Fix check_model: LLM_MODELS is a list of str, not Enum (#317)

[idevasena]

Problem

Running mlpstorage checkpointing datasize with any valid model (e.g. --model llama3-8b) fails with:

Error running check check_model: 'str' object has no attribute 'value'

Root Cause

In mlpstorage_py/rules/run_checkers/checkpointing.py, line 34 treats LLM_MODELS entries as Enum members:

valid_models = [m.value for m in self.supported_models]

However, LLM_MODELS (defined in config.py) is a plain list of strings, not an Enum. Calling .value on a str raises an AttributeError.

Fix

Replace the list comprehension with a direct reference since the values are already strings:

valid_models = list(self.supported_models)

Testing

• Verified mlpstorage checkpointing datasize --model llama3-8b ... no longer crashes.

:~/Storage_Repo_Tests/storage_prfix317$ ./mlpstorage checkpointing datasize --client-host-memory-in-gb 256 --model "llama3-8b" --num-processes 8 --checkpoint-folder ~/checkpoint_temp --file --allow-run-as-root
Hosts is: ['127.0.0.1']
Hosts is: ['127.0.0.1']
⠋ Validating environment... 0:00:002026-04-07 21:51:06|INFO: Environment validation passed
2026-04-07 21:51:06|STATUS: Benchmark results directory: /tmp/mlperf_storage_results/checkpointing/llama3-8b/20260407_215106
2026-04-07 21:51:06|INFO: Created benchmark run: checkpointing_datasize_llama3-8b_20260407_215106
2026-04-07 21:51:06|STATUS: Verifying benchmark run for checkpointing_datasize_llama3-8b_20260407_215106
2026-04-07 21:51:06|STATUS: Benchmark run qualifies for CLOSED category ([RunID(program='checkpointing', command='datasize', model='llama3-8b', run_datetime='20260407_215106')])
2026-04-07 21:51:06|WARNING: Running the benchmark without verification for open or closed configurations. These results are not valid for submission. Use --open or --closed to specify a configuration.
2026-04-07 21:51:06|STATUS: Instantiated the Checkpointing Benchmark...
⠦ Collecting cluster info... ━━━━━━━━━━━━━━━━━━━━╺━━━━━━━━━━━━━━━━━━━ 2/4 0:00:002026-04-07 21:51:07|RESULT: Total GiB required per rank:
	Rank 0: 13.12GiB
	Rank 1: 13.12GiB
	Rank 2: 13.12GiB
	Rank 3: 13.12GiB
	Rank 4: 13.12GiB
	Rank 5: 13.12GiB
	Rank 6: 13.12GiB
	Rank 7: 13.12GiB
2026-04-07 21:51:07|RESULT: Total GiB required for all ranks: 105.00GiB
2026-04-07 21:51:08|STATUS: Writing metadata for benchmark to: /tmp/mlperf_storage_results/checkpointing/llama3-8b/20260407_215106/checkpointing_20260407_215106_metadata.json

• Verified that passing an invalid model name still produces the expected validation error.

:~/Storage_Repo_Tests/storage_prfix317$ ./mlpstorage checkpointing datasize \
     --client-host-memory-in-gb 256 \
     --model "bogus-model" \
     --num-processes 8 \
     --checkpoint-folder ~/checkpoint_temp \
     --file --allow-run-as-root
Invalid LLM model. Supported models are: llama3-70b, llama3-405b, llama3-1t, llama3-8b

[github-actions]

MLCommons CLA bot All contributors have signed the MLCommons CLA ✍️ ✅

[russfellows]

Looks like this has been tested, great.