Merge pull request #680 from mv1388/docs_improvement

Docs improvement

aitoolbox/experiment/experiment_saver.py

@@ -21,7 +21,7 @@ def save_experiment(self, model, result_package, training_history, experiment_ti
             training_history (aitoolbox.experiment.training_history.TrainingHistory):
             experiment_timestamp (str): time stamp of the training start
             save_true_pred_labels (bool): should ground truth labels also be saved
-            separate_files (bool): should the results be saved in separate pickle files or should all of the results
+            separate_files (bool): should the results be saved in separate pickle files or should all the results
                 be batched together in a single results file
             protect_existing_folder (bool): can override potentially already existing folder or not
 
@@ -60,7 +60,7 @@ def save_experiment(self, model, result_package, training_history, experiment_ti
             training_history (aitoolbox.experiment.training_history.TrainingHistory):
             experiment_timestamp (str): time stamp at the start of training
             save_true_pred_labels (bool): should ground truth labels also be saved
-            separate_files (bool): should the results be saved in separate pickle files or should all of the results
+            separate_files (bool): should the results be saved in separate pickle files or should all the results
                 be batched together in a single results file
             protect_existing_folder (bool): can override potentially already existing folder or not
 

aitoolbox/experiment/local_experiment_saver.py

@@ -41,7 +41,7 @@ def save_experiment(self, model, result_package, training_history, experiment_ti
             training_history (aitoolbox.experiment.training_history.TrainingHistory):
             experiment_timestamp (str): time stamp at the start of training
             save_true_pred_labels (bool): should ground truth labels also be saved
-            separate_files (bool): should the results be saved in separate pickle files or should all of the results
+            separate_files (bool): should the results be saved in separate pickle files or should all the results
                 be batched together in a single results file
             protect_existing_folder (bool): can override potentially already existing folder or not
 

aitoolbox/experiment/result_package/abstract_result_packages.py

@@ -18,7 +18,7 @@ def __init__(self, pkg_name=None, strict_content_check=False, np_array=True, **k
         Args:
             pkg_name (str or None): result package name used just for clarity
             strict_content_check (bool): should just print warning or raise the error and crash
-            np_array (bool or str): how the inputs should be handled. Should the package try to automatically guess or
+            np_array (bool or str): how the inputs should be handled. Should the package try to automatically guess, or
                 you want to manually decide whether to leave the inputs as they are or convert them to np.array.
                 Possible options: True, False, 'auto'
                 Be slightly careful with 'auto' as it sometimes doesn't work so it is preferable to explicitly use True/False
@@ -151,9 +151,9 @@ def get_additional_results_dump_paths(self):
         return self.additional_results_dump_paths
 
     def list_additional_results_dump_paths(self):
-        """Specify the list of meta data files you also want to save & upload to s3 during the experiment saving procedure
+        """Specify the list of metadata files you also want to save & upload to s3 during the experiment saving procedure
 
-        By default there are no additional files that are saved as the return is None. If you want to save your
+        By default, there are no additional files that are saved as the return is None. If you want to save your
         specific additional files produced during the training procedure, then override this method specifying
         the file paths.
 

aitoolbox/experiment/result_reporting/hyperparam_reporter.py

@@ -52,8 +52,10 @@ def copy_to_cloud_storage(self, local_hyperparams_file_path, cloud_saver, file_n
         """Copy saved text local file into cloud storage
 
         Args:
-            local_hyperparams_file_path (str): path to hyperparams file stored on local disk. File to be uploaded to cloud
+            local_hyperparams_file_path (str): path to hyperparams file stored on local disk.
+                File to be uploaded to cloud
             cloud_saver (BaseModelSaver or BaseResultsSaver or BaseModelGoogleStorageSaver or BaseResultsGoogleStorageSaver):
+                cloud saver object
             file_name (str or None): manually specify the file name to be saved to the cloud instead of taking
                 the default from self.file_name
 

aitoolbox/experiment/result_reporting/report_generator.py

@@ -128,7 +128,7 @@ def generate_report(self, training_history, epoch, file_name, results_folder_nam
             epoch (int): current epoch
             file_name (str): output text file name
             results_folder_name (str): results folder path where the report file will be located
-            file_format (str): output file format. Can be either 'txt' human readable output or
+            file_format (str): output file format. Can be either 'txt' human-readable output or
                 'tsv' for a tabular format or 'csv' for comma separated format.
 
