server/api.py

from flask import Flask, request, jsonify
from transformers import pipeline
from transformers import GPTNeoXForCausalLM, GPTNeoXTokenizerFast

model = GPTNeoXForCausalLM.from_pretrained(
    "Mauricio-100/agent-ai",
    trust_remote_code=True,
    device_map="auto"
)

tokenizer = GPTNeoXTokenizerFast.from_pretrained("Mauricio-100/agent-ai")

app = Flask(__name__)
print("🧠 Chargement du modèle Gopu Agent pour l'API...")
bot = pipeline("text-generation", model="Mauricio-100/agent-ai")

@app.route("/")
def home():
    return "🤖 Gopu Agent API prête !"

@app.route("/ask", methods=["POST"])
def ask():
    data = request.get_json()
    prompt = data.get("prompt", "")
    if not prompt:
        return jsonify({"error": "prompt manquant"}), 400
    response = bot(prompt, max_new_tokens=200)[0]["generated_text"]
    return jsonify({"response": response})

if __name__ == "__main__":
    app.run(host="0.0.0.0", port=7860)