nodes/city96-ComfyUI_ExtraModels
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Update README.md

Browse Source
This commit is contained in:
City
2024-05-20 11:52:48 +02:00
committed by GitHub
parent bad1b65c09
commit ea705d06e8
1 changed files with 21 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
README.md
+21
View File
@@ -93,6 +93,27 @@ Everything else can be the same the same as in the example above.
## HunYuan DiT
WIP implementation of [HunYuan DiT by Tencent](https://github.com/Tencent/HunyuanDiT)
> [!WARNING]
> Only a proof of concept, most things don't work yet and only 1024x1024 is supported.
>
> The text encoder device/dtype selection also doesn't work.
The initial work on this was done by [chaojie](https://github.com/chaojie) in [this PR](https://github.com/city96/ComfyUI_ExtraModels/pull/37).
Instructions:
- Download the [first text encoder from here](https://huggingface.co/Tencent-Hunyuan/HunyuanDiT/blob/main/t2i/clip_text_encoder/pytorch_model.bin) and place it in `ComfyUI/models/clip` - rename to "chinese-roberta-wwm-ext-large.bin"
- Download the [second text encoder from here](https://huggingface.co/Tencent-Hunyuan/HunyuanDiT/blob/main/t2i/mt5/pytorch_model.bin) and place it in `ComfyUI/models/T5` - rename it to "mT5.bin"
- Download the [model file from here](https://huggingface.co/Tencent-Hunyuan/HunyuanDiT/blob/main/t2i/model/pytorch_model_module.pt) and place it in `ComfyUI/checkpoints` - rename it to "HunYuanDiT.pt"
- Download/use any SDXL VAE, for example [this one](https://huggingface.co/madebyollin/sdxl-vae-fp16-fix)
![image](https://github.com/city96/ComfyUI_ExtraModels/assets/125218114/7a9d6e34-d3f4-4f67-a17f-4f2d6795e54e)
## DiT
[Original Repo](https://github.com/facebookresearch/DiT)