Update app.py

app.py

@@ -1,10 +1,9 @@
 import gradio as gr
-from huggingface_hub import InferenceClient
+from bn_llm_wrapper import generate
 
 """
 For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
 """
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
 
 
 def respond(
@@ -15,29 +14,19 @@ def respond(
     temperature,
     top_p,
 ):
-    messages = [{"role": "system", "content": system_message}]
+    context = ""
+    for chat in history:
+        context += f"রোগী: {chat[0]}\nথেরাপিস্ট: {chat[1]}\n"
 
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
+    answer = generate(message.text, context, 
+                      temperature=temperature,
+                      top_p=top_p,
+                      max_new_tokens=max_tokens).split('### ANSWER:')[1]
 
-    messages.append({"role": "user", "content": message})
+    if '</s>' in answer:
+        answer = answer.split('</s>')[0].strip()
 
-    response = ""
-
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
-
-        response += token
-        yield response
+    return answer
 
 """
 For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
@@ -46,12 +35,12 @@ demo = gr.ChatInterface(
     respond,
     additional_inputs=[
         gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
+        gr.Slider(minimum=1, maximum=2048, value=256, step=1, label="Max new tokens"),
         gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
         gr.Slider(
             minimum=0.1,
             maximum=1.0,
-            value=0.95,
+            value=0.7,
             step=0.05,
             label="Top-p (nucleus sampling)",
         ),