fix GQA compat PyTorch 2.4 and strip _orig_mod. on checkpoint load ef31fa4 Juan S Santillana commited on 26 days ago
Upload training_v2/train/utils.py with huggingface_hub c32c1c8 verified jsantillana commited on 28 days ago
Upload training_v2/train/pretrain_750m.py with huggingface_hub 56ecf73 verified jsantillana commited on 28 days ago