Add top_p

modules/text_generation.py

@@ -86,11 +86,11 @@ def generate_reply(question, max_new_tokens, do_sample, temperature, top_p, typi
 
     if shared.is_RWKV:
         if shared.args.no_stream:
-            reply = shared.model.generate(question, token_count=max_new_tokens, temperature=temperature)
+            reply = shared.model.generate(question, token_count=max_new_tokens, temperature=temperature, top_p=top_p)
             yield formatted_outputs(reply, None)
         else:
             for i in range(max_new_tokens//8):
-                reply = shared.model.generate(question, token_count=8, temperature=temperature)
+                reply = shared.model.generate(question, token_count=8, temperature=temperature, top_p=top_p)
                 yield formatted_outputs(reply, None)
                 question = reply
         return formatted_outputs(reply, None)