torchmanager - v1.4.1

Updates: - Minor bugs fixed - PyTorch compatibility fixed

- Python
Published by kisonho 11 months ago

torchmanager - v1.4

API updates: - Add dynamic parallel type for losses.ParallelLoss - Add mapping model support for checkpoint - Add specific data loader type to data.batched wrapper function - Deprecated Python 3.9 - Introducing abstract base configurations configs.BaseConfigs, base managers including BaseTestingManager (abstract test_step) and BaseTrainingManager (abstract train_step), and abstract base losses and metrics (losses.BaseLoss and metrics.BaseMetric) - Introducing metrics.AccumulativeFeatureMetric as an abstract metrics for accumulative support metrics - Introducing metrics.CosineSimilarity, metrics.KID, and metrics.MS_SSIM - Introducing CkptConvertable protocol - Introduce optimize function to optimize the model - Metrics now accepts non-dictionary input and target when target is given - Moved zero gradients to optimize function in training manager - The metrics.FID now inherits metrics.AccumulativeFeatureMetric for accumulative support

Other updates: - Logger improved - Minor bugs fixed - Performance improved - Typing improvement

- Python
Published by kisonho about 1 year ago

torchmanager - v1.4 (Release Candidate 2)

API updates: - Add dynamic parallel type for losses.ParallelLoss - Add mapping model support for checkpoint - Add specific data loader type to data.batched wrapper function - Deprecated Python 3.9 - Introducing abstract base configurations configs.BaseConfigs, base managers including BaseTestingManager (abstract test_step) and BaseTrainingManager (abstract train_step), and abstract base losses and metrics (losses.BaseLoss and metrics.BaseMetric) - Introducing metrics.CosineSimilarity and metrics.KID - Introducing CkptConvertable protocol - Introducing metrics.AccumulativeFeatureMetric for accumulative support metrics - Introduce optimize function to optimize the model - Metrics now accepts non-dictionary input and target when target is given - Moved zero gradients to optimize function in training manager

Other updates: - Logger improved - Minor bugs fixed - Performance improved - Typing improvement

Release candidate updates: - Introducing metrics.MS_SSIM for multi-scale SSIM - Minor bugs fixed - Switched default accumulative flag metrics.KID to False - The metrics.FID now inherits metrics.AccumulativeFeatureMetric for accumulative support

- Python
Published by kisonho about 1 year ago

torchmanager - v1.4 (Beta 5)

API updates: - Add dynamic parallel type for losses.ParallelLoss - Add mapping model support for checkpoint - Add specific data loader type to data.batched wrapper function - Deprecated Python 3.9 - Implement BaseManager as a context manager for training mode - Introducing metrics.CosineSimilarity for cosine similarity metrics - Introducing CkptConvertable protocol - Introduce optimize function to optimize the model - Metrics now accepts non-dictionary input and target when target is given - Moved zero gradients to optimize function in training manager

Other updates: - Logger improved - Minor bugs fixed - Performance improved - Typing improvement

Beta updates: - Accumulative support for KID - Add backward and optimize method as abstract methods in base training managers. - Add logging training finished - Introducing abstract base configurations configs.BaseConfigs - Introducing abstract base managers, including BaseTestingManager (abstract test_step), and BaseTrainingManager (abstract train_step) - Introducing abstract base losses and metrics (losses.BaseLoss and metrics.BaseMetric) - Introducing metrics.AccumulativeFeatureMetric for accumulative support metrics - Minor bugs fixed

- Python
Published by kisonho about 1 year ago

torchmanager - v1.3.4

Updates: - Minor and compatibility bugs fixed - Printing out Python system version in configs

- Python
Published by kisonho over 1 year ago

torchmanager - v1.3.3

Updates: - Manager will be reset to CPU after initialized - Minor bugs fixed

- Python
Published by kisonho over 1 year ago

torchmanager - v1.3.2

Updates: - Add max value control for metrics.PSNR, default max value is 1 - Minor bugs fixed

