Hello! thanks for sharing your great work!
For better result, I think, using Wan2.1 or 2.2-14B model is a must.
Can your code/model support it? or should I retrain it? or just some code editings are required?
I tried just download ./Wan2.1-T2V-14B-Diffusers and then set model_path: "./Wan2.1-T2V-14B-Diffusers" #Wan2.1-T2V-1.3B-Diffusers" in the b2 config but it still uses 32GB GPU memory and takes about 1 min to generate 256x256x49 frames video same as Wan 2.1 1.3B video model.
Please give me some guide.
See you soon
Thank you!
Seunghyun Lee.
Hello! thanks for sharing your great work!
For better result, I think, using Wan2.1 or 2.2-14B model is a must.
Can your code/model support it? or should I retrain it? or just some code editings are required?
I tried just download ./Wan2.1-T2V-14B-Diffusers and then set model_path: "./Wan2.1-T2V-14B-Diffusers" #Wan2.1-T2V-1.3B-Diffusers" in the b2 config but it still uses 32GB GPU memory and takes about 1 min to generate 256x256x49 frames video same as Wan 2.1 1.3B video model.
Please give me some guide.
See you soon
Thank you!
Seunghyun Lee.