app.py

import gradio as gr
import spaces 
import time 
import os 
import transformers 
from transformers import pipeline
import torch 

key = (os.getenv('API_KEY'))
model_id = "meta-llama/Meta-Llama-3-8B-Instruct"



pipeline = transformers.pipeline(
    "text-generation",
    model=model_id,
    model_kwargs={"torch_dtype": torch.bfloat16},
    device_map="auto",
    token = key
)


messages = [
    {"role": "system", "content": "You are a pirate chatbot who always responds in pirate speak!"},
    {"role": "user", "content": "Who are you?"},
]

terminators = [
    pipeline.tokenizer.eos_token_id,
    pipeline.tokenizer.convert_tokens_to_ids("<|eot_id|>")
]


outputs = pipeline(
    messages,
    max_new_tokens=256,
    eos_token_id=terminators,
    do_sample=True,
    temperature=0.6,
    top_p=0.9,
)

@spaces.GPU(duration=240)


# Fonction de génération de texte
def generate_text(prompt):
    
    inputs = tokenizer(prompt, return_tensors="pt")
    
    response_ids = model.generate(inputs.input_ids)
    
    response_text = tokenizer.decode(response_ids[0], skip_special_tokens=True)

    return response_text 
# Définir une fonction pour l'interface de chat

def chatbot(message, history):
    return generate_text(message)
   
gr.ChatInterface(chatbot).launch()