def log_run(split: str, epoch: int, writer: tf.summary.SummaryWriter,
label_names: Sequence[str], metrics: MutableMapping[str, float],
heaps: Mapping[str,
Mapping[int,
List[HeapItem]]], cm: np.ndarray) -> None:
"""Logs the outputs (metrics, confusion matrix, tp/fp/fn images) from a
single epoch run to Tensorboard.
Args:
metrics: dict, keys already prefixed with {split}/
"""
per_class_recall = recall_from_confusion_matrix(cm, label_names)
metrics.update(prefix_all_keys(per_class_recall, f'{split}/label_recall/'))
# log metrics
for metric, value in metrics.items():
tf.summary.scalar(metric, value, epoch)
# log confusion matrix
cm_fig = plot_utils.plot_confusion_matrix(cm,
classes=label_names,
normalize=True)
cm_fig_img = tf.convert_to_tensor(fig_to_img(cm_fig)[np.newaxis, ...])
tf.summary.image(f'confusion_matrix/{split}', cm_fig_img, step=epoch)
# log tp/fp/fn images
for heap_type, heap_dict in heaps.items():
log_images_with_confidence(heap_dict,
label_names,
epoch=epoch,
tag=f'{split}/{heap_type}')
writer.flush()
def create_summary(writer: tf.summary.SummaryWriter,
optimizer_name: str,
nb_img_utilisees,
optimizer_parameters: Dict,
loss: str,
metriques_utilisees: List[str],
but_essai: str,
informations_additionnelles: str,
id: str,
dataset_name: str = "",
taille_x_img: int = 1600,
taille_y_img: int = 900,
taille_x_img_redim: int = 400,
taille_y_img_redim: int = 225,
batch_size=10,
nb_img_tot=173959,
nb_epochs=1):
markdown = f"""# Résumé de l'entrainement du {id}
Entrainement sur {dataset_name} ({nb_img_tot} images ; {min(nb_img_utilisees,nb_img_tot)} utilisées) avec des images de taille {taille_x_img} px par {taille_y_img} px redimensionnées à {taille_x_img_redim} px x {taille_y_img_redim} px
Batch size de {batch_size}
## Paramètres d'entrainement
Entrainement sur {nb_epochs} epochs
Optimisateur {optimizer_name} avec les paramètres :\n"""
for k, v in optimizer_parameters.items():
markdown += f"{k} : {v}"
markdown += f"""\nLoss : {loss}
Métriques : """
markdown += ", ".join([f"{metrique}" for metrique in metriques_utilisees])
markdown += f"""\n## Description de l'essai\n\n{but_essai}\n\n{informations_additionnelles}"""
with writer.as_default():
tf.summary.text("Resume", markdown, step=0)
writer.flush()
def save_dict_to_tensorboard(event_writer: tf.summary.SummaryWriter,
dict: Dict, step: int):
for key, val in dict.items():
with event_writer.as_default():
tf.summary.scalar(name=key, data=val, step=step)
event_writer.flush()
def write_scalars(writer: tf.summary.SummaryWriter, scalar_dict, step):
with writer.as_default():
for (k, v) in scalar_dict.items():
tf.summary.scalar(k, v, step=step)
writer.flush()