app.py

import os
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM
from huggingface_hub import login

# Directly assign your Hugging Face token here
hf_token = "your_hugging_face_api_token"

# Log in to Hugging Face
login(token=hf_token)

# Load the Mixtral-8x7B-Instruct model and tokenizer with authorization header
model_name = 'mistralai/Mistral-7B-Instruct-v0.3'
headers = {"Authorization": f"Bearer {hf_token}"}

# Ensure sentencepiece is installed
try:
    import sentencepiece
except ImportError:
    raise ImportError("The sentencepiece library is required for this tokenizer. Please install it with `pip install sentencepiece`.")

tokenizer = AutoTokenizer.from_pretrained(model_name, use_auth_token=hf_token)
model = AutoModelForCausalLM.from_pretrained(model_name, use_auth_token=hf_token)

# Check if a GPU is available and if not, fall back to CPU
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model.to(device)

# Example text input
text_input = "How did Tesla perform in Q1 2024?"

# Tokenize the input text
inputs = tokenizer(text_input, return_tensors="pt").to(device)

# Generate a response
outputs = model.generate(**inputs, max_length=150, temperature=0.7, top_p=0.9, top_k=50)

# Decode the generated tokens to a readable string
response = tokenizer.decode(outputs[0], skip_special_tokens=True)

# Print the response
print(response)