Test & Score

Evaluate model

void eddl::evaluate(model m, const vector<Tensor*> &in, const vector<Tensor*> &out, int bs = -1)

Returns the loss value & metrics values for the model in test mode.

Parameters

• m – Model to train

• in – Input data (features)

• out – Output data (labels)

• bs – Batch size (size [100])

Returns

(void) Evaluates the model

Example:

evaluate(mymodel, {X_test}, {Y_test});

Make predictions

vector<Tensor*> eddl::predict(model m, const vector<Tensor*> &in)

Performs a prediction with input data.

Parameters

• m – Model

• in – Input data (features)

Returns

vector of output tensors.

Example:

preds = predict(mymodel, {X_test});