search

junjiehe96 / FastInst

[CVPR2023] FastInst: A Simple Query-Based Model for Real-Time Instance Segmentation
MIT License
175 stars 16 forks source link

How to test models with "coco_2017_test-dev" #19

Closed xlim1996 closed 10 months ago

xlim1996 commented 10 months ago

I download the "image_info_test-dev2017.json" in COCO. And use the "coco_2017_test-dev" as the TEST DATASET. When I run the train_net.py to evaluate the model, I can't get the accuracy and the result shows "Annotations are not available for evaluation."

If you can help me to solve the problem,I really appreciate it.

junjiehe96 commented 10 months ago
  1. set DATASET.TEST as "coco_2017_test-dev" and run train_net.py --eval-only to get results.
  2. upload results to evaluation server. For details you can refer to this.
xlim1996 commented 10 months ago

thanks a lot

xlim1996 commented 10 months ago

I'm sorry to bother you again. I got the coco_instances_results.json from the model, which is 1.2GB. I can only compress it to 380MB with zip. And when I submitted it to https://codalab.lisn.upsaclay.fr/competitions/7383. Submissions are constantly failing with error elapsed time.

If you can help me to solve the problem, I will appreciate it.

error detail: elapsedTime: 174.19653415679932 exitCode: 137 ingestionExitCode: 0

junjiehe96 commented 10 months ago

@xlim1996 Hi, the size is right. I tried some times in such case and it finally processed successfully.

xlim1996 commented 10 months ago

Thanks a lot. I will try to figure it out.