Hi,
I used TensorBoard to visualize the training logs of RTMPose. The event file is created as:
events.out.tfevents.1721752053.sheida-Vec.15612.0
I have added visualizer to my config as: #Visualizer
visualizer = dict(vis_backends=[
dict(type=‘LocalVisBackend’),
dict(type=‘TensorboardVisBackend’),
], name = ‘visualizer’)
and also I have this hooks:
default_hooks = dict(
logger=dict(type=‘LoggerHook’, interval=10),
checkpoint=dict(save_best=‘coco/AP’, rule=‘greater’, max_keep_ckpts=1),
visualization=dict(type=‘PoseVisualizationHook’)
)
log_config = dict(
interval=10,
hooks=[
dict(type=‘TextLoggerHook’),
dict(type=‘TensorboardLoggerHook’)
]
)
Still I get ’ No scalar data was found.’ when clicking on the link given by running this code:
tensorboard --logdir work_dirs/path, where the event file is stored.
I read some comments that TensorBoard has a problem with understanding directories in Windows, but my system is Ubuntu, and now I have no idea how I can resolve this issue.