Update app.py

app.py

@@ -2,6 +2,13 @@ from transformers import AutoTokenizer, AutoModelForSequenceClassification, pipe
 import torch
 import pickle
 import streamlit as st
+from huggingface_hub import InferenceClient
+
+client = InferenceClient(
+    "mistralai/Mistral-7B-Instruct-v0.1"
+)
+
+
 device = torch.device("cuda") if torch.cuda.is_available() else torch.device("cpu")
 
 from translate import Translator
@@ -10,6 +17,31 @@ def init_session_state():
     if 'history' not in st.session_state:
         st.session_state.history = ""
 
+
+temperature=0.9
+max_new_tokens=256
+top_p=0.95
+repetition_penalty=1.0
+
+
+generate_kwargs = dict(
+        temperature=temperature,
+        max_new_tokens=max_new_tokens,
+        top_p=top_p,
+        repetition_penalty=repetition_penalty,
+        do_sample=True,
+        seed=42,
+    )
+
+def format_prompt(message, history):
+  prompt = "<s>"
+  for user_prompt, bot_response in history:
+    prompt += f"[INST] {user_prompt} [/INST]"
+    prompt += f" {bot_response}</s> "
+  prompt += f"[INST] {message} [/INST]"
+  return prompt
+
+
 # Initialize session state
 init_session_state()
 
@@ -79,14 +111,26 @@ if text:
 
     
 
-    st.session_state.history += "Based on this info only:" + answer +" ,answer this question, by reasoning step by step:" + text  # Add new text to history
-    out = pipe(st.session_state.history, max_new_tokens=256)  # Generate output based on history
+    # st.session_state.history += "Based on this info only:" + answer +" ,answer this question, by reasoning step by step:" + text  # Add new text to history
+    # out = pipe(st.session_state.history, max_new_tokens=256)  # Generate output based on history
+
+    history = st.session_state.history
+    prompt =  "Based on this info only:" + answer +" ,answer this question, by reasoning step by step:" + text
+    formatted_prompt = format_prompt(prompt, history)
+
+    stream = client.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
+    output = ""
+
+    for response in stream:
+        output += response.token.text
+        yield output
+    return output
 
 
     # st.text(st.session_state.history)
     
-    translated_text2 = translator2.translate(out[0]['generated_text'])
-    
+    # translated_text2 = translator2.translate(out[0]['generated_text'])
+    translated_text2 = translator2.translate(output)
 
     st.text(translated_text2)