latihsekarang.py

import os
import uuid
import shutil
import json
import yaml
import torch
from PIL import Image
from slugify import slugify
from transformers import AutoProcessor, AutoModelForCausalLM
import pinggy as pg

# Ensure the current working directory is in sys.path
import sys
sys.path.insert(0, os.getcwd())
sys.path.insert(0, "ai-toolkit")
from toolkit.job import get_job
from huggingface_hub import whoami

MAX_IMAGES = 150

def load_captioning(uploaded_files, concept_sentence):
    uploaded_images = [file for file in uploaded_files if not file.endswith('.txt')]
    txt_files = [file for file in uploaded_files if file.endswith('.txt')]
    txt_files_dict = {os.path.splitext(os.path.basename(txt_file))[0]: txt_file for txt_file in txt_files}
    updates = []

    if len(uploaded_images) <= 1:
        raise pg.Error("Please upload at least 2 images to train your model (the ideal number with default settings is between 4-30)")
    elif len(uploaded_images) > MAX_IMAGES:
        raise pg.Error(f"For now, only {MAX_IMAGES} or less images are allowed for training")

    updates.append(pg.Update(visible=True))
    for i in range(1, MAX_IMAGES + 1):
        visible = i <= len(uploaded_images)
        updates.append(pg.Update(visible=visible))

        image_value = uploaded_images[i - 1] if visible else None
        updates.append(pg.Update(value=image_value, visible=visible))

        corresponding_caption = False
        if image_value:
            base_name = os.path.splitext(os.path.basename(image_value))[0]
            if base_name in txt_files_dict:
                with open(txt_files_dict[base_name], 'r') as file:
                    corresponding_caption = file.read()

        text_value = corresponding_caption if visible and corresponding_caption else "[trigger]" if visible and concept_sentence else None
        updates.append(pg.Update(value=text_value, visible=visible))

    updates.append(pg.Update(visible=True))
    updates.append(pg.Update(placeholder=f'A portrait of person in a bustling cafe {concept_sentence}', value=f'A person in a bustling cafe {concept_sentence}'))
    updates.append(pg.Update(placeholder=f"A mountainous landscape in the style of {concept_sentence}"))
    updates.append(pg.Update(placeholder=f"A {concept_sentence} in a mall"))
    updates.append(pg.Update(visible=True))

    return updates

def hide_captioning():
    return pg.Update(visible=False), pg.Update(visible=False), pg.Update(visible=False) 

def create_dataset(images, *captions):
    destination_folder = f"datasets/{uuid.uuid4()}"
    os.makedirs(destination_folder, exist_ok=True)

    jsonl_file_path = os.path.join(destination_folder, "metadata.jsonl")
    with open(jsonl_file_path, "a") as jsonl_file:
        for index, image in enumerate(images):
            new_image_path = shutil.copy(image, destination_folder)
            original_caption = captions[index]
            file_name = os.path.basename(new_image_path)
            data = {"file_name": file_name, "prompt": original_caption}
            jsonl_file.write(json.dumps(data) + "\n")

    return destination_folder

def run_captioning(images, concept_sentence, *captions):
    device = "cuda" if torch.cuda.is_available() else "cpu"
    torch_dtype = torch.float16
    model = AutoModelForCausalLM.from_pretrained(
        "multimodalart/Florence-2-large-no-flash-attn", torch_dtype=torch_dtype, trust_remote_code=True
    ).to(device)
    processor = AutoProcessor.from_pretrained("multimodalart/Florence-2-large-no-flash-attn", trust_remote_code=True)

    captions = list(captions)
    for i, image_path in enumerate(images):
        if isinstance(image_path, str):
            image = Image.open(image_path).convert("RGB")

        prompt = "<DETAILED_CAPTION>"
        inputs = processor(text=prompt, images=image, return_tensors="pt").to(device, torch_dtype)

        generated_ids = model.generate(
            input_ids=inputs["input_ids"], pixel_values=inputs["pixel_values"], max_new_tokens=1024, num_beams=3
        )

        generated_text = processor.batch_decode(generated_ids, skip_special_tokens=False)[0]
        parsed_answer = processor.post_process_generation(
            generated_text, task=prompt, image_size=(image.width, image.height)
        )
        caption_text = parsed_answer["<DETAILED_CAPTION>"].replace("The image shows ", "")
        if concept_sentence:
            caption_text = f"{caption_text} [trigger]"
        captions[i] = caption_text

