NViz
Training Data Format
The training data you upload must be in the format of a JSON file with several pairs of input and output vectors stored in an array with key “data”. Outputs must be normalized (scaled 0-1).
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
{
data: [
{
input: [1.24, -0.53, 0.72],
output: [0.91, 0.65]
},
{
input: [-0.66, 0.74, -0.04],
output: [1, 0]
},
{
input: [-1.26, -0.43, 0.72],
output: [0.88, 0.36]
},
...
]
}Several sample datasets can be found on the GitHub repository.
Input Format
The input data you upload for prediction testing must be in the format of a JSON file with several input arrays (each of the size corresponding to your network's architecture) stored in an array with key “inputs”.
1
2
3
4
5
6
7
8
{
inputs: [
[1.05, 0.74, 0.24],
[0.66, 0.40, 0.89],
[0.76, 0.54, 0.19],
...
]
}Sample input files for each of the sample training datasets can also be found on the GitHub repository.