/7119 | [`Docs`] Invalid link | ### Describe the bug
[`mask_image`'s link](https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/diffusers/ip_adapter_mask.png) is not valid at [the Inpainting section of General Tasks](https://huggingface.co/docs/diffusers/main/en/using-diffusers/ip_adapter?tasks=Inpainting#general-tasks).
... | closed | completed | false | 2 | [
"bug"
] | [] | 2024-02-27T12:30:15Z | 2024-02-27T19:37:10Z | 2024-02-27T19:26:40Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | tolgacangoz | 46,008,593 | MDQ6VXNlcjQ2MDA4NTkz | User | false |
huggingface/diffusers | 1,395,526,803 | I_kwDOHa8MBc5TLgyT | 712 | https://github.com/huggingface/diffusers/issues/712 | https://api.github.com/repos/huggingface/diffusers/issues/712 | Dreambooth broken, possibly because of ADAM optimizer, possibly more. | I think Huggingface's Dreambooth is the only popular SD implementation that also uses Prior Preservation Loss, so I've been motivated to get it working, but the results have been terrible, and the entire model degrades, regardless of: # timesteps, learning rate, PPL turned on/off, # instance samples, # class regulariza... | closed | completed | false | 41 | [] | [
"patil-suraj"
] | 2022-10-04T00:57:37Z | 2023-07-18T05:49:51Z | 2022-10-21T21:49:13Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | affableroots | 114,898,764 | U_kgDOBtk3TA | User | false |
huggingface/diffusers | 1,395,790,156 | I_kwDOHa8MBc5TMhFM | 713 | https://github.com/huggingface/diffusers/issues/713 | https://api.github.com/repos/huggingface/diffusers/issues/713 | Error during training | Hey... I keep receiving this error every time i try training. Could you please help me resolve it. The only changes i make to the code is changing the input and output directory and removing the 8bit adam flag as i get assigned an a100
Caching latents: 100% 12/12 [00:01<00:00, 6.91it/s]
Steps: 0% 0/1000 [00:00<... | closed | completed | false | 0 | [] | [] | 2022-10-04T07:27:53Z | 2022-10-04T07:29:03Z | 2022-10-04T07:29:03Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Hans120791 | 93,925,794 | U_kgDOBZkxog | User | false |
huggingface/diffusers | 2,158,429,863 | I_kwDOHa8MBc6Apwan | 7,130 | https://github.com/huggingface/diffusers/issues/7130 | https://api.github.com/repos/huggingface/diffusers/issues/7130 | from stable_diffusion_pipeline import StableDiffusionPipeline | ### Model/Pipeline/Scheduler description
how can i load my own llm,like Chillout and real?
### Open source status
- [X] The model implementation is available.
- [X] The model weights are available (Only relevant if addition is not a scheduler).
### Provide useful links for the implementation
_No response_ | closed | completed | false | 12 | [] | [] | 2024-02-28T08:51:11Z | 2024-11-20T00:32:25Z | 2024-11-20T00:32:22Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | henbucuoshanghai | 37,647,094 | MDQ6VXNlcjM3NjQ3MDk0 | User | false |
huggingface/diffusers | 2,159,550,680 | I_kwDOHa8MBc6AuCDY | 7,133 | https://github.com/huggingface/diffusers/issues/7133 | https://api.github.com/repos/huggingface/diffusers/issues/7133 | SDXL Dreambooth LoRA - size mismatch in text encoders when resuming from checkpoint | ### Describe the bug
When resuming sdxl dreambooth lora training from a checkpoint, it results in a size mismatch error in the text encoder.
