uber-research / coordconv Goto Github PK

Hi,

I'm wondering where exactly did you put a coordconv layer for a2c algorithm. Is it before the input?
The paper states 'Adding a CoordConv layer to an actor network within A2C', does this mean the critic network does not use coordconv? but actor and critic network share parameters.

Thanks

I get zero accuracies by running the instructed command:

# coordconv version
python train.py --arch coordconv_rendering -mb 16 -E 100 -L 0.005 --opt adam --l2 0.001 -mul 1 --use_mse_loss

I am using the instructed tensorflow 1.14 environment as well.
Since there are obvious runtime errors (e.g., variable not declared here and here), I suspect the codes are released without even testing if the codes runnable (not to mention the numbers in the paper).

Can the authors make some clarifications on what is the status of this repo?

Hi, in your code you normalize xx_channel and yy_channel as following:

Hi,
This is a very minor issue--nevertheless, the rest of the repository working, I really wanted to have this fixed.
This repository code at generating data is not python3 compatible.