input shape (ML)
On the first layer of a network, you need to define the inputShape. The inputShape is [1] when there is 1 number as the input.
Python Example
For this data:
| X: | -1 | 0 | 1 | 2 | 3 | 4 |
|---|---|---|---|---|---|---|
| Y: | -2 | 1 | 4 | 7 | 10 | 13 |
The input shape is [1]. I don’t really understand why.
JavaScript Example
// this might be incorrect…
const inputs = [ [1, 2], [3, 4], [5, 6], [7, 8]];
const inputTensor = tf.tensor2d(inputs, [4, 2]);
But shouldn’t the input shape of the Python example then be [6, 1]?