Fix training difference test result handling

[mrhimanshu]

Summary

• isolate JAX and PyTorch result pickles per workload invocation
• propagate framework subprocess failures before reading result files
• document full and single-workload commands and runtime requirements
• add unit coverage for temporary-file cleanup and failure propagation

Fixes #808

Validation

• uvx ruff==0.12.0 check .
• uvx ruff==0.12.0 format --check .
• focused Python 3.11 unit tests: 3 passed

The full training-difference suite was not run locally because it requires eight visible GPUs, both framework stacks, and workload data.

Tooling disclosure: agent assistance was used during implementation and validation.

[github-actions]

MLCommons CLA bot:
Thank you very much for your submission; we really appreciate it. Before we can accept your contribution,
we ask that you sign the MLCommons CLA (Apache 2). Please submit your GitHub ID to our onboarding form to initiate
authorization. If you are from a MLCommons member organization, we will request that you be added to the CLA.
If you are not from a member organization, we will email you a CLA to sign. For any questions, please contact
support@mlcommons.org.
0 out of 1 committers have signed the MLCommons CLA.
❌ @mrhimanshu
_{You can retrigger this bot by commenting recheck in this Pull Request}

[mrhimanshu]

recheck