peachyland / concept_removal Goto Github PK

Textual inversion is a method to personalize text2image models like stable diffusion on your own images using just 3-5 examples. The textual_inversion.py script shows how to implement the training procedure and adapt it for stable diffusion.

Colab for training

Colab for inference

Before running the scripts, make sure to install the library's training dependencies:

Important

To make sure you can successfully run the latest versions of the example scripts, we highly recommend installing from source and keeping the install up to date as we update the example scripts frequently and install some example-specific requirements. To do this, execute the following steps in a new virtual environment:

git clone https://github.com/huggingface/diffusers
cd diffusers
pip install .

Then cd in the example folder and run:

pip install -r requirements.txt

And initialize an 🤗 Accelerate environment with:

accelerate config

First, let's login so that we can upload the checkpoint to the Hub during training:

huggingface-cli login

Now let's get our dataset. For this example we will use some cat images: https://huggingface.co/datasets/diffusers/cat_toy_example .

Let's first download it locally:

from huggingface_hub import snapshot_download

local_dir = "./cat"
snapshot_download("diffusers/cat_toy_example", local_dir=local_dir, repo_type="dataset", ignore_patterns=".gitattributes")

This will be our training data. Now we can launch the training using:

Note: Change the resolution to 768 if you are using the stable-diffusion-2 768x768 model.

Note: Please follow the README_sdxl.md if you are using the stable-diffusion-xl.

export MODEL_NAME="runwayml/stable-diffusion-v1-5"
export DATA_DIR="./cat"

accelerate launch textual_inversion.py \
  --pretrained_model_name_or_path=$MODEL_NAME \
  --train_data_dir=$DATA_DIR \
  --learnable_property="object" \
  --placeholder_token="<cat-toy>" \
  --initializer_token="toy" \
  --resolution=512 \
  --train_batch_size=1 \
  --gradient_accumulation_steps=4 \
  --max_train_steps=3000 \
  --learning_rate=5.0e-04 \
  --scale_lr \
  --lr_scheduler="constant" \
  --lr_warmup_steps=0 \
  --push_to_hub \
  --output_dir="textual_inversion_cat"

A full training run takes ~1 hour on one V100 GPU.

Note: As described in the official paper only one embedding vector is used for the placeholder token, e.g. "<cat-toy>". However, one can also add multiple embedding vectors for the placeholder token to increase the number of fine-tuneable parameters. This can help the model to learn more complex details. To use multiple embedding vectors, you should define --num_vectors to a number larger than one, e.g.:

--num_vectors 5

The saved textual inversion vectors will then be larger in size compared to the default case.

Once you have trained a model using above command, the inference can be done simply using the StableDiffusionPipeline. Make sure to include the placeholder_token in your prompt.

from diffusers import StableDiffusionPipeline
import torch

model_id = "path-to-your-trained-model"
pipe = StableDiffusionPipeline.from_pretrained(model_id,torch_dtype=torch.float16).to("cuda")

prompt = "A <cat-toy> backpack"

image = pipe(prompt, num_inference_steps=50, guidance_scale=7.5).images[0]

image.save("cat-backpack.png")

For faster training on TPUs and GPUs you can leverage the flax training example. Follow the instructions above to get the model and dataset before running the script.

Before running the scripts, make sure to install the library's training dependencies:

pip install -U -r requirements_flax.txt

export MODEL_NAME="duongna/stable-diffusion-v1-4-flax"
export DATA_DIR="path-to-dir-containing-images"

python textual_inversion_flax.py \
  --pretrained_model_name_or_path=$MODEL_NAME \
  --train_data_dir=$DATA_DIR \
  --learnable_property="object" \
  --placeholder_token="<cat-toy>" \
  --initializer_token="toy" \
  --resolution=512 \
  --train_batch_size=1 \
  --max_train_steps=3000 \
  --learning_rate=5.0e-04 \
  --scale_lr \
  --output_dir="textual_inversion_cat"

It should be at least 70% faster than the PyTorch script with the same configuration.

You can enable memory efficient attention by installing xFormers and padding the --enable_xformers_memory_efficient_attention argument to the script. This is not available with the Flax/JAX implementation.