1.7K views · 13 reactions | Turned @karpathy's baby GPT into a character-level text diffusion model, entropy-based training objective. Run it in Google Colab: - https://colab.research.google.com/github/ash80/diffusion-gpt/blob/master/The_Annotated_Discrete_Diffusion_Models.ipynb - https://arxiv.org/abs/2310.16834 And the source of Andrej Karpathy nanochat:-https://www.facebook.com/share/p/1JcR2V7maB/ | Md Ismail Sojal | Facebook