Wan video 14b 480p lora trained for i2v (but also supports t2v)
- trained for 24 hours using 3090
- trained in resolution: fp8 304x304 (33 and 49 frames) batch 1 (31 GB VRAM)
- Dataset: 28 photos and 51 videos
prompt: hand_grab, woman standing, camera zooms in, man's right hand is grabbing her butt in shorts, back view, woman is standing at kitchen
1.3b lora wasn't looking good, so i dropped training it.
Inference Providers
NEW
This model isn't deployed by any Inference Provider.
🙋
Ask for provider support
Model tree for Ftfyhh/wan_hand_grab_lora_14b
Base model
Wan-AI/Wan2.1-I2V-14B-480P