Buckets:

hf-doc-build
/

doc-dev

Files

xet

hf-doc-build/doc-dev / diffusers /main /en /training /dreambooth.html

rtrm

21 days ago

download

raw

75.9 kB

	<meta charset="utf-8" /><meta name="hf:doc:metadata" content="{"title":"DreamBooth","local":"dreambooth","sections":[{"title":"Script parameters","local":"script-parameters","sections":[{"title":"Min-SNR weighting","local":"min-snr-weighting","sections":[],"depth":3},{"title":"Prior preservation loss","local":"prior-preservation-loss","sections":[],"depth":3},{"title":"Train text encoder","local":"train-text-encoder","sections":[],"depth":3}],"depth":2},{"title":"Training script","local":"training-script","sections":[],"depth":2},{"title":"Launch the script","local":"launch-the-script","sections":[],"depth":2},{"title":"LoRA","local":"lora","sections":[],"depth":2},{"title":"Stable Diffusion XL","local":"stable-diffusion-xl","sections":[],"depth":2},{"title":"DeepFloyd IF","local":"deepfloyd-if","sections":[{"title":"Training tips","local":"training-tips","sections":[],"depth":3}],"depth":2},{"title":"Next steps","local":"next-steps","sections":[],"depth":2}],"depth":1}">
	<link href="/docs/diffusers/main/en/_app/immutable/assets/0.e3b0c442.css" rel="modulepreload">
	<link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/entry/start.21e27d66.js">
	<link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/chunks/scheduler.8c3d61f6.js">
	<link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/chunks/singletons.1db06f6d.js">
	<link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/chunks/index.0997d446.js">
	<link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/chunks/paths.085e8bc8.js">
	<link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/entry/app.de4fb612.js">
	<link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/chunks/index.da70eac4.js">
	<link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/nodes/0.f6117ae5.js">
	<link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/chunks/each.e59479a4.js">
	<link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/nodes/187.a887dc8c.js">
	<link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/chunks/Tip.1d9b8c37.js">
	<link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/chunks/CodeBlock.00a903b3.js">
	<link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/chunks/EditOnGithub.1e64e623.js">
	<link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/chunks/HfOption.c1483eb1.js">
	<link rel="modulepreload" href="/docs/diffusers/main/en/_app/immutable/chunks/stores.d6eecc38.js"><!-- HEAD_svelte-u9bgzb_START --><meta name="hf:doc:metadata" content="{"title":"DreamBooth","local":"dreambooth","sections":[{"title":"Script parameters","local":"script-parameters","sections":[{"title":"Min-SNR weighting","local":"min-snr-weighting","sections":[],"depth":3},{"title":"Prior preservation loss","local":"prior-preservation-loss","sections":[],"depth":3},{"title":"Train text encoder","local":"train-text-encoder","sections":[],"depth":3}],"depth":2},{"title":"Training script","local":"training-script","sections":[],"depth":2},{"title":"Launch the script","local":"launch-the-script","sections":[],"depth":2},{"title":"LoRA","local":"lora","sections":[],"depth":2},{"title":"Stable Diffusion XL","local":"stable-diffusion-xl","sections":[],"depth":2},{"title":"DeepFloyd IF","local":"deepfloyd-if","sections":[{"title":"Training tips","local":"training-tips","sections":[],"depth":3}],"depth":2},{"title":"Next steps","local":"next-steps","sections":[],"depth":2}],"depth":1}"><!-- HEAD_svelte-u9bgzb_END --> <p></p> <h1 class="relative group"><a id="dreambooth" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#dreambooth"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>DreamBooth</span></h1> <p data-svelte-h="svelte-yz6jri"><a href="https://huggingface.co/papers/2208.12242" rel="nofollow">DreamBooth</a> is a training technique that updates the entire diffusion model by training on just a few images of a subject or style. It works by associating a special word in the prompt with the example images.</p> <p data-svelte-h="svelte-en2tmp">If you’re training on a GPU with limited vRAM, you should try enabling the <code>gradient_checkpointing</code> and <code>mixed_precision</code> parameters in the training command. You can also reduce your memory footprint by using memory-efficient attention with <a href="../optimization/xformers">xFormers</a>. JAX/Flax training is also supported for efficient training on TPUs and GPUs, but it doesn’t support gradient checkpointing or xFormers. You should have a GPU with >30GB of memory if you want to train faster with Flax.</p> <p data-svelte-h="svelte-1ol1kd9">This guide will explore the <a href="https://github.com/huggingface/diffusers/blob/main/examples/dreambooth/train_dreambooth.py" rel="nofollow">train_dreambooth.py</a> script to help you become more familiar with it, and how you can adapt it for your own use-case.</p> <p data-svelte-h="svelte-l7dm2q">Before running the script, make sure you install the library from source:</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START -->git <span class="hljs-built_in">clone</span> https://github.com/huggingface/diffusers
	<span class="hljs-built_in">cd</span> diffusers
	pip install .<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1nancx8">Navigate to the example folder with the training script and install the required dependencies for the script you’re using:</p> <div class="flex space-x-2 items-center my-1.5 mr-8 h-7 !pl-0 -mx-3 md:mx-0"><div class="flex items-center border rounded-lg px-1.5 py-1 leading-none select-none text-smd border-gray-800 bg-black dark:bg-gray-700 text-white">PyTorch </div><div class="flex items-center border rounded-lg px-1.5 py-1 leading-none select-none text-smd text-gray-500 cursor-pointer opacity-90 hover:text-gray-700 dark:hover:text-gray-200 hover:shadow-sm">Flax </div></div> <div class="language-select"><div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-built_in">cd</span> examples/dreambooth
	pip install -r requirements.txt<!-- HTML_TAG_END --></pre></div> </div> <div class="course-tip bg-gradient-to-br dark:bg-gradient-to-r before:border-green-500 dark:before:border-green-800 from-green-50 dark:from-gray-900 to-white dark:to-gray-950 border border-green-50 text-green-700 dark:text-gray-400"><p data-svelte-h="svelte-1qbiqsn">🤗 Accelerate is a library for helping you train on multiple GPUs/TPUs or with mixed-precision. It’ll automatically configure your training setup based on your hardware and environment. Take a look at the 🤗 Accelerate <a href="https://huggingface.co/docs/accelerate/quicktour" rel="nofollow">Quick tour</a> to learn more.</p></div> <p data-svelte-h="svelte-60q53m">Initialize an 🤗 Accelerate environment:</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START -->accelerate config<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-di6juu">To setup a default 🤗 Accelerate environment without choosing any configurations:</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START -->accelerate config default<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-tsz4qp">Or if your environment doesn’t support an interactive shell, like a notebook, you can use:</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-keyword">from</span> accelerate.utils <span class="hljs-keyword">import</span> write_basic_config

