justinpinkney / stable-diffusion

MIT License
1.45k stars 266 forks source link

The loss is 0.1 #69

Open pokameng opened 1 year ago

pokameng commented 1 year ago

@justinpinkney When I train with the image as a condition, the loss is always at 0.1, why is this? image and the sample is not good. image