Textual Inversion

Textual Inversion is a technique for capturing novel concepts from a small number of example images. While the technique was originally demonstrated with a latent diffusion model, it has since been applied to other model variants like Stable Diffusion. The learned concepts can be used to better control the images generated from text-to-image pipelines. It learns new “words” in the text encoder’s embedding space, which are used within text prompts for personalized image generation.

This guide will show you how to train a runwayml/stable-diffusion-v1-5 model with Textual Inversion. All the training scripts for Textual Inversion used in this guide can be found here if you’re interested in taking a closer look at how things work under the hood.

Before you begin, make sure you install the library’s training dependencies:

pip install diffusers accelerate transformers

After all the dependencies have been set up, initialize a 🤗Accelerate environment with:

accelerate config

To setup a default 🤗 Accelerate environment without choosing any configurations:

accelerate config default

Or if your environment doesn’t support an interactive shell like a notebook, you can use:

from accelerate.utils import write_basic_config

write_basic_config()

Finally, you try and install xFormers to reduce your memory footprint with xFormers memory-efficient attention. Once you have xFormers installed, add the --enable_xformers_memory_efficient_attention argument to the training script. xFormers is not supported for Flax.

Upload model to Hub

If you want to store your model on the Hub, add the following argument to the training script:

--push_to_hub

Save and load checkpoints

It is often a good idea to regularly save checkpoints of your model during training. This way, you can resume training from a saved checkpoint if your training is interrupted for any reason. To save a checkpoint, pass the following argument to the training script to save the full training state in a subfolder in output_dir every 500 steps:

--checkpointing_steps=500

To resume training from a saved checkpoint, pass the following argument to the training script and the specific checkpoint you’d like to resume from:

--resume_from_checkpoint="checkpoint-1500"

Finetuning

For your training dataset, download these images of a cat statue and store them in a directory.

Set the MODEL_NAME environment variable to the model repository id, and the DATA_DIR environment variable to the path of the directory containing the images. Now you can launch the training script:

export MODEL_NAME="runwayml/stable-diffusion-v1-5"
export DATA_DIR="path-to-dir-containing-images"

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 \
  --output_dir="textual_inversion_cat"

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"

Intermediate logging

If you’re interested in following along with your model training progress, you can save the generated images from the training process. Add the following arguments to the training script to enable intermediate logging:

• validation_prompt, the prompt used to generate samples (this is set to None by default and intermediate logging is disabled)
• num_validation_images, the number of sample images to generate
• validation_steps, the number of steps before generating num_validation_images from the validation_prompt

--validation_prompt="A <cat-toy> backpack"
--num_validation_images=4
--validation_steps=100

Inference

Once you have trained a model, you can use it for inference with the StableDiffusionPipeline.

The textual inversion script will by default only save the textual inversion embedding vector(s) that have been added to the text encoder embedding matrix and consequently been trained.

from diffusers import StableDiffusionPipeline
import torch

model_id = "runwayml/stable-diffusion-v1-5"
pipe = StableDiffusionPipeline.from_pretrained(model_id, torch_dtype=torch.float16).to("cuda")

pipe.load_textual_inversion("sd-concepts-library/cat-toy")

prompt = "A <cat-toy> backpack"

image = pipe(prompt, num_inference_steps=50).images[0]
image.save("cat-backpack.png")

pipe.load_textual_inversion("./charturnerv2.pt")

import jax
import numpy as np
from flax.jax_utils import replicate
from flax.training.common_utils import shard
from diffusers import FlaxStableDiffusionPipeline

model_path = "path-to-your-trained-model"
pipe, params = FlaxStableDiffusionPipeline.from_pretrained(model_path, dtype=jax.numpy.bfloat16)

prompt = "A <cat-toy> backpack"
prng_seed = jax.random.PRNGKey(0)
num_inference_steps = 50

num_samples = jax.device_count()
prompt = num_samples * [prompt]
prompt_ids = pipeline.prepare_inputs(prompt)

# shard inputs and rng
params = replicate(params)
prng_seed = jax.random.split(prng_seed, jax.device_count())
prompt_ids = shard(prompt_ids)

images = pipeline(prompt_ids, params, prng_seed, num_inference_steps, jit=True).images
images = pipeline.numpy_to_pil(np.asarray(images.reshape((num_samples,) + images.shape[-3:])))
image.save("cat-backpack.png")

How it works

Usually, text prompts are tokenized into an embedding before being passed to a model, which is often a transformer. Textual Inversion does something similar, but it learns a new token embedding, v*, from a special token S* in the diagram above. The model output is used to condition the diffusion model, which helps the diffusion model understand the prompt and new concepts from just a few example images.

To do this, Textual Inversion uses a generator model and noisy versions of the training images. The generator tries to predict less noisy versions of the images, and the token embedding v* is optimized based on how well the generator does. If the token embedding successfully captures the new concept, it gives more useful information to the diffusion model and helps create clearer images with less noise. This optimization process typically occurs after several thousand steps of exposure to a variety of prompt and image variants.