HongChuang/ultralytics-ascend
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 3

Set workers=0 for MPS Train and Val modes (#4697)

Browse source
This commit is contained in:
Glenn Jocher 2023-09-02 04:25:03 +02:00 • committed by GitHub
parent 2bc6e647c7
commit a4fabfdacf
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
14 changed files with 88 additions and 74 deletions
Download patch file Download diff file Expand all files Collapse all files

27
ultralytics/utils/callbacks/comet.py
View file

@ -1,10 +1,6 @@
# Ultralytics YOLO 🚀, AGPL-3.0 license
import os
from pathlib import Path
from ultralytics.utils import LOGGER, RANK, SETTINGS, TESTS_RUNNING, ops
from ultralytics.utils.torch_utils import model_info_for_loggers
try:
assert not TESTS_RUNNING # do not log pytest
@ -12,18 +8,22 @@ try:
import comet_ml
assert hasattr(comet_ml, '__version__') # verify package is not directory
import os
from pathlib import Path
# Ensures certain logging functions only run for supported tasks
COMET_SUPPORTED_TASKS = ['detect']
# Names of plots created by YOLOv8 that are logged to Comet
EVALUATION_PLOT_NAMES = 'F1_curve', 'P_curve', 'R_curve', 'PR_curve', 'confusion_matrix'
LABEL_PLOT_NAMES = 'labels', 'labels_correlogram'
_comet_image_prediction_count = 0
except (ImportError, AssertionError):
comet_ml = None
# Ensures certain logging functions only run for supported tasks
COMET_SUPPORTED_TASKS = ['detect']
# Names of plots created by YOLOv8 that are logged to Comet
EVALUATION_PLOT_NAMES = 'F1_curve', 'P_curve', 'R_curve', 'PR_curve', 'confusion_matrix'
LABEL_PLOT_NAMES = 'labels', 'labels_correlogram'
_comet_image_prediction_count = 0
def _get_comet_mode():
return os.getenv('COMET_MODE', 'online')
@ -327,6 +327,7 @@ def on_fit_epoch_end(trainer):
experiment.log_metrics(trainer.metrics, step=curr_step, epoch=curr_epoch)
experiment.log_metrics(trainer.lr, step=curr_step, epoch=curr_epoch)
if curr_epoch == 1:
from ultralytics.utils.torch_utils import model_info_for_loggers
experiment.log_metrics(model_info_for_loggers(trainer), step=curr_step, epoch=curr_epoch)
if not save_assets: