| | |
- builtins.object
-
- TrainingAnimator
class TrainingAnimator(builtins.object) |
| |
TrainingAnimator(figure_size=(18, 10), dpi=100)
A utility class to create and manage training animations.
This class provides callback functions that can be used during model training.
|
| |
Methods defined here:
- __init__(self, figure_size=(18, 10), dpi=100)
- Initialize the animator with given figure size and DPI.
Args:
figure_size: (tuple) - Size of the figure (width, height)
dpi: (int) - DPI for rendering
- add_training_frame(self)
- Add a frame to the training video.
- animate_training_metrics(self, interval=200, blit=True, save_path=None, save_format='mp4', fps=10, dpi=300)
- Create an animation of the training metrics.
Args:
interval: (int) - Delay between frames in milliseconds
blit: (bool) - Whether to use blitting for efficient animation
save_path: (str - optional): Path to save the animation
save_format: (str) - Format to save animation ('mp4', 'gif', etc.)
fps: (int) - Frames per second for the saved video
dpi: (int) - DPI for the saved animation
Returns:
animation.FuncAnimation: Animation object
- finish_training_video(self, print_message=True)
- Finish and save the training video.
- initialize(self, metrics_to_track, has_validation=False)
- Initialize the animation with specified metrics.
Args:
metrics_to_track: (list) - List of metrics to track
has_validation: (bool) - Whether validation metrics are available
- setup_training_video(self, filepath, fps=10, dpi=None)
- Set up a video writer to capture training progress in real-time.
Args:
filepath: (str) - Path to save the video
fps: (int) - Frames per second
dpi: (int, optional) - DPI for rendering
- update_metrics(self, epoch_metrics, validation=False)
- Update the stored metrics with new values.
Args:
epoch_metrics (dict): Dictionary containing metric values
validation (bool): Whether these are validation metrics
Data descriptors defined here:
- __dict__
- dictionary for instance variables
- __weakref__
- list of weak references to the object
| |