def on_epoch_end(self, epoch, logs):
if self.every_epoch:
met_dict = {'epoch': epoch}
for key, value in logs.items():
logs[key] = round(value, 4)
met_dict.update(logs)
log_metrics(met_dict, verbose=False)
elif epoch == self.params['epochs'] - 1:
met_dict = {}
for key, value in logs.items():
logs[key] = round(value, 4)
met_dict.update(logs)
log_metrics(met_dict, verbose=False)
if self.notify_slack:
result = {
'message':
'Training complete after ' + str(self.params['epochs']) +
' epochs.',
'metrics':
met_dict
}
if self.hyperparams:
result['hyperparams'] = self.hyperparams
notify(json.dumps(result, indent=2, cls=Encoder),
verbose=False,
safe=True)
def test_log_metrics_verbose(mock_api_post_block, capsys):
data = {'acc': 0.89, 'val_acc': 0.86}
log_metrics(data)
captured = capsys.readouterr()
expected_result = '[jovian] Metrics logged.'
assert expected_result == captured.out.strip()
def test_log_metrics(self, mock_api_post_block):
data = {'acc': 0.89, 'val_acc': 0.86}
expected_result = [('fake_slug_metrics_1', 'metrics', {}),
('fake_slug_metrics_2', 'metrics', {}),
('fake_slug_hyperparams_1', 'hyperparams', {}),
('fake_slug_hyperparams_2', 'hyperparams', {}),
('fake_slug_3', 'metrics', data)]
log_metrics(acc=0.89, val_acc=0.86)
self.assertEqual(jovian.utils.records._data_blocks, expected_result)
def on_epoch_end(self, epoch: int, smooth_loss: Tensor, last_metrics: list, **ka):
met_values = [epoch,
smooth_loss.item()] # smoothened avg. train loss for the epoch
if self.valid_set:
# last_metrics is a list with first elem as valid_loss followed by all
# the metrics of the learner
met_values.extend([str(last_metrics[0])] + [i.item()
for i in last_metrics[1:]])
log_metrics(dict(zip(self.met_names, met_values)))