### Reproduction
Command being run each time is:
```
accelerate launch train_dreambooth_lora_sdxl.py --pretrained_model_name_or_path=stabilityai/stable-diffusion-xl-bas... | closed | completed | false | 4 | [
"bug"
] | [] | 2024-02-28T18:14:08Z | 2024-03-01T14:00:49Z | 2024-03-01T14:00:49Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | shawnrushefsky | 6,685,047 | MDQ6VXNlcjY2ODUwNDc= | User | false |
huggingface/diffusers | 2,160,022,896 | I_kwDOHa8MBc6Av1Vw | 7,136 | https://github.com/huggingface/diffusers/issues/7136 | https://api.github.com/repos/huggingface/diffusers/issues/7136 | adding `callback_on_step_end` for `StableDiffusionLDM3DPipeline` | Just realize the StableDiffusionLDM3DPipeline is still not using the new callback API
https://github.com/huggingface/diffusers/blob/main/src/diffusers/pipelines/stable_diffusion_ldm3d/pipeline_stable_diffusion_ldm3d.py
Is anyone interested in refactoring it? you can use this PR as a reference
https://github.com/... | closed | completed | false | 2 | [
"good first issue",
"help wanted",
"contributions-welcome"
] | [] | 2024-02-28T23:02:26Z | 2024-03-04T04:44:15Z | 2024-03-04T04:44:15Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | yiyixuxu | 12,631,849 | MDQ6VXNlcjEyNjMxODQ5 | User | false |
huggingface/diffusers | 2,160,164,324 | I_kwDOHa8MBc6AwX3k | 7,138 | https://github.com/huggingface/diffusers/issues/7138 | https://api.github.com/repos/huggingface/diffusers/issues/7138 | Structure of unet in SDXL mismatch the structure of the official model | ### Describe the bug
I'm trying to train the SDXL with pokemon-blip-captions dataset on the basis of pretrained weight. I have downloaded the whole repo form [stabilityai/stable-diffusion-xl-base-1.0](https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0/tree/main), while the ValueError caused by structure... | closed | completed | false | 3 | [
"bug"
] | [] | 2024-02-29T01:22:55Z | 2024-03-02T02:09:57Z | 2024-03-02T02:09:57Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | MengqingCao | 52,243,582 | MDQ6VXNlcjUyMjQzNTgy | User | false |
huggingface/diffusers | 1,395,961,400 | I_kwDOHa8MBc5TNK44 | 714 | https://github.com/huggingface/diffusers/issues/714 | https://api.github.com/repos/huggingface/diffusers/issues/714 | Parallel dataloader for unconditional diffusion example | **Is your feature request related to a problem? Please describe.**
The loading of the samples is sequential and without prefetching n the `train_unconditional.py` script.
This results in an underutilization of the GPU if large images and/or "slow" IO are used.
**Describe the solution you'd like**
Adding a paramet... | closed | completed | false | 3 | [] | [
"anton-l"
] | 2022-10-04T09:43:09Z | 2022-10-27T13:55:37Z | 2022-10-27T13:55:37Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Flova | 15,075,613 | MDQ6VXNlcjE1MDc1NjEz | User | false |
huggingface/diffusers | 2,160,283,182 | I_kwDOHa8MBc6Aw04u | 7,142 | https://github.com/huggingface/diffusers/issues/7142 | https://api.github.com/repos/huggingface/diffusers/issues/7142 | [SDXL-lightning Inference] diffusers inference result is different with webui | ### Describe the bug
I use diffuser to generate the image with dreamshaper_lightning model, but the result is really strange.
Model url:https://civitai.com/models/112902?modelVersionId=354657
### ... | closed | completed | false | 5 | [
"bug"
] | [] | 2024-02-29T03:32:16Z | 2024-03-01T07:02:31Z | 2024-03-01T07:02:11Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | ApolloRay | 36,093,263 | MDQ6VXNlcjM2MDkzMjYz | User | false |
huggingface/diffusers | 2,160,827,438 | I_kwDOHa8MBc6Ay5wu | 7,146 | https://github.com/huggingface/diffusers/issues/7146 | https://api.github.com/repos/huggingface/diffusers/issues/7146 | CUDA OOM Error when fine tuning SDXL using train_text_to_image_sdxl.py | ### Describe the bug
Hello, I'm trying to fine tune SDXL using the train_text_to_image_sdxl.py file with the following parameters and I'm getting a CUDA Out of Memory error even before any training begins. I've also tried to train it on a custom dataset of 70 images (with captions) and I'm facing the same error.
