Training run fails to start with error "TypeError: openssl_md5() takes at most 1 argument (2 given)"

[vmay-chegg]

Please check that this issue hasn't been reported before.

• I searched previous Bug Reports didn't find any similar reports.

Expected Behavior

The training run should start.

Current behaviour

Getting the error:
TypeError: openssl_md5() takes at most 1 argument (2 given)

Steps to reproduce

Follow the Quickstart steps from Readme.md:

git clone https://github.com/OpenAccess-AI-Collective/axolotl

pip3 install -e .
pip3 install -U git+https://github.com/huggingface/peft.git
accelerate launch scripts/finetune.py examples/openllama-3b/lora.yml

Env info:
Ubuntu
8x A100 GPU
Python 3.10.12

Possible solution

No response

Which Operating Systems are you using?

• Linux
• macOS
• Windows

Python Version

3.10.12

axolotl branch-commit

main/c01015f33f257badd05fbf8e680f22c294e9628e

Acknowledgements

• My issue title is concise, descriptive, and in title casing.
• I have searched the existing issues to make sure this bug has not been reported yet.
• I am using the latest version of axolotl.
• I have provided enough information for the maintainers to reproduce and diagnose the issue.

[NanoCode012]

Could you provide stack trace please?

[winglian]

Caused by these calls here
https://github.com/OpenAccess-AI-Collective/axolotl/blob/f806e86a6ef253937ec6433884ca1c0ac47d4066/src/axolotl/utils/data.py#L471-L476

The linter required I add usedforsecurity=False since we don't need cryptographic hashes. Might be related to the version of openssl

[winglian]

perhaps implement something like fabianvf/python@1ac54ef

[NanoCode012]

Weirdly, it says it was added in py3.9 in that linked Issue, but the author uses python 3.10

[adarshxs]

Any fix? I encountered the same error

[NanoCode012]

@adarshxs , could you provide your env and stack trace? Are you running in docker or conda?