|
| 1 | +# Stable Diffusion text-to-image fine-tuning |
| 2 | + |
| 3 | +The `train_text_to_image.py` script shows how to fine-tune stable diffusion model on your own dataset. |
| 4 | + |
| 5 | +___Note___: |
| 6 | + |
| 7 | +___This script is experimental. The script fine-tunes the whole model and often times the model overifits and runs into issues like catastrophic forgetting. It's recommended to try different hyperparamters to get the best result on your dataset.___ |
| 8 | + |
| 9 | + |
| 10 | +## Running locally |
| 11 | +### Installing the dependencies |
| 12 | + |
| 13 | +Before running the scripts, make sure to install the library's training dependencies: |
| 14 | + |
| 15 | +```bash |
| 16 | +pip install git+https://github.com/huggingface/diffusers.git |
| 17 | +pip install -U -r requirements.txt |
| 18 | +``` |
| 19 | + |
| 20 | +And initialize an [🤗Accelerate](https://github.com/huggingface/accelerate/) environment with: |
| 21 | + |
| 22 | +```bash |
| 23 | +accelerate config |
| 24 | +``` |
| 25 | + |
| 26 | +### Pokemon example |
| 27 | + |
| 28 | +You need to accept the model license before downloading or using the weights. In this example we'll use model version `v1-4`, so you'll need to visit [its card](https://huggingface.co/CompVis/stable-diffusion-v1-4), read the license and tick the checkbox if you agree. |
| 29 | + |
| 30 | +You have to be a registered user in 🤗 Hugging Face Hub, and you'll also need to use an access token for the code to work. For more information on access tokens, please refer to [this section of the documentation](https://huggingface.co/docs/hub/security-tokens). |
| 31 | + |
| 32 | +Run the following command to authenticate your token |
| 33 | + |
| 34 | +```bash |
| 35 | +huggingface-cli login |
| 36 | +``` |
| 37 | + |
| 38 | +If you have already cloned the repo, then you won't need to go through these steps. |
| 39 | + |
| 40 | +<br> |
| 41 | + |
| 42 | +#### Hardware |
| 43 | +With `gradient_checkpointing` and `mixed_precision` it should be possible to fine tune the model on a single 24GB GPU. For higher `batch_size` and faster training it's better to use GPUs with >30GB memory. |
| 44 | + |
| 45 | +```bash |
| 46 | +export MODEL_NAME="CompVis/stable-diffusion-v1-4" |
| 47 | +export dataset_name="lambdalabs/pokemon-blip-captions" |
| 48 | + |
| 49 | +accelerate launch train_text_to_image.py \ |
| 50 | + --pretrained_model_name_or_path=$MODEL_NAME \ |
| 51 | + --dataset_name=$dataset_name \ |
| 52 | + --use_ema \ |
| 53 | + --resolution=512 --center_crop --random_flip \ |
| 54 | + --train_batch_size=1 \ |
| 55 | + --gradient_accumulation_steps=4 \ |
| 56 | + --gradient_checkpointing \ |
| 57 | + --mixed_precision="fp16" \ |
| 58 | + --max_train_steps=15000 \ |
| 59 | + --learning_rate=1e-05 \ |
| 60 | + --max_grad_norm=1 \ |
| 61 | + --lr_scheduler="constant" --lr_warmup_steps=0 \ |
| 62 | + --output_dir="sd-pokemon-model" |
| 63 | +``` |
| 64 | + |
| 65 | + |
| 66 | +To run on your own training files prepare the dataset according to the format required by `datasets`, you can find the instructions for how to do that in this [document](https://huggingface.co/docs/datasets/v2.4.0/en/image_load#imagefolder-with-metadata). |
| 67 | +If you wish to use custom loading logic, you should modify the script, we have left pointers for that in the training script. |
| 68 | + |
| 69 | +```bash |
| 70 | +export MODEL_NAME="CompVis/stable-diffusion-v1-4" |
| 71 | +export TRAIN_DIR="path_to_your_dataset" |
| 72 | + |
| 73 | +accelerate launch train_text_to_image.py \ |
| 74 | + --pretrained_model_name_or_path=$MODEL_NAME \ |
| 75 | + --train_data_dir=$TRAIN_DIR \ |
| 76 | + --use_ema \ |
| 77 | + --resolution=512 --center_crop --random_flip \ |
| 78 | + --train_batch_size=1 \ |
| 79 | + --gradient_accumulation_steps=4 \ |
| 80 | + --gradient_checkpointing \ |
| 81 | + --mixed_precision="fp16" \ |
| 82 | + --max_train_steps=15000 \ |
| 83 | + --learning_rate=1e-05 \ |
| 84 | + --max_grad_norm=1 \ |
| 85 | + --lr_scheduler="constant" --lr_warmup_steps=0 \ |
| 86 | + --output_dir="sd-pokemon-model" |
| 87 | +``` |
| 88 | + |
| 89 | +Once the training is finished the model will be saved in the `output_dir` specified in the command. In this example it's `sd-pokemon-model`. To load the fine-tuned model for inference just pass that path to `StableDiffusionPipeline` |
| 90 | + |
| 91 | + |
| 92 | +```python |
| 93 | +from diffusers import StableDiffusionPipeline |
| 94 | + |
| 95 | +model_path = "path_to_saved_model" |
| 96 | +pipe = StableDiffusionPipeline.from_pretrained(model_path, torch_dtype=torch.float16) |
| 97 | +pipe.to("cuda") |
| 98 | + |
| 99 | +image = pipe(prompt="yoda").images[0] |
| 100 | +image.save("yoda-pokemon.png") |
| 101 | +``` |
0 commit comments