Training + Experimentation

[YounesBensafia]

Files to create and modify

Create:

• docs/training_loops.md: Summarize custom and framework-based training loops
• docs/distributed_training.md: Document distributed training strategies
• docs/hyperparameter_search.md: Summarize search methods and tuning best practices
• docs/evaluation.md: Document evaluation strategies and metrics
• docs/experiment_tracking.md: Summarize MLFlow/W&B usage for versioning and tracking experiments
• docs/model_compression.md: Summarize quantization, distillation, and other compression techniques

Acceptance Criteria

• Training loop strategies are clearly documented
• Distributed training techniques and considerations are summarized
• Hyperparameter search methods and pros/cons are documented
• Evaluation strategies and metrics are defined
• Experiment tracking and versioning practices are outlined
• Model compression approaches are documented and explained