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 | 1,402,545,660 | I_kwDOHa8MBc5TmSX8 | 791 | https://github.com/huggingface/diffusers/issues/791 | https://api.github.com/repos/huggingface/diffusers/issues/791 | 8007000E Not enough memory resources are available to complete this operation. | when I run the dml_onnx.py file in (amd_venv) C:\amd-stable-diffusion\difusers-dml\examples\inference>dml_onnx.py
I get an error like this:
(amd_venv) C:\amd-stable-diffusion\difusers-dml\examples\inference>dml_onnx.py
Fetching 19 files: 100%|███████████████████████████████████████████ | 19/19 [00:00<00:00, 1966.1... | closed | completed | false | 7 | [
"stale"
] | [
"anton-l"
] | 2022-10-10T03:36:21Z | 2022-12-09T15:04:22Z | 2022-12-09T15:04:22Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | muhademan | 60,034,198 | MDQ6VXNlcjYwMDM0MTk4 | User | false |
huggingface/diffusers | 2,290,586,903 | I_kwDOHa8MBc6Ih5UX | 7,914 | https://github.com/huggingface/diffusers/issues/7914 | https://api.github.com/repos/huggingface/diffusers/issues/7914 | device_map='balanced' pipe.load_textual_inversion() function error | hi
https://github.com/huggingface/diffusers/blob/b934215d4c376ea2e08e28103443686b95ea772c/src/diffusers/loaders/textual_inversion.py#L458
The code at this reference does not seem to handle device_map='balanced' and will report an error
error msg:
```
diffusers\src\diffusers\pipelines\pipeline_utils.py", lin... | closed | completed | false | 3 | [] | [] | 2024-05-11T01:46:59Z | 2024-05-13T19:35:24Z | 2024-05-13T19:35:23Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Honey-666 | 76,274,942 | MDQ6VXNlcjc2Mjc0OTQy | User | false |
huggingface/diffusers | 2,291,160,852 | I_kwDOHa8MBc6IkFcU | 7,920 | https://github.com/huggingface/diffusers/issues/7920 | https://api.github.com/repos/huggingface/diffusers/issues/7920 | training example for instruct pix2pix doesn't zero out embeds | ### Describe the bug
When running inference on SDXL, the config specifies to zero out the embedding when the prompt is empty.
### Reproduction
```py
# Get null conditioning
def compute_null_conditioning():
null_conditioning_list = []
for a_tokenizer, a_text_encoder in zip(tokenizers... | open | null | false | 9 | [
"bug",
"stale"
] | [] | 2024-05-12T04:26:58Z | 2024-09-14T15:12:47Z | null | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | bghira | 59,658,056 | MDQ6VXNlcjU5NjU4MDU2 | User | false |
huggingface/diffusers | 2,291,212,366 | I_kwDOHa8MBc6IkSBO | 7,921 | https://github.com/huggingface/diffusers/issues/7921 | https://api.github.com/repos/huggingface/diffusers/issues/7921 | how can I use control_guidance_start, end in ip adapter ? | I want to generate image with controlnet pose + ip adapter
I want to use Ip adapter with start 0.0 ~ 0.4 like this
<img width="910" alt="image" src="https://github.com/huggingface/diffusers/assets/8098724/59fd4d38-600b-4860-920f-f6f2b6c94e86">
but I cant find any params to apply control_guidance_start, end fo... | closed | completed | false | 1 | [] | [] | 2024-05-12T07:51:08Z | 2024-05-13T12:18:01Z | 2024-05-13T12:18:01Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | nullarch | 8,098,724 | MDQ6VXNlcjgwOTg3MjQ= | User | false |
huggingface/diffusers | 2,291,231,684 | I_kwDOHa8MBc6IkWvE | 7,922 | https://github.com/huggingface/diffusers/issues/7922 | https://api.github.com/repos/huggingface/diffusers/issues/7922 | AttributeError: module diffusers has no attribute ORTStableDiffusionXLPipeline. Did you mean: 'StableDiffusionXLPipeline | ### Describe the bug
AttributeError: module diffusers has no attribute ORTStableDiffusionXLPipeline. Did you mean: 'StableDiffusionXLPipeline
### Reproduction
even after using it remain same issue
what ... | closed | completed | false | 4 | [
"bug"
] | [] | 2024-05-12T08:52:16Z | 2024-11-02T20:23:59Z | 2024-05-12T17:35:24Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | neonarc4 | 140,290,288 | U_kgDOCFyo8A | User | false |
huggingface/diffusers | 2,291,334,568 | I_kwDOHa8MBc6Ikv2o | 7,925 | https://github.com/huggingface/diffusers/issues/7925 | https://api.github.com/repos/huggingface/diffusers/issues/7925 | About training controlnet with lora and meet problems | ### Describe the bug
I have tried to train a controlnet with lora, but after I merge the controlnet model and lora, the weight of controlnet seems not changed. (maybe inaccurate as i did not check the tensor)
Known:
1. I trained controlnet with lora (actually finetuning unet encoder and some other layers with lo... | closed | completed | false | 2 | [
"bug"
] | [] | 2024-05-12T13:28:16Z | 2024-07-05T02:45:36Z | 2024-06-14T13:55:19Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | lavinal712 | 98,888,959 | U_kgDOBeTs_w | User | false |
huggingface/diffusers | 2,291,671,027 | I_kwDOHa8MBc6ImB_z | 7,928 | https://github.com/huggingface/diffusers/issues/7928 | https://api.github.com/repos/huggingface/diffusers/issues/7928 | support Kandinsky 3.1 | Kandinsky 3.1 has updated a few new features(including Kandinsky Flash) a few weeks ago which is yet supported in diffusers.
Please check out https://github.com/ai-forever/Kandinsky-3
| closed | completed | false | 3 | [] | [] | 2024-05-13T02:23:42Z | 2024-10-16T00:15:39Z | 2024-10-16T00:15:39Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | EYcab | 11,568,206 | MDQ6VXNlcjExNTY4MjA2 | User | false |
huggingface/diffusers | 2,292,595,703 | I_kwDOHa8MBc6Ipjv3 | 7,931 | https://github.com/huggingface/diffusers/issues/7931 | https://api.github.com/repos/huggingface/diffusers/issues/7931 | `StableDiffusionXLControlNetInpaintPipeline` not working with IP-Adapter when using `ip_adapter_image_embeds` parameter. | ### Describe the bug
`StableDiffusionXLControlNetInpaintPipeline` not working with IP-Adapter when using `ip_adapter_image_embeds` parameter.
This PR solves the issue: #7924
### Reproduction
```python
from diffusers import StableDiffusionXLControlNetInpaintPipeline, ControlNetModel
from diffusers.utils import ... | closed | completed | false | 1 | [
"bug"
] | [] | 2024-05-13T11:50:09Z | 2024-05-16T03:16:47Z | 2024-05-16T03:16:47Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | detkov | 34,102,070 | MDQ6VXNlcjM0MTAyMDcw | User | false |
huggingface/diffusers | 2,292,923,049 | I_kwDOHa8MBc6Iqzqp | 7,935 | https://github.com/huggingface/diffusers/issues/7935 | https://api.github.com/repos/huggingface/diffusers/issues/7935 | stabilityai/stable-diffusion-2-1 does not appear to have a file named config.json. | ### Describe the bug
/home/pcl/Minconda/conda3/envs/DIFT/bin/python3.10 /home/pcl/DETR/SDAseg/others/DIFT-main/demo.py
Traceback (most recent call last):
File "/home/pcl/Minconda/conda3/envs/DIFT/lib/python3.10/site-packages/diffusers/configuration_utils.py", line 337, in load_config
config_file = hf_hub_d... | open | null | false | 5 | [
"bug",
"stale"
] | [] | 2024-05-13T14:02:45Z | 2024-11-14T15:06:11Z | null | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | HaoqianSong | 99,889,442 | U_kgDOBfQxIg | User | false |
huggingface/diffusers | 2,294,004,203 | I_kwDOHa8MBc6Iu7nr | 7,936 | https://github.com/huggingface/diffusers/issues/7936 | https://api.github.com/repos/huggingface/diffusers/issues/7936 | Load local safetensors file raised invalid json format | ### Describe the bug
I have this Dockerfile to download the checkpoint from Dreamshaper XL Turbo but when I tried to load the checkpoint with `AutoPipelineForText2Image.from_pretrained` I got the following error:
```
#12 35.68 Traceback (most recent call last):
#12 35.68 File "/app/src/download_weights.py", lin... | closed | completed | false | 7 | [
"bug",
"single_file"
] | [
"DN6"
] | 2024-05-13T23:18:47Z | 2024-05-30T23:58:18Z | 2024-05-30T23:58:18Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | thatlamguy | 683,245 | MDQ6VXNlcjY4MzI0NQ== | User | false |
huggingface/diffusers | 2,294,384,842 | I_kwDOHa8MBc6IwYjK | 7,939 | https://github.com/huggingface/diffusers/issues/7939 | https://api.github.com/repos/huggingface/diffusers/issues/7939 | fail to load modal from same folder | ### Describe the bug
it cant load model just loading it from link not from disk
### Reproduction
```py
from diffusers import OnnxRuntimeModel, DiffusionPipeline
import torch
pipeline = DiffusionPipeline.from_pretrained(
"anything-v3-fp16-pruned.safetensors",
use_safetensors=True,
revision="f... | closed | completed | false | 2 | [
"bug"
] | [] | 2024-05-14T04:37:59Z | 2024-05-14T07:30:50Z | 2024-05-14T07:24:39Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | neonarc4 | 140,290,288 | U_kgDOCFyo8A | User | false |
huggingface/diffusers | 1,403,406,524 | I_kwDOHa8MBc5Tpki8 | 794 | https://github.com/huggingface/diffusers/issues/794 | https://api.github.com/repos/huggingface/diffusers/issues/794 | Better error message for unsupported classes | Case in point: attempting to load a pipeline that uses the k-LMS scheduler using `from_pretrained` without `scipy` installed fails [here](https://github.com/huggingface/diffusers/blob/v0.4.1/src/diffusers/pipeline_utils.py#L457) because `load_method_name` is `None`. The error the user sees is:
```
Traceback (most r... | closed | completed | false | 0 | [] | [] | 2022-10-10T16:04:44Z | 2022-10-12T12:41:17Z | 2022-10-12T12:41:17Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | pcuenca | 1,177,582 | MDQ6VXNlcjExNzc1ODI= | User | false |
huggingface/diffusers | 2,294,449,699 | I_kwDOHa8MBc6IwoYj | 7,940 | https://github.com/huggingface/diffusers/issues/7940 | https://api.github.com/repos/huggingface/diffusers/issues/7940 | [ko] Translating docs(Conceptual Guides) to Korean. | Hello!
