-
Hi guys,
I'm working on the segmentation of road network in aerial RGB images (TIFF, 1024 x 1024 px). I intend to use a CNN (implemented using DL4J) with supervised learning to detect which pixels …
-
Spark Cluster Size: 3-node with each node having 32GB RAM and 8-core CPU
Total Spark Core: 20
Data Size: 45 million rows with size 9GB
Data is read and split pretty fast (within 1 minute)
After th…
-
We are considering creating another base representation of hetnets. One of the main goals is to facilitate faster network loading, which at present can take over a minute and a half to load a graph.
…
-
I just train an LSTM, save and load that and print again the evaluation on the same test set.
However the results are different
```
import com.beust.jcommander.JCommander;
import com.beust.jcomm…
-
I want to draw a Multi layer network by using Python. The expected graph likes this:
![oky9v](https://user-images.githubusercontent.com/13978535/37772812-c0117274-2e16-11e8-8aeb-f1d652939e6e.png)
T…
-
#### I have a rather small example (like a toy example) which had a really unexpected memory allocation problem. After a few epochs the memory required was too much to handle.
The code is a modifie…
-
As mentioned on the options of `Infomap` program, we have the following useful, but unexplored options of Infomap algorithm.
* two level / multi-level clustering
* clustering of multilayer (multip…
-
It would be great if the snap library could support multilayer networks. Is there any plans in this direction?
For example, other libraries that support multilayer networks include [graph-tool](htt…
-
Il y a une anomalie avec N2D2 car j'obtiens des résultats différents avec deux versions de N2D2.
Version actuelle "21/03/2018": accuracy = 77.92 %
Version antérieure "16/03/2017": acc…
-
I don't know the correct answer to this question, or whether Neural networks are best left to other courses. Given that deep networks are a particularly hot field in Machine Learning right now, I am …