search

Zhendong-Wang / Prompt-Diffusion

Official PyTorch implementation of the paper "In-Context Learning Unlocked for Diffusion Models"
Apache License 2.0
364 stars 10 forks source link

Training #8

Open Jasonhyw opened 8 months ago

Jasonhyw commented 8 months ago

Hi, how do I retrain using train.py? I would like to reproduce the experimental results first. I followed the scrpits/.sh file for training. f2e87a95ba27167c67569769cb65ea2 Does the above code need to be commented?

Zhendong-Wang commented 8 months ago

No. The resume path is for resuming from the pretrained Stable Diffusion v1.5 checkpoint. Our model is finetuned based on that ckpt. Note you need to first run python tool_add_control.py to create the control_sd15_ini.ckpt.