Ahmet Cakir comments

Repositories
Issues
Comments

Results 5 comments of


                                            Ahmet Cakir

`relu` propagates NaN on CPU but not on GPU

Just wanted to add that the if no activation function is used, you get `NaNs` as expected. ``` gpu(BatchNorm(32))(gpu(NaN32*zeros(Float32, (32,1)))) 32×1 CuArray{Float32, 2, CUDA.Mem.DeviceBuffer}: NaN NaN NaN NaN NaN NaN...

Error: curl_multi_add_handle: 8

Yes, it was in `asyncmap`. No other parallelism was used.

Error: curl_multi_add_handle: 8

I did also ctrl-c but I cannot fully recall if I started seeing this error only after the ctrl-c unfortunately.

Error: curl_multi_add_handle: 8

The same error again and it only started appearing after Ctrl-C. And your suggestion above seems to fix it.

Add more testing for `ComposedSchedule`

I would not mind taking a crack at this, however, I will probably need some handholding as I am relatively new to the Julia world.