Pretraining details?
#1
by
devingulliver
- opened
Your model performs astonishingly well for its size. It would be of great use to the open-source LLM community to know the dataset and hyperparameters used to train it.