Update app.py

app.py

@@ -83,9 +83,7 @@ def search(query):
 client_gemma = InferenceClient("mistralai/Mistral-7B-Instruct-v0.3")
 client_mixtral = InferenceClient("NousResearch/Nous-Hermes-2-Mixtral-8x7B-DPO")
 client_llama = InferenceClient("meta-llama/Meta-Llama-3-8B-Instruct")
-
-
-func_caller = []
+client_yi = InferenceClient("01-ai/Yi-1.5-34B-Chat")
 
 # Define the main chat function
 def respond(message, history):
@@ -176,15 +174,15 @@ def respond(message, history):
                     buffer += new_text
                     yield buffer
             else:
-                messages = f"<|start_header_id|>system\nYou are OpenCHAT mini a helpful assistant made by KingNish. You answers users query like human friend. You are also Expert in every field and also learn and try to answer from contexts related to previous question. Try your best to give best response possible to user. You also try to show emotions using Emojis and reply like human, use short forms, friendly tone and emotions.<|end_header_id|>"
+                messages = f"<|im_start|>system\nYou are OpenCHAT mini a helpful assistant made by KingNish. You answers users query like human friend. You are also Expert in every field and also learn and try to answer from contexts related to previous question. Try your best to give best response possible to user. You also try to show emotions using Emojis and reply like human, use short forms, friendly tone and emotions.<|im_end|>"
                 for msg in history:
-                    messages += f"\n<|start_header_id|>user\n{str(msg[0])}<|end_header_id|>"
-                    messages += f"\n<|start_header_id|>assistant\n{str(msg[1])}<|end_header_id|>"
-                messages+=f"\n<|start_header_id|>user\n{message_text}<|end_header_id|>\n<|start_header_id|>assistant\n"
-                stream = client_llama.text_generation(messages, max_new_tokens=2000, do_sample=True, stream=True, details=True, return_full_text=False)
+                    messages += f"\n<|im_start|>user\n{str(msg[0])}<|im_end|>"
+                    messages += f"\n<|im_start|>assistant\n{str(msg[1])}<|im_end|>"
+                messages+=f"\n<|im_start|>user\n{message_text}<|im_end|>\n<|im_start|>assistant\n"
+                stream = client_yi.text_generation(messages, max_new_tokens=2000, do_sample=True, stream=True, details=True, return_full_text=False)
                 output = ""
                 for response in stream:
-                    if not response.token.text == "<|eot_id|>":
+                    if not response.token.text == "<|endoftext|>":
                         output += response.token.text
                         yield output
         except: