# PyTesting

:::info
You can use this space to add any PyTest ideas, PyTests you've created or any problems you're having.
:::

## :bulb: Ideas

Ideas for PyTests that could improve ML/AI pipelines:
- [ ] Check that data augmentation techniques are producing the variations desired.
- [ ] Test the training process by verifying that the loss decreases over time.
- [ ] Test that the gradients are computed correctly, and weights are being updated.
- [ ] Validate the model's architecture by asserting that the model's layers output shapes are as expected.
- [ ] Verify that the evaluation metrics (e.g. accuracy, AUC, precision...) are being calculated correctly.
- [ ] Check the model predicts outcomes we expect from test data.
- [ ] Check the model is saving and loading correctly without loss in data.
- [ ] Check plots (e.g. confusion matrix, saliency maps).
- [ ] Check the memory usage during training and/or inference to ensure it stays within acceptable limits.
- [ ] Error handling - check that error handling is being handled appropriately, e.g. in case of wrong input data.

***

## :wrench: Shared Tests

:::success
Let's see how many different tests we can come up with! Share your Pytests here:
:::

***

## :cry: Problems

:::warning
What ails you? Put your problems here and lets see if we can work together to help you find an answer.

Copy the template table below and add your problems in the dedicated section beneath.
:::

**Template**
| **Problem**| <Fill here> |
|---------- |---------- |
|**Code (if applicable)**| <`Fill here`> |
|**Potential Solution(s)**| <Fill here> |

### Problems

| **Problem**| <Fill here> |
|---------- |---------- |
|**Code (if applicable)**| <`Fill here`> |
|**Potential Solution(s)**| <Fill here> |

| **Problem**| <Fill here> |
|---------- |---------- |
|**Code (if applicable)**| <`Fill here`> |
|**Potential Solution(s)**| <Fill here> |