ai-training

A minimal fake-training project with:

• Hydra config-driven argument passing
• loguru-based training logs
• Weights & Biases metric logging
• model checkpoint saving
• Optuna-based hyperparameter tuning

Run

python train.py

Override config values via Hydra-style CLI args, for example:

python train.py training.steps=50 training.learning_rate=0.05

Test

python -m unittest discover -s tests -q