I'm ... | closed | completed | false | 6 | [
"bug"
] | [] | 2024-02-29T09:41:06Z | 2024-03-01T04:20:19Z | 2024-03-01T04:20:19Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | DarshSamyak | 146,829,267 | U_kgDOCMBv0w | User | false |
huggingface/diffusers | 2,160,972,517 | I_kwDOHa8MBc6AzdLl | 7,147 | https://github.com/huggingface/diffusers/issues/7147 | https://api.github.com/repos/huggingface/diffusers/issues/7147 | Cannot Import Name 'StableCascadeDecoderPipeline' | ### Describe the bug
I'm running the following command in Google Colab but an error occurs.
### Reproduction
!pip install diffusers
import torch
from diffusers import StableCascadeDecoderPipeline, StableCascadePriorPipeline
### Logs
```shell
ImportError Traceback (most re... | closed | completed | false | 4 | [
"bug"
] | [] | 2024-02-29T10:54:00Z | 2024-03-01T06:47:13Z | 2024-03-01T06:47:12Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | HashingTag | 24,617,729 | MDQ6VXNlcjI0NjE3NzI5 | User | false |
huggingface/diffusers | 2,161,463,306 | I_kwDOHa8MBc6A1VAK | 7,148 | https://github.com/huggingface/diffusers/issues/7148 | https://api.github.com/repos/huggingface/diffusers/issues/7148 | Issues with Model Loading During Inference | I created a custom pipeline and can successfully resume training from a checkpoint, achieving satisfactory validation results. However, upon loading a saved model for inference, the results were incorrect. It seems that the model was not properly saved or loaded. Below are the steps I followed to save the model in the ... | closed | completed | false | 2 | [] | [] | 2024-02-29T15:00:03Z | 2024-03-04T15:56:08Z | 2024-03-04T15:56:08Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | hanweikung | 643,825 | MDQ6VXNlcjY0MzgyNQ== | User | false |
huggingface/diffusers | 2,162,020,524 | I_kwDOHa8MBc6A3dCs | 7,152 | https://github.com/huggingface/diffusers/issues/7152 | https://api.github.com/repos/huggingface/diffusers/issues/7152 | Support torch._subclksses.EulerDiscreteScheduler scheduler | ### Describe the bug
Slightly related to https://github.com/huggingface/accelerate/issues/2494 in the sense it applies the suggested workaround to avoid the big modeling feature (not supported by `FakeTensorMode`).
Once the model is initialized, it hits an issue within `EulerDiscreteScheduler` when `numpy.array` is... | closed | completed | false | 11 | [
"bug"
] | [] | 2024-02-29T20:11:10Z | 2024-03-04T19:20:01Z | 2024-03-04T19:20:00Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | thiagocrepaldi | 5,469,809 | MDQ6VXNlcjU0Njk4MDk= | User | false |
huggingface/diffusers | 2,162,457,038 | I_kwDOHa8MBc6A5HnO | 7,157 | https://github.com/huggingface/diffusers/issues/7157 | https://api.github.com/repos/huggingface/diffusers/issues/7157 | [LCM-LoRA] unexpected keys while loading LCM LoRA with `peft` | Code to reproduce:
```python
from diffusers import LCMScheduler, StableDiffusionXLPipeline
import torch
pipeline = StableDiffusionXLPipeline.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0", torch_dtype=torch.float16).to("cuda")
pipeline.load_lora_weights("latent-consistency/lcm-lora-sdxl")
pipeline... | closed | completed | false | 7 | [] | [] | 2024-03-01T02:23:26Z | 2024-11-19T10:17:11Z | 2024-11-19T10:17:11Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | sayakpaul | 22,957,388 | MDQ6VXNlcjIyOTU3Mzg4 | User | false |
huggingface/diffusers | 2,162,468,781 | I_kwDOHa8MBc6A5Ket | 7,159 | https://github.com/huggingface/diffusers/issues/7159 | https://api.github.com/repos/huggingface/diffusers/issues/7159 | [LoRA] torch compile is failing even when LoRA is fused. | Code to reproduce:
