Finished training at step 279303

eval_results.json

@@ -0,0 +1,4 @@
+{
+    "eval_loss": 2.79034161567688,
+    "eval_perplexity": 16.28658260350435
+}