[DRAFT] Fix issue #220

[simone99n]

Description

Draft implementation for the issue #220 (Initialization of parallelization on LSF Scheduler)
The main idea is to have the Python code read a set of possible environment variables (e.g., from a JSON configuration) that define the rank and size of each process. Since each HPC job scheduler exposes different environment variables (SLURM exposes SLURM_PROCID/SLURM_NTASKS, but LSF that use mpirun exposes PMI_RANK and PMI_SIZE), this approach avoids hardcoding scheduler-specific names. Instead, the model will iterate through a list of known variable names.

Since the issue #437 (DeepSpeed Support) has been opened, its resolution will overlap with the current issue. . In particular, init_torch() and init_ddp() are not needed, as they are managed under the hood by DeepSpeed.

Type of Change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Documentation update

Issue Number

Issue #220

Code Compatibility

• I have performed a self-review of my code

Code Performance and Testing

• I ran the uv run train and (if necessary) uv run evaluate on a least one GPU node and it works
• If the new feature introduces modifications at the config level, I have made sure to have notified the other software developers through Mattermost and updated the paths in the $WEATHER_GENERATOR_PRIVATE directory

Dependencies

• I have ensured that the code is still pip-installable after the changes and runs
• I have tested that new dependencies themselves are pip-installable.
• I have not introduced new dependencies in the inference portion of the pipeline

Documentation

• My code follows the style guidelines of this project
• I have updated the documentation and docstrings to reflect the changes
• I have added comments to my code, particularly in hard-to-understand areas

Additional Notes