app.py

import gradio as gr
from huggingface_hub import HfApi
import spaces
import shutil
import logging
import subprocess
from pathlib import Path

@spaces.GPU
def write_repo(base_model, model_to_merge):
    with open("repo.txt", "w") as repo:
        repo.write(base_model + "\n" + model_to_merge)

def merge_and_upload(base_model, model_to_merge, scaling_factor, weight_drop_prob, repo_name, token):
    # Define a fixed output path
    outpath = Path('/tmp/output')
    if outpath.exists() and outpath.is_dir():
        shutil.rmtree(outpath)

    write_repo(base_model, model_to_merge)

    # Construct the command to run hf_merge.py
    command = [
        "python3", "hf_merge.py",
        "-p", str(weight_drop_prob),
        "-lambda", str(scaling_factor),
        "repo.txt", str(outpath)
    ]

    # Set up logging
    logging.basicConfig(level=logging.INFO)
    log_output = ""

    # Run the command and capture the output
    result = subprocess.run(command, capture_output=True, text=True)

    # Log the output
    log_output += result.stdout + "\n"
    log_output += result.stderr + "\n"
    logging.info(result.stdout)
    logging.error(result.stderr)

    # Check if the merge was successful
    if result.returncode != 0:
        return None, f"Error in merging models: {result.stderr}", log_output

    # Update progress bar
    yield 0.5, "Merging completed. Uploading to Hugging Face Hub..."

    # Upload the result to Hugging Face Hub
    api = HfApi(token=token)
    try:
        # Get the username of the user who is logged in
        user = api.whoami(token=token)["name"]

        # Autofill the repo name if none is provided
        if not repo_name:
            repo_name = f"{user}/default-repo"

        # Create a new repo or update an existing one
        api.create_repo(repo_id=repo_name, token=token, exist_ok=True)

        # Upload the file
        api.upload_folder(
            folder_path=str(outpath),
            repo_id=repo_name,
            repo_type="model",
            token=token
        )
        repo_url = f"https://huggingface.co/{repo_name}"
        yield 1.0, "Upload completed."
        return repo_url, "Model merged and uploaded successfully!", log_output
    except Exception as e:
        return None, f"Error uploading to Hugging Face Hub: {str(e)}", log_output

# Define the Gradio interface
with gr.Blocks(theme="Ytheme/Minecraft", fill_width=True, delete_cache=(60, 3600)) as demo:
    gr.Markdown("# Model Merger and Uploader")
    gr.Markdown("Combine any two models using a Super Mario merge(DARE) as described in the linked whitepaper.")
    gr.Markdown("Works with:")
    gr.Markdown("* Stable Diffusion (1.5, XL/XL Turbo)")
    gr.Markdown("* LLMs (Mistral, Llama, etc)")
    gr.Markdown("* LoRas (must be same size)")
    gr.Markdown("* Any two homologous models")

    with gr.Column():
        with gr.Row():
            token = gr.Textbox(label="Your HF write token", placeholder="hf_...", value="", max_lines=1)
        with gr.Row():
            base_model = gr.Textbox(label="Base Model", placeholder=".safetensors")
        with gr.Row():
            model_to_merge = gr.Textbox(label="Merge Model", placeholder=".bin/.safetensors")
        with gr.Row():
            repo_name = gr.Textbox(label="New Model", placeholder="SDXL-", info="If empty, auto-complete", value="", max_lines=1)
        with gr.Row():
            scaling_factor = gr.Slider(minimum=0, maximum=10, value=3.0, label="Scaling Factor")
        with gr.Row():
            weight_drop_prob = gr.Slider(minimum=0, maximum=1, value=0.3, label="Weight Drop Probability")

        repo_url = gr.Markdown(label="Repository URL")

        gr.Button("Merge").click(
            merge_and_upload,
            inputs=[base_model, model_to_merge, scaling_factor, weight_drop_prob, repo_name, token],
            outputs=[repo_url]
        )

demo.launch()