Thank you for creating a wonderful repository. I would like to start translation about these four documents in Korean.
- [x] [Philosophy](https://github.com/huggingface/diffusers/blob/main/docs/source/en/conceptual/philosophy.md)
- [x] [Diffusers’ Ethical Guidelines](https://github.com/huggingface/diffusers... | closed | completed | false | 11 | [
"stale"
] | [] | 2024-05-14T05:37:40Z | 2024-09-15T07:22:30Z | 2024-09-15T07:22:30Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | mreraser | 33,192,762 | MDQ6VXNlcjMzMTkyNzYy | User | false |
huggingface/diffusers | 2,295,191,924 | I_kwDOHa8MBc6Izdl0 | 7,945 | https://github.com/huggingface/diffusers/issues/7945 | https://api.github.com/repos/huggingface/diffusers/issues/7945 | Combining community pipeline for image generation | ### Describe the bug
I cannot use both stable diffusion XL [reference ](https://github.com/Mikubill/sd-webui-controlnet/discussions/1236) and [Instant ID](https://github.com/InstantID/InstantID) in the same pipeline. I get `'FrozenDict' object has no attribute 'block_out_channels'"`
### Reproduction
```
from stable... | open | null | false | 4 | [
"bug",
"stale"
] | [] | 2024-05-14T11:36:59Z | 2024-09-14T15:12:25Z | null | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Fqlox | 33,659,731 | MDQ6VXNlcjMzNjU5NzMx | User | false |
huggingface/diffusers | 2,295,689,739 | I_kwDOHa8MBc6I1XIL | 7,946 | https://github.com/huggingface/diffusers/issues/7946 | https://api.github.com/repos/huggingface/diffusers/issues/7946 | custom_pipeline not being cached | ### Describe the bug
Components of custom pipelines (defined via custom_pipeline parameter) do not seem to be cached.
I have a multistep system where I'm loading all my components and then caching them explicitly using transformers.utils.move_cache.
I have an sdxl pipeline with `custom_pipeline="lpw_stable_diffusi... | open | null | false | 1 | [
"bug",
"stale"
] | [] | 2024-05-14T14:45:58Z | 2024-09-14T15:12:22Z | null | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | John0x | 5,455,173 | MDQ6VXNlcjU0NTUxNzM= | User | false |
huggingface/diffusers | 2,296,720,433 | I_kwDOHa8MBc6I5Swx | 7,948 | https://github.com/huggingface/diffusers/issues/7948 | https://api.github.com/repos/huggingface/diffusers/issues/7948 | Support MuLan, a plug-and-play language adapter to adapt existing diffusion model for up to 110+ languages without additional training | ### Model/Pipeline/Scheduler description
```diff
# pip install mulankit
from diffusers import StableDiffusionPipeline
+ import mulankit
pipe = StableDiffusionPipeline.from_pretrained('Lykon/dreamshaper-8')
+ pipe = mulankit.transform(pipe, 'mulanai/mulan-lang-adapter::sd15_aesthetic.pth')
image = pipe('一只蓝色的🐶... | open | null | false | 2 | [
"community-examples",
"stale",
"contributions-welcome"
] | [] | 2024-05-15T02:28:38Z | 2024-12-13T15:06:37Z | null | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Zeqiang-Lai | 26,198,430 | MDQ6VXNlcjI2MTk4NDMw | User | false |
huggingface/diffusers | 2,297,287,730 | I_kwDOHa8MBc6I7dQy | 7,950 | https://github.com/huggingface/diffusers/issues/7950 | https://api.github.com/repos/huggingface/diffusers/issues/7950 | PyTorch 2.3.0 Incompatibility with Current Diffusers Library | ### Describe the bug
When running the StableDiffusionXLPipeline with a specific model file (Clay_SDXL.safetensors), inference works correctly on CPU but results in a segmentation fault when run on GPU. Below are the specific steps and configurations that lead to this issue.
### Reproduction
import torch
from diffus... | closed | completed | false | 8 | [
"bug"
] | [] | 2024-05-15T09:03:50Z | 2024-10-20T23:06:13Z | 2024-10-20T23:06:13Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | JethroChow | 81,962,026 | MDQ6VXNlcjgxOTYyMDI2 | User | false |
huggingface/diffusers | 2,298,209,218 | I_kwDOHa8MBc6I--PC | 7,951 | https://github.com/huggingface/diffusers/issues/7951 | https://api.github.com/repos/huggingface/diffusers/issues/7951 | Add support for custom CLIPTextModelWithProjection in SDXL for new language training | **Is your feature request related to a problem? Please describe.**
I built a Sanskrit CLIP which works great with StableDiffusionPipeline. But when I try to train a new model for StableDiffusionXLPipeline, I am facing challenges for shape mismatch. I guess that there is very little documentation on adding the CLIPText... | closed | completed | false | 0 | [] | [] | 2024-05-15T15:16:50Z | 2024-05-15T19:58:22Z | 2024-05-15T19:58:22Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | humanely | 6,862,868 | MDQ6VXNlcjY4NjI4Njg= | User | false |
huggingface/diffusers | 2,298,817,925 | I_kwDOHa8MBc6JBS2F | 7,954 | https://github.com/huggingface/diffusers/issues/7954 | https://api.github.com/repos/huggingface/diffusers/issues/7954 | how can i know what the model base are | from transformers import DiffusionPipeline
base_model_id = "anything-v3-fp16-pruned.safetensors"
pipeline = DiffusionPipeline.from_pretrained(
base_model_id,
model_dir=model_dir,
use_safetensors=True,
)
| closed | completed | false | 2 | [] | [] | 2024-05-15T20:49:27Z | 2024-05-17T22:51:15Z | 2024-05-17T22:51:15Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | xalteropsx | 103,671,642 | U_kgDOBi3nWg | User | false |
huggingface/diffusers | 2,298,954,868 | I_kwDOHa8MBc6JB0R0 | 7,955 | https://github.com/huggingface/diffusers/issues/7955 | https://api.github.com/repos/huggingface/diffusers/issues/7955 | ConsistentID: Portrait Generation with Multimodal Fine-Grained Identity Preservation | ### Model/Pipeline/Scheduler description
Existing methods for facial identity transfer for diffusion denoising image generation models face challenges in achieving high fidelity and detailed identity (ID) consistency, primarily due to insufficient fine-grained control over facial areas and the lack of a comprehensive ... | open | null | false | 2 | [
"community-examples",
"stale",
"contributions-welcome"
] | [] | 2024-05-15T22:25:49Z | 2024-12-14T15:06:09Z | null | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | clarencechen | 8,482,341 | MDQ6VXNlcjg0ODIzNDE= | User | false |
huggingface/diffusers | 2,299,773,979 | I_kwDOHa8MBc6JE8Qb | 7,957 | https://github.com/huggingface/diffusers/issues/7957 | https://api.github.com/repos/huggingface/diffusers/issues/7957 | How to implement `IPAdapterAttnProcessor2_0` with xformers | I want to fine-tune IP-adapter model with xformers, but I did not find the implementation of the xformers version corresponding to IPAdapterAttnProcessor2_0. I want to implement attention processor in xformers, are the following two lines of code the only difference between the two versions?
In `XFormersAttnProcesso... | closed | completed | false | 1 | [] | [] | 2024-05-16T08:54:07Z | 2024-05-23T13:03:42Z | 2024-05-23T13:03:42Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | JWargrave | 65,226,618 | MDQ6VXNlcjY1MjI2NjE4 | User | false |
huggingface/diffusers | 2,300,272,710 | I_kwDOHa8MBc6JG2BG | 7,959 | https://github.com/huggingface/diffusers/issues/7959 | https://api.github.com/repos/huggingface/diffusers/issues/7959 | [training examples] reduce complexity by running final validations before export | I was thinking about Sayak's suggestion lately that the training examples are too long, and went through looking for redundant/unnecessary code sections that can be reduced or eliminated for readability.
The main thing that stands out is how the validations occur during the trainer unwind stage.
During training, ... | closed | not_planned | false | 3 | [] | [
"sayakpaul"
] | 2024-05-16T12:28:47Z | 2024-06-05T10:54:54Z | 2024-06-05T10:54:54Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | bghira | 59,658,056 | MDQ6VXNlcjU5NjU4MDU2 | User | false |
huggingface/diffusers | 2,300,837,682 | I_kwDOHa8MBc6JI_8y | 7,960 | https://github.com/huggingface/diffusers/issues/7960 | https://api.github.com/repos/huggingface/diffusers/issues/7960 | SDXL Training Fails for Multi GPU Machine | ### Describe the bug
While the training script `**train_text_to_image_lora_sdxl.py**` runs perfectly fine on A100 1 GPU machine, it fails to complete the data mapping on machines with multiple GPUs.
I have tried following:
export NCCL_P2P_DISABLE=1
export NCCL_IB_DISABLE=1
```
05/16/2024 13:59:48 - INFO -... | closed | completed | false | 6 | [
"bug"
] | [] | 2024-05-16T16:23:00Z | 2024-05-16T16:59:57Z | 2024-05-16T16:59:57Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | humanely | 6,862,868 | MDQ6VXNlcjY4NjI4Njg= | User | false |
huggingface/diffusers | 2,301,374,606 | I_kwDOHa8MBc6JLDCO | 7,962 | https://github.com/huggingface/diffusers/issues/7962 | https://api.github.com/repos/huggingface/diffusers/issues/7962 | Instruct-pix2pix pipeline: add ability to pass `cross_attention_kwargs` in call method | **Is your feature request related to a problem? Please describe.**
I have a use-case where I need to use custom adapter together with instruct-pix2pix model, which requires passing custom `cross_attention_kwargs` into unet.
