aevilorz

I also have 2 questions in these attention calcution codes. BTW, i use pytorch 0.4. **1. Why using `Variable(torch.zeros(max_len,this_batch_size)` to store `attn_energies` ?** When i first read this line, i...

I think there could be a solution to attention calculation in matrix form in #107 .

1. I ran the config file @6lancmange provided and found that the order of the configures is the key: setting the **dawn/dust-time** following **fade** and **adjustment-method** following **location-provider** made the...

- For 2, `sudo apt-get install x11proto-xf86vidmode-dev libxext-dev` worked for me. The difference apt source might be the reason. Anyways, it's cool that you make it work! - For 4,...

Finally, i got a term **multi-head X** refered to the multi-screen case. Setting the **xorg.conf** should be the correct way. I followed the instructions but met another problem. When using...