mlcommons / mlcommons/cm4mlops

Can't set output location for model download

Open
#658 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
20
Forks
21
PR merge metrics
No merged PRs in 30d

Description

I have a script which pulls a dataset and model for the Llama2-70b workload to my server. After the user enters the HF token, it is expected that the model will download to a specified location, not merely to the 'cache'. The dataset does this successfully by setting a parameter 'outdirname', but the model does not. From past experience I attempted the following steps, but in all cases, the model remains only in the cache:

  • Setting the '--to' flag
  • Setting the '--outdirname' flag
  • Setting these environmental variables: LLAMA2_CHECKPOINT_PATH and CM_ML_MODEL_PATH

This is the script I'm trying to pull with: https://github.com/mlcommons/cm4mlops/tree/mlperf-inference/script/get-ml-model-llama2

Is there any way to direct the final output location?

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start with script/get-ml-model-llama2 and trace how the HF token, --to, --outdirname, LLAMA2_CHECKPOINT_PATH, and CM_ML_MODEL_PATH are handled. Reproduce the download and verify that the final model files are placed in the requested output directory rather than only in the cache.

Written by the indexing model from the issue text.

Assessment

Tech stack
huggingface, python
Domain
machine-learning, tooling
Issue type
Bug
Difficulty
3/5
Estimated time
1-2 days
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.