	write_basic_config()<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1fkfdql">Lastly, if you want to train a model on your own dataset, take a look at the <a href="create_dataset">Create a dataset for training</a> guide to learn how to create a dataset that works with the training script.</p> <div class="course-tip bg-gradient-to-br dark:bg-gradient-to-r before:border-green-500 dark:before:border-green-800 from-green-50 dark:from-gray-900 to-white dark:to-gray-950 border border-green-50 text-green-700 dark:text-gray-400"><p data-svelte-h="svelte-ag3huc">The following sections highlight parts of the training script that are important for understanding how to modify it, but it doesn’t cover every aspect of the script in detail. If you’re interested in learning more, feel free to read through the <a href="https://github.com/huggingface/diffusers/blob/main/examples/dreambooth/train_dreambooth.py" rel="nofollow">script</a> and let us know if you have any questions or concerns.</p></div> <h2 class="relative group"><a id="script-parameters" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#script-parameters"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Script parameters</span></h2> <div class="course-tip course-tip-orange bg-gradient-to-br dark:bg-gradient-to-r before:border-orange-500 dark:before:border-orange-800 from-orange-50 dark:from-gray-900 to-white dark:to-gray-950 border border-orange-50 text-orange-700 dark:text-gray-400"><p data-svelte-h="svelte-1gadrly">DreamBooth is very sensitive to training hyperparameters, and it is easy to overfit. Read the <a href="https://huggingface.co/blog/dreambooth" rel="nofollow">Training Stable Diffusion with Dreambooth using 🧨 Diffusers</a> blog post for recommended settings for different subjects to help you choose the appropriate hyperparameters.</p></div> <p data-svelte-h="svelte-gb0mrm">The training script offers many parameters for customizing your training run. All of the parameters and their descriptions are found in the <a href="https://github.com/huggingface/diffusers/blob/072e00897a7cf4302c347a63ec917b4b8add16d4/examples/dreambooth/train_dreambooth.py#L228" rel="nofollow"><code>parse_args()</code></a> function. The parameters are set with default values that should work pretty well out-of-the-box, but you can also set your own values in the training command if you’d like.</p> <p data-svelte-h="svelte-1aa3zni">For example, to train in the bf16 format:</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START -->accelerate launch train_dreambooth.py \
	--mixed_precision=<span class="hljs-string">"bf16"</span><!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1ug35mp">Some basic and important parameters to know and specify are:</p> <ul data-svelte-h="svelte-pohxvr"><li><code>--pretrained_model_name_or_path</code>: the name of the model on the Hub or a local path to the pretrained model</li> <li><code>--instance_data_dir</code>: path to a folder containing the training dataset (example images)</li> <li><code>--instance_prompt</code>: the text prompt that contains the special word for the example images</li> <li><code>--train_text_encoder</code>: whether to also train the text encoder</li> <li><code>--output_dir</code>: where to save the trained model</li> <li><code>--push_to_hub</code>: whether to push the trained model to the Hub</li> <li><code>--checkpointing_steps</code>: frequency of saving a checkpoint as the model trains; this is useful if for some reason training is interrupted, you can continue training from that checkpoint by adding <code>--resume_from_checkpoint</code> to your training command</li></ul> <h3 class="relative group"><a id="min-snr-weighting" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#min-snr-weighting"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Min-SNR weighting</span></h3> <p data-svelte-h="svelte-isitbv">The <a href="https://huggingface.co/papers/2303.09556" rel="nofollow">Min-SNR</a> weighting strategy can help with training by rebalancing the loss to achieve faster convergence. The training script supports predicting <code>epsilon</code> (noise) or <code>v_prediction</code>, but Min-SNR is compatible with both prediction types. This weighting strategy is only supported by PyTorch and is unavailable in the Flax training script.</p> <p data-svelte-h="svelte-tp3kp">Add the <code>--snr_gamma</code> parameter and set it to the recommended value of 5.0:</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START -->accelerate launch train_dreambooth.py \