However, having ability to pass this parameter would allow other use-cases as well, like LoRAs... | closed | completed | false | 2 | [] | [] | 2024-05-16T21:24:20Z | 2024-05-20T17:16:40Z | 2024-05-20T17:16:40Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | AlexeyZhuravlev | 8,759,774 | MDQ6VXNlcjg3NTk3NzQ= | User | false |
huggingface/diffusers | 2,302,163,914 | I_kwDOHa8MBc6JODvK | 7,968 | https://github.com/huggingface/diffusers/issues/7968 | https://api.github.com/repos/huggingface/diffusers/issues/7968 | Attention in Motion Module of UNetMotionModel | ### Describe the bug
As far as I know, UNetMotionModel is adopted for AnimateDiff. Hence, I look into the original implementation of AnimateDiff, it is noticed that they use cross-attention in Motion Module (https://github.com/guoyww/AnimateDiff/blob/main/animatediff/models/unet_blocks.py#L382). On the contrary, Diffu... | closed | completed | false | 3 | [
"bug"
] | [] | 2024-05-17T08:49:03Z | 2024-06-03T04:00:44Z | 2024-06-03T04:00:44Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | huyduong7101 | 93,563,638 | U_kgDOBZOq9g | User | false |
huggingface/diffusers | 2,302,540,725 | I_kwDOHa8MBc6JPfu1 | 7,970 | https://github.com/huggingface/diffusers/issues/7970 | https://api.github.com/repos/huggingface/diffusers/issues/7970 | Increasing RAM usage with enable_model_cpu_offload | ### Describe the bug
When using enable_model_cpu_offload on StableDiffusionXLPipeline, each consecutive __call__ takes more and more RAM. Also, after deleting pipe not all memory is freed
### Reproduction
```python
import gc
import torch
from diffusers import StableDiffusionXLPipeline
import psutil
de... | open | null | false | 9 | [
"bug",
"stale"
] | [] | 2024-05-17T11:53:40Z | 2024-12-14T15:06:02Z | null | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | max-fofanov | 89,916,160 | MDQ6VXNlcjg5OTE2MTYw | User | false |
huggingface/diffusers | 2,303,152,055 | I_kwDOHa8MBc6JR0-3 | 7,972 | https://github.com/huggingface/diffusers/issues/7972 | https://api.github.com/repos/huggingface/diffusers/issues/7972 | Type mismatch for LEDITS++ | ### Describe the bug
When running the code snippet from [https://huggingface.co/docs/diffusers/api/pipelines/ledits_pp#diffusers.LEditsPPPipelineStableDiffusionXL](https://huggingface.co/docs/diffusers/api/pipelines/ledits_pp#diffusers.LEditsPPPipelineStableDiffusionXL) I get a RuntimeError for as mismatch between t... | closed | completed | false | 3 | [
"bug"
] | [] | 2024-05-17T16:30:11Z | 2024-05-23T12:59:55Z | 2024-05-23T12:59:55Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | vdelale | 69,793,390 | MDQ6VXNlcjY5NzkzMzkw | User | false |
huggingface/diffusers | 2,303,391,355 | I_kwDOHa8MBc6JSvZ7 | 7,974 | https://github.com/huggingface/diffusers/issues/7974 | https://api.github.com/repos/huggingface/diffusers/issues/7974 | Bug: Inpaint with IP Adapter (face+style) | ### Describe the bug
Inpaint with IP Adapter(face+style) doesn't seem to work. I'm getting this error, if I run the following code:
`RuntimeError: mat1 and mat2 shapes cannot be multiplied (5140x1664 and 1280x1280)`
### Reproduction
```
import torch
from diffusers import DDIMScheduler, AutoPipelineForInpainti... | closed | completed | false | 2 | [
"bug"
] | [] | 2024-05-17T18:56:22Z | 2024-05-18T04:25:03Z | 2024-05-18T04:14:47Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | tin2tin | 1,322,593 | MDQ6VXNlcjEzMjI1OTM= | User | false |
huggingface/diffusers | 2,304,515,781 | I_kwDOHa8MBc6JXB7F | 7,978 | https://github.com/huggingface/diffusers/issues/7978 | https://api.github.com/repos/huggingface/diffusers/issues/7978 | RuntimeError: operator torchvision::nms does not exist | ### Describe the bug
I get this error when I use the xformers parameter.
```bash
File "/root/projects/vton_train/train_text_to_image_sdxl.py", line 43, in <module>
from torchvision import transforms
File "/usr/local/lib/python3.10/dist-packages/torchvision/__init__.py", line 6, in <module>
from torc... | closed | completed | false | 4 | [
"bug"
] | [] | 2024-05-19T09:18:59Z | 2024-05-21T13:52:50Z | 2024-05-21T13:52:49Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | kadirnar | 36,204,372 | MDQ6VXNlcjM2MjA0Mzcy | User | false |
huggingface/diffusers | 2,304,528,885 | I_kwDOHa8MBc6JXFH1 | 7,979 | https://github.com/huggingface/diffusers/issues/7979 | https://api.github.com/repos/huggingface/diffusers/issues/7979 | config file is not a valid json file | File "Z:\software\python11\Lib\site-packages\diffusers\configuration_utils.py", line 436, in load_config
raise EnvironmentError(f"It looks like the config file at '{config_file}' is not a valid JSON file.")
OSError: It looks like the config file at 'X:\test\dreamshaper_8.safetensors' is not a valid JSON file.
... | closed | completed | false | 10 | [] | [] | 2024-05-19T09:54:52Z | 2024-06-15T07:00:50Z | 2024-06-15T07:00:50Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | neonarc4 | 140,290,288 | U_kgDOCFyo8A | User | false |
huggingface/diffusers | 2,304,631,958 | I_kwDOHa8MBc6JXeSW | 7,980 | https://github.com/huggingface/diffusers/issues/7980 | https://api.github.com/repos/huggingface/diffusers/issues/7980 | A way to know from the documentation when a given class was introduced | **Is your feature request related to a problem? Please describe.**
While working with various codebase that rely on different diffuser versions I find it challenging to pinpoint when a given sheduler/ base model was introduced.
For now I basically manually increment the version in the URL at a given doc page -> ctrl ... | closed | not_planned | false | 2 | [
"stale"
] | [] | 2024-05-19T14:14:34Z | 2024-09-14T15:25:46Z | 2024-09-14T15:25:46Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | melMass | 7,041,726 | MDQ6VXNlcjcwNDE3MjY= | User | false |
huggingface/diffusers | 2,304,668,599 | I_kwDOHa8MBc6JXnO3 | 7,982 | https://github.com/huggingface/diffusers/issues/7982 | https://api.github.com/repos/huggingface/diffusers/issues/7982 | `from_pipe` method is not working with `MultiControlNetModel` pipeline component | ### Describe the bug
When trying to init a `StableDiffusionXLControlNetInpaintPipeline` with `from_pipe` method from the `StableDiffusionXLControlNetPipeline` with multiple controlnets, an error, which basically says that only `ControlNet` class is supported, occurs.
If using `from_pipe(pipeline, controlnet=pipel... | closed | completed | false | 2 | [
"bug"
] | [] | 2024-05-19T15:44:05Z | 2024-05-20T16:25:09Z | 2024-05-20T16:24:50Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | detkov | 34,102,070 | MDQ6VXNlcjM0MTAyMDcw | User | false |
huggingface/diffusers | 2,304,709,872 | I_kwDOHa8MBc6JXxTw | 7,985 | https://github.com/huggingface/diffusers/issues/7985 | https://api.github.com/repos/huggingface/diffusers/issues/7985 | IPAdapter not compatible to torch.compile | ### Describe the bug
When use IPAdapter with torch.compile will return error message below:
```
torch._dynamo.exc.Unsupported: UNPACK_SEQUENCE NNModuleVariable()
```
Due to the use of this line
https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/embeddings.py#L879
Which can be fixed w... | closed | completed | false | 3 | [
"bug"
] | [] | 2024-05-19T17:36:25Z | 2024-05-29T14:21:04Z | 2024-05-29T14:21:04Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | rootonchair | 23,548,268 | MDQ6VXNlcjIzNTQ4MjY4 | User | false |
huggingface/diffusers | 1,403,700,650 | I_kwDOHa8MBc5TqsWq | 799 | https://github.com/huggingface/diffusers/issues/799 | https://api.github.com/repos/huggingface/diffusers/issues/799 | Define a metadata format for textual inversion concepts | **Is your feature request related to a problem? Please describe.**
The textual inversion script currently creates a dict with one term and its associated embedding. Usual scripts load one or more embeddings using the keys as token for prompt.
This format does not allow for metadata associated with the key(s). The l... | closed | completed | false | 11 | [] | [
"patil-suraj"
] | 2022-10-10T20:52:12Z | 2022-12-07T18:20:59Z | 2022-11-28T15:03:40Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | allo- | 1,482,341 | MDQ6VXNlcjE0ODIzNDE= | User | false |
huggingface/diffusers | 2,306,954,746 | I_kwDOHa8MBc6JgVX6 | 7,991 | https://github.com/huggingface/diffusers/issues/7991 | https://api.github.com/repos/huggingface/diffusers/issues/7991 | Bug in diffusers basic training tutorial | ### Describe the bug
In the diffusers [basic training tutorial](https://huggingface.co/docs/diffusers/tutorials/basic_training), setting the manual seed argument (`generator=torch.manual_seed(config.seed)`) in the pipeline call inside evaluate() function rewinds the dataloader shuffling, leading to overfitting due to ... | closed | completed | false | 5 | [
"bug"
] | [] | 2024-05-20T23:16:11Z | 2024-05-21T21:24:14Z | 2024-05-21T21:06:10Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | yue-here | 14,996,988 | MDQ6VXNlcjE0OTk2OTg4 | User | false |
huggingface/diffusers | 2,308,612,189 | I_kwDOHa8MBc6JmqBd | 7,997 | https://github.com/huggingface/diffusers/issues/7997 | https://api.github.com/repos/huggingface/diffusers/issues/7997 | Full SDXL DreamBooth finetuning, LoRA extraction | In the dreambooth community, it has been empirically shown that extracting a LoRA results in better performance than directly training a LoRA. Enabling a full dreambooth finetune of SDXL would not only enable this functionality, but also further enable users to extract and test out LoRAs of multiple different network r... | open | null | false | 7 | [
"stale"
] | [] | 2024-05-21T15:51:19Z | 2024-09-14T15:11:52Z | null | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | neuron-party | 96,799,331 | U_kgDOBcUKYw | User | false |
huggingface/diffusers | 1,266,387,205 | I_kwDOHa8MBc5Le4kF | 8 | https://github.com/huggingface/diffusers/issues/8 | https://api.github.com/repos/huggingface/diffusers/issues/8 | Schedulers - what code should go into a "Scheduler" class? | At the moment, the scheduler class has very little logic -> see: https://github.com/huggingface/diffusers/blob/2852c80540ed9a9a435bb8b2ce9e393811af519c/src/diffusers/schedulers/gaussian_ddpm.py#L26
where as the example of the unrolled denoising process is getting quite complicated (copied from the README):
```py... | closed | completed | false | 0 | [] | [] | 2022-06-09T16:28:35Z | 2022-06-12T22:18:01Z | 2022-06-12T22:18:01Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | patrickvonplaten | 23,423,619 | MDQ6VXNlcjIzNDIzNjE5 | User | false |
huggingface/diffusers | 1,404,006,370 | I_kwDOHa8MBc5Tr2_i | 801 | https://github.com/huggingface/diffusers/issues/801 | https://api.github.com/repos/huggingface/diffusers/issues/801 | In-painting & Img-to-Img seemingly do not abide `enable_attention_slicing()` | ### Describe the bug
I'm able to run (seemingly) all text-to-image examples without any sort of issue, including [the one in the ReadMe](https://github.com/huggingface/diffusers#text-to-image-generation-with-stable-diffusion). Memory gets a bit tight, but stays under my 8gb limit.
