Update README.md

README.md

@@ -91,13 +91,13 @@ llm = LLM(
     gpu_memory_utilization=0.7
 )
 
-instance_end_token_id = llm.get_tokenizer().convert_token_to_ids('<|im_end|>')
+turn_end_token_id = 61876 # <|im_end|>
 params = SamplingParams(
-    temperature=0.01,
-    top_p=0.01,
+    temperature=0.001,
+    top_p=0.001,
     max_tokens=4096,
     repetition_penalty=1.1,
-    stop_token_ids=[instance_end_token_id]
+    stop_token_ids=[turn_end_token_id]
 )
 
 def _inference(prompt, llm, params):