         Returns:

aitoolbox/torchtrain/callbacks/abstract.py

@@ -103,7 +103,7 @@ def on_after_gradient_update(self, optimizer_idx):
         """Logic executed after the model gradients are updated
 
         To ensure the execution of this callback enable the `self.train_loop_obj.grad_cb_used = True` option in the
-        on_train_loop_registration(). Otherwise logic implemented here will not be executed by the TrainLoop.
+        on_train_loop_registration(). Otherwise, logic implemented here will not be executed by the TrainLoop.
 
         Args:
             optimizer_idx (int): index of the current optimizer. Mostly useful when using multiple optimizers.
@@ -118,7 +118,7 @@ def on_after_optimizer_step(self):
         """Logic executed after the optimizer does a new step and updates the model weights
 
         To ensure the execution of this callback enable the `self.train_loop_obj.grad_cb_used = True` option in the
-        on_train_loop_registration(). Otherwise logic implemented here will not be executed by the TrainLoop.
+        on_train_loop_registration(). Otherwise, logic implemented here will not be executed by the TrainLoop.
 
         Returns:
             None

aitoolbox/torchtrain/callbacks/basic.py

@@ -389,7 +389,7 @@ def __init__(self, fn_to_execute,
             after_gradient_update (bool): should execute after the gradient update
             after_optimizer_step (bool): should execute after the optimizer step
             execution_order (int): order of the callback execution. If all the used callbacks have the orders set to 0,
-                than the callbacks are executed in the order they were registered.
+                then the callbacks are executed in the order they were registered.
         """
         AbstractCallback.__init__(self, 'Execute given function as a callback', execution_order, device_idx_execution)
         self.fn_to_execute = fn_to_execute

aitoolbox/torchtrain/callbacks/gradient.py

@@ -15,14 +15,14 @@ class GradientCallbackBase(AbstractCallback):
     def __init__(self, callback_name, execution_order=0):
         """Base abstract class for gradient related callbacks
 
-        It has not implemented logic except for the the turning enabling of the grad_cb_used inside TrainLoop as part of
+        It has not implemented logic except for the turning enabling of the grad_cb_used inside TrainLoop as part of
         the on_train_loop_registration(). Consequently, this potentially repeated task in every gradient calculation
         callback doesn't need to be done for every implemented callback.
 
         Args:
             callback_name (str): name of the callback
             execution_order (int): order of the callback execution. If all the used callbacks have the orders set to 0,
-                than the callbacks are executed in the order they were registered.
+                then the callbacks are executed in the order they were registered.
         """
         AbstractCallback.__init__(self, callback_name, execution_order)
 
@@ -59,7 +59,7 @@ def __init__(self, max_grad_norm, **kwargs):
 
         Args:
             max_grad_norm (int or float): max norm of the gradients
-            **kwargs: torch.nn.utils.clip_grad_norm_ additional arguemnts
+            **kwargs: torch.nn.utils.clip_grad_norm_ additional arguments
         """
         GradientCallbackBase.__init__(self, 'Gradient norm clipping')
         self.max_grad_norm = max_grad_norm

aitoolbox/torchtrain/callbacks/performance_eval.py

@@ -136,7 +136,7 @@ def __init__(self, metrics, on_each_epoch=True, report_frequency=None,
         """Print the model performance to the console
 
         Best used in combination with the callback which actually calculates some performance evaluation metrics, such
-        as ModelPerformanceEvaluation. Otherwise we are limited only to automatic loss calculation reporting.
+        as ModelPerformanceEvaluation. Otherwise, we are limited only to automatic loss calculation reporting.
 
