Zengtudor/Wan2.1
1
0
Fork 0
mirror of https://github.com/Wan-Video/Wan2.1.git synced 2025-11-04 22:26:36 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Fixed typos

Browse Source
This commit is contained in:
DeepBeepMeep 2025-03-02 18:43:45 +01:00
parent f3b365a5da
commit 0085a0b0c6
2 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
README.md
View File

@ -119,7 +119,7 @@ To run the application while loading entirely the diffusion model in VRAM (sligh
```bash
python gradio_server.py --profile 3
```
Please note that diffusion model of Wan2.1GP is extremely VRAM optimized and this will greatly benefit low VRAM systems since the diffusion / denoising step is the longest part of the generation process. However, the VAE encoder (at the beginning of a image 2 video process) and the VAE decoder (at the end of any video process) is only 20% lighter and it will require temporarly 22 GB of VRAM for a 720p generation and 12 GB of VRAM for a 480p generation. Therefore if you have less than these numbers, you may experience slow down at the begining and at the end of the generation process due to pytorch VRAM offloading.
Please note that diffusion model of Wan2.1GP is extremely VRAM optimized and this will greatly benefit low VRAM systems since the diffusion / denoising step is the longest part of the generation process. However, the VAE encoder (at the beginning of a image 2 video process) and the VAE decoder (at the end of any video process) is only 20% lighter and it will require temporarly 22 GB of VRAM for a 720p generation and 12 GB of VRAM for a 480p generation. Therefore if you have less than these numbers, you may experience slow down at the beginning and at the end of the generation process due to pytorch VRAM offloading.
### Loras support

4
gradio_server.py
View File

@ -458,7 +458,7 @@ def generate_header(model_filename, compile, attention_mode):
if "image" in model_filename:
model_name = "Wan2.1 image2video"
model_name += "720p" if "720p" in model_filename else "480"
model_name += "720p" if "720p" in model_filename else "480p"
else:
model_name = "Wan2.1 text2video"
model_name += "14B" if "14B" in model_filename else "1.3B"
@ -977,7 +977,7 @@ def create_demo():
gr.Markdown("- 848 x 480 with the 1.3B model: 80 frames (5s) : 5 GB of VRAM")
gr.Markdown("- 1280 x 720 with a 14B model: 80 frames (5s): 11 GB of VRAM")
gr.Markdown("Note that the VAE stages (encoding / decoding at image2video ) or just the decoding at text2video will create a temporary VRAM peaks (up to 12GB for 420P and 22 GB for 720P)")
gr.Markdown("It is not recommmended to generate a video longer than 8s (128 frames) even if there is still some VRAM left as some artifact may appear")
gr.Markdown("It is not recommmended to generate a video longer than 8s (128 frames) even if there is still some VRAM left as some artifacts may appear")
gr.Markdown("Please note that if your turn on compilation, the first generation step of the first video generation will be slow due to the compilation. Therefore all your tests should be done with compilation turned off.")