app.py

import gradio as gr
from transformers import AutoTokenizer, AutoModelForCausalLM

# تحميل النموذج والـ Tokenizer
model_path = "inceptionai/jais-13b"
tokenizer = AutoTokenizer.from_pretrained(model_path)
model = AutoModelForCausalLM.from_pretrained(model_path, device_map="auto", trust_remote_code=True)

# دالة للحصول على الإجابة
def get_response(text):
    input_ids = tokenizer(text, return_tensors="pt").input_ids
    inputs = input_ids.to("cuda" if torch.cuda.is_available() else "cpu")
    input_len = inputs.shape[-1]
    generate_ids = model.generate(
        inputs,
        top_p=0.9,
        temperature=0.3,
        max_length=200 - input_len,
        min_length=input_len + 4,
        repetition_penalty=1.2,
        do_sample=True,
    )
    response = tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=True)[0]
    return response

# واجهة Gradio
iface = gr.Interface(
    fn=get_response,
    inputs="text",
    outputs="text",
    title="Jais-13b Demo",
    description="تجربة نموذج Jais-13b للغة العربية."
)

iface.launch()