A.py

import os
import requests
from safetensors.torch import load_file, save_file
import torch
torch.cuda.empty_cache()
import torch.nn.functional as F
from tqdm import tqdm

def download_file(url, dest_path):
    print(f"Downloading {url} to {dest_path}")
    response = requests.get(url, stream=True)
    if response.status_code == 200:
        with open(dest_path, 'wb') as f:
            for chunk in response.iter_content(1024):
                f.write(chunk)
    else:
        raise Exception(f"Failed to download file from {url}")

def load_model(file_path):
    return load_file(file_path)

def save_model(merged_model, output_file):
    print(f"Saving merged model to {output_file}")
    save_file(merged_model, output_file)

def resize_tensor_shapes(tensor1, tensor2):
    if tensor1.size() == tensor2.size():
        return tensor1, tensor2

    max_shape = [max(s1, s2) for s1, s2 in zip(tensor1.shape, tensor2.shape)]
    tensor1_resized = F.pad(tensor1, (0, max_shape[-1] - tensor1.size(-1)))
    tensor2_resized = F.pad(tensor2, (0, max_shape[-1] - tensor2.size(-1)))

    return tensor1_resized, tensor2_resized

def merge_checkpoints(ckpt1, ckpt2, blend_ratio=0.5):
    print(f"Merging checkpoints with blend ratio: {blend_ratio}")
    merged = {}
    all_keys = set(ckpt1.keys()).union(set(ckpt2.keys()))

    for key in tqdm(all_keys, desc="Merging Checkpoints", unit="layer"):
        t1, t2 = ckpt1.get(key), ckpt2.get(key)
        if t1 is not None and t2 is not None:
            t1, t2 = resize_tensor_shapes(t1, t2)
            merged[key] = blend_ratio * t1 + (1 - blend_ratio) * t2
        elif t1 is not None:
            merged[key] = t1
        else:
            merged[key] = t2

    return merged

def cleanup_files(*file_paths):
    for file_path in file_paths:
        if os.path.exists(file_path):
            os.remove(file_path)
            print(f"Deleted {file_path}")

if __name__ == "__main__":
    try:
        model1_url = "https://huggingface.co/multimodalart/FLUX.1-dev2pro-full/resolve/main/flux1-dev.safetensors"
        model2_url = "https://huggingface.co/datasets/John6666/flux1-backup-202409/resolve/main/brainflux_v10.safetensors"
        
        model1_path = "flux1-dev.safetensors"
        model2_path = "brainflux_v10.safetensors"
        blend_ratio = 0.4
        output_file = "output_checkpoint.safetensors"

        # Downloading files
        download_file(model1_url, model1_path)
        download_file(model2_url, model2_path)

        # Loading models
        model1 = load_model(model1_path)
        model2 = load_model(model2_path)

        # Merging models
        merged_model = merge_checkpoints(model1, model2, blend_ratio)

        # Saving merged model
        save_model(merged_model, output_file)

        # Cleaning up downloaded files
        cleanup_files(model1_path, model2_path)
        
    except Exception as e:
        print(f"An error occurred: {e}")