Hunyuan text2video Lora. Anime, Dragon Ball Z ~ドラゴンボールZ. v1

I think model is overfitted with Goku. I wanted to experiment with "Dragon Ball Z" tag but for some reason Goku appears everywhere even when his name is not in the prompt. Outputs funny but little bit cursed xd.

Description

Hunyuan Lora model trained on the short clips of Dragon Ball Z from the first episode of the anime, 34 clips in total with avg length: 2.78 sec.

Trained using Diffusion-pipe repo.

Inference params.

• lora_strength: 1.0

• dtype: bfloat16

• resolution: [[704,576]] (width, heigth)

• num_frames: 93

• steps: 20

• embedded_guidance_scale: 9.00 *note I found that this value was good for my other lora so I used same here, I think it worth to experiment with;

• enhance video weight: 4.0 *note I think this parameter also can be adjusted and there are some other params in enhance video node.

Data

• amount: 34 clips from 0.33 to 5.91 sec.

• avg_length: 2.76 sec