I notice in ingestion.py, it takes an argument of "hidden_dir", which defaults to "AutoML3_sample_ref";
What is this hidden_dir for? And how can we find these "ref" data for "AutoML3_input_data_1"?
Thanks.Posted by: xiayunsun @ Aug. 8, 2018, 7:18 p.m.
This is a directory "hidden" for participants so, you should not have access in the CodaLab platform., and you should not attempt to have access to (doing so is a form of cheating)
Although you can have access to it in the starting kit (AutoML3_sample_Ref)
HJPosted by: hugo.jair @ Aug. 8, 2018, 7:28 p.m.
Thanks for reply.
But without it how can I evaluate my performance on AutoML3_input_data_1 locally? Or I can keep using AutoML3_sample_ref?Posted by: xiayunsun @ Aug. 8, 2018, 7:32 p.m.
Edit: I don't think I can use "AutoML3_sample_ref" together with "AutoML3_input_data_1" though; because their directory layout have different data names (one with AA/B/C/D/E, one with ada/rl).Posted by: xiayunsun @ Aug. 8, 2018, 7:33 p.m.