However, when I switch to trying o... | closed | not_planned | false | 3 | [
"bug"
] | [] | 2022-10-11T04:59:56Z | 2022-10-12T02:12:48Z | 2022-10-12T01:45:53Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | rschristian | 33,403,762 | MDQ6VXNlcjMzNDAzNzYy | User | false |
huggingface/diffusers | 1,405,049,117 | I_kwDOHa8MBc5Tv1kd | 805 | https://github.com/huggingface/diffusers/issues/805 | https://api.github.com/repos/huggingface/diffusers/issues/805 | [ONNX] Loading textual inversion .bin/.pt embeddings | I tried to follow this guide:
https://colab.research.google.com/github/huggingface/notebooks/blob/main/diffusers/stable_conceptualizer_inference.ipynb
But OnnxRuntimeModel does not have any of the necessary functions that CLIPTextModel has and since I'm no dev I couldn't figure out how to get this working.
If this i... | closed | completed | false | 3 | [
"stale"
] | [
"anton-l"
] | 2022-10-11T18:17:01Z | 2022-11-20T15:02:51Z | 2022-11-20T15:02:51Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | GreenLandisaLie | 80,653,235 | MDQ6VXNlcjgwNjUzMjM1 | User | false |
huggingface/diffusers | 1,405,386,197 | I_kwDOHa8MBc5TxH3V | 807 | https://github.com/huggingface/diffusers/issues/807 | https://api.github.com/repos/huggingface/diffusers/issues/807 | Dreambooth doesn't train on 8GB | ### Describe the bug
Per the example featured in the repo, it goes OOM when DeepSpeed is loading the optimizer, tested on a 3080 10GB + 64GB RAM in WSL2 and native Linux.
### Reproduction
Follow the pastebin for setup purposes (on WSL2), or just try it yourself https://pastebin.com/0NHA5YTP
### Logs
```shell
The f... | closed | completed | false | 50 | [
"bug",
"stale"
] | [
"patil-suraj"
] | 2022-10-12T00:33:15Z | 2022-12-02T16:56:06Z | 2022-11-27T15:03:14Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | devilismyfriend | 87,043,616 | MDQ6VXNlcjg3MDQzNjE2 | User | false |
huggingface/diffusers | 1,295,818,512 | I_kwDOHa8MBc5NPJ8Q | 81 | https://github.com/huggingface/diffusers/issues/81 | https://api.github.com/repos/huggingface/diffusers/issues/81 | Usage of init_git_repo is not very clear | init_git_repo expects some arguments which are not very clear (e.g. you need to pass `local_rank` which is not very obvious when the functions is creating a git repo).
cc @anton-l | closed | completed | false | 10 | [
"stale"
] | [
"anton-l"
] | 2022-07-06T13:08:22Z | 2023-01-05T01:28:46Z | 2022-10-27T13:17:00Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | osanseviero | 7,246,357 | MDQ6VXNlcjcyNDYzNTc= | User | false |
huggingface/diffusers | 1,406,158,414 | I_kwDOHa8MBc5T0EZO | 811 | https://github.com/huggingface/diffusers/issues/811 | https://api.github.com/repos/huggingface/diffusers/issues/811 | Dreambooth Prior Preservation Loss Example | **What API design would you like to have changed or added to the library? Why?**
What happens when a not safe for work image is generated by Stable Diffusion when I am trying to use it to create class images? It returns a black image. I want to know how does this Image affect the training process. Are the black images... | closed | completed | false | 4 | [
"stale"
] | [] | 2022-10-12T12:46:58Z | 2022-11-16T21:07:15Z | 2022-11-16T21:07:15Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | ghpkishore | 26,677,859 | MDQ6VXNlcjI2Njc3ODU5 | User | false |
huggingface/diffusers | 1,406,169,515 | I_kwDOHa8MBc5T0HGr | 812 | https://github.com/huggingface/diffusers/issues/812 | https://api.github.com/repos/huggingface/diffusers/issues/812 | Inpaint at full resolution | **Is your feature request related to a problem? Please describe.**
It would be excellent if StableDiffusionInpaintPipeline has the option of inpainting at full resolution since most of the time, if your initial image includes human face, and you want to keep human face unchanged, the output image of the face part i... | closed | completed | false | 5 | [
"stale"
] | [] | 2022-10-12T12:54:37Z | 2022-11-27T15:03:13Z | 2022-11-27T15:03:13Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | arandomgoodguy | 94,698,515 | U_kgDOBaT8Ew | User | false |
huggingface/diffusers | 1,406,560,399 | I_kwDOHa8MBc5T1miP | 817 | https://github.com/huggingface/diffusers/issues/817 | https://api.github.com/repos/huggingface/diffusers/issues/817 | Mixed precision is not working on dreambooth example | ### Describe the bug
Whenever I run the code with the `mixed_precision="fp16"` flag it crashes.
- It appears that something with the Grad scaler in accelerate is not working
- It's weird for me calling `optimizer.zero_grad()` after the backwards, but maybe accelerate is doing some magic I am not aware.
```
... | closed | completed | false | 3 | [
"bug"
] | [
"patil-suraj"
] | 2022-10-12T17:25:19Z | 2022-10-13T13:41:21Z | 2022-10-13T13:41:20Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | tcapelle | 18,441,985 | MDQ6VXNlcjE4NDQxOTg1 | User | false |
huggingface/diffusers | 2,309,033,471 | I_kwDOHa8MBc6JoQ3_ | 8,182 | https://github.com/huggingface/diffusers/issues/8182 | https://api.github.com/repos/huggingface/diffusers/issues/8182 | AttributeError: 'Logger' object has no attribute 'warninging'. Did you mean: 'warning'? | ### Describe the bug
AttributeError: 'Logger' object has no attribute 'warninging'. Did you mean: 'warning'?
There is a spelling error in the logger object.
### Reproduction
loading a model using .from_single_file()
### Logs
```shell
Traceback (most recent call last):
File "/workspace/hbot-backend-anime/src... | closed | completed | false | 0 | [
"bug"
] | [] | 2024-05-21T20:09:10Z | 2024-05-22T06:59:12Z | 2024-05-22T06:59:12Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | AMohamedAakhil | 78,485,654 | MDQ6VXNlcjc4NDg1NjU0 | User | false |
huggingface/diffusers | 1,296,047,369 | I_kwDOHa8MBc5NQB0J | 82 | https://github.com/huggingface/diffusers/issues/82 | https://api.github.com/repos/huggingface/diffusers/issues/82 | latent_diffusion not found | Hi :)
I was trying to recreate the example of [text-to-image](https://github.com/milyiyo/diffusers/tree/main#text-to-image-generation-with-latent-diffusion) in Colab, but I am getting the following error:
`OSError: file /root/.cache/huggingface/hub/models--fusing--latent-diffusion-text2im-large/snapshots/d5eab561... | closed | completed | false | 1 | [] | [] | 2022-07-06T15:56:42Z | 2022-07-13T15:55:30Z | 2022-07-13T15:55:29Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | milyiyo | 8,120,990 | MDQ6VXNlcjgxMjA5OTA= | User | false |
huggingface/diffusers | 1,407,369,107 | I_kwDOHa8MBc5T4r-T | 820 | https://github.com/huggingface/diffusers/issues/820 | https://api.github.com/repos/huggingface/diffusers/issues/820 | Pull request #793 broke negative prompts for IMG2IMG pipeline | ### Describe the bug
The commit #793 broke negative prompts. By changing the type of prompt before changing the type of negPrompt, but comparing their types in between, it is now impossible to use single negative prompts.
Would be great if you could resolve this, @patrickvonplaten.
### Reproduction
Use a single ... | closed | completed | false | 5 | [
"bug"
] | [
"patrickvonplaten"
] | 2022-10-13T08:26:58Z | 2022-10-14T17:54:17Z | 2022-10-13T13:45:44Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | OrionFive | 13,980,562 | MDQ6VXNlcjEzOTgwNTYy | User | false |
huggingface/diffusers | 2,309,487,768 | I_kwDOHa8MBc6Jp_yY | 8,226 | https://github.com/huggingface/diffusers/issues/8226 | https://api.github.com/repos/huggingface/diffusers/issues/8226 | VectorQuantizer Loss | ### Describe the bug
I'm using the VQModel in diffusers for my generation project and I got a little confusion on calculating the vector quantizer loss:
https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/autoencoders/vae.py#L731
The first part "torch.mean((z_q.detach() - z) ** 2)" should be ... | closed | completed | false | 2 | [
"bug"
] | [] | 2024-05-22T03:20:27Z | 2024-05-22T05:04:00Z | 2024-05-22T05:04:00Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | bochent | 17,314,925 | MDQ6VXNlcjE3MzE0OTI1 | User | false |
huggingface/diffusers | 2,309,694,109 | I_kwDOHa8MBc6JqyKd | 8,228 | https://github.com/huggingface/diffusers/issues/8228 | https://api.github.com/repos/huggingface/diffusers/issues/8228 | 0.28.0 from_single_file Failed to load CLIPTextModel | ### Describe the bug
train model with OneTrainer and Everydream2, both same error
this one with onetrainer
```
Fetching 11 files: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████... | closed | completed | false | 13 | [
"bug",
"single_file"
] | [] | 2024-05-22T06:35:15Z | 2024-06-06T02:17:13Z | 2024-06-03T03:59:23Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | crapthings | 1,147,704 | MDQ6VXNlcjExNDc3MDQ= | User | false |
huggingface/diffusers | 2,311,179,257 | I_kwDOHa8MBc6Jwcv5 | 8,233 | https://github.com/huggingface/diffusers/issues/8233 | https://api.github.com/repos/huggingface/diffusers/issues/8233 | Depth pipeline gives error at fp16 | ### Describe the bug
When using the StableDiffusionDepth2ImgPipeline, I get ``RuntimeError: Input type (MPSFloatType) and weight type (MPSHalfType) should be the same`` if I'm using fp16 with MPS on my M2.