         When listing callbacks for the TrainLoop it is important to list the ModelPerformanceEvaluation before
         this ModelPerformancePrintReport. This ensures that the calculated results are present in the
@@ -309,7 +309,7 @@ def __init__(self, callback_name, execution_order=0,
         Args:
             callback_name (str): name of the callback
             execution_order (int): order of the callback execution. If all the used callbacks have the orders set to 0,
-                than the callbacks are executed in the order they were registered.
+                then the callbacks are executed in the order they were registered.
             epoch_end (bool): should plot after every epoch
             train_end (bool): should plot at the end of the training
             file_format (str): output file format
@@ -443,7 +443,7 @@ def __init__(self, epoch_end=True, train_end=False, file_format='txt',
         Args:
             epoch_end (bool): should plot after every epoch
             train_end (bool): should plot at the end of the training
-            file_format (str): output file format. Can be either 'txt' human readable output or
+            file_format (str): output file format. Can be either 'txt' human-readable output or
                 'tsv' for a tabular format or 'csv' for comma separated format.
             project_name (str or None): root name of the project
             experiment_name (str or None): name of the particular experiment

aitoolbox/torchtrain/data/batch_model_feed_defs.py

@@ -13,7 +13,7 @@ class AbstractModelFeedDefinition(ABC):
     The primary way of defining the model for TrainLoop training is to utilize:
         aitoolbox.torchtrain.model.TTModel
 
-    Use of the ModelFeedDefinition is the legacy way of defining the model. However in certain scenarios where the
+    Use of the ModelFeedDefinition is the legacy way of defining the model. However, in certain scenarios where the
     TTModel might prove to increase complexity, ModelFeedDefinition still is useful for augmenting the nn.Module with
     the logic to calculate loss and predictions.
     """

aitoolbox/torchtrain/multi_loss_optim.py

@@ -33,7 +33,7 @@ def backward(self, optimizer_idx, iteration, amp_grad_scaler):
             iteration (int): Current iteration index. Not used in the most simple setup but provided in case of more
                 elaborate loss backward logic is devised.
             amp_grad_scaler (torch.cuda.amp.GradScaler): AMP GradScaler. If scaler ``enabled`` parameter is set to False
-                the loss is still passed to it but it gets returned unscaled so the behaviour is as it is in the case
+                the loss is still passed to it, but it gets returned unscaled so the behaviour is as it is in the case
                 of non-AMP training.
 
         Returns:
@@ -79,7 +79,7 @@ def step(self, optimizer_idx, iteration, amp_grad_scaler):
             iteration (int): Current iteration index. Not used in the most simple setup but provided in case of more
                 elaborate loss backward logic is devised.
             amp_grad_scaler (torch.cuda.amp.GradScaler): AMP GradScaler. If scaler ``enabled`` parameter is set to False
-                the optimizer have it's normal step() method called withoud applying the AMP mandated unscaling
+                the optimizer have it's normal step() method called without applying the AMP mandated unscaling
                 beforehand. In this respect the behaviour will be the same as in the non-AMP training.
 
         Returns:

aitoolbox/torchtrain/parallel.py

@@ -50,7 +50,7 @@ def __init__(self, module,
                  default_model_methods=('get_loss', 'get_loss_eval', 'get_predictions'), **kwargs):
         """torchtrain enabled DataParallel
 
-        This DataParallel wrapper works in the same way as the original PyTorch nn.DataParallel. Furthermore it exposes
+        This DataParallel wrapper works in the same way as the original PyTorch nn.DataParallel. Furthermore, it exposes
         TTModel batch data feeding definitions (additional abstract methods) to the TrainLoop while still enabling
         multi GPU training.
 

aitoolbox/torchtrain/train_loop/components/callback_handler.py

@@ -20,7 +20,7 @@ def __init__(self, train_loop_obj):
     def register_callbacks(self, callbacks, cache_callbacks=False):
         """Register TrainLoop object reference inside the listed callbacks when the TrainLoop is created
 
-        Normally, this is called from inside of the train loop by the TrainLoop itself. Basically train loop "registers"
+        Normally, this is called from inside the train loop by the TrainLoop itself. Basically train loop "registers"
         itself.
 
         Args:

aitoolbox/torchtrain/train_loop/train_loop.py

@@ -773,7 +773,7 @@ def _spawn_fit(self, gpu, ddp_args, num_epochs, num_iterations, callbacks, grad_
             grad_accumulation (int): number of batches the gradients are accumulated before updating weights
             in_process_data_load (list or None): in-process data loading logic implemented as a torchtrain callback.
                 The logic should be placed inside the on_multiprocess_start() callback function.
-                When using this data loading option bare in mind that loaded dataset will be replicated in memory for
+                When using this data loading option bear in mind that loaded dataset will be replicated in memory for
                 every spawned training process. This can in turn in cause extensive overall memory consumption.
         """
         rank = ddp_args['node_rank'] * ddp_args['num_gpus'] + gpu