-
hi, hologerry~
I'm currently run your code on 4 V100 32G. I found it took about 1.3s for each iteration (batch size =128/GPU), thus the total training time for 100 epochs is about 7 days.
Does …
-
Hey there,
I recently tested out your model and it seems to me that there is a major efficiency problem here. The fact that you use a 1 hot encoded vector to represent each character is a huge waste…
-
@scott-saunders putting in a thread here rather than the one on londonmapbot discussion.
You mentioned the setup of Actions taking a while, I've noticed this a lot with the `setup-r` helper in one of…
-
First off, this framework works great and it is extremely easy to use. Thank you creating it and making it open source! Although - I am experiencing fairly considerable decrease in performance though.…
-
```
The generation of vertices and mipmaps in FFTOceanSurface needs an overhaul
as the current implementation is a little ineffecient.
At the moment a vertex array is created to store all of the ver…
-
Since it is pretty hard for new player to catch up.
What if.
More professions player knew, the less efficient he would be at each of them.
Meaning, if he only focused on one thing, for example wood…
-
Uploaded the first image. Will upload the remaining images in the next couple of minutes.
Nikon Z6 _3 Lossless compressed FX.NEF
Nikon Z6 _3 High Efficiency 1_1.NEF
Nikon Z6 _3 High Efficiency 16…
-
The function has an efficiency problem. As of now, the priorities are in order:
1. RAM utilization
2. Computation time
This issue has been opened to incorporate others that propose methods to i…
-
I believe it's worth it to make image efficiency scanning (e.g. `dive`) a step in CI to make sure we don't bloat the images later down the road.
WDYT?
-
Oneplus 3T, Android 8.0, install SSRR Android 3.5.1, and SSR Android 3.4.0.5, Two android version link servers are normal & followed by youtube link, open the same video which to Select detailed stat…