qwerrwe / examples /phi /README.md
winglian's picture
make phi training work with Loras (#588)
62eaee7 unverified
|
raw
history blame
276 Bytes
# Phi
Due to some nuances with the phi code, please use deepspeed when training phi for full finetune.
```shell
accelerate launch -m axolotl.cli.train examples/phi/phi-ft.yml --deepspeed deepspeed/zero1.json
# OR
python -m axolotl.cli.train examples/phi/phi-qlora.yml
```