FoundationVision / VAR

[GPT beats diffusion🔥] [scaling laws in visual generation📈] Official impl. of "Visual Autoregressive Modeling: Scalable Image Generation via Next-Scale Prediction". An *ultra-simple, user-friendly yet state-of-the-art* codebase for autoregressive image generation!
MIT License
3.8k stars 285 forks source link

question about 2 stage training #15

Closed plutoyuxie closed 2 months ago

plutoyuxie commented 2 months ago

Is a multi-scale VQVAE trained in stage 1 used for all VAR transformers with different parameters ?

keyu-tian commented 2 months ago

@plutoyuxie yes