        yield captions
    model.to("cpu")
    del model
    del processor

def recursive_update(d, u):
    for k, v in u.items():
        if isinstance(v, dict) and v:
            d[k] = recursive_update(d.get(k, {}), v)
        else:
            d[k] = v
    return d

def start_training(
    lora_name,
    concept_sentence,
    steps,
    lr,
    rank,
    model_to_train,
    low_vram,
    dataset_folder,
    sample_1,
    sample_2,
    sample_3,
    use_more_advanced_options,
    more_advanced_options,
):
    push_to_hub = True
    if not lora_name:
        raise pg.Error("You forgot to insert your LoRA name! This name has to be unique.")
    try:
        if whoami()["auth"]["accessToken"]["role"] == "write" or "repo.write" in whoami()["auth"]["accessToken"]["fineGrained"]["scoped"][0]["permissions"]:
            pg.Info(f"Starting training locally {whoami()['name']}. Your LoRA will be available locally and in Hugging Face after it finishes.")
        else:
            push_to_hub = False
            pg.Warning("Started training locally. Your LoRa will only be available locally because you didn't login with a `write` token to Hugging Face")
    except:
        push_to_hub = False
        pg.Warning("Started training locally. Your LoRa will only be available locally because you didn't login with a `write` token to Hugging Face")

    slugged_lora_name = slugify(lora_name)

    with open("config/examples/train_lora_flux_24gb.yaml", "r") as f:
        config = yaml.safe_load(f)

    config["config"]["name"] = slugged_lora_name
    config["config"]["process"][0]["model"]["low_vram"] = low_vram
    config["config"]["process"][0]["train"]["skip_first_sample"] = True
    config["config"]["process"][0]["train"]["steps"] = int(steps)
    config["config"]["process"][0]["train"]["lr"] = float(lr)
    config["config"]["process"][0]["network"]["linear"] = int(rank)
    config["config"]["process"][0]["network"]["linear_alpha"] = int(rank)
    config["config"]["process"][0]["datasets"][0]["folder_path"] = dataset_folder
    config["config"]["process"][0]["save"]["push_to_hub"] = push_to_hub
    if push_to_hub:
        try:
            username = whoami()["name"]
        except:
            raise pg.Error("Error trying to retrieve your username. Are you sure you are logged in with Hugging Face?")
        config["config"]["process"][0]["save"]["hf_repo_id"] = f"{username}/{slugged_lora_name}"
        config["config"]["process"][0]["save"]["hf_private"] = True
    if concept_sentence:
        config["config"]["process"][0]["trigger_word"] = concept_sentence

    if sample_1 or sample_2 or sample_3:
        config["config"]["process"][0]["train"]["disable_sampling"] = False
        config["config"]["process"][0]["sample"]["sample_every"] = steps
        config["config"]["process"][0]["sample"]["sample_steps"] = 28
        config["config"]["process"][0]["sample"]["prompts"] = []
        if sample_1:
            config["config"]["process"][0]["sample"]["prompts"].append(sample_1)
        if sample_2:
            config["config"]["process"][0]["sample"]["prompts"].append(sample_2)
        if sample_3:
            config["config"]["process"][0]["sample"]["prompts"].append(sample_3)
    else:
        config["config"]["process"][0]["train"]["disable_sampling"] = True
    if model_to_train == "schnell":
        config["config"]["process"][0]["model"]["name_or_path"] = "black-forest-labs/FLUX.1-schnell"
        config["config"]["process"][0]["model"]["assistant_lora_path"] = "ostris/FLUX.1-schnell-training-adapter"
        config["config"]["process"][0]["sample"]["sample_steps"] = 4
    if use_more_advanced_options:
        more_advanced_options_dict = yaml.safe_load(more_advanced_options)
        config["config"]["process"][0] = recursive_update(config["config"]["process"][0], more_advanced_options_dict)

    random_config_name = str(uuid.uuid4())
    os.makedirs("tmp", exist_ok=True)
    config_path = f"tmp/{random_config_name}-{slugged_lora_name}.yaml"
    with open(config_path, "w") as f:
        yaml.dump(config, f)

