type float32_tensor = (float, Bigarray.float32_elt) Tensorflow_core.Tensor.ttype t = {train_images : float32_tensor;train_labels : float32_tensor;test_images : float32_tensor;test_labels : float32_tensor;}val read_files :
?train_image_file:string ->
?train_label_file:string ->
?test_image_file:string ->
?test_label_file:string ->
unit ->
tval train_batch :
t ->
batch_size:int ->
batch_idx:int ->
float32_tensor * float32_tensorval batch_accuracy :
?samples:int ->
t ->
[ `train | `test ] ->
batch_size:int ->
predict:(float32_tensor -> float32_tensor) ->
floatbatch_accuracy ?samples t ~batch_size ~predict computes the accuracy of the predict function on test images using batches of size at most batch_size. The average is computed on samples images.