search

JoePenna / Dreambooth-Stable-Diffusion

Implementation of Dreambooth (https://arxiv.org/abs/2208.12242) by way of Textual Inversion (https://arxiv.org/abs/2208.01618) for Stable Diffusion (https://arxiv.org/abs/2112.10752). Tweaks focused on training faces, objects, and styles.
MIT License
3.2k stars 554 forks source link

StableDiffusionPipeline load pretrained models and learned_embeds_dict from different directories #156

Closed ghost closed 1 year ago

ghost commented 1 year ago

I want to save text_encoder, vae, unet, tokenizer and scheduler in a separate folder and learned_embeds_dict in a different directory and then load these models from their respective directory while calling StableDiffusionPipeline.from_pretrained. The reason I want to do is to train models on different styles for each user and making sure we store text_encoder, vae, unet, tokenizer and scheduler only once and use it rather than downloading again in each training session

djbielejeski commented 1 year ago

You'll have to fork and modify on your own.