Hey,

I’m performing a binary classification on a 3d MRI scan dataset while using nn.CrossEntropy as my loss function.

In my training loop i calculate the loss as

```
loss = loss_fn(pred, y)
```

which is pretty standard.

What i’m trying to learn is what can be inferred from the loss value. i.e

- What do these loss values represent?
- What classifies as a good value?
- What do these values tell me about my model, data…?

The results i’m currently getting are

```
loss: 530.178589 [ 0/ 474]
loss: 231.394638 [ 70/ 474]
loss: 61.357143 [ 140/ 474]
loss: 482.464294 [ 210/ 474]
loss: 452.214294 [ 280/ 474]
loss: 173.785721 [ 350/ 474]
loss: 117.030411 [ 420/ 474]
```

Which intuitively does not feel satisfactory.

Would really appreciate if someone can help me understand what is a good score (if any).

Thanks