About the correspondence between experimental results and tables of the paper

leokarlin / LaSO

LaSO: Label-Set Operations networks for multi-label few-shot learning - official implementation

BSD 3-Clause "New" or "Revised" License

86 stars 14 forks source link

About the correspondence between experimental results and tables of the paper #18

Closed hyqyoung closed 7 months ago

hyqyoung commented 2 years ago

Sorry that I am new to the few-shot learning, I can now run the three files 1. train_setops_stripped.py, 2. test_precision.py and 3. test_augmentation.py under ./scripts_coco, but may I ask how the results of these scripts correspond to the tables in the paper? Thanks!

alfassy commented 7 months ago

The results of those scripts should be very close to the tables in the paper, see the README file for updated numbers expected from these tables. Sorry for the super late answer :)