dlr CHANGE: Add optional log(cosh(dice_loss))

Ref https://doi.org/10.1109/cibcb48159.2020.9277638
This commit is contained in:
Starbeamrainbowlabs 2023-03-10 20:24:13 +00:00
parent f25d1b5b1a
commit c5fc62c411
Signed by: sbrl
GPG key ID: 1BE5172E637709C2
3 changed files with 14 additions and 6 deletions

View file

@ -37,6 +37,7 @@ show_help() {
echo -e " NO_REMOVE_ISOLATED_PIXELS Set to any value to avoid the engine from removing isolated pixels - that is, water pixels with no other surrounding pixels, either side to side to diagonally." >&2; echo -e " NO_REMOVE_ISOLATED_PIXELS Set to any value to avoid the engine from removing isolated pixels - that is, water pixels with no other surrounding pixels, either side to side to diagonally." >&2;
echo -e " EPOCHS The number of epochs to train for." >&2; echo -e " EPOCHS The number of epochs to train for." >&2;
echo -e " LOSS The loss function to use. Default: cross-entropy (possible values: cross-entropy, cross-entropy-dice)." >&2; echo -e " LOSS The loss function to use. Default: cross-entropy (possible values: cross-entropy, cross-entropy-dice)." >&2;
echo -e " DICE_LOG_COSH When in cross-entropy-dice mode, in addition do loss = cel + log(cosh(dice_loss)) instead of just loss = cel + dice_loss." >&2;
echo -e " PATH_CHECKPOINT The path to a checkcpoint to load. If specified, a model will be loaded instead of being trained." >&2; echo -e " PATH_CHECKPOINT The path to a checkcpoint to load. If specified, a model will be loaded instead of being trained." >&2;
echo -e " LEARNING_RATE The learning rate to use. Default: 0.001." >&2; echo -e " LEARNING_RATE The learning rate to use. Default: 0.001." >&2;
echo -e " PREDICT_COUNT The number of items from the (SCRAMBLED) dataset to make a prediction for." >&2; echo -e " PREDICT_COUNT The number of items from the (SCRAMBLED) dataset to make a prediction for." >&2;
@ -69,7 +70,7 @@ echo -e ">>> DIR_OUTPUT: ${DIR_OUTPUT}";
echo -e ">>> Additional args: ${ARGS}"; echo -e ">>> Additional args: ${ARGS}";
export PATH=$HOME/software/bin:$PATH; export PATH=$HOME/software/bin:$PATH;
export IMAGE_SIZE BATCH_SIZE DIR_RAINFALLWATER PATH_HEIGHTMAP PATH_COLOURMAP STEPS_PER_EPOCH DIR_OUTPUT PATH_CHECKPOINT EPOCHS PREDICT_COUNT NO_REMOVE_ISOLATED_PIXELS LOSS LEARNING_RATE; export IMAGE_SIZE BATCH_SIZE DIR_RAINFALLWATER PATH_HEIGHTMAP PATH_COLOURMAP STEPS_PER_EPOCH DIR_OUTPUT PATH_CHECKPOINT EPOCHS PREDICT_COUNT NO_REMOVE_ISOLATED_PIXELS LOSS LEARNING_RATE DICE_LOG_COSH;
echo ">>> Installing requirements"; echo ">>> Installing requirements";
conda run -n py38 pip install -q -r requirements.txt; conda run -n py38 pip install -q -r requirements.txt;