```python
from diffusers import LCMScheduler, StableDiffusionXLPipeline
import torch
pipeline = StableDiffusionXLPipeline.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0", torch_dtype=torch.float16).to("cuda")
pipeline.load_lora_weights("latent-consistency/lcm-lora-sdxl")
pipeline... | closed | completed | false | 2 | [] | [] | 2024-03-01T02:35:19Z | 2024-03-22T06:58:31Z | 2024-03-01T03:34:16Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | sayakpaul | 22,957,388 | MDQ6VXNlcjIyOTU3Mzg4 | User | false |
huggingface/diffusers | 2,162,841,617 | I_kwDOHa8MBc6A6lgR | 7,165 | https://github.com/huggingface/diffusers/issues/7165 | https://api.github.com/repos/huggingface/diffusers/issues/7165 | Wrong scale of dynamic thresholding in Jax implementation | https://github.com/huggingface/diffusers/blob/f4fc75035f2d567dfe00fc6b8671ecd38dd144b2/src/diffusers/schedulers/scheduling_dpmsolver_multistep_flax.py#L164
Apparently, `q` in [jnp.percentile](https://jax.readthedocs.io/en/latest/_autosummary/jax.numpy.percentile.html) should between [0, 100] instead of [0, 1] as in ... | open | null | false | 3 | [
"stale",
"jax/flax"
] | [
"pcuenca"
] | 2024-03-01T08:03:16Z | 2024-04-25T15:03:38Z | null | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | alex8937 | 14,084,517 | MDQ6VXNlcjE0MDg0NTE3 | User | false |
huggingface/diffusers | 2,162,922,675 | I_kwDOHa8MBc6A65Sz | 7,168 | https://github.com/huggingface/diffusers/issues/7168 | https://api.github.com/repos/huggingface/diffusers/issues/7168 | How can I get correct ip adapter image embeds? I got 4D tensors and I cannnot use it. | ### Describe the bug
IP Adapter image embed should be 3D tensors. But I got 4D tensors.
### Reproduction
```python
import torch
from diffusers import AutoPipelineForText2Image, DDIMScheduler
from diffusers.utils import load_image
pipeline = AutoPipelineForText2Image.from_pretrained(
"stabilityai/sta... | closed | completed | false | 6 | [
"bug"
] | [] | 2024-03-01T08:56:39Z | 2024-05-06T23:26:12Z | 2024-03-03T07:37:01Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | dai-ichiro | 53,703,225 | MDQ6VXNlcjUzNzAzMjI1 | User | false |
huggingface/diffusers | 2,164,627,599 | I_kwDOHa8MBc6BBZiP | 7,177 | https://github.com/huggingface/diffusers/issues/7177 | https://api.github.com/repos/huggingface/diffusers/issues/7177 | error in StableCascadeDecoderPipeline.from_pretrained | ### Describe the bug
When I used "StableCascadeDecoderPipeline.from_pretrained", the error accured "AttributeError: module diffusers.pipelines.stable_cascade has no attribute StableCascadeUNet". Can someone help me? @DN6 @
### Reproduction
from diffusers import StableCascadeDecoderPipeline
StableCascadeDecoderPipel... | closed | completed | false | 27 | [
"bug",
"stale"
] | [] | 2024-03-02T06:07:25Z | 2024-05-26T15:03:44Z | 2024-04-15T14:45:12Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | hahaha-cc | 110,004,067 | U_kgDOBo6HYw | User | false |
huggingface/diffusers | 2,164,695,653 | I_kwDOHa8MBc6BBqJl | 7,178 | https://github.com/huggingface/diffusers/issues/7178 | https://api.github.com/repos/huggingface/diffusers/issues/7178 | export_to_video does not work properly | ### Describe the bug
The ```export_to_video``` function in the [export_utils](https://github.com/huggingface/diffusers/blob/main/src/diffusers/utils/export_utils.py) file does not work properly. I think the problem is the dimensions.
