March 2019
Intermediate to advanced
196 pages
4h 50m
English
With the following arguments, our dataset will consist of two items taken from each row of the filename file, both of the float type, with the first line of the file ignored and columns 1 and 2 used (column numbering is, of course, 0-based):
filename = ["./size_1000.csv"]record_defaults = [tf.float32] * 2 # two required float columnsdataset = tf.data.experimental.CsvDataset(filename, record_defaults, header=True, select_cols=[1,2])for item in dataset: print(item)
Read now
Unlock full access