JingyeChen

It is recommended to re-download it. Thanks!

``` After I redownload the metadata, should I put the maion_laion images into the corresponding directorys with their urls as the keys and resize them into 512*512 ? ``` Yes,...

Thansk for your attention to TextDiffuser. [x1, y1, x2, y2] denotes the coords of top-left and bottom-right points, which belong to the minimum horizontal rectangle of the 4 point annotation....

Did you specify the path of the checkpoints?

Thanks for your interest in TextDiffuser. Could you print the size of sample, feature_mask, masked_feature?

Hello, thanks for your interest in TextDiffuser-2! May I ask what is the version of the Transformer library?

https://github.com/microsoft/unilm/blob/master/textdiffuser/inference.py#L434C33-L434C33

Thanks for your interest in TextDiffuser-2. Please check args.checkpointing_steps and verify if the training steps meet the saving trigger.

Please follow the code of TextDiffuser-1 for evaluation. Thanks!

Thanks for your interest in TextDiffuser-2. Basically it is simple for adapting. You need to prepare a text image dataset in other language and add corresponding tokens to the tokenizer....