accelerator |
Create an accelerator |
as_dataloader |
Creates a dataloader from its input |
as_dataloader.array |
Creates a dataloader from its input |
as_dataloader.dataloader |
Creates a dataloader from its input |
as_dataloader.dataset |
Creates a dataloader from its input |
as_dataloader.list |
Creates a dataloader from its input |
as_dataloader.matrix |
Creates a dataloader from its input |
as_dataloader.numeric |
Creates a dataloader from its input |
as_dataloader.torch_tensor |
Creates a dataloader from its input |
context |
Context object |
ctx |
Context object |
evaluate |
Evaluates a fitted model on a dataset |
fit.luz_module_generator |
Fit a 'nn_module' |
get_metrics |
Get metrics from the object |
get_metrics.luz_module_fitted |
Get metrics from the object |
lr_finder |
Learning Rate Finder |
luz_callback |
Create a new callback |
luz_callback_auto_resume |
Resume training callback |
luz_callback_csv_logger |
CSV logger callback |
luz_callback_early_stopping |
Early stopping callback |
luz_callback_gradient_clip |
Gradient clipping callback |
luz_callback_interrupt |
Interrupt callback |
luz_callback_keep_best_model |
Keep the best model |
luz_callback_lr_scheduler |
Learning rate scheduler callback |
luz_callback_metrics |
Metrics callback |
luz_callback_mixup |
Mixup callback |
luz_callback_model_checkpoint |
Checkpoints model weights |
luz_callback_profile |
Profile callback |
luz_callback_progress |
Progress callback |
luz_callback_resume_from_checkpoint |
Allow resume model training from a specific checkpoint |
luz_callback_tfevents |
tfevents callback |
luz_callback_train_valid |
Train-eval callback |
luz_load |
Load trained model |
luz_load_checkpoint |
Loads a checkpoint |
luz_load_model_weights |
Loads model weights into a fitted object. |
luz_metric |
Creates a new luz metric |
luz_metric_accuracy |
Accuracy |
luz_metric_binary_accuracy |
Binary accuracy |
luz_metric_binary_accuracy_with_logits |
Binary accuracy with logits |
luz_metric_binary_auroc |
Computes the area under the ROC |
luz_metric_mae |
Mean absolute error |
luz_metric_mse |
Mean squared error |
luz_metric_multiclass_auroc |
Computes the multi-class AUROC |
luz_metric_rmse |
Root mean squared error |
luz_metric_set |
Creates a metric set |
luz_save |
Saves luz objects to disk |
luz_save_model_weights |
Loads model weights into a fitted object. |
nnf_mixup |
Mixup logic |
nn_mixup_loss |
Loss to be used with 'callbacks_mixup()'. |
predict.luz_module_fitted |
Create predictions for a fitted model |
setup |
Set's up a 'nn_module' to use with luz |
set_hparams |
Set hyper-parameter of a module |
set_opt_hparams |
Set optimizer hyper-parameters |