A new approach for personalizing text-to-image diffusion models to user needs has been proposed, enabling the synthesis of novel images of a specific subject in different contexts. This method involves fine-tuning a pretrained model with a few input images of the subject and using a unique identifier to embed the subject in the output domain. As a result, fully-novel photorealistic images of the subject can be synthesized in diverse scenes, poses, views, and lighting conditions that are not present in the reference images.

You can use cURL or any other http client to run inferences:

```bash
curl -X POST \
    -H "Authorization: bearer $(deepctl auth token)"  \
    -F 'instance_prompt=a photo of nkb person'  \
    -F 'class_prompt=a photo of a person'  \
    -F instance_data=@my_archive.zip  \
    -F output_model_name=my_username/my_model_name  \
    'https://api.deepinfra.com/v1/inference/deepinfra/dreambooth'
```

which will give you back something similar to:

```json
{
  "model_name": "my_user_name/my_cool_model",
  "version": "90lkjfslkf5lkdjflskjfsdkl",
  "final_loss": 0.123,
  "request_id": null,
  "inference_status": {
    "status": "unknown",
    "runtime_ms": 0,
    "cost": 0.0,
    "tokens_generated": 0,
    "tokens_input": 0
  }
}

```


You can use our command-line tool [deepctl](/docs/getting-started) to run
inferences:

```bash
deepctl infer \
    -m 'deepinfra/dreambooth'  \
    -i 'instance_prompt=a photo of nkb person'  \
    -i 'class_prompt=a photo of a person'  \
    -i instance_data=@my_archive.zip  \
    -i output_model_name=my_username/my_model_name
```

which will give you back something similar to:

```json
{
  "model_name": "my_user_name/my_cool_model",
  "version": "90lkjfslkf5lkdjflskjfsdkl",
  "final_loss": 0.123,
  "request_id": null,
  "inference_status": {
    "status": "unknown",
    "runtime_ms": 0,
    "cost": 0.0,
    "tokens_generated": 0,
    "tokens_input": 0
  }
}

```


The prompt used to describe your training images. Should be something like:'a photo of [identifier] [class noun]'. where [identifier] should be something unique

instance_prompt

The prompt used to describe your class. Should be something like:'a photo of a [class noun]'. where [class_noun] describe a class of instance

class_prompt

zip file containing images of instance you want to train on

instance_data

The name of the model to be saved. Must start with your github username

output_model_name

The maximum number of training steps to run

max_train_steps

train_batch_size

Whether to center crop the images before resizing to resolution

center_crop

Initial learning rate (after the potential warmup period) to use.

learning_rate

The scheduler type to use during training.

lr_scheduler

Number of steps for the warmup in the lr scheduler

lr_warmup_steps

Whether or not to use 8-bit Adam from bitsandbytes.

use_8bit_adam

The beta1 parameter for the Adam optimizer.

adam_beta1

The beta2 parameter for the Adam optimizer.

adam_beta2

adam_weight_decay

adam_epsilon

max_grad_norm

Whether or not to use gradient checkpointing to save memory at the expense of slower backward pass.

gradient_checkpointing

The webhook to call when inference is done, by default you will get the output in the response of your inference request