- Python
Published by kisonho over 1 year ago

torchmanager - v1.3.1

Updates: - Minor bugs fixed - Typing improvement

- Python
Published by kisonho over 1 year ago

torchmanager - v1.3

API Updates - A wrapped loss function will be excluded from manager when saving into checkpoints - Add confgs.JSONConfigs to directly load a json file - Add configs.YAMLConfigs to directly load a yaml file, PyYAML package is required - Add dictionary support during unpack_data for managers - Add torchmanager_core.view.logging.add_console method to add logger console display - Introducing callbacks.LambdaCallback and callbacks.MultiCallbacks - Introducing data.reversed_sliding_window method to restore a torch.Tensor for the sliced windows - Introducing eval function to evaluate via metrics - Introducing metrics.PSNR for PSNR metric - Introducing torchmanager_core.backward package for backward hooking - Introducing torchmanager_core.errors.ConfigsFormatError, torchmanager_core.errors.TransformError, and torchmanager_core.errors.VersionError for better exception handling - Move forward method into BaseManager - Managers will now save API version instead of current version - The callbacks.Experiment callback now extends callbacks.MultiCallbacks

Other updates: - Deprecated Python 3.8 Support - Minor bugs fixed - Performance Improved - Typing improved

- Python
Published by kisonho almost 2 years ago

torchmanager - v1.2.6

Updates: - Compatibility fixed for PyTorch < 2.1

- Python
Published by kisonho almost 2 years ago

torchmanager - v1.2.5

Updates: - Minor bugs fixed - Typing improvement

- Python
Published by kisonho almost 2 years ago

torchmanager - v1.2.4

Updates: - Minor bugs fixed

- Python
Published by kisonho about 2 years ago

torchmanager - v1.2.3

Updates: - Deprecated the old metrics.ConfusionMetrics for name corrected metrics.ConfusionMatrix method. - Minor bugs fixed - Move and deprecate package metrics.conf_met into package metrics.conf_mat for name correction

- Python
Published by kisonho over 2 years ago

torchmanager - v1.2.2

Updates: - Introducing torchmanager_core.version package - Minor bugs fixed - Typing improvement

- Python
Published by kisonho over 2 years ago

torchmanager - v1.2.1

Updates: - Minor bugs fixed - Typing improvement

- Python
Published by kisonho over 2 years ago

torchmanager - v1.2

API Updates: - Add class index control for metrics.BinaryConfusionMetric - Add EPOCH_START and BATCH_START to torchmanager_core.protocols.Frequency - Add number of workers support for data.Dataset - Add return_summary parameter for fit method to return training summary after training finished - Basic configurations implementation with configs.Configs class - Change metrics.ConfusionMetrics to abstract class - Deprecated train package - Implements losses.MAE, metrics.FeatureMetric, metrics.ExtractorScore, metrics.FID, metrics.F1, metrics.Precision, metrics.Recall, metrics.SSIM, metrics.Dice and metrics.PartialDice - Introducing backward, forward function to do backward propagation and forward pass in Manager - Introducing configs package - Introducing random package with freezing/unfreezing random seeds - Introducing a new version convert mechanism using torchmanager_core.protocols.VersionConvertible protocol - Introducing data_parallel and reset method in Manager for multi-GPUs control - Use forward method to calculate TP, FP, and FN and further calculate metric in forward_metric method in metrics.BinaryConfusionMetric

Other updates: - Enables update on batch or epoch start for callbacks.FrequencyCallback - Improved Exception handling during prediction - Minor bugs fixed - Removed deprecated methods

- Python
Published by kisonho over 2 years ago

torchmanager - v1.2 (Release Candidate 8)

