Textural Inversion Support for Stable Diffusion 3

[circle345]

I am using webui under the sd3_textural_inversion branch and got an error when training embeddings with textural inversion on stable diffusion 3. I would like to know if this version supports textural inversion training and if the latest version of the code supports modification in batch size.

The resolution of the dataset images I used for training is 1024 x 1024. After starting training, cmd shows the following error:
RuntimeError: Sizes of tensors must match except in dimension 0. Expected size 768 but got size 2048 for tensor number 1 in the list.

[hyao1]

have you achiveved the support？