	--snr_gamma=5.0<!-- HTML_TAG_END --></pre></div> <h3 class="relative group"><a id="prior-preservation-loss" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#prior-preservation-loss"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Prior preservation loss</span></h3> <p data-svelte-h="svelte-1oc83h6">Prior preservation loss is a method that uses a model’s own generated samples to help it learn how to generate more diverse images. Because these generated sample images belong to the same class as the images you provided, they help the model retain what it has learned about the class and how it can use what it already knows about the class to make new compositions.</p> <ul data-svelte-h="svelte-1e6kqrs"><li><code>--with_prior_preservation</code>: whether to use prior preservation loss</li> <li><code>--prior_loss_weight</code>: controls the influence of the prior preservation loss on the model</li> <li><code>--class_data_dir</code>: path to a folder containing the generated class sample images</li> <li><code>--class_prompt</code>: the text prompt describing the class of the generated sample images</li></ul> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START -->accelerate launch train_dreambooth.py \
	--with_prior_preservation \
	--prior_loss_weight=1.0 \
	--class_data_dir=<span class="hljs-string">"path/to/class/images"</span> \
	--class_prompt=<span class="hljs-string">"text prompt describing class"</span><!-- HTML_TAG_END --></pre></div> <h3 class="relative group"><a id="train-text-encoder" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#train-text-encoder"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Train text encoder</span></h3> <p data-svelte-h="svelte-n5xd96">To improve the quality of the generated outputs, you can also train the text encoder in addition to the UNet. This requires additional memory and you’ll need a GPU with at least 24GB of vRAM. If you have the necessary hardware, then training the text encoder produces better results, especially when generating images of faces. Enable this option by:</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START -->accelerate launch train_dreambooth.py \
	--train_text_encoder<!-- HTML_TAG_END --></pre></div> <h2 class="relative group"><a id="training-script" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#training-script"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Training script</span></h2> <p data-svelte-h="svelte-76mqwg">DreamBooth comes with its own dataset classes:</p> <ul data-svelte-h="svelte-1sge1kf"><li><a href="https://github.com/huggingface/diffusers/blob/072e00897a7cf4302c347a63ec917b4b8add16d4/examples/dreambooth/train_dreambooth.py#L604" rel="nofollow"><code>DreamBoothDataset</code></a>: preprocesses the images and class images, and tokenizes the prompts for training</li> <li><a href="https://github.com/huggingface/diffusers/blob/072e00897a7cf4302c347a63ec917b4b8add16d4/examples/dreambooth/train_dreambooth.py#L738" rel="nofollow"><code>PromptDataset</code></a>: generates the prompt embeddings to generate the class images</li></ul> <p data-svelte-h="svelte-c9697n">If you enabled <a href="https://github.com/huggingface/diffusers/blob/072e00897a7cf4302c347a63ec917b4b8add16d4/examples/dreambooth/train_dreambooth.py#L842" rel="nofollow">prior preservation loss</a>, the class images are generated here:</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START -->sample_dataset = PromptDataset(args.class_prompt, num_new_images)
	sample_dataloader = torch.utils.data.DataLoader(sample_dataset, batch_size=args.sample_batch_size)