I've already figured out a fix: https://github.com/jonahclarsen/diffusers-bugfixes/commit/d738833a9ddbd95426... | closed | completed | false | 1 | [
"bug"
] | [] | 2024-05-22T18:10:13Z | 2024-05-30T23:59:51Z | 2024-05-30T23:59:51Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | jonahclarsen | 97,376,233 | U_kgDOBc3X6Q | User | false |
huggingface/diffusers | 2,311,522,260 | I_kwDOHa8MBc6JxwfU | 8,234 | https://github.com/huggingface/diffusers/issues/8234 | https://api.github.com/repos/huggingface/diffusers/issues/8234 | When setting ip_adapter_scale with a dict and ipadapter plus TypeError: unsupported operand type(s) for *: 'dict' and 'Tensor' | ### Describe the bug
This code works fine when I use pipeline.set_ip_adapter_scale(1.0) but fails as below when I set the scale witha dict
File \"/usr/local/lib/python3.11/dist-packages/diffusers/models/transformers/transformer_2d.py\", line 397, in forward
hidden_states = block(
^^^^^^
... | closed | completed | false | 4 | [
"bug"
] | [] | 2024-05-22T21:33:28Z | 2024-05-24T02:09:04Z | 2024-05-24T02:09:04Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | barber5 | 1,551,200 | MDQ6VXNlcjE1NTEyMDA= | User | false |
huggingface/diffusers | 2,312,158,053 | I_kwDOHa8MBc6J0Ltl | 8,236 | https://github.com/huggingface/diffusers/issues/8236 | https://api.github.com/repos/huggingface/diffusers/issues/8236 | Wrong learning rate scheduler training step count for examples with multi-gpu when setting `--num_train_epochs` | ### Describe the bug
I think there are still some problems with the learning rate scheduler. This is resolved when you set `--max_train_steps`, as discussed in #3954 , but not completely.
For example, the code snippet https://github.com/huggingface/diffusers/blob/main/examples/text_to_image/train_text_to_image.py#L... | closed | completed | false | 7 | [
"bug"
] | [] | 2024-05-23T07:24:46Z | 2024-05-30T09:53:21Z | 2024-05-30T09:53:21Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | geniuspatrick | 74,176,172 | MDQ6VXNlcjc0MTc2MTcy | User | false |
huggingface/diffusers | 2,312,247,295 | I_kwDOHa8MBc6J0hf_ | 8,237 | https://github.com/huggingface/diffusers/issues/8237 | https://api.github.com/repos/huggingface/diffusers/issues/8237 | SDXL LoRA key not loaded "lora_te2_text_projection.*" (LoRA from OneTrainer) | ### Describe the bug
When trying to use a LoRA in Diffusers, created in OneTrainer:
These keys are not loaded:
```
Python\Python311\site-packages\diffusers\loaders\lora_conversion_utils.py", line 273, in _convert_kohya_lora_to_diffusers
raise ValueError(f"The following keys have not been correctly be renam... | closed | completed | false | 5 | [
"bug"
] | [] | 2024-05-23T08:07:35Z | 2024-05-29T06:09:17Z | 2024-05-28T18:25:55Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | tin2tin | 1,322,593 | MDQ6VXNlcjEzMjI1OTM= | User | false |
huggingface/diffusers | 2,312,929,630 | I_kwDOHa8MBc6J3IFe | 8,241 | https://github.com/huggingface/diffusers/issues/8241 | https://api.github.com/repos/huggingface/diffusers/issues/8241 | Abnormal size for LEDITS++ model ? | ### Describe the bug
I tried to run the diffusers pipeline for LEDITS++ with `LEditsPPPipelineStableDiffusionXL` but I encounter a Cuda Out of memory error, which I find abnormal, since the error states that it tried allocating other 136 GiB.
Here are all the information needed
### Reproduction
- Created an... | closed | completed | false | 3 | [
"bug"
] | [] | 2024-05-23T13:26:12Z | 2024-06-18T13:59:22Z | 2024-06-18T13:58:31Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | vdelale | 69,793,390 | MDQ6VXNlcjY5NzkzMzkw | User | false |
huggingface/diffusers | 2,313,372,631 | I_kwDOHa8MBc6J40PX | 8,264 | https://github.com/huggingface/diffusers/issues/8264 | https://api.github.com/repos/huggingface/diffusers/issues/8264 | Train stopped at 0% | ### Describe the bug
I used this script for training lora model with command line parameters at this link:
https://huggingface.co/docs/diffusers/training/lora
However, training stops at 0% without any error in the cmd.
Through some print() I discovered that the script stops at the first iteration of the For at line... | closed | completed | false | 8 | [
"bug"
] | [] | 2024-05-23T16:30:08Z | 2024-05-29T12:27:44Z | 2024-05-29T12:27:44Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | adinoolfi | 104,144,179 | U_kgDOBjUdMw | User | false |
huggingface/diffusers | 2,317,337,973 | I_kwDOHa8MBc6KH8V1 | 8,274 | https://github.com/huggingface/diffusers/issues/8274 | https://api.github.com/repos/huggingface/diffusers/issues/8274 | FIFO-Diffusion: Generating Infinite Videos from Text without Training through Rolling Video Denoising | ### Model/Pipeline/Scheduler description
The authors propose a novel inference technique based on a pretrained diffusion model for text-conditional video generation. Their approach, called FIFO-Diffusion, is conceptually capable of generating infinitely long videos without training. This is achieved by iteratively per... | open | null | false | 6 | [
"community-examples",
"Good second issue",
"contributions-welcome"
] | [] | 2024-05-26T00:15:43Z | 2025-03-17T17:35:58Z | null | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | clarencechen | 8,482,341 | MDQ6VXNlcjg0ODIzNDE= | User | false |
huggingface/diffusers | 2,317,578,365 | I_kwDOHa8MBc6KI3B9 | 8,277 | https://github.com/huggingface/diffusers/issues/8277 | https://api.github.com/repos/huggingface/diffusers/issues/8277 | Consistency training fails to converge | **Consistency training fails to converge**
When I use examples/research_projects/consistency_training/train_cm_ct_unconditional.py to train the consistency model, I try to follow the experimental parameters of the original paper, but I cannot obtain normal generation results.
**Script**
```bash
python -m accelera... | open | null | false | 6 | [
"stale"
] | [] | 2024-05-26T08:20:57Z | 2024-09-28T15:05:44Z | null | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | yhx-03 | 58,554,991 | MDQ6VXNlcjU4NTU0OTkx | User | false |
huggingface/diffusers | 2,317,753,509 | I_kwDOHa8MBc6KJhyl | 8,279 | https://github.com/huggingface/diffusers/issues/8279 | https://api.github.com/repos/huggingface/diffusers/issues/8279 | do we need internet everytime to run single_file? | StableDiffusion.x_single_file( )
without internet it wont work is their way u provide an cache ? to run it offline | closed | completed | false | 5 | [] | [] | 2024-05-26T15:00:59Z | 2024-05-26T17:03:49Z | 2024-05-26T16:59:09Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | xalteropsx | 103,671,642 | U_kgDOBi3nWg | User | false |
huggingface/diffusers | 2,317,896,943 | I_kwDOHa8MBc6KKEzv | 8,281 | https://github.com/huggingface/diffusers/issues/8281 | https://api.github.com/repos/huggingface/diffusers/issues/8281 | Potentially wrong scheduler in train_text_to_image_sdxl.py? | ### Describe the bug
I see that in [train_text_to_image_sdxl.py](https://github.com/huggingface/diffusers/blob/main/examples/text_to_image/train_text_to_image_sdxl.py) the scheduler being used is `DDPMScheduler`, which -- as the name suggests -- is the specific diffusion formulation in [this paper](https://arxiv.org... | closed | completed | false | 13 | [
"bug"
] | [] | 2024-05-26T19:18:23Z | 2024-05-30T20:15:05Z | 2024-05-30T20:15:05Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | christopher-beckham | 2,417,792 | MDQ6VXNlcjI0MTc3OTI= | User | false |
huggingface/diffusers | 2,319,500,752 | I_kwDOHa8MBc6KQMXQ | 8,287 | https://github.com/huggingface/diffusers/issues/8287 | https://api.github.com/repos/huggingface/diffusers/issues/8287 | StableDiffusionPipeline fails when text_encoder=None | ### Describe the bug
When we've pre-generated prompt embeds, the expectation is that we can reuse these indefinitely and that `encode_prompts` will not try to access the non-existent text_encoder.
### Reproduction
```py
from diffusers import StableDiffusionPipeline
from compel import Compel
from transformer... | closed | completed | false | 3 | [
"bug"
] | [] | 2024-05-27T16:36:06Z | 2024-07-05T05:15:46Z | 2024-07-05T05:15:46Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | bghira | 59,658,056 | MDQ6VXNlcjU5NjU4MDU2 | User | false |
huggingface/diffusers | 2,319,642,311 | I_kwDOHa8MBc6KQu7H | 8,288 | https://github.com/huggingface/diffusers/issues/8288 | https://api.github.com/repos/huggingface/diffusers/issues/8288 | DPM solver ++ third order with SDE | I see this in A1111... is it possible to get the missing implementation here:
https://github.com/huggingface/diffusers/blob/main/src/diffusers/schedulers/scheduling_dpmsolver_multistep.py#L887
Thanks
(better results than 2nd order https://www.reddit.com/r/StableDiffusion/comments/16amqso/i_think_dpm_3m_sde_karras_is... | open | null | false | 8 | [
"stale"
] | [
"yiyixuxu"
] | 2024-05-27T18:47:07Z | 2024-09-14T15:11:36Z | null | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | christopher5106 | 6,875,375 | MDQ6VXNlcjY4NzUzNzU= | User | false |
huggingface/diffusers | 2,320,001,708 | I_kwDOHa8MBc6KSGqs | 8,293 | https://github.com/huggingface/diffusers/issues/8293 | https://api.github.com/repos/huggingface/diffusers/issues/8293 | is diffusers pipline support emphasis? | curious to know | closed | completed | false | 2 | [] | [] | 2024-05-28T02:22:42Z | 2024-05-29T12:40:24Z | 2024-05-28T17:22:45Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | xalteropsx | 103,671,642 | U_kgDOBi3nWg | User | false |
huggingface/diffusers | 2,320,276,020 | I_kwDOHa8MBc6KTJo0 | 8,296 | https://github.com/huggingface/diffusers/issues/8296 | https://api.github.com/repos/huggingface/diffusers/issues/8296 | [Training] Resume checkpoint global step inconsistent/confusion across scripts | ### Describe the bug
Hi,
I have been working on training scripts for multiple models (T2I, IP2P) and found the different logic to calculate `step` and `epoch` while resuming training different across scripts.