    job = get_job(config_path)
    job.run()
    job.cleanup()

    return f"Training completed successfully. Model saved as {slugged_lora_name}"

config_yaml = '''
device: cuda:0
model:
  is_flux: true
  quantize: true
network:
  linear: 16
  linear_alpha: 16
  type: lora
sample:
  guidance_scale: 3.5
  height: 1024
  neg: ''
  sample_every: 1000
  sample_steps: 28
  sampler: flowmatch
  seed: 42
  walk_seed: true
  width: 1024
save:
  dtype: float16
  hf_private: true
  max_step_saves_to_keep: 4
  push_to_hub: true
  save_every: 10000
train:
  batch_size: 1
  dtype: bf16
  ema_config:
    ema_decay: 0.99
    use_ema: true
  gradient_accumulation_steps: 1
  gradient_checkpointing: true
  noise_scheduler: flowmatch 
  optimizer: adamw8bit
  train_text_encoder: false
  train_unet: true
'''

def main():
    with pg.App() as app:
        app.add_page(title="LoRA Ease for FLUX", description="Train a high quality FLUX LoRA in a breeze")

        app.add_textbox(
            id="lora_name",
            label="The name of your LoRA",
            placeholder="e.g.: Persian Miniature Painting style, Cat Toy",
        )
        app.add_textbox(
            id="concept_sentence",
            label="Trigger word/sentence",
            placeholder="uncommon word like p3rs0n or trtcrd, or sentence like 'in the style of CNSTLL'",
        )

        image_upload = app.add_file_upload(
            id="images",
            label="Upload your images",
            file_types=["image", ".txt"],
            multiple=True,
        )

        captioning_area = app.add_container(id="captioning_area", visible=False)
        captioning_area.add_text("Custom captioning")
        do_captioning = app.add_button("Add AI captions with Florence-2", id="do_captioning")

        for i in range(1, MAX_IMAGES + 1):
            with captioning_area.add_row(id=f"captioning_row_{i}", visible=False) as row:
                row.add_image(id=f"image_{i}", width=111, height=111, visible=False)
                row.add_textbox(id=f"caption_{i}", label=f"Caption {i}")

        app.add_accordion(title="Advanced options", open=False)
        app.add_number(id="steps", label="Steps", value=1000, min=1, max=10000)
        app.add_number(id="lr", label="Learning Rate", value=4e-4, min=1e-6, max=1e-3)
        app.add_number(id="rank", label="LoRA Rank", value=16, min=4, max=128)
        app.add_radio(id="model_to_train", options=["dev", "schnell"], value="dev", label="Model to train")
        app.add_checkbox(id="low_vram", label="Low VRAM", value=True)

        with app.add_accordion(title="Even more advanced options", open=False):
            app.add_checkbox(id="use_more_advanced_options", label="Use more advanced options", value=False)
            app.add_code(id="more_advanced_options", value=config_yaml, language="yaml")

        app.add_accordion(title="Sample prompts (optional)", visible=False)
        app.add_textbox(id="sample_1", label="Test prompt 1")
        app.add_textbox(id="sample_2", label="Test prompt 2")
        app.add_textbox(id="sample_3", label="Test prompt 3")

        start = app.add_button("Start training", id="start", visible=False)
        progress_area = app.add_text("")

        app.on_upload(id="images", fn=load_captioning, inputs=["images", "concept_sentence"], outputs=["captioning_area", "sample", "start"])
        app.on_click(id="do_captioning", fn=run_captioning, inputs=["images", "concept_sentence"] + [f"caption_{i}" for i in range(1, MAX_IMAGES + 1)], outputs=[f"caption_{i}" for i in range(1, MAX_IMAGES + 1)])
        app.on_click(id="start", fn=create_dataset, inputs=["images"] + [f"caption_{i}" for i in range(1, MAX_IMAGES + 1)], outputs=["dataset_folder"])
        app.on_click(id="start", fn=start_training, inputs=[
            "lora_name",
            "concept_sentence",
            "steps",
            "lr",
            "rank",
            "model_to_train",
            "low_vram",
            "dataset_folder",
            "sample_1",
            "sample_2",
            "sample_3",
            "use_more_advanced_options",
            "more_advanced_options"
        ], outputs=["progress_area"])

        app.run()

if __name__ == "__main__":
    main()