mnist_sandbox package

Submodules

mnist_sandbox.eval module

mnist_sandbox.eval.model_evaluate(model, test_loader)

Evaluate model with specified test dataset.

• Evaluation without torch.grad calculations.

Parameters:

• model (pl.LightningModule) – Trained model in inference mode.

• test_loader (DataLoader) – Test dataset to be evaluated on.

Returns:

• correct (int) – Count of correct predictions.

• total (int) – Count of all items.

Return type:

Tuple[int, int, ndarray]

mnist_sandbox.model module

class mnist_sandbox.model.MNISTNet(*args, **kwargs)

Bases: LightningModule

MNIST CNN

Model architecture

Layer	Name	Type	Params
0	conv1	Conv2d	320
1	conv2	Conv2d	18.5K
2	conv3	Conv2d	55.4K
3	fc1	Linear	442K
4	fc2	Linear	65.7K
5	fc3	Linear	1.3K

• 584 K Trainable params

• 0 Non-trainable params

• 584 K Total params

• 2.336 Total estimated model params size (MB)

Methods

__call__(*args, **kwargs)
configure_optimizers()	Optimizer and learning rate schedulers
forward(x)	Model process
on_train_epoch_end()	Model training epoch end
on_validation_epoch_end()	Model validation epoch end
training_step(batch, batch_idx)	Model training step
validation_step(batch, batch_idx)	Model validation step

configure_optimizers()

Optimizer and learning rate schedulers

Returns:

output – Tuple with Adam optimizer and ReduceLROnPlateau scheduler

Return type:

Tuple[torch.optim.Optimizer, torch.optim.lr_scheduler._LRScheduler]

forward(x)

Model process

Parameters:

x (torch.Tensor) – input data of size [batch_size, 3, 28, 28]

Returns:

x – Tensor of size [batch_size, 10] of probabilities logits

Return type:

torch.Tensor

on_train_epoch_end()

Model training epoch end

• Calculate average loss & accuracy scores;

• Log them & pring.

on_validation_epoch_end()

Model validation epoch end

• Calculate average loss & accuracy scores;

• Log them & pring.

training_step(batch, batch_idx)

Model training step

Parameters:

• batch (Tuple[torch.Tensor, torch.Tensor[]) – __Input tensor__ and __Target tensor__(one hot encoded)

• batch_idx (int) – Hasn’t used

Returns:

output – Dictionary with loss and accuracy scores

Return type:

dict

validation_step(batch, batch_idx)

Model validation step

Parameters:

• batch (Tuple[torch.Tensor, torch.Tensor[]) – __Input tensor__ and __Target tensor__(one hot encoded)

• batch_idx (int) – Hasn’t used

Returns:

output – Dictionary with loss and accuracy scores

Return type:

dict

mnist_sandbox.train module

mnist_sandbox.train.train_m(model, n_epochs, train_loader, val_loader, save=False, logging_url='file:./.logs/my-mlflow-logs')

Training model on train_loader and validation part with val_loader.

• Save trained model weights into sota_mnist_cnn.pth

Parameters:

• model (pl.LightningModule) – Model to be trained.

• n_epochs (int) – Count of training epochs.

• train_loader (DataLoader) – Training dataset.

• val_loader (DataLoader) – Validation dataset.

• save (bool) – Save result model. Default is False, because DVC is used

• logging_url (str) – Where to store mlflow logs

Returns:

model – Trained model.

Return type:

pl.LightningModule