	sample_dataloader = accelerator.prepare(sample_dataloader)
	pipeline.to(accelerator.device)

	<span class="hljs-keyword">for</span> example <span class="hljs-keyword">in</span> tqdm(
	sample_dataloader, desc=<span class="hljs-string">"Generating class images"</span>, disable=<span class="hljs-keyword">not</span> accelerator.is_local_main_process
	):
	images = pipeline(example[<span class="hljs-string">"prompt"</span>]).images<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-e2s2vm">Next is the <a href="https://github.com/huggingface/diffusers/blob/072e00897a7cf4302c347a63ec917b4b8add16d4/examples/dreambooth/train_dreambooth.py#L799" rel="nofollow"><code>main()</code></a> function which handles setting up the dataset for training and the training loop itself. The script loads the <a href="https://github.com/huggingface/diffusers/blob/072e00897a7cf4302c347a63ec917b4b8add16d4/examples/dreambooth/train_dreambooth.py#L898" rel="nofollow">tokenizer</a>, <a href="https://github.com/huggingface/diffusers/blob/072e00897a7cf4302c347a63ec917b4b8add16d4/examples/dreambooth/train_dreambooth.py#L912C1-L912C1" rel="nofollow">scheduler and models</a>:</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-comment"># Load the tokenizer</span>
	<span class="hljs-keyword">if</span> args.tokenizer_name:
	tokenizer = AutoTokenizer.from_pretrained(args.tokenizer_name, revision=args.revision, use_fast=<span class="hljs-literal">False</span>)
	<span class="hljs-keyword">elif</span> args.pretrained_model_name_or_path:
	tokenizer = AutoTokenizer.from_pretrained(
	args.pretrained_model_name_or_path,
	subfolder=<span class="hljs-string">"tokenizer"</span>,
	revision=args.revision,
	use_fast=<span class="hljs-literal">False</span>,
	)

	<span class="hljs-comment"># Load scheduler and models</span>
	noise_scheduler = DDPMScheduler.from_pretrained(args.pretrained_model_name_or_path, subfolder=<span class="hljs-string">"scheduler"</span>)
	text_encoder = text_encoder_cls.from_pretrained(
	args.pretrained_model_name_or_path, subfolder=<span class="hljs-string">"text_encoder"</span>, revision=args.revision
	)

	<span class="hljs-keyword">if</span> model_has_vae(args):
	vae = AutoencoderKL.from_pretrained(
	args.pretrained_model_name_or_path, subfolder=<span class="hljs-string">"vae"</span>, revision=args.revision
	)
	<span class="hljs-keyword">else</span>:
	vae = <span class="hljs-literal">None</span>

	unet = UNet2DConditionModel.from_pretrained(
	args.pretrained_model_name_or_path, subfolder=<span class="hljs-string">"unet"</span>, revision=args.revision
	)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-s7uwdj">Then, it’s time to <a href="https://github.com/huggingface/diffusers/blob/072e00897a7cf4302c347a63ec917b4b8add16d4/examples/dreambooth/train_dreambooth.py#L1073" rel="nofollow">create the training dataset</a> and DataLoader from <code>DreamBoothDataset</code>:</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START -->train_dataset = DreamBoothDataset(
	instance_data_root=args.instance_data_dir,
	instance_prompt=args.instance_prompt,
	class_data_root=args.class_data_dir <span class="hljs-keyword">if</span> args.with_prior_preservation <span class="hljs-keyword">else</span> <span class="hljs-literal">None</span>,
	class_prompt=args.class_prompt,
	class_num=args.num_class_images,
	tokenizer=tokenizer,
	size=args.resolution,
	center_crop=args.center_crop,
	encoder_hidden_states=pre_computed_encoder_hidden_states,
	class_prompt_encoder_hidden_states=pre_computed_class_prompt_encoder_hidden_states,
	tokenizer_max_length=args.tokenizer_max_length,
	)

