Experiments for Downstream OpenCL DevMap Dataset

[vrjmath]

❓ Questions and Help

I was wondering if you had the code for running experiments on the downstream dataset? I see a create.py file that creates the DevMap dataset but I don't see the dataset used anywhere.

I also had three questions about the method for performing experiments.

1. How is the train-test-validation split done for the DevMap dataset. Is it the same as the DeepDataFlow dataset here?
2. In the paper you achieve F1 scores of 0.88 and 0.80. How many trials were averaged to calculate these scores?
3. Is there a seed used for the downstream experiments?

Thank you for your help!