
Cartoon characters often hit trees, poles, rocks, mountain sides, fake tunnels, and various other things at high speeds. I wanted a lora that could do that. This probably would be better as a FLUX lora, but those are very expensive.
描述:
Samples weren't promising
Currently this seems to recognize the concept when a tree is involved. Characters often face the "wrong way", with their backs against the object in the output.
Probably needs more images in training if I ever update this. This was trained on 10 images only
训练词语: objectimpact
名称: 1101966_training_data.zip
大小 (KB): 4580
类型: Training Data
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success
名称: Colliding_against_object-000019.safetensors
大小 (KB): 223101
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success