push_to_hf.py

import os
from pathlib import Path
from huggingface_hub import create_repo, upload_folder
from PIL import Image

def save_model_card(repo_id: str, images=None, base_model=str, dataset_name=str, repo_folder=None):
    img_str = ""
    for i, image in enumerate(images):
        image.save(os.path.join(repo_folder, f"image_{i}.png"))
        img_str += f"![img_{i}](./image_{i}.png)\n"

    yaml = f"""
---
license: creativeml-openrail-m
base_model: {base_model}
tags:
- stable-diffusion
- stable-diffusion-diffusers
- text-to-image
- diffusers
- lora
inference: true
---
    """
    model_card = f"""
# LoRA text2image fine-tuning - {repo_id}
These are LoRA adaption weights for {base_model}. The weights were fine-tuned on the {dataset_name} dataset. You can find some example images in the following. \n
{img_str}
"""
    with open(os.path.join(repo_folder, "README.md"), "w") as f:
        f.write(yaml + model_card)


if __name__ == "__main__":

    REPOS = {
        "tom_cruise_plain": {"hub_model_id": "person-thumbs-up-plain-lora", "output_dir": "/l/vision/v5/sragas/easel_ai/models_plain/"},
        "tom_cruise": {"hub_model_id": "person-thumbs-up-lora", "output_dir": "/l/vision/v5/sragas/easel_ai/models/"},
        "tom_cruise_no_cap": {"hub_model_id": "person-thumbs-up-lora-no-cap", "output_dir": "/l/vision/v5/sragas/easel_ai/models_no_cap/"}
    }

    current_repo_id = "tom_cruise"
    current_repo = REPOS[current_repo_id]

    print(f"{'-'*20} CURRENT REPO: {current_repo_id} {'-'*20}")

    hub_model_id = current_repo["hub_model_id"]
    output_dir = current_repo["output_dir"]
    pretrained_model_name_or_path = "runwayml/stable-diffusion-v1-5"

    images = [Image.open(output_dir+file) for file in os.listdir(output_dir) if ".png" in file]
    
    repo_id = create_repo(
        repo_id=hub_model_id or Path(output_dir).name, exist_ok=True, token=None
    ).repo_id
    
    save_model_card(
        repo_id,
        images=images,
        base_model=pretrained_model_name_or_path,
        dataset_name="Custom dataset",
        repo_folder=output_dir,
    )

    upload_folder(
        repo_id=repo_id,
        folder_path=output_dir,
        commit_message="End of training",
        ignore_patterns=["step_*", "epoch_*"],
    )