Update no trainer scripts for language modeling and image classification examples (#18443)

* Update no_trainer script for image-classification

* Update no_trainer scripts for language-modeling examples

* Remove unused variable

* Removing truncation from losses array for language modeling examples

examples/pytorch/image-classification/run_image_classification_no_trainer.py

@@ -508,19 +508,11 @@ def main():
                 break
 
         model.eval()
-        samples_seen = 0
         for step, batch in enumerate(eval_dataloader):
             with torch.no_grad():
                 outputs = model(**batch)
             predictions = outputs.logits.argmax(dim=-1)
-            predictions, references = accelerator.gather((predictions, batch["labels"]))
-            # If we are in a multiprocess environment, the last batch has duplicates
-            if accelerator.num_processes > 1:
-                if step == len(eval_dataloader) - 1:
-                    predictions = predictions[: len(eval_dataloader.dataset) - samples_seen]
-                    references = references[: len(eval_dataloader.dataset) - samples_seen]
-                else:
-                    samples_seen += references.shape[0]
+            predictions, references = accelerator.gather_for_metrics((predictions, batch["labels"]))
             metric.add_batch(
                 predictions=predictions,
                 references=references,

examples/pytorch/language-modeling/run_clm_no_trainer.py

@@ -597,10 +597,9 @@ def main():
                 outputs = model(**batch)
 
             loss = outputs.loss
-            losses.append(accelerator.gather(loss.repeat(args.per_device_eval_batch_size)))
+            losses.append(accelerator.gather_for_metrics(loss.repeat(args.per_device_eval_batch_size)))
 
         losses = torch.cat(losses)
-        losses = losses[: len(eval_dataset)]
         try:
             eval_loss = torch.mean(losses)
             perplexity = math.exp(eval_loss)

examples/pytorch/language-modeling/run_mlm_no_trainer.py

@@ -642,10 +642,9 @@ def main():
                 outputs = model(**batch)
 
             loss = outputs.loss
-            losses.append(accelerator.gather(loss.repeat(args.per_device_eval_batch_size)))
+            losses.append(accelerator.gather_for_metrics(loss.repeat(args.per_device_eval_batch_size)))
 
         losses = torch.cat(losses)
-        losses = losses[: len(eval_dataset)]
         try:
             eval_loss = torch.mean(losses)
             perplexity = math.exp(eval_loss)