Reproducibility on TON dataset experiments
Chasexj opened this issue · 1 comments
Chasexj commented
Can you people release the configuration and driver scripts for evaluating the model accuracy when using the full TON dataset?
This should include both training/testing on the real dataset as well as training on synthetic and testing on the real dataset.
sniperyyc commented
Hi, this is one of our downstream tasks and if time permits, we will try to include it after finishing the current refactoring of the PyTorch version model.