Fix math_verify_func for multithreaded environment

[SurbhiJainUSC]

Description

The existing math_verify_func() would fail during training because the underlying Math-Verify 'parse' function uses signal.alarm(), which is incompatible with threaded environments. This results in zero reward score even when the gold and predictions are same.

Error:

Math-Verify 'parse' function doesn't support threaded environment due to usage of signal.alarm() in timeout mechanism. If you need to run in multithreaded environment it's recommended to set the parsing_timeout=None, which will run without timeout (and signal handling). In this case you need to handle the timeouting yourself.

This PR implements a manual math verification function with a timeout (default=5, same as parse() in math_verify), ensuring it can run safely in a multithreaded context.

Tests

python3 -m maxtext.trainers.post_train.rl.train_rl \
model_name=llama3.1-8b-Instruct \
load_parameters_path=$MAXTEXT_CKPT_PATH \
run_name=$RUN_NAME \
base_output_directory=$BASE_OUTPUT_DIRECTORY \
dataset_name=huggingface:nvidia/OpenMathInstruct-2 train_split=train_1M steps=10 \
rollout_tensor_parallelism=1 \
hf_train_files=hf://datasets/nvidia/OpenMathInstruct-2/data/train_1M-*.parquet \
skip_jax_distributed_system=True

Checklist

Before submitting this PR, please make sure (put X in square brackets):

• I have performed a self-review of my code. For an optional AI review, add the gemini-review label.
• I have necessary comments in my code, particularly in hard-to-understand areas.
• I have run end-to-end tests tests and provided workload links above if applicable.
• I have made or will make corresponding changes to the doc if needed, including adding new documentation pages to the relevant Table of Contents (toctree directive) as explained in our documentation.

[codecov]

Codecov Report

❌ Patch coverage is 66.66667% with 9 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/maxtext/trainers/post_train/rl/utils_rl.py	65.38%	9 Missing ⚠️

📢 Thoughts on this report? Let us know!