	train_dataloader = torch.utils.data.DataLoader(
	train_dataset,
	batch_size=args.train_batch_size,
	shuffle=<span class="hljs-literal">True</span>,
	collate_fn=<span class="hljs-keyword">lambda</span> examples: collate_fn(examples, args.with_prior_preservation),
	num_workers=args.dataloader_num_workers,
	)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-4ixpqk">Lastly, the <a href="https://github.com/huggingface/diffusers/blob/072e00897a7cf4302c347a63ec917b4b8add16d4/examples/dreambooth/train_dreambooth.py#L1151" rel="nofollow">training loop</a> takes care of the remaining steps such as converting images to latent space, adding noise to the input, predicting the noise residual, and calculating the loss.</p> <p data-svelte-h="svelte-6gmbd2">If you want to learn more about how the training loop works, check out the <a href="../using-diffusers/write_own_pipeline">Understanding pipelines, models and schedulers</a> tutorial which breaks down the basic pattern of the denoising process.</p> <h2 class="relative group"><a id="launch-the-script" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#launch-the-script"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Launch the script</span></h2> <p data-svelte-h="svelte-1fq45yi">You’re now ready to launch the training script! 🚀</p> <p data-svelte-h="svelte-191fegd">For this guide, you’ll download some images of a <a href="https://huggingface.co/datasets/diffusers/dog-example" rel="nofollow">dog</a> and store them in a directory. But remember, you can create and use your own dataset if you want (see the <a href="create_dataset">Create a dataset for training</a> guide).</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-keyword">from</span> huggingface_hub <span class="hljs-keyword">import</span> snapshot_download

	local_dir = <span class="hljs-string">"./dog"</span>
	snapshot_download(
	<span class="hljs-string">"diffusers/dog-example"</span>,
	local_dir=local_dir,
	repo_type=<span class="hljs-string">"dataset"</span>,
	ignore_patterns=<span class="hljs-string">".gitattributes"</span>,
	)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-bekdku">Set the environment variable <code>MODEL_NAME</code> to a model id on the Hub or a path to a local model, <code>INSTANCE_DIR</code> to the path where you just downloaded the dog images to, and <code>OUTPUT_DIR</code> to where you want to save the model. You’ll use <code>sks</code> as the special word to tie the training to.</p> <p data-svelte-h="svelte-196wtq5">If you’re interested in following along with the training process, you can periodically save generated images as training progresses. Add the following parameters to the training command:</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START -->--validation_prompt=<span class="hljs-string">"a photo of a sks dog"</span>
	--num_validation_images=4
	--validation_steps=100<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-7wz93i">One more thing before you launch the script! Depending on the GPU you have, you may need to enable certain optimizations to train DreamBooth.</p> <div class="flex space-x-2 items-center my-1.5 mr-8 h-7 !pl-0 -mx-3 md:mx-0"><div class="flex items-center border rounded-lg px-1.5 py-1 leading-none select-none text-smd border-gray-800 bg-black dark:bg-gray-700 text-white">16GB </div><div class="flex items-center border rounded-lg px-1.5 py-1 leading-none select-none text-smd text-gray-500 cursor-pointer opacity-90 hover:text-gray-700 dark:hover:text-gray-200 hover:shadow-sm">12GB </div><div class="flex items-center border rounded-lg px-1.5 py-1 leading-none select-none text-smd text-gray-500 cursor-pointer opacity-90 hover:text-gray-700 dark:hover:text-gray-200 hover:shadow-sm">8GB </div></div> <div class="language-select"><p data-svelte-h="svelte-fd1675">On a 16GB GPU, you can use bitsandbytes 8-bit optimizer and gradient checkpointing to help you train a DreamBooth model. Install bitsandbytes:</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START -->pip install bitsandbytes<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1rkm1uq">Then, add the following parameter to your training command:</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START -->accelerate launch train_dreambooth.py \
	--gradient_checkpointing \
	--use_8bit_adam \<!-- HTML_TAG_END --></pre></div> </div> <div class="flex space-x-2 items-center my-1.5 mr-8 h-7 !pl-0 -mx-3 md:mx-0"><div class="flex items-center border rounded-lg px-1.5 py-1 leading-none select-none text-smd border-gray-800 bg-black dark:bg-gray-700 text-white">PyTorch </div><div class="flex items-center border rounded-lg px-1.5 py-1 leading-none select-none text-smd text-gray-500 cursor-pointer opacity-90 hover:text-gray-700 dark:hover:text-gray-200 hover:shadow-sm">Flax </div></div> <div class="language-select"><div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-built_in">export</span> MODEL_NAME=<span class="hljs-string">"runwayml/stable-diffusion-v1-5"</span>
	<span class="hljs-built_in">export</span> INSTANCE_DIR=<span class="hljs-string">"./dog"</span>
	<span class="hljs-built_in">export</span> OUTPUT_DIR=<span class="hljs-string">"path_to_saved_model"</span>

