Evaluate model accuracy using Torchmetrics
Evaluating how well your model performs is essential - especially when preparing it for deployment! Let's smoothly integrate accuracy calculation using Torchmetrics right into the validation_step(). Don't forget to log the results, so you can easily monitor your model's progress.
This exercise is part of the course
Scalable AI Models with PyTorch Lightning
Exercise instructions
- Import
Accuracyfromtorchmetrics. - Instantiate the accuracy metric inside
__init__(). - Calculate accuracy within
validation_step()and log it as'val_acc'.
Hands-on interactive exercise
Have a go at this exercise by completing this sample code.
# Import relevant metric
from torchmetrics import ____
import lightning.pytorch as pl
class ClassifierModel(pl.LightningModule):
def __init__(self):
super().__init__()
# Instantiate accuracy metric
self.accuracy = ____()
def validation_step(self, batch, batch_idx):
x, y = batch
preds = self(x)
# Calculate accuracy and log it as val_acc
acc = self.____(preds, y)
self.log(____, acc)