In `train_text_to_image.py` script [link](https://github.com/huggingface/diffusers/blob/fe5f035f797a5fa66... | open | null | false | 3 | [
"bug",
"stale"
] | [] | 2024-05-28T06:54:42Z | 2024-09-14T15:11:33Z | null | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | vinm007 | 136,594,182 | U_kgDOCCRDBg | User | false |
huggingface/diffusers | 2,321,750,746 | I_kwDOHa8MBc6KYxra | 8,305 | https://github.com/huggingface/diffusers/issues/8305 | https://api.github.com/repos/huggingface/diffusers/issues/8305 | `TCDScheduler` : `eta` out of range? | ### Describe the bug
Hi,
I am starting to play with the relatively recent scheduler [TCDScheduler](https://github.com/huggingface/diffusers/blob/v0.28.0/src/diffusers/schedulers/scheduling_tcd.py#L133) and encountered a bug yesterday.
The thing is, I'm embarrassed to say I am unable to provide the exact detai... | closed | completed | false | 2 | [
"bug"
] | [
"yiyixuxu"
] | 2024-05-28T19:14:55Z | 2024-06-18T15:59:37Z | 2024-06-18T15:59:37Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Clement-Lelievre | 70,368,164 | MDQ6VXNlcjcwMzY4MTY0 | User | false |
huggingface/diffusers | 2,321,757,461 | I_kwDOHa8MBc6KYzUV | 8,306 | https://github.com/huggingface/diffusers/issues/8306 | https://api.github.com/repos/huggingface/diffusers/issues/8306 | Diffusion Transformers License | ### Describe the bug
From this [HN thread](https://news.ycombinator.com/item?id=40478921): the Diffusion Transformer pipeline code includes a license `Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)` at the top:
https://github.com/huggingface/diffusers/blob/b69fd990ad8026f21893499ab396d969b62bb8cc/src/diff... | open | null | false | 4 | [
"bug",
"stale"
] | [] | 2024-05-28T19:19:36Z | 2024-09-14T15:11:30Z | null | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | andrewkchan | 8,591,901 | MDQ6VXNlcjg1OTE5MDE= | User | false |
huggingface/diffusers | 2,322,221,937 | I_kwDOHa8MBc6Kaktx | 8,307 | https://github.com/huggingface/diffusers/issues/8307 | https://api.github.com/repos/huggingface/diffusers/issues/8307 | AttributeError: module diffusers has no attribute StableCascadeUNet | ### Describe the bug
Receive the error:
```AttributeError: module diffusers has no attribute StableCascadeUNet```
when attempting to generate in Forge.
### Reproduction
I don't have any code. Just opened a clean install in Forge, and attempted to generate.
### Logs
_No response_
### System Info
- `diffusers` v... | closed | completed | false | 7 | [
"bug"
] | [] | 2024-05-29T02:26:44Z | 2024-10-15T21:34:23Z | 2024-10-15T21:34:23Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Vektor369 | 124,237,282 | U_kgDOB2e14g | User | false |
huggingface/diffusers | 2,322,384,225 | I_kwDOHa8MBc6KbMVh | 8,309 | https://github.com/huggingface/diffusers/issues/8309 | https://api.github.com/repos/huggingface/diffusers/issues/8309 | Add Conditional Diffusion Distillation | ### Model/Pipeline/Scheduler description
Conditional Diffusion Distillation (CoDi) is a new diffusion generation method recently proposed by Google Research and Johns Hopkins University. Accepted by CVPR24, CoDi is based on consistency models and offers a significant advancement in accelerating latent diffusion mode... | open | null | false | 1 | [
"stale"
] | [] | 2024-05-29T05:25:13Z | 2024-09-14T15:11:24Z | null | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | MKFMIKU | 13,622,651 | MDQ6VXNlcjEzNjIyNjUx | User | false |
huggingface/diffusers | 2,323,980,628 | I_kwDOHa8MBc6KhSFU | 8,315 | https://github.com/huggingface/diffusers/issues/8315 | https://api.github.com/repos/huggingface/diffusers/issues/8315 | [`Docs`] JAX/Flax Installation | ### Describe the bug
[Installation gives errors in docs](https://huggingface.co/docs/diffusers/main/en/using-diffusers/stable_diffusion_jax_how_to).
### Reproduction
```sh
pip install -q jax==0.3.25 jaxlib==0.3.25 flax transformers ftfy
```
### Logs
```
ERROR: Could not find a version that satisfies ... | closed | completed | false | 2 | [
"bug"
] | [] | 2024-05-29T18:29:16Z | 2024-06-09T16:36:32Z | 2024-06-09T15:44:56Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | tolgacangoz | 46,008,593 | MDQ6VXNlcjQ2MDA4NTkz | User | false |
huggingface/diffusers | 2,325,137,495 | I_kwDOHa8MBc6KlshX | 8,317 | https://github.com/huggingface/diffusers/issues/8317 | https://api.github.com/repos/huggingface/diffusers/issues/8317 | `AsymmetricAutoencoderKL`: missing `generator` argument in `decode()` called from `forward()` | ### Describe the bug
`AsymmetricAutoencoderKL.decode()` has parameter [`generator`](https://github.com/huggingface/diffusers/blob/main/src/diffusers/models/autoencoders/autoencoder_asym_kl.py#L144),
```python
def decode(
self,
z: torch.FloatTensor,
generator: Optional[torch.Generator]... | closed | completed | false | 2 | [
"bug"
] | [] | 2024-05-30T09:02:51Z | 2024-06-06T02:00:59Z | 2024-06-06T02:00:59Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | townwish4git | 143,256,262 | U_kgDOCInqxg | User | false |
huggingface/diffusers | 2,325,849,626 | I_kwDOHa8MBc6KoaYa | 8,320 | https://github.com/huggingface/diffusers/issues/8320 | https://api.github.com/repos/huggingface/diffusers/issues/8320 | [0.27.2]: from_single_file: 'AutoencoderKL' object has no attribute '__name__'. Did you mean: '__ne__'? | ### Describe the bug
finetune a dreambooth model with ED2 but can't load with 0.27.2
the model work with a1111
### Reproduction
pipe.from_single_file
https://huggingface.co/crapthings/diffusers-issue/resolve/main/base-ed2.safetensors
### Logs
_No response_
### System Info
pip install diffuser... | open | null | false | 10 | [
"bug",
"stale"
] | [] | 2024-05-30T14:40:34Z | 2024-09-14T15:11:14Z | null | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | crapthings | 1,147,704 | MDQ6VXNlcjExNDc3MDQ= | User | false |
huggingface/diffusers | 2,326,089,711 | I_kwDOHa8MBc6KpU_v | 8,323 | https://github.com/huggingface/diffusers/issues/8323 | https://api.github.com/repos/huggingface/diffusers/issues/8323 | scheduler文件无法识别 | ### Describe the bug
发生异常: ValueError
Pipeline <class 'diffusers.pipelines.latent_consistency_models.pipeline_latent_consistency_text2img.LatentConsistencyModelPipeline'> expected {'tokenizer', 'text_encoder', 'unet', 'vae', 'feature_extractor', 'safety_checker', 'scheduler'}, but only {'tokenizer', 'text_encoder', '... | open | null | false | 3 | [
"bug",
"stale"
] | [] | 2024-05-30T16:30:24Z | 2024-09-14T15:11:11Z | null | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | wx331406 | 71,815,782 | MDQ6VXNlcjcxODE1Nzgy | User | false |
huggingface/diffusers | 1,408,375,662 | I_kwDOHa8MBc5T8htu | 834 | https://github.com/huggingface/diffusers/issues/834 | https://api.github.com/repos/huggingface/diffusers/issues/834 | [Community] CPU Under Utilized | ### Describe the bug
When using CPU as device, the CPU is very under utilized. I get about 50-66% utilization. The fans don't even need to turn on. Should I be running a special version of PyTorch, or?
... | closed | completed | false | 8 | [
"bug",
"help wanted",
"Good second issue"
] | [] | 2022-10-13T20:19:47Z | 2026-01-11T22:14:03Z | 2026-01-09T17:58:33Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | WASasquatch | 1,151,589 | MDQ6VXNlcjExNTE1ODk= | User | false |
huggingface/diffusers | 1,408,669,359 | I_kwDOHa8MBc5T9pav | 835 | https://github.com/huggingface/diffusers/issues/835 | https://api.github.com/repos/huggingface/diffusers/issues/835 | numpy_to_pil fails on flax pipeline output | ### Describe the bug
When I try to `numpy_to_pil` the output of FlaxStableDiffusionPipeline, it crashes with
> 'DeviceArray' object has no attribute '__array_interface__'
(maybe possibly related to safety_checker=None)
### Reproduction
```py
pipe, params = diffusers.FlaxStableDiffusionPipeline.from_pretra... | closed | completed | false | 8 | [
"bug",
"stale"
] | [
"pcuenca"
] | 2022-10-14T02:38:51Z | 2022-11-28T15:03:38Z | 2022-11-28T15:03:38Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | keturn | 83,819 | MDQ6VXNlcjgzODE5 | User | false |
huggingface/diffusers | 2,326,843,506 | I_kwDOHa8MBc6KsNBy | 8,356 | https://github.com/huggingface/diffusers/issues/8356 | https://api.github.com/repos/huggingface/diffusers/issues/8356 | EDMEulerScheduler fails when fineturning CosXL_edit model. | When I try to fineturning CosXL model. As they stated in [CosXL](https://huggingface.co/stabilityai/cosxl), CosXL employ Cosine-Continuous EDM VPred scheduler to turn the model. I follow the instruction of [train_dreambooth_lora_sdxl](https://github.com/huggingface/diffusers/blob/main/examples/dreambooth/train_dreamboo... | closed | completed | false | 2 | [] | [] | 2024-05-31T02:24:11Z | 2024-07-05T05:13:04Z | 2024-07-05T05:13:04Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | leeruibin | 26,408,630 | MDQ6VXNlcjI2NDA4NjMw | User | false |
huggingface/diffusers | 2,326,926,182 | I_kwDOHa8MBc6KshNm | 8,358 | https://github.com/huggingface/diffusers/issues/8358 | https://api.github.com/repos/huggingface/diffusers/issues/8358 | Bug in FlaxDiffusionPipeline's from_pt option in Google Collab's TPUv2 | ### Describe the bug
I get an error when I try to load a model with the from_pt option of FlaxDiffusionPipeline in GoogleCollab TPU v2.
Incidentally, it sometimes succeeds in loading.
