search

Weifeng-Chen / control-a-video

Official Implementation of "Control-A-Video: Controllable Text-to-Video Generation with Diffusion Models"
GNU General Public License v3.0
359 stars 26 forks source link

Is Controlnet and Unet trained jointly or seperately? #3

Closed Semolce closed 1 year ago

Semolce commented 1 year ago

Hi, great work! I wonder do you train the controlnet and unet jointly or seperately (like train unet first and then controlnet or vice versa)? Thanks in advance.

Weifeng-Chen commented 1 year ago

Hi, great work! I wonder do you train the controlnet and unet jointly or seperately (like train unet first and then controlnet or vice versa)? Thanks in advance.

jointly training