	accelerate launch train_dreambooth.py \
	--pretrained_model_name_or_path=<span class="hljs-variable">$MODEL_NAME</span> \
	--instance_data_dir=<span class="hljs-variable">$INSTANCE_DIR</span> \
	--output_dir=<span class="hljs-variable">$OUTPUT_DIR</span> \
	--instance_prompt=<span class="hljs-string">"a photo of sks dog"</span> \
	--resolution=512 \
	--train_batch_size=1 \
	--gradient_accumulation_steps=1 \
	--learning_rate=5e-6 \
	--lr_scheduler=<span class="hljs-string">"constant"</span> \
	--lr_warmup_steps=0 \
	--max_train_steps=400 \
	--push_to_hub<!-- HTML_TAG_END --></pre></div> </div> <p data-svelte-h="svelte-1x6in5t">Once training is complete, you can use your newly trained model for inference!</p> <div class="course-tip bg-gradient-to-br dark:bg-gradient-to-r before:border-green-500 dark:before:border-green-800 from-green-50 dark:from-gray-900 to-white dark:to-gray-950 border border-green-50 text-green-700 dark:text-gray-400"><p data-svelte-h="svelte-fduxd">Can’t wait to try your model for inference before training is complete? 🤭 Make sure you have the latest version of 🤗 Accelerate installed.</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> DiffusionPipeline, UNet2DConditionModel
	<span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> CLIPTextModel
	<span class="hljs-keyword">import</span> torch

	unet = UNet2DConditionModel.from_pretrained(<span class="hljs-string">"path/to/model/checkpoint-100/unet"</span>)

	<span class="hljs-comment"># if you have trained with `--args.train_text_encoder` make sure to also load the text encoder</span>
	text_encoder = CLIPTextModel.from_pretrained(<span class="hljs-string">"path/to/model/checkpoint-100/checkpoint-100/text_encoder"</span>)

	pipeline = DiffusionPipeline.from_pretrained(
	<span class="hljs-string">"runwayml/stable-diffusion-v1-5"</span>, unet=unet, text_encoder=text_encoder, dtype=torch.float16,
	).to(<span class="hljs-string">"cuda"</span>)

	image = pipeline(<span class="hljs-string">"A photo of sks dog in a bucket"</span>, num_inference_steps=<span class="hljs-number">50</span>, guidance_scale=<span class="hljs-number">7.5</span>).images[<span class="hljs-number">0</span>]
	image.save(<span class="hljs-string">"dog-bucket.png"</span>)<!-- HTML_TAG_END --></pre></div></div> <div class="flex space-x-2 items-center my-1.5 mr-8 h-7 !pl-0 -mx-3 md:mx-0"><div class="flex items-center border rounded-lg px-1.5 py-1 leading-none select-none text-smd border-gray-800 bg-black dark:bg-gray-700 text-white">PyTorch </div><div class="flex items-center border rounded-lg px-1.5 py-1 leading-none select-none text-smd text-gray-500 cursor-pointer opacity-90 hover:text-gray-700 dark:hover:text-gray-200 hover:shadow-sm">Flax </div></div> <div class="language-select"><div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> DiffusionPipeline
	<span class="hljs-keyword">import</span> torch

