Discrepancy between architecture in paper and architecture in deepsea.beluga.2002.cpu

[AvantiShri]

Thank you for making your code and model available! I just wanted to bring it to your attention that there appears to be a discrepancy between the architecture specified in the supplement of the Expecto paper and the architecture provided in the file deepsea.beluga.2002.cpu; the last two convolutional layers appear to have 640 filters rather than 960, as illustrated below:

[jzthree]

Thanks for reporting this! You are right - this is a mistake in the paper supplement. I will contact the journal to correct this.