Finetuning

class eztorch.models.finetuning.FinetuningModel(trunk, classifier, optimizer, pretrained_trunk_path, trunk_pattern='^(trunk\\\\.)', two_groups=False, freeze_trunk=False, train_transform=None, val_transform=None, test_transform=None, val_time_augmentation=None, test_time_augmentation=None, update_bn_momentum=False, freeze_bn_layers=False)

Fine-tuning training.

Parameters:

• trunk (

  Default:) – Config to build a trunk.

• classifier (

  Default:) – Config to build a classifier.

• optimizer (

  Default:) – Config to build an optimizer for trunk.

• pretrained_trunk_path (

  Default:) – Path to the pretrained trunk file.

• trunk_pattern (

  Default:, optional) – Pattern to retrieve the trunk model in checkpoint state_dict and delete the key.

  Default: '^(trunk\.)'

• two_groups (

  Default:, optional) – If True, use two groups of parameters for optimizer, the trunk and the head.

  Default: False

• freeze_trunk (

  Default:, optional) – If True, freeze the trunk.

  Default: False

• train_transform (

  Default:, optional) – Config to perform transformation on train input.

  Default: None

• val_transform (

  Default:, optional) – Config to perform transformation on val input.

  Default: None

• test_transform (

  Default:, optional) – Config to perform transformation on test input.

  Default: None

• val_time_augmentation (

  Default:, optional) – Ensembling method for test time augmentation used at validation.

  Default: None

• test_time_augmentation (

  Default:, optional) – Ensembling method for test time augmentation used at test.

  Default: None

• update_bn_momentum (

  Default:, optional) – If True update batch norm statistics according to \(max(1 - 10/steps\_per\_epoch, 0.9)\).

  Default: False

• freeze_bn_layers (

  Default:, optional) – If True, freeze the batch norm layers.

  Default: False

Example:

trunk = {...} # config to build a trunk
classifier = {...} # config to build a classifier
optimizer = {...} # config to build an optimizer
pretrained_trunk_path = ... # path where the trunk has been saved

model = FinetuningModel(trunk, classifier, optimizer, pretrained_trunk_path)

class eztorch.models.soccernet_spotting.SoccerNetSpottingModel(trunk, head_class, optimizer, pretrained_path=None, pretrained_trunk_path=None, loss_fn_args={}, prediction_args={}, eval_step_timestamp=1.0, trunk_pattern='^(trunk\\\\.)', freeze_trunk=False, train_transform=None, val_transform=None, test_transform=None, save_val_preds_path='val_preds/', save_test_preds_path='test_preds/', NMS_args={'window': 10, 'threshold': 0.001}, evaluation_args={}, do_compile=False)

Model to perform action spotting.

Parameters:

• trunk (

  Default:) – Config to build a trunk.

• head_class (

  Default:) – Config to build a head for classification.

• optimizer (

  Default:) – Config to build an optimizer for trunk.

• pretrained_trunk_path (

  Default:, optional) – Path to the pretrained trunk file.

  Default: None

• pretrained_path (

  Default:, optional) – Path to the pretrained model.

  Default: None

• prediction_args (

  Default:, optional) – Arguments to configure predictions.

  Default: {}

• loss_fn_args (

  Default:, optional) – Arguments for the loss function.

  Default: {}

• eval_step_timestamp (

  Default:, optional) – Step between each timestamp.

  Default: 1.0

• trunk_pattern (

  Default:, optional) – Pattern to retrieve the trunk model in checkpoint state_dict and delete the key.

  Default: '^(trunk\.)'

• freeze_trunk (

  Default:, optional) – Whether to freeze the trunk.

  Default: False

• train_transform (

  Default:, optional) – Config to perform transformation on train input.

  Default: None

• val_transform (

  Default:, optional) – Config to perform transformation on val input.

  Default: None

• test_transform (

  Default:, optional) – Config to perform transformation on test input.

  Default: None

• save_val_preds_path (

  Default:, optional) – Path to store the validation predictions.

  Default: 'val_preds/'

• save_test_preds_path (

  Default:, optional) – Path to store the test predictions.

  Default: 'test_preds/'

• NMS_args (

  Default:, optional) – Arguments to configure the NMS.

  Default: {'window': 10, 'threshold': 0.001}

• evaluation_args (

  Default:, optional) – Arguments to configure the evaluation.

  Default: {}

Example:

trunk = {...} # config to build a trunk
head_class = {...} # config to build a head for classification
optimizer = {...} # config to build an optimizer
pretrained_trunk_path = ... # path where the trunk has been saved

model = SoccerNetSpottingModel(trunk, head_class, optimizer, pretrained_trunk_path)

class eztorch.models.spotting.SpottingModel(trunk, head_class, optimizer, pretrained_path=None, pretrained_trunk_path=None, loss_fn_args={}, prediction_args={}, trunk_pattern='^(trunk\\\\.)', freeze_trunk=False, train_transform=None, val_transform=None, test_transform=None, save_val_preds_path='val_preds/', save_test_preds_path='test_preds/', NMS_args={'window': 10, 'threshold': 0.001}, evaluation_args={}, do_compile=False)

Model to perform spotting.

Parameters:

• trunk (

  Default:) – Config to build a trunk.

• head_class (

  Default:) – Config to build a head for classification.

• optimizer (

  Default:) – Config to build an optimizer for trunk.

• pretrained_trunk_path (

  Default:, optional) – Path to the pretrained trunk file.

  Default: None

• pretrained_path (

  Default:, optional) – Path to the pretrained model.

  Default: None

• prediction_args (

  Default:, optional) – Arguments to configure predictions.

  Default: {}

• loss_fn_args (

  Default:, optional) – Arguments for the loss function.

  Default: {}

• trunk_pattern (

  Default:, optional) – Pattern to retrieve the trunk model in checkpoint state_dict and delete the key.

  Default: '^(trunk\.)'

• freeze_trunk (

  Default:, optional) – Whether to freeze the trunk.

  Default: False

• train_transform (

  Default:, optional) – Config to perform transformation on train input.

  Default: None

• val_transform (

  Default:, optional) – Config to perform transformation on val input.

  Default: None

• test_transform (

  Default:, optional) – Config to perform transformation on test input.

  Default: None

• save_val_preds_path (

  Default:, optional) – Path to store the validation predictions.

  Default: 'val_preds/'

• save_test_preds_path (

  Default:, optional) – Path to store the test predictions.

  Default: 'test_preds/'

• NMS_args (

  Default:, optional) – Arguments to configure the NMS.

  Default: {'window': 10, 'threshold': 0.001}

• evaluation_args (

  Default:, optional) – Arguments to configure the evaluation.

  Default: {}

Example:

trunk = {...} # config to build a trunk
head_class = {...} # config to build a head for classification
optimizer = {...} # config to build an optimizer
pretrained_trunk_path = ... # path where the trunk has been saved

model = SpottingModel(trunk, head_class, optimizer, pretrained_trunk_path)