### Reproduction
```
import torch
from diffusers import DiffusionPipeline, DPMSol... | closed | completed | false | 1 | [
"bug"
] | [] | 2024-03-02T08:12:19Z | 2024-03-02T08:28:30Z | 2024-03-02T08:28:07Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | harshvardhan2707 | 72,443,636 | MDQ6VXNlcjcyNDQzNjM2 | User | false |
huggingface/diffusers | 2,164,737,557 | I_kwDOHa8MBc6BB0YV | 7,180 | https://github.com/huggingface/diffusers/issues/7180 | https://api.github.com/repos/huggingface/diffusers/issues/7180 | SDXL Image2Image Controlnet pipeline fails with cpu offload | ### Describe the bug
I am running SDXL-lightning with a canny edge controlnet. Running on a T4 (16G VRAM).
I am using `enable_model_cpu_offload` to reduce memory usage, but I am running into the following error:
`mat1 and mat2 must have the same dtype`
I am guessing something is wrong with the offloading of som... | closed | completed | false | 1 | [
"bug"
] | [] | 2024-03-02T09:54:55Z | 2024-03-03T04:54:48Z | 2024-03-03T04:54:48Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | odusseys | 7,863,870 | MDQ6VXNlcjc4NjM4NzA= | User | false |
huggingface/diffusers | 2,164,784,637 | I_kwDOHa8MBc6BB_39 | 7,183 | https://github.com/huggingface/diffusers/issues/7183 | https://api.github.com/repos/huggingface/diffusers/issues/7183 | scheduler.config._class_name displays wrong class name if the scheduler is changed | ### Describe the bug
```py
print(pipe.scheduler._class_name)
print(pipe.scheduler.config._class_name)
```
Both print the wrong class name if the scheduler is changed using the from_config class method, they print the original scheduler class name if you change the scheduler a few times
change the scheduler ... | closed | completed | false | 2 | [
"bug"
] | [] | 2024-03-02T12:16:44Z | 2024-03-04T17:19:00Z | 2024-03-04T17:19:00Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Vargol | 62,868 | MDQ6VXNlcjYyODY4 | User | false |
huggingface/diffusers | 2,164,825,266 | I_kwDOHa8MBc6BCJyy | 7,185 | https://github.com/huggingface/diffusers/issues/7185 | https://api.github.com/repos/huggingface/diffusers/issues/7185 | Does diffusers does not support pytorch 2.2.0? | ### Describe the bug
When I install diffusers in a environment with torch 2.2.0, I see it degrade torch into 2.1.2, and mixed my environments.
### Reproduction
When I install diffusers in a environment with torch 2.2.0, I see it degrade torch into 2.1.2, and mixed my environments.
### Logs
```shell
When I install ... | closed | completed | false | 4 | [
"bug"
] | [] | 2024-03-02T14:07:54Z | 2024-03-03T09:24:21Z | 2024-03-03T06:37:09Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Luciennnnnnn | 20,135,317 | MDQ6VXNlcjIwMTM1MzE3 | User | false |
huggingface/diffusers | 2,164,912,932 | I_kwDOHa8MBc6BCfMk | 7,188 | https://github.com/huggingface/diffusers/issues/7188 | https://api.github.com/repos/huggingface/diffusers/issues/7188 | "Nan" in StableDiffusionXLPipeline's vae encoder | When I use the vae encoder of StableDiffusionXLPipeline to encode an image of [B, 3, 1024, 1024], the latents output is "Nan" value. Is there any solutions?
