This is a pre-trained Score Entropy Discrete Diffusion (SEDD) model. Vocab size is 50256 and uses the GPT2 Tokenizer. Trained on OpenWebText.
