A highly-wrapped PyTorch training and testing manager
BSD-2-CLAUSE License
Bot releases are hidden (Show)
Updates:
Published by kisonho 11 months ago
API Updates:
metrics.BinaryConfusionMetric
EPOCH_START
and BATCH_START
to torchmanager_core.protocols.Frequency
data.Dataset
return_summary
parameter for fit
method to return training summary after training finishedconfigs.Configs
classmetrics.ConfusionMetrics
to abstract classtrain
packagelosses.MAE
, metrics.FeatureMetric
, metrics.ExtractorScore
, metrics.FID
, metrics.F1
, metrics.Precision
, metrics.Recall
, metrics.SSIM
, metrics.Dice
and metrics.PartialDice
backward
, forward
function to do backward propagation and forward pass in Manager
configs
packagerandom
package with freezing/unfreezing random seedstorchmanager_core.protocols.VersionConvertible
protocoldata_parallel
and reset
method in Manager
for multi-GPUs controlforward
method to calculate TP, FP, and FN and further calculate metric in forward_metric
method in metrics.BinaryConfusionMetric
Other updates:
callbacks.FrequencyCallback
Exception
handling during predictionPublished by kisonho 12 months ago
API Updates:
metrics.BinaryConfusionMetric
EPOCH_START
and BATCH_START
to torchmanager_core.protocols.Frequency
data.Dataset
configs.Configs
classmetrics.ConfusionMetrics
to abstract classtrain
packagelosses.MAE
, metrics.FeatureMetric
, metrics.ExtractorScore
, metrics.FID
, metrics.F1
, metrics.Precision
, metrics.Recall
, metrics.SSIM
, metrics.Dice
and metrics.PartialDice
backward
, forward
function to do backward propagation and forward pass in Manager
configs
packagerandom
package with freezing/unfreezing random seedstorchmanager_core.protocols.VersionConvertible
protocoldata_parallel
and reset
method in Manager
for multi-GPUs controlforward
method to calculate TP, FP, and FN and further calculate metric in forward_metric
method in metrics.BinaryConfusionMetric
Other updates:
callbacks.FrequencyCallback
Exception
handling during predictionRC Updates:
return_summary
parameter for fit
method to return training summary after training finishedPublished by kisonho about 1 year ago
Updates:
Published by kisonho about 1 year ago
API Updates:
EPOCH_START
and BATCH_START
to torchmanager_core.protocols.Frequency
data.Dataset
configs.Configs
classmetrics.ConfusionMetrics
to abstract classtrain
packagelosses.MAE
metrics.FeatureMetric
, metrics.ExtractorScore
, metrics.FID
, metrics.F1
, metrics.Precision
, and metrics.Recall
, metrics.SSIM
, metrics.Dice
and metrics.PartialDice
backward
, forward
function to do backward propagation and forward pass in Manager
configs
packagerandom
package with freezing/unfreezing random seedstorchmanager_core.protocols.VersionConvertible
protocoldata_parallel
and reset
method in Manager
for multi-GPUs controlforward
method to calculate TP, FP, and FN and further calculate metric in forward_metric
method in metrics.BinaryConfusionMetric
Other updates:
callbacks.FrequencyCallback
Exception
handling during predictionRC Updates:
Published by kisonho about 1 year ago
API Updates:
EPOCH_START
and BATCH_START
to torchmanager_core.protocols.Frequency
data.Dataset
configs.Configs
classmetrics.ConfusionMetrics
to abstract classlosses.MAE
metrics.FeatureMetric
, metrics.ExtractorScore
, metrics.FID
, metrics.F1
, metrics.Precision
, and metrics.Recall
, metrics.SSIM
.backward
, forward
function to do backward propagation and forward pass in Manager
configs
packagerandom
package with freezing/unfreezing random seedstorchmanager_core.protocols.VersionConvertible
protocoldata_parallel
and reset
method in Manager
for multi-GPUs controlforward
method to calculate TP, FP, and FN and further calculate metric in forward_metric
method in metrics.BinaryConfusionMetric
Other updates:
callbacks.FrequencyCallback
Exception
handling during predictionRC Updates:
_compile
function in BasicManager
train
packagemetrics.Dice
and metrics.PartialDice
Published by kisonho about 1 year ago
Updates:
Published by kisonho over 1 year ago
Updates:
torchmanager_core.Version
Published by kisonho over 1 year ago
Updates:
data.Dataset
can now handle unsupervised dataset for reconstruction or dataset with packed dictionaries without further implementation of unpack_data
callbacks._Checkpoint
to read-only publicsetup.py
to pyproject.toml
Published by kisonho over 1 year ago
Updates:
Version
class for better version checkingmetrics.ConfusionMetrics
to abstract classdata.utils.sliding_window
to data.sliding_window
data.utils
package and redundant codePublished by kisonho over 1 year ago
Updates:
losses.mse._ReductableLoss
for losses.MSE
sliding_window
function in data.utils
packageTypeError
when loading data.Dataset
in a data.batched
function instead of warningPublished by kisonho over 1 year ago
API updates:
torchmanager.losses.KLDiv
lossMSE
loss with NAN replaced to zeroerrors
package in torchmanager_core
Resulting
protocoltorchmanager.callbacks.Experiment
callback that combines torchmanager.callbacks.LastCheckpoint
callback, multiple torchmanager.callbacks.BestCheckpoint
callbacks, and torchmanager.callbacks.TensorBoard
callback with formatted logs, which all saved in a wrapped .exp
formatted foldertorchmanager.data
packagetorchmanager.losses.ParallelLoss
for loss functions multi-GPUs supportKLDiv
loss.exp
folder using torchmanager.train.list_checkpoints
methodtorchmanager.train.Checkpoint
by .exp
folder directory and checkpoint name using torchmanager.train.load
methodOther updates:
Published by kisonho almost 2 years ago
Updates:
_metrics
in EarlyStop
callback_update
method implementation requirement for abstract class FrequencyCallback
Published by kisonho almost 2 years ago
Updates:
results
property for the metrics to return all current results as a torch.Tensor
MeanIoU
metricFrequencyCallback
with frequency supportDynamicWeight
callback that dynamically change the weight
property in an objectpredict
method in TestingManager
to predict a dataset and returns a list
of predictionsdevice
in fit
and test
as a list of specified devicesPublished by kisonho about 2 years ago
Updates:
torchmanager_core.devices.DeviceMovable
and train.StateDictLoadable
protocolStateDictLoadable
protocolEarlyStop
callback implementation to stop the training when monitored metrics not improvement for several epochsLoss
now supports weight coefficient valuePublished by kisonho over 2 years ago
Updates:
Manager
into basic.BaseManager
, basic.DataManager
, testing.Manager
, and training.Manager
unpatch
method (private) to unpatch dataPublished by kisonho over 2 years ago
Updates:
current_epoch
index implementationManager
can be passed into Checkpoint
as well as related callbacks to be savedLrScheduler
callback now updates lr to fit the initial_epoch
indexPublished by kisonho over 2 years ago
Updates:
Published by kisonho over 2 years ago
Updates:
Published by kisonho over 2 years ago
Main Release