The model is based on v1.2 with further training. V1.5 is released in Oct 2022 by Runway ML, a partner of Stability AI. Below is a list of models that can be used for general purposes. There are thousands of fine-tuned Stable Diffusion models. I will cover the v1 models in this section and the v2 models in the next section. There are two groups of models: v1 and v2. In layman’s terms, it’s like using existing words to describe a new concept.
Only the text embedding network is fine-tuned while keeping the rest of the model unchanged. A new keyword is created specifically for the new object. The goal is similar to Dreambooth: Inject a custom subject into the model with only a few examples. There’s another less popular fine-tuning technique called textual inversion (sometimes called embedding). A model trained with Dreambooth requires a special keyword to condition the model. You can take a few pictures of yourself and use Dreambooth to put yourself into the model. It works with as few as 3-5 custom images. For example, you can train Stable Diffusion v1.5 with an additional dataset of vintage cars to bias the aesthetic of cars towards the sub-genre.ĭreambooth, initially developed by Google, is a technique to inject custom subjects into text-to-image models. They both start with a base model like Stable Diffusion v1.4 or v1.5.Īdditional training is achieved by training a base model with an additional dataset you are interested in. Two main fine-tuning methods are (1) Additional training and (2) Dreambooth. Instead of tinkering with the prompt, you can fine-tune the model with images of that sub-genre. But it could be difficult to generate images of a sub-genre of anime. For example, it can and will generate anime-style images with the keyword “anime” in the prompt. Stable diffusion is great but is not good at everything. It takes a model that is trained on a wide dataset and trains a bit more on a narrow dataset.Ī fine-tuned model will be biased toward generating images similar to your dataset while maintaining the versatility of the original model. Fine-tuning is a common technique in machine learning.
0 kommentar(er)
