Change use_cache to true

This speeds up how quickly text is outputted by the model

config.json

@@ -18,6 +18,6 @@
   "tie_word_embeddings": false,
   "torch_dtype": "float32",
   "transformers_version": "4.28.1",
-  "use_cache": false,
+  "use_cache": true,
   "vocab_size": 32001
 }