	pipeline = DiffusionPipeline.from_pretrained(<span class="hljs-string">"path_to_saved_model"</span>, torch_dtype=torch.float16, use_safetensors=<span class="hljs-literal">True</span>).to(<span class="hljs-string">"cuda"</span>)
	image = pipeline(<span class="hljs-string">"A photo of sks dog in a bucket"</span>, num_inference_steps=<span class="hljs-number">50</span>, guidance_scale=<span class="hljs-number">7.5</span>).images[<span class="hljs-number">0</span>]
	image.save(<span class="hljs-string">"dog-bucket.png"</span>)<!-- HTML_TAG_END --></pre></div> </div> <h2 class="relative group"><a id="lora" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#lora"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>LoRA</span></h2> <p data-svelte-h="svelte-1bstso5">LoRA is a training technique for significantly reducing the number of trainable parameters. As a result, training is faster and it is easier to store the resulting weights because they are a lot smaller (~100MBs). Use the <a href="https://github.com/huggingface/diffusers/blob/main/examples/dreambooth/train_dreambooth_lora.py" rel="nofollow">train_dreambooth_lora.py</a> script to train with LoRA.</p> <p data-svelte-h="svelte-1sxi0tl">The LoRA training script is discussed in more detail in the <a href="lora">LoRA training</a> guide.</p> <h2 class="relative group"><a id="stable-diffusion-xl" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#stable-diffusion-xl"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Stable Diffusion XL</span></h2> <p data-svelte-h="svelte-1stipnz">Stable Diffusion XL (SDXL) is a powerful text-to-image model that generates high-resolution images, and it adds a second text-encoder to its architecture. Use the <a href="https://github.com/huggingface/diffusers/blob/main/examples/dreambooth/train_dreambooth_lora_sdxl.py" rel="nofollow">train_dreambooth_lora_sdxl.py</a> script to train a SDXL model with LoRA.</p> <p data-svelte-h="svelte-6dxtva">The SDXL training script is discussed in more detail in the <a href="sdxl">SDXL training</a> guide.</p> <h2 class="relative group"><a id="deepfloyd-if" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#deepfloyd-if"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>DeepFloyd IF</span></h2> <p data-svelte-h="svelte-187eih0">DeepFloyd IF is a cascading pixel diffusion model with three stages. The first stage generates a base image and the second and third stages progressively upscales the base image into a high-resolution 1024x1024 image. Use the <a href="https://github.com/huggingface/diffusers/blob/main/examples/dreambooth/train_dreambooth_lora.py" rel="nofollow">train_dreambooth_lora.py</a> or <a href="https://github.com/huggingface/diffusers/blob/main/examples/dreambooth/train_dreambooth.py" rel="nofollow">train_dreambooth.py</a> scripts to train a DeepFloyd IF model with LoRA or the full model.</p> <p data-svelte-h="svelte-3y6ob9">DeepFloyd IF uses predicted variance, but the Diffusers training scripts uses predicted error so the trained DeepFloyd IF models are switched to a fixed variance schedule. The training scripts will update the scheduler config of the fully trained model for you. However, when you load the saved LoRA weights you must also update the pipeline’s scheduler config.</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-keyword">from</span> diffusers <span class="hljs-keyword">import</span> DiffusionPipeline

	pipe = DiffusionPipeline.from_pretrained(<span class="hljs-string">"DeepFloyd/IF-I-XL-v1.0"</span>, use_safetensors=<span class="hljs-literal">True</span>)

	pipe.load_lora_weights(<span class="hljs-string">"<lora weights path>"</span>)

	<span class="hljs-comment"># Update scheduler config to fixed variance schedule</span>
	pipe.scheduler = pipe.scheduler.__class__.from_config(pipe.scheduler.config, variance_type=<span class="hljs-string">"fixed_small"</span>)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-19tur1m">The stage 2 model requires additional validation images to upscale. You can download and use a downsized version of the training images for this.</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-keyword">from</span> huggingface_hub <span class="hljs-keyword">import</span> snapshot_download

	local_dir = <span class="hljs-string">"./dog_downsized"</span>
	snapshot_download(
	<span class="hljs-string">"diffusers/dog-example-downsized"</span>,
	local_dir=local_dir,
	repo_type=<span class="hljs-string">"dataset"</span>,
	ignore_patterns=<span class="hljs-string">".gitattributes"</span>,
	)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-5uztdo">The code samples below provide a brief overview of how to train a DeepFloyd IF model with a combination of DreamBooth and LoRA. Some important parameters to note are:</p> <ul data-svelte-h="svelte-vlxh2r"><li><code>--resolution=64</code>, a much smaller resolution is required because DeepFloyd IF is a pixel diffusion model and to work on uncompressed pixels, the input images must be smaller</li> <li><code>--pre_compute_text_embeddings</code>, compute the text embeddings ahead of time to save memory because the <a href="https://huggingface.co/docs/transformers/main/en/model_doc/t5#transformers.T5Model" rel="nofollow">T5Model</a> can take up a lot of memory</li> <li><code>--tokenizer_max_length=77</code>, you can use a longer default text length with T5 as the text encoder but the default model encoding procedure uses a shorter text length</li> <li><code>--text_encoder_use_attention_mask</code>, to pass the attention mask to the text encoder</li></ul> <div class="flex space-x-2 items-center my-1.5 mr-8 h-7 !pl-0 -mx-3 md:mx-0"><div class="flex items-center border rounded-lg px-1.5 py-1 leading-none select-none text-smd border-gray-800 bg-black dark:bg-gray-700 text-white">Stage 1 LoRA DreamBooth </div><div class="flex items-center border rounded-lg px-1.5 py-1 leading-none select-none text-smd text-gray-500 cursor-pointer opacity-90 hover:text-gray-700 dark:hover:text-gray-200 hover:shadow-sm">Stage 2 LoRA DreamBooth </div><div class="flex items-center border rounded-lg px-1.5 py-1 leading-none select-none text-smd text-gray-500 cursor-pointer opacity-90 hover:text-gray-700 dark:hover:text-gray-200 hover:shadow-sm">Stage 1 DreamBooth </div><div class="flex items-center border rounded-lg px-1.5 py-1 leading-none select-none text-smd text-gray-500 cursor-pointer opacity-90 hover:text-gray-700 dark:hover:text-gray-200 hover:shadow-sm">Stage 2 DreamBooth </div></div> <div class="language-select"><p data-svelte-h="svelte-1tn6pyi">Training stage 1 of DeepFloyd IF with LoRA and DreamBooth requires ~28GB of memory.</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-built_in">export</span> MODEL_NAME=<span class="hljs-string">"DeepFloyd/IF-I-XL-v1.0"</span>
	<span class="hljs-built_in">export</span> INSTANCE_DIR=<span class="hljs-string">"dog"</span>
	<span class="hljs-built_in">export</span> OUTPUT_DIR=<span class="hljs-string">"dreambooth_dog_lora"</span>

