INNER CODE UNIT · Python

evaluate

barissayil/SentimentAnalysis · analyzer.py:58

    def evaluate(self, val_loader, criterion):
        # Set model to evaluation mode.
        self.model.eval()
        # Initialize batch accuracy summation, loss, and number of batches.
        batch_accuracy_summation, loss, num_batches = 0, 0, 0
        # Don't track gradient.
        with torch.no_grad():
            # Go through validation set in batches.
            for input_ids, attention_mask, labels in tqdm(
                val_loader, desc="Evaluating"
            ):
                # Put input IDs, attention mask, and labels to device.
                input_ids, attention_mask, labels = (
                    input_ids.to(self.device),
                    attention_mask.to(self.device),
                    labels.to(self.device),
                )
                # Get logits.

View source record →

📰 Research Paper
Loading…
⏳ Fetching content…