```
pipe = StableDiffusionXLPipeline.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0", torch_dtype=torch.float16, use_safetensors=True, variant... | closed | completed | false | 7 | [] | [] | 2024-03-02T18:08:48Z | 2025-04-03T04:11:44Z | 2024-03-03T12:29:03Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | tuning12 | 134,306,342 | U_kgDOCAFaJg | User | false |
huggingface/diffusers | 2,165,050,453 | I_kwDOHa8MBc6BDAxV | 7,190 | https://github.com/huggingface/diffusers/issues/7190 | https://api.github.com/repos/huggingface/diffusers/issues/7190 | Trajectory Consistency Distillation | ### Model/Pipeline/Scheduler description
Latent Consistency Model (LCM) extends the Consistency Model to the latent space and leverages the guided consistency distillation technique to achieve impressive performance in accelerating text-to-image synthesis. However, LCM struggles to generate images with both clarity ... | closed | completed | false | 3 | [
"stale"
] | [] | 2024-03-03T00:34:54Z | 2024-04-02T15:03:37Z | 2024-04-02T15:03:36Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | clarencechen | 8,482,341 | MDQ6VXNlcjg0ODIzNDE= | User | false |
huggingface/diffusers | 2,165,110,396 | I_kwDOHa8MBc6BDPZ8 | 7,191 | https://github.com/huggingface/diffusers/issues/7191 | https://api.github.com/repos/huggingface/diffusers/issues/7191 | Strange import problem | ### Describe the bug
<img width="890" alt="image" src="https://github.com/huggingface/diffusers/assets/20135317/60a73583-4790-4975-989b-49ca19c890da">
I just can not import this loader
### Reproduction
from diffusers.loaders.controlnet import FromOriginalControlnetMixin
### Logs
_No response_
### System Info
... | closed | completed | false | 2 | [
"bug"
] | [] | 2024-03-03T03:35:35Z | 2024-03-03T04:52:13Z | 2024-03-03T04:37:40Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Luciennnnnnn | 20,135,317 | MDQ6VXNlcjIwMTM1MzE3 | User | false |
huggingface/diffusers | 2,165,293,675 | I_kwDOHa8MBc6BD8Jr | 7,194 | https://github.com/huggingface/diffusers/issues/7194 | https://api.github.com/repos/huggingface/diffusers/issues/7194 | Train IP Adapter | There isn't an example of how to train or finetune your own IP Adapter model currently.
It would be helpful to have an implementation of the how to train/finetune the IP Adapter face id, sdxl and plus versions
I'd like to work on this issue
| closed | completed | false | 7 | [
"stale"
] | [] | 2024-03-03T10:35:17Z | 2024-11-28T05:42:29Z | 2024-11-28T05:42:29Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | AMohamedAakhil | 78,485,654 | MDQ6VXNlcjc4NDg1NjU0 | User | false |
huggingface/diffusers | 2,165,557,737 | I_kwDOHa8MBc6BE8np | 7,197 | https://github.com/huggingface/diffusers/issues/7197 | https://api.github.com/repos/huggingface/diffusers/issues/7197 | Train Text-image.py KeyError: Shortest Edge when resuming training from local SD 1.5 | ### Describe the bug
I am unable to resume training of a locally saved SD 1.5 model (trained using the same text-image.py script, albeit likely an earlier revision). There is an error when loading the CLIP component of the model pipeline. Here is the CLIP configuration:
![image](https://github.com/huggingface/diffu... | closed | completed | false | 3 | [
"bug",
"stale",
"training"
] | [] | 2024-03-03T21:08:50Z | 2024-11-28T05:41:59Z | 2024-11-28T05:41:59Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | suggestiondiabolique | 63,476,827 | MDQ6VXNlcjYzNDc2ODI3 | User | false |
huggingface/diffusers | 2,165,678,496 | I_kwDOHa8MBc6BFaGg | 7,198 | https://github.com/huggingface/diffusers/issues/7198 | https://api.github.com/repos/huggingface/diffusers/issues/7198 | Sanster/anything-4.0-inpainting is missing fp16 files | got this error:
