# Copyright (c) MONAI Consortium
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
# http://www.apache.org/licenses/LICENSE-2.0
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
import datetime
import filecmp
import json
import logging
import os
import shutil
import time
from typing import Any, Dict
import torch
from monai.engines.workflow import Engine, Events
logger = logging.getLogger(__name__)
[docs]def prepare_stats(start_ts, trainer, evaluator):
def tensor_to_list(d):
r = dict()
for dk, dv in d.items():
r[dk] = dv.tolist() if torch.is_tensor(dv) else dv
return r
stats: Dict[str, Any] = dict()
stats.update(trainer.get_stats())
stats["epoch"] = trainer.state.epoch
stats["start_ts"] = int(start_ts)
if trainer.state.epoch == trainer.state.max_epochs:
stats["total_time"] = str(datetime.timedelta(seconds=int(time.time() - start_ts)))
else:
stats["current_time"] = str(datetime.timedelta(seconds=int(time.time() - start_ts)))
for k, v in {"train": trainer, "eval": evaluator}.items():
if not v:
continue
stats["best_metric"] = v.state.best_metric
stats[k] = {
"metrics": tensor_to_list(v.state.metrics),
"key_metric_name": v.state.key_metric_name,
"best_metric": v.state.best_metric,
"best_metric_epoch": v.state.best_metric_epoch,
}
return stats
[docs]class PublishStatsAndModel:
def __init__(self, stats_path, publish_path, key_metric_filename, start_ts, run_id, output_dir, trainer, evaluator):
self._stats_path = stats_path
self._publish_path = publish_path
self._key_metric_filename = key_metric_filename
self.start_ts = start_ts
self.run_id = run_id
self.output_dir = output_dir
self.trainer = trainer
self.evaluator = evaluator
[docs] def iteration_completed(self):
filename = datetime.datetime.now().strftime(f"stats_{self.run_id}.json")
filename = os.path.join(self.output_dir, filename)
stats = prepare_stats(self.start_ts, self.trainer, self.evaluator)
with open(filename, "w") as f:
json.dump(stats, f, indent=2)
if self._stats_path:
shutil.copy(filename, self._stats_path)
publish_path = self._publish_path
if publish_path:
final_model = os.path.join(self.output_dir, self._key_metric_filename)
if os.path.exists(final_model):
if not os.path.exists(publish_path) or not filecmp.cmp(publish_path, final_model):
shutil.copy(final_model, publish_path)
logger.info(f"New Model published: {final_model} => {publish_path}")
return stats
[docs] def attach(self, engine: Engine) -> None:
if not engine.has_event_handler(self.iteration_completed, Events.EPOCH_COMPLETED):
engine.add_event_handler(Events.EPOCH_COMPLETED, self.iteration_completed)
def __call__(self, engine: Engine) -> None:
self.iteration_completed()