| name | model-verification-unit-tests |
| description | Verifies model components like loss functions using unit tests and saves the results as NumPy archives. |
Model Verification with Unit Tests
Verifying individual components of a training pipeline, such as a loss function, ensures mathematical correctness before full-scale training.
Running Tests
Use the unittest framework or pytest. For a script like unit_test_1.py:
python unit_test/unit_test_1.py
Data Persistence
To allow external verification of results, save computed tensors to a .npz file:
import numpy as np
np.savez(
"/path/to/loss.npz",
losses=losses.detach().cpu().numpy(),
)
Fixed Tensor Inputs
When verifying a loss function, use fixed tensors to ensure deterministic output:
policy_chosen_logps = torch.load("path/to/tensor.pt")