ArcShot-Wan2.2/2.1-I2V-A14B

UPDATED Wan-2.2-HighNoise

ArcShot-Wan2.2-I2V-A14B. It is trained using film material, but I found both anime and comic can used it to generate a good result by adding prompts like anime style or comic style. And it's trained with direction caption , e.g. camera moving from left to right/right to left around the subject. However, it seems not that work on inferencing. You can try. The result is quite good during my testing. You can use MediaInfo to check the generation info in the video.If you like my job , you can donate me here: style="background-color:rgb(26, 27, 30);color:rgb(193, 194, 197);font-family:-apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, Helvetica, Arial, sans-serif, "Apple Color Emoji", "Segoe UI Emoji";font-size:16px">. Just paying the electricity bill costs me 1,000 yuan a month...This LoRA is not for commercial use. Please lease dm me, If you have any business cooperation inquiries.This LoRA is trained for Wan2.2-high noise. Please make sure add it to the high noise Wan2.2 model.

Suggest Strength : 0.8

Prompts: arc shot, camera moving from left to right around the subject,

Wan-2.1

ArcShot-Wan2.1-I2V-720P. It is trained using film material, but I found both anime and comic can used it to generate a good result by adding prompts like anime style or comic style. And the background will switch to forest usually, you can add some prompts like indoor background to keep the background not changed. Suggest Strength : 0.9-1.0

I uploaded an article of my experience of training Wan2.1 LoRA (Both I2V & T2V): My Personally Training Experience of WAN:Starting with Data.