Update app.py

app.py

@@ -1,43 +1,63 @@
-from flask import Flask, request, render_template
-from transformers import pipeline, AutoTokenizer
-import torch
-
-app = Flask(__name__)
-
-# Load a lightweight model (e.g., Zephyr-7B, Mistral-7B)
-model_name = "mistralai/Mistral-7B-Instruct-v0.2"
-tokenizer = AutoTokenizer.from_pretrained(model_name)
-chatbot = pipeline(
-    "text-generation",
-    model=model_name,
-    tokenizer=tokenizer,
-    torch_dtype=torch.float16,
-    device_map="auto"  # Uses GPU if available
+import streamlit as st
+import speech_recognition as sr
+from langchain.llms import HuggingFaceHub
+from langchain.prompts import PromptTemplate
+from langchain.chains import LLMChain
+
+# Initialize Hugging Face model (you can use other models based on your needs)
+llm = HuggingFaceHub(
+    repo_id="gpt2",  # You can replace with other models on Hugging Face
+    model_kwargs={"temperature": 0.7}
 )
 
-@app.route("/", methods=["GET", "POST"])
-def home():
-    if request.method == "POST":
-        user_input = request.form["user_input"]
-        response = generate_response(user_input)
-        return render_template("index.html", user_input=user_input, bot_response=response)
-    return render_template("index.html")
-
-def generate_response(prompt):
-    # Format prompt for instruction-following models
-    messages = [{"role": "user", "content": prompt}]
-    prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
-    
-    # Generate response
-    outputs = chatbot(
-        prompt,
-        max_new_tokens=256,
-        do_sample=True,
-        temperature=0.7,
-        top_k=50,
-        top_p=0.95
-    )
-    return outputs[0]["generated_text"][len(prompt):]  # Extract only the bot's reply
-
-if __name__ == "__main__":
-    app.run(host="0.0.0.0", port=5000, debug=True)
+# Define the prompt template for translation or text generation
+prompt_template = "Translate the following text to Urdu: {text}"
+prompt = PromptTemplate(input_variables=["text"], template=prompt_template)
+
+# Set up the chain
+llm_chain = LLMChain(prompt=prompt, llm=llm)
+
+# Function to process text using LangChain
+def process_text(input_text):
+    result = llm_chain.run(input_text)
+    return result
+
+# Function to recognize speech input
+def recognize_speech():
+    recognizer = sr.Recognizer()
+    with sr.Microphone() as source:
+        st.write("Please speak now...")
+        audio = recognizer.listen(source)
+    try:
+        text = recognizer.recognize_google(audio, language="en-US")
+        st.write(f"Recognized Text: {text}")
+        return text
+    except sr.UnknownValueError:
+        st.error("Sorry, I could not understand the audio.")
+        return ""
+    except sr.RequestError:
+        st.error("Could not request results from Google Speech Recognition service.")
+        return ""
+
+# Streamlit UI
+st.title("Voice and Text Input for English and Urdu")
+st.write("You can either type your text or speak to input in English or Urdu!")
+
+# Choose language input type
+input_type = st.selectbox("Select Input Type", ["Text", "Voice"])
+
+# Process based on input type
+if input_type == "Text":
+    text_input = st.text_area("Enter text:")
+    if st.button("Submit"):
+        if text_input:
+            result = process_text(text_input)
+            st.write(f"Processed Result: {result}")
+        else:
+            st.error("Please enter some text.")
+elif input_type == "Voice":
+    if st.button("Start Recording"):
+        speech_text = recognize_speech()
+        if speech_text:
+            result = process_text(speech_text)
+            st.write(f"Processed Result: {result}")