C:\Users\mdmen\anaconda3\envs\myiopaintenv\lib\site-packages\diffusers\pipelines\pipeline_utils.py:275: FutureWarning: You are loading the variant fp16 from Sanster/anything-4.0-inpainting via `revision='fp16'`. This behavior is deprecated and will be removed in diffusers v1. One should use `variant=... | closed | completed | false | 3 | [] | [] | 2024-03-04T00:50:05Z | 2024-11-20T00:30:15Z | 2024-11-20T00:30:13Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | offmybach | 137,804,812 | U_kgDOCDa8DA | User | false |
huggingface/diffusers | 2,165,784,801 | I_kwDOHa8MBc6BF0Dh | 7,199 | https://github.com/huggingface/diffusers/issues/7199 | https://api.github.com/repos/huggingface/diffusers/issues/7199 | ViewFusion: Towards Multi-View Consistency via Interpolated Denoising | ### Model/Pipeline/Scheduler description
Novel-view synthesis through diffusion models has demonstrated remarkable potential for generating diverse and high-quality images. Yet, the independent process of image generation in these prevailing methods leads to challenges in maintaining multiple view consistency. To addr... | closed | completed | false | 6 | [
"community-examples",
"contributions-welcome"
] | [] | 2024-03-04T02:33:52Z | 2024-11-28T05:49:59Z | 2024-11-28T05:49:59Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | clarencechen | 8,482,341 | MDQ6VXNlcjg0ODIzNDE= | User | false |
huggingface/diffusers | 1,396,486,648 | I_kwDOHa8MBc5TPLH4 | 720 | https://github.com/huggingface/diffusers/issues/720 | https://api.github.com/repos/huggingface/diffusers/issues/720 | traced UNet from documentation fails with RuntimeError: expected scalar type Half but found Float | ### Describe the bug
when running the traced UNet example from
https://github.com/huggingface/diffusers/blob/main/docs/source/optimization/fp16.mdx
the error - RuntimeError: expected scalar type Half but found Float
occurs in the section
```
# warmup
for _ in range(3):
with torch.inference_mode():
... | closed | completed | false | 3 | [
"bug"
] | [] | 2022-10-04T15:38:32Z | 2022-10-08T19:02:52Z | 2022-10-04T15:45:55Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Thomas-MMJ | 112,830,596 | U_kgDOBrmohA | User | false |
huggingface/diffusers | 2,166,713,296 | I_kwDOHa8MBc6BJWvQ | 7,207 | https://github.com/huggingface/diffusers/issues/7207 | https://api.github.com/repos/huggingface/diffusers/issues/7207 | 30%+ Speedup for AMD RDNA3/ROCm using Flash Attention w/ SDP Fallback | > ### Discussed in https://github.com/huggingface/diffusers/discussions/7172
>
> <div type='discussions-op-text'>
>
> <sup>Originally posted by **Beinsezii** March 1, 2024</sup>
> Then, to actually *use* flash attention in Diffusers you need to implement it in an attention processor *and* have a fallback for un... | closed | completed | false | 1 | [] | [] | 2024-03-04T12:17:54Z | 2024-03-05T11:06:55Z | 2024-03-05T11:06:54Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | vire70 | 1,501,414 | MDQ6VXNlcjE1MDE0MTQ= | User | false |
huggingface/diffusers | 2,167,032,267 | I_kwDOHa8MBc6BKknL | 7,209 | https://github.com/huggingface/diffusers/issues/7209 | https://api.github.com/repos/huggingface/diffusers/issues/7209 | HTTPSConnectionPool Error When Loading Local safetensor Model Offline with from_single_file function | ### Describe the bug
I greatly appreciate your assistance and look forward to your help.
As the title suggests, I encounter an HTTPSConnectionPool error when attempting to load a local safetensor model offline using the following code:
`sdxl_pipeline = StableDiffusionXLPipeline.from_single_file(model_path, use_saf... | closed | completed | false | 5 | [
"bug"
] | [] | 2024-03-04T14:50:32Z | 2024-03-05T07:56:55Z | 2024-03-05T07:56:55Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | howyoungchen | 57,206,159 | MDQ6VXNlcjU3MjA2MTU5 | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.