added some comments to app.py

app.py

@@ -25,7 +25,8 @@ filename = "ggml-vicuna-7b-1.1-q4_1.bin"
 
 download_file(ggml_model_path, filename)
 
-
+# n_ctx is max length of prompt
+# n_batch is no. of tokens fed to model at a time
 llm = Llama(model_path=filename, n_ctx=512, n_batch=126)