### Reproduction
pip install -q diffusers
from diffusers import FlaxDiffusionPipeline
import jax
pipe = FlaxDiffusionPipeline.f... | open | null | false | 15 | [
"bug",
"stale"
] | [] | 2024-05-31T04:12:40Z | 2025-02-10T15:06:43Z | null | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | suzukimain | 131,413,573 | U_kgDOB9U2RQ | User | false |
huggingface/diffusers | 2,327,164,488 | I_kwDOHa8MBc6KtbZI | 8,359 | https://github.com/huggingface/diffusers/issues/8359 | https://api.github.com/repos/huggingface/diffusers/issues/8359 | [0.28.0]: from_single_file doesn't work with custom checkpoints | ### Describe the bug
Can't load custom SDXL checkpoint with from_single_file with custom checkpoints since 0.28.0 version. All SDXL pipelines don't work, with or without controlnet.
Triedsetting low_cpu_mem_usage=False and ignore_mismatched_sizes=True, it doesn't change anything.
### Reproduction
pipe = StableDiffu... | closed | completed | false | 8 | [
"bug"
] | [
"DN6"
] | 2024-05-31T07:39:40Z | 2024-11-07T19:34:28Z | 2024-06-03T07:40:09Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | putdanil | 75,547,456 | MDQ6VXNlcjc1NTQ3NDU2 | User | false |
huggingface/diffusers | 1,408,759,277 | I_kwDOHa8MBc5T9_Xt | 836 | https://github.com/huggingface/diffusers/issues/836 | https://api.github.com/repos/huggingface/diffusers/issues/836 | Congrats! | Please allow me to say that the work and performance in this repository is really impressing! 👍
Unbelievable how many improvements are included in the last updates in that short time intervals. Brilliant! I cannot remember ever have seen a repository with comparable power.
(Yes, this "issue" can be closed if d... | closed | completed | false | 1 | [
"stale"
] | [] | 2022-10-14T04:47:29Z | 2022-11-21T15:03:02Z | 2022-11-21T15:03:02Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Marcophono2 | 22,599,855 | MDQ6VXNlcjIyNTk5ODU1 | User | false |
huggingface/diffusers | 2,327,535,661 | I_kwDOHa8MBc6Ku2At | 8,371 | https://github.com/huggingface/diffusers/issues/8371 | https://api.github.com/repos/huggingface/diffusers/issues/8371 | prs-eth/marigold-depth-lcm-v1-0 is not available on huggingface | `prs-eth/marigold-depth-lcm-v1-0` is not available on huggingface
So I think that the path of marigold in [Releases](https://github.com/huggingface/diffusers/releases) [v0.28.0](https://github.com/huggingface/diffusers/releases/tag/v0.28.0) is wrong.
![image](https://github.com/huggingface/diffusers/assets/4391... | closed | completed | false | 5 | [] | [] | 2024-05-31T10:48:07Z | 2024-06-01T07:13:30Z | 2024-06-01T06:50:26Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | YoucanBaby | 43,912,540 | MDQ6VXNlcjQzOTEyNTQw | User | false |
huggingface/diffusers | 2,328,289,543 | I_kwDOHa8MBc6KxuEH | 8,373 | https://github.com/huggingface/diffusers/issues/8373 | https://api.github.com/repos/huggingface/diffusers/issues/8373 | "Input is in incorrect format." incorrect error message | ### Describe the bug
https://github.com/huggingface/diffusers/blob/983dec3bf787c064ed57f2621c9b7375d443f746/src/diffusers/image_processor.py#L536
`f"Input is in incorrect format. Currently, we only support {', '.join(supported_formats)}"` fails because [`supported_formats`](https://github.com/huggingface/diffusers/... | closed | completed | false | 2 | [
"bug"
] | [] | 2024-05-31T17:27:17Z | 2024-07-05T05:12:42Z | 2024-07-05T05:12:42Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | LWprogramming | 13,173,037 | MDQ6VXNlcjEzMTczMDM3 | User | false |
huggingface/diffusers | 1,409,030,404 | I_kwDOHa8MBc5T_BkE | 838 | https://github.com/huggingface/diffusers/issues/838 | https://api.github.com/repos/huggingface/diffusers/issues/838 | Flax pipeline: width and height are ignored | ### Describe the bug
`width` and `height` are passed to the generate function, and we do check that they are multiples of `8`, but then we always create the latents using the standard size.
### Reproduction
```Python
images = pipeline(prompt_ids, p_params, rng, width=448, height=832, jit=True)[0]
images.shape
# o... | closed | completed | false | 3 | [
"bug",
"stale"
] | [] | 2022-10-14T09:11:22Z | 2022-11-21T15:03:01Z | 2022-11-21T15:03:01Z | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | pcuenca | 1,177,582 | MDQ6VXNlcjExNzc1ODI= | User | false |
huggingface/diffusers | 2,330,094,897 | I_kwDOHa8MBc6K4m0x | 8,380 | https://github.com/huggingface/diffusers/issues/8380 | https://api.github.com/repos/huggingface/diffusers/issues/8380 | [0.28.0.dev0] In ControlNetModel forward function, added_cond_kwargs arg's default value should be {} rather than None | ### Describe the bug
**Version**: [0.28.0.dev0]
**Bug description**: In ControlNetModel forward function, added_cond_kwargs arg's default value should be {} rather than None
**Reason**: in the source code(controlnet.py):
elif self.config.addition_embed_type == "text_time":
if "text_embeds"... | closed | completed | false | 5 | [
"bug"
] | [] | 2024-06-03T03:37:47Z | 2024-06-03T09:05:12Z | 2024-06-03T09:05:11Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | CatLoves | 7,496,455 | MDQ6VXNlcjc0OTY0NTU= | User | false |
huggingface/diffusers | 2,330,273,902 | I_kwDOHa8MBc6K5Shu | 8,382 | https://github.com/huggingface/diffusers/issues/8382 | https://api.github.com/repos/huggingface/diffusers/issues/8382 | Training ControlNet-Xs | Thank you @UmerHA for contributing the ControlNet demo. I tried to create a training script so I could train the model on my data but had some problems using the UNetControlNetXSModel.
Can you provide training or give me some suggestions? | closed | completed | true | 3 | [] | [] | 2024-06-03T06:22:44Z | 2024-06-03T12:56:13Z | 2024-06-03T12:56:13Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | nguyendung1612 | 43,613,837 | MDQ6VXNlcjQzNjEzODM3 | User | false |
huggingface/diffusers | 2,330,449,819 | I_kwDOHa8MBc6K59eb | 8,384 | https://github.com/huggingface/diffusers/issues/8384 | https://api.github.com/repos/huggingface/diffusers/issues/8384 | [Community] Help us fix the LR schedulers when `num_train_epochs` is passed in a distributed training env | ## Context
Refer to https://github.com/huggingface/diffusers/pull/8312 for the full context. The changes introduced in the PR should be propagated to the following scripts, too:
- advanced_diffusion_training
- [x] train_dreambooth_lora_sd15_advanced.py
- [x] train_dreambooth_lora_sdxl_advanced.py
- consisten... | open | reopened | false | 5 | [
"Good second issue",
"contributions-welcome"
] | [] | 2024-06-03T08:03:07Z | 2025-05-13T02:05:11Z | null | MEMBER | null | 20260407T133413Z | 2026-04-07T13:34:13Z | sayakpaul | 22,957,388 | MDQ6VXNlcjIyOTU3Mzg4 | User | false |
huggingface/diffusers | 2,331,575,515 | I_kwDOHa8MBc6K-QTb | 8,390 | https://github.com/huggingface/diffusers/issues/8390 | https://api.github.com/repos/huggingface/diffusers/issues/8390 | Stable Cascade Controlnet training | **Is your feature request related to a problem? Please describe.**
Last year I wrote [a long article on how to train controlnets](https://civitai.com/articles/2078) using diffusers, and trained [two](https://civitai.com/models/137638) [controlnets](civitai.com/models/149740) using diffusers. A few months ago Stable Ca... | closed | completed | false | 6 | [
"stale",
"contributions-welcome"
] | [] | 2024-06-03T16:24:47Z | 2024-11-09T15:25:25Z | 2024-11-09T15:25:25Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | geroldmeisinger | 112,266,044 | U_kgDOBrELPA | User | false |
huggingface/diffusers | 2,331,733,119 | I_kwDOHa8MBc6K-2x_ | 8,392 | https://github.com/huggingface/diffusers/issues/8392 | https://api.github.com/repos/huggingface/diffusers/issues/8392 | StableDiffusionControlNetpaint destroying original image contrast and sharpness | why controlnet paint destroy the original state of color contrast | closed | completed | false | 11 | [] | [] | 2024-06-03T17:57:10Z | 2024-06-03T23:46:02Z | 2024-06-03T23:46:02Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | xalteropsx | 103,671,642 | U_kgDOBi3nWg | User | false |
huggingface/diffusers | 2,332,488,529 | I_kwDOHa8MBc6LBvNR | 8,393 | https://github.com/huggingface/diffusers/issues/8393 | https://api.github.com/repos/huggingface/diffusers/issues/8393 | Take too much time twhen loading checkpoint shards | ### Describe the bug
I tried to use `from_pretrained` function in [Latte](https://github.com/Vchitect/Latte) to automatically download related models from `hugginface`, but it took a long time to load the T5 model:
https://github.com/Vchitect/Latte/blob/main/sample/sample_t2v.py#L36
![image](https://github.com/h... | closed | completed | true | 1 | [
"bug"
] | [] | 2024-06-04T03:59:55Z | 2024-06-04T06:57:30Z | 2024-06-04T06:57:30Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | maxin-cn | 38,418,898 | MDQ6VXNlcjM4NDE4ODk4 | User | false |
huggingface/diffusers | 2,333,129,422 | I_kwDOHa8MBc6LELrO | 8,398 | https://github.com/huggingface/diffusers/issues/8398 | https://api.github.com/repos/huggingface/diffusers/issues/8398 | Make the input `UNet2DConditionModel` and the returned `UNetMotionModel` instances share weights | **What API design would you like to have changed or added to the library? Why?**
[diffusers.UNetMotionModel](https://huggingface.co/docs/diffusers/main/en/api/models/unet-motion#diffusers.UNetMotionModel).from_unet2d
https://github.com/huggingface/diffusers/blob/6be43bd855772f5ffb064bc7c6049f578b3856f8/src/diffus... | open | null | false | 2 | [
"stale"
] | [] | 2024-06-04T10:14:39Z | 2024-09-14T15:10:50Z | null | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | ernestchu | 51,432,514 | MDQ6VXNlcjUxNDMyNTE0 | User | false |
huggingface/diffusers | 1,302,647,111 | I_kwDOHa8MBc5NpNFH | 84 | https://github.com/huggingface/diffusers/issues/84 | https://api.github.com/repos/huggingface/diffusers/issues/84 | [Feature] Add Langevin Dynamics Sampler | The score-based generative models sometimes us a Langevin Dynamics step to de-noise samples; see [NCSN](https://arxiv.org/abs/2006.09011). In the diffusion model for generating molecule conformations, [GeoDiff](https://github.com/MinkaiXu/GeoDiff), this step function had the best results (even though it is simpler than... | closed | not_planned | false | 7 | [
"Good second issue"
] | [] | 2022-07-12T21:37:00Z | 2024-11-21T11:19:50Z | 2024-11-21T11:19:50Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | natolambert | 10,695,622 | MDQ6VXNlcjEwNjk1NjIy | User | false |
huggingface/diffusers | 2,333,611,005 | I_kwDOHa8MBc6LGBP9 | 8,400 | https://github.com/huggingface/diffusers/issues/8400 | https://api.github.com/repos/huggingface/diffusers/issues/8400 | how can we load model to lora from singlefile ? | pipe.load_lora_weights("lora/aesthetic_anime_v1s.safetensors")
File "Z:\software\python11\Lib\site-packages\diffusers\loaders\lora.py", line 1230, in load_lora_weights
raise ValueError("PEFT backend is required for this method.")
