# Evidence: A sharded training step

Copy this template into your own portfolio and fill it after running your modified exercise.

## Evidence required

Each local gradient already divides by its local count. Summing four equal local means omits division by four. For unequal counts, an unweighted mean also misrepresents examples; sum count-weighted local means and divide by total count.

- Command: `python3 phases/09-distributed/02-a-sharded-training-step/code/main.py`
- Working directory: repository root
- Python / JAX / NumPy versions: 
- Hardware and device: 
- My prediction before running: 
- My change to the exercise: 
- Exit code and meaningful output: 
- Why the result supports or contradicts my prediction: 
- Checkpoint result: 
- Artifact link or saved filename: 
- Portfolio status: saved in local workspace