app.py

import streamlit as st
from transformers import AutoTokenizer, AutoModelForCausalLM

# Load model and tokenizer
tokenizer = AutoTokenizer.from_pretrained("namannn/llama2-13b-hyperbolic-cluster-pruned")
model = AutoModelForCausalLM.from_pretrained("namannn/llama2-13b-hyperbolic-cluster-pruned")

# Streamlit UI components
st.title("Text Generation with LLaMa2-13b Hyperbolic Model")
st.write("Enter a prompt below and the model will generate text.")

# User input for prompt
prompt = st.text_area("Input Prompt", "Once upon a time, in a land far away")

# Slider for controlling the length of the output
max_length = st.slider("Max Length of Generated Text", min_value=50, max_value=200, value=100)

# Button to trigger text generation
if st.button("Generate Text"):
    if prompt:
        # Encode the prompt text
        inputs = tokenizer(prompt, return_tensors="pt")

        # Generate text with the model
        outputs = model.generate(
            inputs["input_ids"],
            max_length=max_length,
            num_return_sequences=1,
            no_repeat_ngram_size=2,  # You can tune this for diversity
            do_sample=True,  # Use sampling for diverse generation
            top_k=50,  # Top-k sampling for diversity
            top_p=0.95,  # Top-p (nucleus) sampling
            temperature=0.7  # Control randomness (lower = more deterministic)
        )

        # Decode and display generated text
        generated_text = tokenizer.decode(outputs[0], skip_special_tokens=True)
        st.subheader("Generated Text:")
        st.write(generated_text)
    else:
        st.warning("Please enter a prompt to generate text.")