Version 2 uses 62 images instead of just 34. This version also has increased "from behind" and multiple girls training.

Gymnasts and Gymnastics

Trained on 62 images instead of just 34 with better behind shot training.

Gymnasts_and_Gymnastics.safetensors

My first lora intended to be a proof of concept more than anything. A better version will likely come later. Intended to produce images of NCAA gymnasts on balance beam, uneven bars, and floor. You might want to keep the lora weight around 0.4 to 0.5

Increased the total number of images trained on from 34 to 62. Added more "from behind" photos.

Files with SHA256: 97e1e6e93084099b0fa1c07046da74607ac252d951c12efd97e21871284c34a9