def test_null_target_selection():
import train
train_dataset, test_dataset = train.get_datasets()
count_null = 0
for idx, data_item in enumerate(train_dataset):
# print("graph_{}:{}".format(idx, data_item.y.item()))
if data_item.y.item() == 0:
count_null += 1
print("{:.2f}% of targets are null".format(
100 * (count_null / len(train_dataset))))
def test_train_one_epoch(self):
train_ds, test_ds = train.get_datasets()
input_rng = onp.random.RandomState(0)
model = train.create_model(random.PRNGKey(0))
optimizer = train.create_optimizer(model, 0.1, 0.9)
optimizer, train_metrics = train.train_epoch(optimizer, train_ds, 128,
0, input_rng)
self.assertLessEqual(train_metrics['loss'], 0.27)
self.assertGreaterEqual(train_metrics['accuracy'], 0.92)
loss, accuracy = train.eval_model(optimizer.target, test_ds)
self.assertLessEqual(loss, 0.06)
self.assertGreaterEqual(accuracy, 0.98)
def test_random_target_selection():
# dataset = RescueDataset("/home/okan/rescuesim/rcrs-server/dataset", "firebrigade", comp="robocup2019",
# scenario="test2", team="ait", node_classification=False, read_info_map=True)
import train
train_dataset, test_dataset = train.get_datasets()
num_random = 0
for data_item in train_dataset:
if 'type' in data_item.info_map:
if data_item.info_map['type'] == 'random':
num_random += 1
print("{:.2f}% of selection is random".format(
100 * (num_random / len(train_dataset))))
def test_single_train_step(self):
train_ds, test_ds = train.get_datasets()
batch_size = 32
model = train.create_model(random.PRNGKey(0))
optimizer = train.create_optimizer(model, 0.1, 0.9)
# test single train step.
optimizer, train_metrics = train.train_step(
optimizer=optimizer,
batch={k: v[:batch_size]
for k, v in train_ds.items()})
self.assertLessEqual(train_metrics['loss'], 2.302)
self.assertGreaterEqual(train_metrics['accuracy'], 0.0625)
# Run eval model.
loss, accuracy = train.eval_model(optimizer.target, test_ds)
self.assertLess(loss, 2.252)
self.assertGreater(accuracy, 0.2597)
import argparse
import numpy as np
from sentiment.inference import SentimentInference
from train import get_datasets
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument('--data_path', default='.cache', type=str)
args = parser.parse_args()
trainset, testset = get_datasets(args.data_path)
inferencer = SentimentInference()
predicted = inferencer.inference_batch(testset['X'])
accuracy = np.mean(predicted == testset['y'])
print(accuracy)