Update app.py

app.py

@@ -41,6 +41,11 @@ def text2text(input_text):
 
     # Does not seem to care if it goes over 512... humm...
     # To make it faster generate 100 tokens at a time
+    # sampling mode.. don't greedily take the highest probability token every time. Helps it chat with some variation
+    # temperature.. how random should the sampling be.
+    # top_p Which set of tokens to sample from. Filters out some low probability tokens before smapling.
+    # 
+    # input_ids should not be over 512 tokens. This method does not break over 512 tokens.. what is it doing?
     outputs = model.generate(input_ids, max_new_tokens=100, do_sample=True, temperature=0.7, top_p=0.8)
 
     # Remove <pad> and </s> eof sequence tokens