app.py

import gradio as gr
import os
from langchain_openai import ChatOpenAI

api_key = os.environ.get("FEATHERLESS_API_KEY")

MODEL_CHOICES = [
    "Qwen/Qwen3-32B",
    "deepseek-ai/DeepSeek-R1-Distill-Llama-70B",
    "meta-llama/Llama-3.3-70B-Instruct",
    "mistralai/Magistral-Small-2506",
    "unsloth/DeepSeek-R1-Distill-Llama-70B",
    "unsloth/Qwen2.5-72B-Instruct",
    "unsloth/Llama-3.3-70B-Instruct",
]

def create_llm(model_name: str):
    return ChatOpenAI(
        base_url="https://api.featherless.ai/v1/",
        api_key=api_key,
        streaming=True,
        model=model_name,
    )

def respond(message, history, system_message, max_tokens, temperature, top_p, model_name):
    llm = create_llm(model_name)
    messages = [{"role":"system","content":system_message}]
    for u, a in history:
        if u: messages.append({"role":"user","content":u})
        if a: messages.append({"role":"assistant","content":a})
    messages.append({"role":"user","content":message})
    llm.max_tokens = max_tokens
    llm.temperature = temperature
    llm.top_p = top_p
    response = ""
    for chunk in llm.stream(messages):
        response += chunk.content
        yield response

with gr.Blocks() as demo:
    with gr.Row():
        model_dropdown = gr.Dropdown(choices=MODEL_CHOICES, value=MODEL_CHOICES[0], label="Pilih model")
    chatbot = gr.Chatbot()
    system_msg = gr.Textbox("You are a friendly Chatbot.", label="System message")
    max_t = gr.Slider(1, 16384, value=2048, step=1, label="Max new tokens")
    temp = gr.Slider(0.1, 2.0, value=0.7, step=0.1, label="Temperature")
    top_p = gr.Slider(0.1, 1.0, value=0.95, step=0.05, label="Top‑p")

    model_dropdown.change(lambda x: None, inputs=model_dropdown, outputs=[])
    demo.launch()