app.py

import gradio as gr
import spaces
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

MODEL_ID = "kadalicious22/snapgate-3B"

tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
model = AutoModelForCausalLM.from_pretrained(
    MODEL_ID,
    torch_dtype=torch.float16,
    device_map="auto"
)

@spaces.GPU
def chat(message, history, system_prompt):
    messages = [{"role": "system", "content": system_prompt}]
    for user, assistant in history:
        messages.append({"role": "user", "content": user})
        messages.append({"role": "assistant", "content": assistant})
    messages.append({"role": "user", "content": message})
    text = tokenizer.apply_chat_template(
        messages,
        tokenize=False,
        add_generation_prompt=True
    )
    inputs = tokenizer(text, return_tensors="pt").to(model.device)
    with torch.no_grad():
        outputs = model.generate(
            **inputs,
            max_new_tokens=512,
            temperature=0.7,
            do_sample=True,
            pad_token_id=tokenizer.eos_token_id
        )
    response = tokenizer.decode(
        outputs[0][inputs["input_ids"].shape[1]:],
        skip_special_tokens=True
    )
    return response

demo = gr.ChatInterface(
    fn=chat,
    additional_inputs=[
        gr.Textbox(
            value="Kamu adalah agen customer service Snapgate. Jawab dengan ramah dan solutif.",
            label="System Prompt",
            lines=3
        )
    ],
    title="🤖 snapgate-3B Demo",
    description="Model customer service, summarization, dan task execution.",
    examples=[
        ["Halo, siapa kamu?"],
        ["Produk saya belum sampai, tolong bantu"],
        ["Rangkum teks berikut: Snapgate adalah platform layanan pelanggan berbasis AI"]
    ]
)

demo.launch()