API Updates: - Add class index control for metrics.BinaryConfusionMetric - Add EPOCH_START and BATCH_START to torchmanager_core.protocols.Frequency - Add number of workers support for data.Dataset - Basic configurations implementation with configs.Configs class - Change metrics.ConfusionMetrics to abstract class - Deprecated train package - Implements losses.MAE, metrics.FeatureMetric, metrics.ExtractorScore, metrics.FID, metrics.F1, metrics.Precision, metrics.Recall, metrics.SSIM, metrics.Dice and metrics.PartialDice - Introducing backward, forward function to do backward propagation and forward pass in Manager - Introducing configs package - Introducing random package with freezing/unfreezing random seeds - Introducing a new version convert mechanism using torchmanager_core.protocols.VersionConvertible protocol - Introducing data_parallel and reset method in Manager for multi-GPUs control - Use forward method to calculate TP, FP, and FN and further calculate metric in forward_metric method in metrics.BinaryConfusionMetric

Other updates: - Enables update on batch or epoch start for callbacks.FrequencyCallback - Improved Exception handling during prediction - Minor bugs fixed - Removed deprecated methods

RC Updates: - Add return_summary parameter for fit method to return training summary after training finished

- Python
Published by kisonho over 2 years ago

torchmanager - v1.1.6

Updates: - Minor bugs fixed

- Python
Published by kisonho over 2 years ago

torchmanager - v1.2 (Release Candidate 6)

API Updates: - Add EPOCH_START and BATCH_START to torchmanager_core.protocols.Frequency - Add number of workers support for data.Dataset - Basic configurations implementation with configs.Configs class - Change metrics.ConfusionMetrics to abstract class - Deprecated train package - Implements losses.MAE - Implements metrics.FeatureMetric, metrics.ExtractorScore, metrics.FID, metrics.F1, metrics.Precision, and metrics.Recall, metrics.SSIM, metrics.Dice and metrics.PartialDice - Introducing backward, forward function to do backward propagation and forward pass in Manager - Introducing configs package - Introducing random package with freezing/unfreezing random seeds - Introducing a new version convert mechanism using torchmanager_core.protocols.VersionConvertible protocol - Introducing data_parallel and reset method in Manager for multi-GPUs control - Use forward method to calculate TP, FP, and FN and further calculate metric in forward_metric method in metrics.BinaryConfusionMetric

Other updates: - Enables update on batch or epoch start for callbacks.FrequencyCallback - Improved Exception handling during prediction - Minor bugs fixed - Removed deprecated methods

RC Updates: - Minor bugs fixed

- Python
Published by kisonho over 2 years ago

torchmanager - v1.2 (Release Candidate 5)

API Updates: - Add EPOCH_START and BATCH_START to torchmanager_core.protocols.Frequency - Add number of workers support for data.Dataset - Basic configurations implementation with configs.Configs class - Change metrics.ConfusionMetrics to abstract class - Implements losses.MAE - Implements metrics.FeatureMetric, metrics.ExtractorScore, metrics.FID, metrics.F1, metrics.Precision, and metrics.Recall, metrics.SSIM. - Introducing backward, forward function to do backward propagation and forward pass in Manager - Introducing configs package - Introducing random package with freezing/unfreezing random seeds - Introducing a new version convert mechanism using torchmanager_core.protocols.VersionConvertible protocol - Introducing data_parallel and reset method in Manager for multi-GPUs control - Use forward method to calculate TP, FP, and FN and further calculate metric in forward_metric method in metrics.BinaryConfusionMetric

Other updates: - Enables update on batch or epoch start for callbacks.FrequencyCallback - Improved Exception handling during prediction - Minor bugs fixed - Removed deprecated methods

RC Updates: - Deprecated _compile function in BasicManager - Deprecated train package - Implement metrics.Dice and metrics.PartialDice - Minor bugs fixed

- Python
Published by kisonho almost 3 years ago

torchmanager - v1.1.5

Updates: - Minor bugs fixed

- Python
Published by kisonho almost 3 years ago

torchmanager - v1.1.4

Updates: - Add release candidate support for torchmanager_core.Version - Minor bugs fixed

