coolneng
/
locimend
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Compare commits

base: coolneng:e9582d08833807f6157a3ec658d093cf72411666
Branches Tags
coolneng:master
coolneng:0.1.1
coolneng:0.1.0
...
compare: coolneng:b7049f5b4710874aca0b72ca837b1eb75850cde5
Branches Tags
coolneng:master
coolneng:0.1.1
coolneng:0.1.0

1 Commits
e9582d0883 ... b7049f5b47

Author SHA1 Message Date
coolneng b7049f5b47
Parallelize dataset transformations 2021-06-24 19:30:46 +02:00
1 changed files with 2 additions and 1 deletions
Show Stats Expand all files Collapse all files

3
src/preprocessing.py
View File

@ -5,6 +5,7 @@ from Bio.SeqIO import parse
from numpy.random import random from numpy.random import random
from tensorflow import Tensor, int64 from tensorflow import Tensor, int64
from tensorflow.data import TFRecordDataset from tensorflow.data import TFRecordDataset
from tensorflow.data import AUTOTUNE, TFRecordDataset
from tensorflow.io import TFRecordWriter, VarLenFeature, parse_single_example from tensorflow.io import TFRecordWriter, VarLenFeature, parse_single_example
from tensorflow.sparse import to_dense from tensorflow.sparse import to_dense
from tensorflow.train import Example, Feature, Features, Int64List from tensorflow.train import Example, Feature, Features, Int64List
@ -104,8 +105,8 @@ def read_dataset(filepath) -> TFRecordDataset:
Read TFRecords files and generate a dataset Read TFRecords files and generate a dataset
""" """
data_input = TFRecordDataset(filenames=filepath) data_input = TFRecordDataset(filenames=filepath)
dataset = data_input.map(map_func=process_input)
shuffled_dataset = dataset.shuffle(buffer_size=10000, seed=42) shuffled_dataset = dataset.shuffle(buffer_size=10000, seed=42)
dataset = data_input.map(map_func=process_input, num_parallel_calls=AUTOTUNE)
batched_dataset = shuffled_dataset.batch(batch_size=BATCH_SIZE).repeat(count=EPOCHS) batched_dataset = shuffled_dataset.batch(batch_size=BATCH_SIZE).repeat(count=EPOCHS)
return batched_dataset return batched_dataset