Implementation / replication of DALL-E (paper), OpenAI's Text to Image Transformer, in Pytorch. It will also contain CLIP for ranking the generations. Kobiso, a research engineer from Naver, has trained on the CUB200 dataset here, using full and deepspeed sparse attention. You can also skip the training of the VAE altogether, using the pretrained model released by OpenAI! The wrapper class should take care of downloading and caching the model for you auto-magically. You can also use the pretrained VAE offered by the authors of Taming Transformers! Currently only the VAE with a codebook size of 1024 is offered, with the hope that it may train a little faster than OpenAI's, which has a size of 8192. In contrast to OpenAI's VAE, it also has an extra layer of downsampling, so the image sequence length is 256 instead of 1024 (this will lead to a 16 reduction in training costs, when you do the math).
Features
- Train DALL-E with pretrained VAE
- The default VQGan is the codebook size 1024 one trained on imagenet
- Adjust text conditioning strength
- Rank the generations
- Deepspeed Sparse Attention
- Train with Microsoft Deepspeed's Sparse Attention
Project Samples
