Getting started using Tensorflow JS models and training. Prediction of solar power production - tensorflow.js

Question

I want to ask a probably very noob question here about my first steps with TensorFlow.js. I want to create a model that can predict the daily solar energy production based on the clouds percentage and temperature during the hours of a day.
I have three arrays for this question:
const production = [["0.00","0.00","0.00","0.00","0.00","0.03","0.20","0.42","0.85","1.51","1.58","1.46","1.68","1.68","0.51","0.24","0.14","0.05","0.00","0.00","0.00","0.00","0.00","0.00"],["0.00","0.00","0.00","0.00","0.00","0.01","0.12","0.29","0.81","1.42","1.62","2.09","2.26","1.77","0.44","0.20","0.11","0.04","0.00","0.00","0.00","0.00","0.00","0.00"]]
const clouds = [["0.90","0.90","0.90","0.90","0.80","0.75","0.75","0.75","0.72","0.27","0.34","0.58","0.35","0.20","0.20","0.20","0.20","0.17","0.20","0.20","0.20","0.20","0.20","0.01"],["0.74","0.20","0.22","0.39","0.79","0.75","0.75","0.75","0.50","0.40","0.40","0.40","0.40","0.40","0.26","0.30","0.20","0.20","0.20","0.33","0.77","0.90","0.90","0.90"]]
const temp = [["15.50","15.22","14.65","14.35","13.84","14.46","15.97","17.08","18.30","19.51","20.39","21.00","21.60","21.94","21.94","21.63","20.89","19.32","17.58","16.40","15.63","14.86","14.16","13.64"],["14.98","14.97","14.89","14.51","14.47","15.14","15.83","16.69","17.89","19.10","20.06","20.84","21.46","21.91","21.58","20.99","19.75","18.06","16.75","15.74","15.49","15.26","15.29","15.45"]]
All these arrays are actual measurements for 2 days.
Production is in kWh (for who's interested) ;-)
All these arrays exist of subarrays of 24 children (one for every hour, starting at 01:00). Because the hour of the day is very important for production of course (day or night).
I want to use Tensorflow.js to predict the production for the day when I ask a weather API for the clouds percentage and temperatures for every hour of a day. Throwing the results into my model as [clouds] [temp] and receiving a new array from the model [production]. Each array consists of 24 numbers representing the hours of the day. I think it is sequential, because normally following equation should work: [production] = a + b*[clouds] + c*[temp]
I experimented a lot using the Boston housing example but I don't seem to be able to fit my parameters to this model. If anyone could fit my measurements in the example, that would be really cool.
For the moment I'm working with this class:
import * as tf from '#tensorflow/tfjs';
/**
* Linear model class
*/
export default class LinearModel {
/**
* Train model
*/
async trainModel(temp, clouds, production){
const layers = tf.layers.dense({
units: 24, // Dimensionality of the output space
inputShape: [24], // Only one param
});
const lossAndOptimizer = {
loss: 'meanSquaredError',
optimizer: 'sgd', // Stochastic gradient descent
};
this.linearModel = tf.sequential();
this.linearModel.add(layers); // Add the layer
this.linearModel.compile(lossAndOptimizer);
// Start the model training!
await this.linearModel.fit(
[tf.tensor2d(temp), tf.tensor2d(clouds)],
tf.tensor2d(production),
);
}
}
And the call is like this:
const model = new LinearModel()
await model.trainModel(temp, clouds, production)
As you can probably notice, the code fails for now. And it has something to do with the 2d-3d dimensions of my values. But I don't really find a clue.
The error logs:
ValueError: Error when checking model input: the Array of Tensors that you are passing to your model is not the size the model expected. Expected to see 1 Tensor(s), but instead got the following list of Tensor(s): Tensor
[['13.14', '12.70', '12.09', ..., '16.22', '16.11', '15.96'],
['15.82', '15.84', '15.73', ..., '15.34', '14.31', '13.30'],
['12.83', '12.36', '11.86', ..., '14.29', '13.87', '13.66'],
