search

showlab / DatasetDM

[NeurIPS2023] DatasetDM:Synthesizing Data with Perception Annotations Using Diffusion Models
https://weijiawu.github.io/DatasetDM_page/
290 stars 12 forks source link

adapt the synthetic data to Mask2Fomer model #30

Closed Liuhm0710 closed 1 month ago

Liuhm0710 commented 2 months ago

Now i have succeeded in generating synthetic data and data augmentation. The result is image with mask in txt mode, but the Mask2Former needs the json mode. I wonder how you solve this problem, you change the mode of the txt file or there are some other methods to link these two models.

weijiawu commented 2 months ago

I rewrote the dataloader for Mask2Former, which can load corresponding annotations in PNG format. Of course, you can also convert the annotations to COCO JSON format.

Liuhm0710 commented 2 months ago

thanks a lot. I will try both methods.

Liuhm0710 commented 2 months ago

i've tried but havent make out yet. May i have a look at your solution? Thanks a lot.

weijiawu commented 2 months ago

链接: https://pan.baidu.com/s/13gLjfpo-YcWKkChgtLHRaQ 提取码: i89e --来自百度网盘超级会员v4的分享

I have uploaded the mask2former code I used at the time to Baidu Netdisk. I hope it will be helpful to you.

Liuhm0710 commented 2 months ago

Thank you very much. I will try this now