Yukang/Llama-2-13b-longlora-16k does not appear to have a file named pytorch_model.bin, tf_model.h5, model.ckpt or flax_model.msgpack.

#1
by GBaker - opened

Missing model files?

This is a lora weight. You need to run merge_lora_weights_and_save_hf_model.py provided in the GitHub repo to get the pytorch_model.bin.

Please refer to this https://github.com/dvlab-research/LongLoRA#merge-lora-weight

Ah, thanks.

Sign up or log in to comment