[Bug]: Could you update the configuration process for LLAMA-7B? EasyLM seems to be updated and the old configuration is not correct anymore.

[Tianshi-Xu]

Issue Type

Build/Install

Modules Involved

SPU runtime

Have you reproduced the bug with SPU HEAD?

Yes

Have you searched existing issues?

Yes

SPU Version

latest

OS Platform and Distribution

Ubuntu 22.04

Python Version

3.10

Compiler Version

GCC 11.3

Current Behavior?

The EasyLM has been updated and the steps in README.md in examples/python/ml/flax_llama7b are incorrect.

Standalone code to reproduce the issue

Lots of errors

Relevant log output

No response

[Tianshi-Xu]

Main errors come from

python convert_hf_to_easylm.py  \
   --hf_model     /home/code/models/Llama-7b    \
   --output_file flax-llama7b-EasyLM.msgpack  \
   --streaming false

[tpppppub]

It seems to be a duplicate of #704

[Tianshi-Xu]

It seems to be a duplicate of #704

Yes, it can work with some tedious modifications, but I recommend updating the readme or giving some hits to prevent future users from these errors. After all, LLAMA is an important model in PPDL.
Thanks a lot!