def new_log_interval_event(self, run_states):
"""
PaddleHub default handler for log_interval_event, it will complete visualization.
Args:
run_states (object): the results in train phase
"""
scores, avg_loss, run_speed = self._calculate_metrics(run_states)
self.tb_writer.add_scalar(tag="Loss_{}".format(self.phase),
scalar_value=avg_loss,
global_step=self._envs['train'].current_step)
log_scores = ""
s = [self.current_step]
for metric in scores:
self.tb_writer.add_scalar(
tag="{}_{}".format(metric, self.phase),
scalar_value=scores[metric],
global_step=self._envs['train'].current_step)
log_scores += "%s=%.5f " % (metric, scores[metric])
s.append(scores[metric])
# train[metric].add_record(self.current_step, scores[metric])
logger.train("step %d / %d: loss=%.5f %s[step/sec: %.2f]" %
(self.current_step, self.max_train_steps, avg_loss,
log_scores, run_speed))
s.append(avg_loss)
# train_loss.add_record(self.current_step, avg_loss)
s = [str(x) for x in s]
with open('./work/log/%s_train%s.txt' % (args.do_model, id),
'a',
encoding='utf-8') as f:
f.write(','.join(s) + '\n')
def _default_log_interval_event(self, run_states):
scores, avg_loss, run_speed = self._calculate_metrics(run_states)
self.tb_writer.add_scalar(tag="Loss_{}".format(self.phase),
scalar_value=avg_loss,
global_step=self._envs['train'].current_step)
log_scores = ""
for metric in scores:
self.tb_writer.add_scalar(
tag="{}_{}".format(metric, self.phase),
scalar_value=scores[metric],
global_step=self._envs['train'].current_step)
log_scores += "%s=%.5f " % (metric, scores[metric])
logger.train("step %d / %d: loss=%.5f %s[step/sec: %.2f]" %
(self.current_step, self.max_train_steps, avg_loss,
log_scores, run_speed))
def _default_log_interval_event(self, run_states):
"""
PaddleHub default handler for log_interval_event, it will complete visualization.
Args:
run_states (object): the results in train phase
"""
scores, avg_loss, run_speed = self._calculate_metrics(run_states)
self.vdl_writer.add_scalar(tag="Loss_{}".format(self.phase),
value=avg_loss,
step=self._envs['train'].current_step)
log_scores = ""
for metric in scores:
self.vdl_writer.add_scalar(tag="{}_{}".format(metric, self.phase),
value=scores[metric],
step=self._envs['train'].current_step)
log_scores += "%s=%.5f " % (metric, scores[metric])
logger.train("step %d / %d: loss=%.5f %s[step/sec: %.2f]" %
(self.current_step, self.max_train_steps, avg_loss,
log_scores, run_speed))
def new_log_interval_event(self, run_states):
scores, avg_loss, run_speed = self._calculate_metrics(run_states)
self.tb_writer.add_scalar(tag="Loss_{}".format(self.phase),
scalar_value=avg_loss,
global_step=self._envs['train'].current_step)
log_scores = ""
log = [self._envs['train'].current_step, avg_loss]
for metric in scores:
self.tb_writer.add_scalar(
tag="{}_{}".format(metric, self.phase),
scalar_value=scores[metric],
global_step=self._envs['train'].current_step)
log_scores += "%s=%.5f " % (metric, scores[metric])
log.append(scores[metric])
logger.train("step %d / %d: loss=%.5f %s[step/sec: %.2f]" %
(self.current_step, self.max_train_steps, avg_loss,
log_scores, run_speed))
log = [str(x) for x in log]
with open('./work/log/cls_log_{}train.txt'.format(train_i),
'a',
encoding='utf-8') as f:
f.write(','.join(log) + '\n')