- Python
Published by kisonho almost 3 years ago

torchmanager - v1.1.3

Updates: - data.Dataset can now handle unsupervised dataset for reconstruction or dataset with packed dictionaries without further implementation of unpack_data - Minor bugs fixed - Move checkpiont property in callbacks._Checkpoint to read-only public - Move from setup.py to pyproject.toml

- Python
Published by kisonho about 3 years ago

torchmanager - v1.1.2

Updates: - Add Version class for better version checking - Change metrics.ConfusionMetrics to abstract class - Minor bugs fixed - Move data.utils.sliding_window to data.sliding_window - Removes data.utils package and redundant code

- Python
Published by kisonho about 3 years ago

torchmanager - v1.1.1

Updates: - Extends losses.mse._ReductableLoss for losses.MSE - Implements sliding_window function in data.utils package - Minor bugs fixed - Raise a TypeError when loading data.Dataset in a data.batched function instead of warning

- Python
Published by kisonho about 3 years ago

torchmanager - v1.1

API updates: - Add softmax input and target control for torchmanager.losses.KLDiv loss - Implement MSE loss with NAN replaced to zero - Introducing errors package in torchmanager_core - Introducing Resulting protocol - Introducing torchmanager.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 folder - Introducing torchmanager.data package - Introducing torchmanager.losses.ParallelLoss for loss functions multi-GPUs support - Option to replace NAN to zero after log calculation in KLDiv loss - Support of listing checkpoints inside an experiment .exp folder using torchmanager.train.list_checkpoints method - Support of loading a torchmanager.train.Checkpoint by .exp folder directory and checkpoint name using torchmanager.train.load method

Other updates: - Performance improvement - Minor bugs fixed

- Python
Published by kisonho over 3 years ago

torchmanager - v1.0.7

Updates: - If a metric has not been called after reset, its result will return a nan. - Introducing protected _metrics in EarlyStop callback - Remove _update method implementation requirement for abstract class FrequencyCallback - The loss weight can now be zero for non-activate

- Python
Published by kisonho over 3 years ago

torchmanager - v1.0.6

Updates: - Add results property for the metrics to return all current results as a torch.Tensor - Add threshold settings to MeanIoU metric - Introducing FrequencyCallback with frequency support - Introducing DynamicWeight callback that dynamically change the weight property in an object - Introducing predict method in TestingManager to predict a dataset and returns a list of predictions - Parsing device in fit and test as a list of specified devices

- Python
Published by kisonho over 3 years ago

torchmanager - v1.0.5

Updates: - Managers now perform to torchmanager_core.devices.DeviceMovable and train.StateDictLoadable protocol - Save weights only now supports any kind of objects that perform to StateDictLoadable protocol - EarlyStop callback implementation to stop the training when monitored metrics not improvement for several epochs - Loss now supports weight coefficient value - Performance improved for multi-GPUs running - Minor bugs fixed

- Python
Published by kisonho almost 4 years ago

torchmanager - v1.0.4

Updates: - Separate Manager into basic.BaseManager, basic.DataManager, testing.Manager, and training.Manager - Add unpatch method (private) to unpatch data - Add target support for multi outputs model in losses and metrics - Minor bugs fixed

- Python
Published by kisonho about 4 years ago

torchmanager - v1.0.3

Updates: - current_epoch index implementation - Manager can be passed into Checkpoint as well as related callbacks to be saved - LrScheduler callback now updates lr to fit the initial_epoch index - Minor bugs fixed

- Python
Published by kisonho about 4 years ago

torchmanager - v1.0.2

Updates: - Performance Improved - Minor bugs fixed

- Python
Published by kisonho about 4 years ago

torchmanager - v1.0.1

Updates: - Multi Losses support - Multi Losses for multiple outputs support - Logging support - Minor bugs fixed

- Python
Published by kisonho over 4 years ago

torchmanager - v1.0

Main Release

- Python
Published by kisonho over 4 years ago