fixes fasttraining

Signed-off-by: Wenqi Li <[email protected]>

Author: wyli

acceleration/fast_training_tutorial.ipynb

@@ -614,7 +614,7 @@
     "                with torch.no_grad():\n",
     "                    val_loader_iterator = iter(val_loader)\n",
     "\n",
-    "                    for val_step in range(len(val_loader)):\n",
+    "                    for _ in range(len(val_loader)):\n",
     "                        # profiling: val dataload\n",
     "                        with nvtx.annotate(\"dataload\", color=\"red\") if profiling else no_profiling:\n",
     "                            val_data = next(val_loader_iterator)\n",