MCPcopy Index your code
hub / github.com/lm-sys/FastChat / train_mem.py

File train_mem.py

fastchat/train/train_mem.py:None–None  ·  view source on GitHub ↗

Source from the content-addressed store, hash-verified

1# Make it more memory efficient by monkey patching the LLaMA model with FlashAttn.
2
3# Need to call this before importing transformers.
4from fastchat.train.llama2_flash_attn_monkey_patch import (

Callers

nothing calls this directly

Calls 2

trainFunction · 0.90

Tested by

no test coverage detected