['13.62', '13.67', '13.68', ..., '14.89', '14.42', '14.05'],
['13.95', '14.00', '14.14', ..., '16.27', '16.24', '16.20'],
['16.23', '16.14', '16.00', ..., '15.39', '15.10', '14.82'],
['14.98', '14.97', '14.89', ..., '15.26', '15.29', '15.45'],
['15.41', '15.35', '15.23', ..., '15.91', '15.65', '15.69'],
['15.50', '15.22', '14.65', ..., '14.86', '14.16', '13.64'],
['13.23', '12.88', '12.55', ..., '16.76', '16.10', '15.50'],
['14.83', '14.22', '13.59', ..., '15.49', '14.84', '14.20'],
['13.70', '13.23', '12.89', ..., '19.00', '19.00', '19.00'],
['19.00', '19.00', '19.00', ..., '16.09', '15.43', '14.77'],
['14.55', '14.09', '19.00', ..., '18.50', '17.10', '16.29'],
['15.78', '15.13', '14.75', ..., '19.00', '17.05', '16.83'],
['16.55', '16.06', '15.70', ..., '19.00', '18.49', '18.50'],
['18.22', '18.27', '18.08', ..., '16.93', '16.27', '16.00'],
['15.98', '16.12', '16.19', ..., '19.00', '19.00', '19.00'],
['19.00', '19.00', '19.00', ..., '19.00', '19.00', '19.00'],
['19.00', '19.00', '19.00', ..., '13.16', '18.64', '17.90']],Tensor
[['0.22', '0.28', '0.17', ..., '0.79', '0.90', '0.80'],
['0.86', '0.87', '0.75', ..., '0.27', '0.20', '0.20'],
['0.20', '0.20', '0.20', ..., '0.20', '0.20', '0.20'],
['0.20', '0.20', '0.20', ..., '0.66', '0.20', '0.13'],
['0.78', '0.29', '0.46', ..., '0.90', '0.82', '0.90'],
['0.90', '0.84', '0.90', ..., '0.25', '0.53', '0.90'],
['0.74', '0.20', '0.22', ..., '0.90', '0.90', '0.90'],
['0.90', '0.90', '0.90', ..., '0.44', '0.68', '0.90'],
['0.90', '0.90', '0.90', ..., '0.20', '0.20', '0.01'],
['0.06', '0.06', '0.06', ..., '0.20', '0.20', '0.20'],
['0.20', '0.20', '0.25', ..., '0.03', '0.02', '0.04'],
['0.10', '0.10', '0.09', ..., '0.20', '0.20', '0.20'],
['0.20', '0.20', '0.20', ..., '0.20', '0.55', '0.59'],
['0.85', '0.69', '0.20', ..., '0.20', '0.20', '0.20'],
['0.20', '0.19', '0.27', ..., '0.20', '0.20', '0.20'],
['0.20', '0.20', '0.20', ..., '0.20', '0.97', '0.96'],
['0.20', '0.20', '0.20', ..., '0.20', '0.20', '0.90'],
['0.90', '0.90', '0.90', ..., '0.20', '0.20', '0.20'],
['0.20', '0.20', '0.20', ..., '0.20', '0.20', '0.20'],
['0.20', '0.20', '0.20', ..., '0.20', '0.54', '0.20']]
at new ValueError (/Users/hacor/regression-test/node_modules/#tensorflow/tfjs-layers/dist/tf-layers.node.js:9827:28)
at standardizeInputData (/Users/hacor/regression-test/node_modules/#tensorflow/tfjs-layers/dist/tf-layers.node.js:18254:19)
at LayersModel.standardizeUserDataXY (/Users/hacor/regression-test/node_modules/#tensorflow/tfjs-layers/dist/tf-layers.node.js:19125:13)
at LayersModel.<anonymous> (/Users/hacor/regression-test/node_modules/#tensorflow/tfjs-layers/dist/tf-layers.node.js:19147:35)
at step (/Users/hacor/regression-test/node_modules/#tensorflow/tfjs-layers/dist/tf-layers.node.js:9745:23)
at Object.next (/Users/hacor/regression-test/node_modules/#tensorflow/tfjs-layers/dist/tf-layers.node.js:9726:53)
at /Users/hacor/regression-test/node_modules/#tensorflow/tfjs-layers/dist/tf-layers.node.js:9719:71
at new Promise (<anonymous>)
at __awaiter (/Users/hacor/regression-test/node_modules/#tensorflow/tfjs-layers/dist/tf-layers.node.js:9715:12)
at LayersModel.standardizeUserData (/Users/hacor/regression-test/node_modules/#tensorflow/tfjs-layers/dist/tf-layers.node.js:19142:16)
Any help would be appreciated!
Best regards,
Hacor

Answer 1

input is always a single tensor and you're sending two - you need to combine temp, clouds into a single tensor
there are number of ways to do that, but most likely you want to have each as a separate dimension, so just stack them:
const inputs = tf.stack([temp, clouds)]);
trainModel(inputs, production)
...
btw, your model has no layers that do anything except a single loss function - so it's not going to do much
take a look at https://github.com/vladmandic/stocks/blob/main/src/model.js to see what im referring to.