	accelerate launch train_dreambooth_lora.py \
	--report_to wandb \
	--pretrained_model_name_or_path=<span class="hljs-variable">$MODEL_NAME</span> \
	--instance_data_dir=<span class="hljs-variable">$INSTANCE_DIR</span> \
	--output_dir=<span class="hljs-variable">$OUTPUT_DIR</span> \
	--instance_prompt=<span class="hljs-string">"a sks dog"</span> \
	--resolution=64 \
	--train_batch_size=4 \
	--gradient_accumulation_steps=1 \
	--learning_rate=5e-6 \
	--scale_lr \
	--max_train_steps=1200 \
	--validation_prompt=<span class="hljs-string">"a sks dog"</span> \
	--validation_epochs=25 \
	--checkpointing_steps=100 \
	--pre_compute_text_embeddings \
	--tokenizer_max_length=77 \
	--text_encoder_use_attention_mask<!-- HTML_TAG_END --></pre></div> </div> <h3 class="relative group"><a id="training-tips" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#training-tips"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Training tips</span></h3> <p data-svelte-h="svelte-1qpzgdo">Training the DeepFloyd IF model can be challenging, but here are some tips that we’ve found helpful:</p> <ul data-svelte-h="svelte-1f59k5d"><li>LoRA is sufficient for training the stage 1 model because the model’s low resolution makes representing finer details difficult regardless.</li> <li>For common or simple objects, you don’t necessarily need to finetune the upscaler. Make sure the prompt passed to the upscaler is adjusted to remove the new token from the instance prompt. For example, if your stage 1 prompt is “a sks dog” then your stage 2 prompt should be “a dog”.</li> <li>For finer details like faces, fully training the stage 2 upscaler is better than training the stage 2 model with LoRA. It also helps to use lower learning rates with larger batch sizes.</li> <li>Lower learning rates should be used to train the stage 2 model.</li> <li>The <a href="/docs/diffusers/main/en/api/schedulers/ddpm#diffusers.DDPMScheduler">DDPMScheduler</a> works better than the DPMSolver used in the training scripts.</li></ul> <h2 class="relative group"><a id="next-steps" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#next-steps"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Next steps</span></h2> <p data-svelte-h="svelte-1lopwzr">Congratulations on training your DreamBooth model! To learn more about how to use your new model, the following guide may be helpful:</p> <ul data-svelte-h="svelte-nczfxk"><li>Learn how to <a href="../using-diffusers/loading_adapters">load a DreamBooth</a> model for inference if you trained your model with LoRA.</li></ul> <a class="!text-gray-400 !no-underline text-sm flex items-center not-prose mt-4" href="https://github.com/huggingface/diffusers/blob/main/docs/source/en/training/dreambooth.md" target="_blank"><span data-svelte-h="svelte-1kd6by1"><</span> <span data-svelte-h="svelte-x0xyl0">></span> <span data-svelte-h="svelte-1dajgef"><span class="underline ml-1.5">Update</span> on GitHub</span></a> <p></p>

	<script>
	{
	__sveltekit_1eu7tzz = {
	assets: "/docs/diffusers/main/en",
	base: "/docs/diffusers/main/en",
	env: {}
	};

	const element = document.currentScript.parentElement;

	const data = [null,null];

	Promise.all([
	import("/docs/diffusers/main/en/_app/immutable/entry/start.21e27d66.js"),
	import("/docs/diffusers/main/en/_app/immutable/entry/app.de4fb612.js")
	]).then(([kit, app]) => {
	kit.start(app, element, {
	node_ids: [0, 187],
	data,
	form: null,
	error: null
	});
	});
	}
	</script>

Xet Storage Details

Size:: 75.9 kB
Xet hash:: f030ef72b5ba765542bf32ff3018d6370b4f72110bc53a3b0ade4549c68e00f6

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.