ValueError: PEFT backend is required for this method.
pipe.load_lora_weigh... | closed | completed | false | 2 | [] | [] | 2024-06-04T13:54:56Z | 2024-06-04T15:53:32Z | 2024-06-04T14:28:37Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | xalteropsx | 103,671,642 | U_kgDOBi3nWg | User | false |
huggingface/diffusers | 2,335,354,938 | I_kwDOHa8MBc6LMrA6 | 8,405 | https://github.com/huggingface/diffusers/issues/8405 | https://api.github.com/repos/huggingface/diffusers/issues/8405 | Use multiple controlnets tensor size error | ### Describe the bug
I get a wrong tensor size error when trying to run multiple controlnets
### Reproduction
```
import argparse
import os
import random
import cv2
import numpy as np
import torch
from PIL import Image
from diffusers import (
AutoencoderKL,
EulerAncestralDiscreteScheduler, )
fro... | open | null | false | 3 | [
"bug",
"stale"
] | [] | 2024-06-05T09:22:26Z | 2024-09-14T15:10:47Z | null | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | iliasprc | 32,684,977 | MDQ6VXNlcjMyNjg0OTc3 | User | false |
huggingface/diffusers | 2,335,770,978 | I_kwDOHa8MBc6LOQli | 8,407 | https://github.com/huggingface/diffusers/issues/8407 | https://api.github.com/repos/huggingface/diffusers/issues/8407 | UniPC scheduler does not support new set_timesteps() kwargs | ### Describe the bug
When using UniPC for SDXL Img2Img pipeline, we run into an error:
```
scheduler.set_timesteps(num_inference_steps, device=device, **kwargs)
TypeError: set_timesteps() got an unexpected keyword argument 'sigmas'
```
The function that is called:
```
def retrieve_timesteps(
sche... | closed | completed | false | 2 | [
"bug"
] | [] | 2024-06-05T12:26:41Z | 2024-06-05T13:12:09Z | 2024-06-05T13:12:09Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | bghira | 59,658,056 | MDQ6VXNlcjU5NjU4MDU2 | User | false |
huggingface/diffusers | 1,409,195,522 | I_kwDOHa8MBc5T_p4C | 841 | https://github.com/huggingface/diffusers/issues/841 | https://api.github.com/repos/huggingface/diffusers/issues/841 | [Community Pipelines] | # Community Pipelines
As of `diffusers==0.4.0`, you can make use of [Community Pipelines](https://huggingface.co/docs/diffusers/using-diffusers/custom_pipelines).
The goal with community pipelines is to have ***a community-driven offering of an exotic variety of features built on top of `diffusers` which is main... | closed | completed | false | 28 | [
"good first issue",
"community-examples"
] | [] | 2022-10-14T11:20:00Z | 2026-01-09T18:00:53Z | 2026-01-09T18:00:53Z | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | patrickvonplaten | 23,423,619 | MDQ6VXNlcjIzNDIzNjE5 | User | false |
huggingface/diffusers | 2,336,346,108 | I_kwDOHa8MBc6LQc_8 | 8,410 | https://github.com/huggingface/diffusers/issues/8410 | https://api.github.com/repos/huggingface/diffusers/issues/8410 | support TextDiffuser pipelines natively in Diffusers | ### Model/Pipeline/Scheduler description
TextDiffuser is an open-source framework and pretrained model series for generating unique depth maps that can be used to guide the model for typography.
It not only... | open | null | false | 5 | [
"stale",
"contributions-welcome"
] | [] | 2024-06-05T16:25:25Z | 2024-12-14T15:05:40Z | null | CONTRIBUTOR | null | 20260407T133413Z | 2026-04-07T13:34:13Z | bghira | 59,658,056 | MDQ6VXNlcjU5NjU4MDU2 | User | false |
huggingface/diffusers | 2,336,653,519 | I_kwDOHa8MBc6LRoDP | 8,413 | https://github.com/huggingface/diffusers/issues/8413 | https://api.github.com/repos/huggingface/diffusers/issues/8413 | sequential crashing the embedding system forward | ``` ^^^^^^^^^^^^^^^^^^^^^
File "Z:\software\python11\Lib\site-packages\compel\compel.py", line 135, in __call__
output = self.build_conditioning_tensor(text_input)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "Z:\software\python11\Lib\site-packages\compel\compel.py", line 112, in b... | closed | completed | false | 5 | [] | [] | 2024-06-05T19:15:42Z | 2024-06-05T20:20:35Z | 2024-06-05T20:19:30Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | xalteropsx | 103,671,642 | U_kgDOBi3nWg | User | false |
huggingface/diffusers | 2,336,896,788 | I_kwDOHa8MBc6LSjcU | 8,414 | https://github.com/huggingface/diffusers/issues/8414 | https://api.github.com/repos/huggingface/diffusers/issues/8414 | [🌟 New Model] ConsistencyTTA: Accelerating Diffusion-Based Text-to-Audio Generation with Consistency Distillation | ### Model/Pipeline/Scheduler description
ConsistencyTTA, introduced in the paper [_Accelerating Diffusion-Based Text-to-Audio Generation
with Consistency Distillation_](https://arxiv.org/abs/2309.10740), is an efficient text-to-audio generation model. Compared to a comparable diffusion-based TTA model, ConsistencyT... | open | null | false | 13 | [
"stale",
"contributions-welcome"
] | [] | 2024-06-05T21:55:23Z | 2024-12-04T15:05:27Z | null | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | Bai-YT | 24,694,069 | MDQ6VXNlcjI0Njk0MDY5 | User | false |
huggingface/diffusers | 2,338,612,306 | I_kwDOHa8MBc6LZGRS | 8,418 | https://github.com/huggingface/diffusers/issues/8418 | https://api.github.com/repos/huggingface/diffusers/issues/8418 | Request to add discriminator based guidance | Would it be possible to get discriminator based guidance from the paper "Refining Generative Process with Discriminator Guidance in Score-based Diffusion Models" that was published at ICML'23.
Paper link: https://proceedings.mlr.press/v202/kim23i/kim23i.pdf
Thanks,
Anubhav | open | null | false | 7 | [
"stale",
"consider-for-modular-diffusers"
] | [] | 2024-06-06T15:58:11Z | 2025-01-11T15:05:52Z | null | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | anubhav1997-sony | 169,923,721 | U_kgDOCiDUiQ | User | false |
huggingface/diffusers | 2,339,432,266 | I_kwDOHa8MBc6LcOdK | 8,421 | https://github.com/huggingface/diffusers/issues/8421 | https://api.github.com/repos/huggingface/diffusers/issues/8421 | sorry for my mistake | null | closed | completed | false | 0 | [] | [] | 2024-06-07T01:51:50Z | 2024-06-07T02:14:06Z | 2024-06-07T02:14:06Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | RobinLiuZX | 80,163,805 | MDQ6VXNlcjgwMTYzODA1 | User | false |
huggingface/diffusers | 2,339,928,761 | I_kwDOHa8MBc6LeHq5 | 8,424 | https://github.com/huggingface/diffusers/issues/8424 | https://api.github.com/repos/huggingface/diffusers/issues/8424 | controlNet training for dstilled models | Is it possible to train controlNet for distilled model (Block-removed Knowledge-distilled Stable Diffusion, https://github.com/Nota-NetsPresso/BK-SDM)?
When I used scripts in example folder for distilled model, it returns error, similar to "unknown mid_block_type ...." , probably because of unet structure changed. A... | open | null | false | 3 | [
"stale"
] | [] | 2024-06-07T08:48:10Z | 2024-09-14T15:10:28Z | null | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | kmpartner | 17,810,734 | MDQ6VXNlcjE3ODEwNzM0 | User | false |
huggingface/diffusers | 2,340,709,684 | I_kwDOHa8MBc6LhGU0 | 8,429 | https://github.com/huggingface/diffusers/issues/8429 | https://api.github.com/repos/huggingface/diffusers/issues/8429 | [I2VGen-XL] The Official Inference Script of I2VGenXLPipline does not Function Well | ### Describe the bug
Hi,
Thank you for your amazing work of `diffusers`, which has brought a whole lot of convenience to researchers of diffusion models and image synthesis.
I downloaded the diffusers model weights of `I2VGen-XL` and was trying to run its official inference script, which is shown below:
```pyth... | closed | completed | false | 4 | [
"bug"
] | [] | 2024-06-07T15:26:09Z | 2024-06-08T07:10:08Z | 2024-06-08T07:08:58Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | AlonzoLeeeooo | 68,578,066 | MDQ6VXNlcjY4NTc4MDY2 | User | false |
huggingface/diffusers | 1,409,336,868 | I_kwDOHa8MBc5UAMYk | 843 | https://github.com/huggingface/diffusers/issues/843 | https://api.github.com/repos/huggingface/diffusers/issues/843 | Prompt Syntax Documentation | **Is your feature request related to a problem? Please describe.**
I do not know what syntax I can use when typing a prompt in, e.g. weighting multiple prompts, negative weights etc.
**Describe the solution you'd like**
I would like documentation and/or tool-tips that explain what syntax is supported for modifying... | closed | not_planned | false | 2 | [] | [] | 2022-10-14T13:13:47Z | 2022-10-14T19:06:57Z | 2022-10-14T13:17:10Z | NONE | null | 20260407T133413Z | 2026-04-07T13:34:13Z | 9of9 | 31,524,061 | MDQ6VXNlcjMxNTI0MDYx | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.