app.py

import os
import gradio as gr
from huggingface_hub import hf_hub_download
from llama_cpp import Llama


# --- Model location ---
REPO_ID = os.getenv("GGUF_REPO_ID", "gladestudio/gladecore")
FILENAME = os.getenv("GGUF_FILENAME", "emotion-run2-best.gguf")
HF_TOKEN = os.getenv("HF_TOKEN")  # not needed for public models

MODEL_PATH = hf_hub_download(
    repo_id=REPO_ID,
    filename=FILENAME,
    repo_type="model",
    token=HF_TOKEN,
)

# --- Llama init ---
N_CTX = int(os.getenv("N_CTX", "4096"))
N_THREADS = int(os.getenv("N_THREADS", str(os.cpu_count() or 4)))
N_GPU_LAYERS = int(os.getenv("N_GPU_LAYERS", "0"))  # >0 only on GPU Space

llm = Llama(
    model_path=MODEL_PATH,
    n_ctx=N_CTX,
    n_threads=N_THREADS,
    n_gpu_layers=N_GPU_LAYERS,
    verbose=False,
    # If GGUF needs a template similar to Llama 2, uncomment:
    # chat_format="llama-2",
)

def respond(message, history: list[dict[str, str]], system_message, max_tokens, temperature, top_p):
    messages = [{"role": "system", "content": system_message}]
    if history:
        messages.extend(history)
    messages.append({"role": "user", "content": message})

    stream = llm.create_chat_completion(
        messages=messages,
        max_tokens=max_tokens,
        temperature=temperature,
        top_p=top_p,
        stream=True,
    )

    partial = ""
    for chunk in stream:
        token = (
            chunk.get("choices", [{}])[0].get("delta", {}).get("content")
            or chunk.get("choices", [{}])[0].get("text", "")
            or ""
        )
        if token:
            partial += token
            yield partial

chatbot = gr.ChatInterface(
    respond,
    type="messages",
    additional_inputs=[
        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
        gr.Slider(minimum=1, maximum=4096, value=512, step=1, label="Max new tokens"),
        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
        gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p (nucleus sampling)"),
    ],
)

with gr.Blocks() as demo:
    chatbot.render()

if __name__ == "__main__":
    demo.launch()