View file

@ -47,6 +47,7 @@ STEPS_PER_EPOCH = int(os.environ["STEPS_PER_EPOCH"]) if "STEPS_PER_EPOCH" in os.
REMOVE_ISOLATED_PIXELS = False if "NO_REMOVE_ISOLATED_PIXELS" in os.environ else True REMOVE_ISOLATED_PIXELS = False if "NO_REMOVE_ISOLATED_PIXELS" in os.environ else True
EPOCHS = int(os.environ["EPOCHS"]) if "EPOCHS" in os.environ else 50 EPOCHS = int(os.environ["EPOCHS"]) if "EPOCHS" in os.environ else 50
LOSS = os.environ["LOSS"] if "LOSS" in os.environ else "cross-entropy-dice" LOSS = os.environ["LOSS"] if "LOSS" in os.environ else "cross-entropy-dice"
DICE_LOG_COSH = True if "DICE_LOG_COSH" in os.environ else False
LEARNING_RATE = float(os.environ["LEARNING_RATE"]) if "LEARNING_RATE" in os.environ else 0.001 LEARNING_RATE = float(os.environ["LEARNING_RATE"]) if "LEARNING_RATE" in os.environ else 0.001
DIR_OUTPUT=os.environ["DIR_OUTPUT"] if "DIR_OUTPUT" in os.environ else f"output/{datetime.utcnow().date().isoformat()}_deeplabv3plus_rainfall_TEST" DIR_OUTPUT=os.environ["DIR_OUTPUT"] if "DIR_OUTPUT" in os.environ else f"output/{datetime.utcnow().date().isoformat()}_deeplabv3plus_rainfall_TEST"
@ -184,7 +185,7 @@ else:
if PATH_CHECKPOINT is None: if PATH_CHECKPOINT is None:
loss_fn = None loss_fn = None
if LOSS == "cross-entropy-dice": if LOSS == "cross-entropy-dice":
loss_fn = LossCrossEntropyDice() loss_fn = LossCrossEntropyDice(log_cosh=DICE_LOG_COSH)
elif LOSS == "cross-entropy": elif LOSS == "cross-entropy":
loss_fn = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True) loss_fn = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True)
else: else:

View file

@ -16,7 +16,6 @@ def dice_loss(y_true, y_pred):
y_pred = tf.math.sigmoid(y_pred) y_pred = tf.math.sigmoid(y_pred)
numerator = 2 * tf.reduce_sum(y_true * y_pred) numerator = 2 * tf.reduce_sum(y_true * y_pred)
denominator = tf.reduce_sum(y_true + y_pred) denominator = tf.reduce_sum(y_true + y_pred)
return 1 - numerator / denominator return 1 - numerator / denominator
class LossCrossEntropyDice(tf.keras.losses.Loss): class LossCrossEntropyDice(tf.keras.losses.Loss):
@ -24,20 +23,27 @@ class LossCrossEntropyDice(tf.keras.losses.Loss):
Combines the two with mean. Combines the two with mean.
The ground truth labels should sparse, NOT one-hot. The predictions should be one-hot, NOT sparse. The ground truth labels should sparse, NOT one-hot. The predictions should be one-hot, NOT sparse.
@source https://lars76.github.io/2018/09/27/loss-functions-for-segmentation.html#9 @source https://lars76.github.io/2018/09/27/loss-functions-for-segmentation.html#9
log_cosh (bool): Whether to do log(cosh(dice_loss)) instead of just dice_loss on its own. Ref https://doi.org/10.1109/cibcb48159.2020.9277638
""" """
def __init__(self, **kwargs): def __init__(self, log_cosh=True, **kwargs):
super(LossCrossEntropyDice, self).__init__(**kwargs) super(LossCrossEntropyDice, self).__init__(**kwargs)
self.param_log_cosh = log_cosh
def call(self, y_true, y_pred): def call(self, y_true, y_pred):
y_true = tf.cast(y_true, tf.float32) y_true = tf.cast(y_true, tf.float32)
y_true = tf.one_hot(tf.cast(y_true, dtype=tf.int32), 2) # Input is sparse y_true = tf.one_hot(tf.cast(y_true, dtype=tf.int32), 2) # Input is sparse
o = tf.nn.sigmoid_cross_entropy_with_logits(y_true, y_pred) + dice_loss(y_true, y_pred)
cel = tf.nn.sigmoid_cross_entropy_with_logits(y_true, y_pred)
dice = dice_loss(y_true, y_pred)
o = cel + dice
return tf.reduce_mean(o) return tf.reduce_mean(o)
def get_config(self): def get_config(self):
config = super(LossCrossEntropyDice, self).get_config() config = super(LossCrossEntropyDice, self).get_config()
config.update({ config.update({
"log_cosh": self.param_log_cosh
}) })
return config return config