Models

time_interpret provides several general deep learning models, as well as a network to use them along with the pytorch lightning framework. These models are listed here:

Summary

tint.models.Bert([...])	Get Bert model for sentence classification, either as a pre-trained model or from scratch.
tint.models.CNN(units, kernel_size[, ...])	Base CNN class.
tint.models.DistilBert([...])	Get DistilBert model for sentence classification, either as a pre-trained model or from scratch.
tint.models.MLP(units[, bias, dropout, ...])	Base MLP class.
tint.models.Net(layers[, loss, optim, lr, ...])	Base Net class.
tint.models.RNN(input_size[, rnn, ...])	A base recurrent model class.
tint.models.Roberta([...])	Get Roberta model for sentence classification, either as a pre-trained model or from scratch.
tint.models.TransformerEncoder(d_model[, ...])	A base transformer encoder model class.

Detailed classes and methods

tint.models.Bert(pretrained_model_name_or_path: Optional[str] = None, config=None, vocab_file=None, cache_dir=None, **kwargs)

Get Bert model for sentence classification, either as a pre-trained model or from scratch.

Parameters:

• pretrained_model_name_or_path – Path of the pre-trained model. If None, return an untrained Bert model. Default to None

• config – Config of the Bert. Required when not loading a pre-trained model, otherwise unused. Default to None

• vocab_file – Path to a vocab file for the tokenizer. Default to None

• cache_dir – Where to save pretrained model. Default to None

• kwargs – Additional arguments for the tokenizer if not pretrained.

Returns:

• Bert Tokenizer (BertTokenizer):

  Bert Tokenizer.

• Bert Model (BertForSequenceClassification):

  Bert model for sentence classification.

Return type:

2-element tuple of Bert Tokenizer, Bert Model

References

https://huggingface.co/docs/transformers/main/en/model_doc/bert

Examples

>>> from tint.models import Bert

 >>> tokenizer, model = Bert("bert-base-uncased")

class tint.models.CNN(units: list, kernel_size: Union[list, int], stride: Union[list, int] = 1, padding: Union[list, int] = 0, dilation: Union[list, int] = 1, groups: Union[list, int] = 1, bias: Union[list, bool] = True, padding_mode: Union[list, str] = 'zeros', dropout: Union[list, float] = 0.0, norm: Optional[Union[list, str]] = None, activations: Union[list, str] = 'relu', pooling: Optional[Union[list, str]] = None, flatten: bool = True)

Base CNN class.

The following batch norms are available:

• BatchNorm2d: 'batch_norm_2d'

the following activations are available:

• CELU: 'celu'

• ELU: 'elu'

• LeakyReLU: 'leaky_relu'

• LogSoftmax: 'log_softmax'

• ReLU: 'relu'

• ReLU6: 'relu6'

• Sigmoid: 'sigmoid'

• Softmax: 'softmax'

• Softplus: 'softplus'

• SoftSign: 'softsign'

• Tanh: 'tanh'

• Tanhshrink: 'tanhshrink'

and the following pooling layers are available:

• MaxPool2d with a kernel size of 2: 'max_pool_2d'

For more insights into specific arguments of the CNN, please refer to Conv2d pytorch documentation.

Parameters:

• units (list) – A list of units, which creates the layers. Default to None

• dropout (list, float) – Dropout rates. Default to 0.0

• norm (list, str) – Normalisation layers. Either a list or a string. Default to None

• activations (list, str) – Activation functions. Either a list or a string. Default to 'relu'

• pooling (list, str) – Pooling module. Either a list or a string. Default to None

• flatten (bool) – Whether to flatten the output of the model or not. Default to True

References

https://pytorch.org/docs/stable/generated/torch.nn.Conv2d.html#torch.nn.Conv2d

Examples

>>> import torch.nn as nn
>>> from tint.models import CNN

>>> cnn = CNN(units=[10, 8, 6], kernel_size=3)  # Simple cnn with relu activations.
>>> cnn = CNN(units=[10, 8, 6], kernel_size=3, dropout=.1)  # Adding dropout.
>>> cnn = CNN(units=[10, 8, 6], kernel_size=3, activations="elu")  # Elu activations.

forward(x: Tensor) → Tensor

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

tint.models.DistilBert(pretrained_model_name_or_path: Optional[str] = None, config=None, vocab_file=None, cache_dir=None, **kwargs)

Get DistilBert model for sentence classification, either as a pre-trained model or from scratch.

Parameters:

• pretrained_model_name_or_path – Path of the pre-trained model. If None, return an untrained DistilBert model. Default to None

• config – Config of the DistilBert. Required when not loading a pre-trained model, otherwise unused. Default to None

• vocab_file – Path to a vocab file for the tokenizer. Default to None

• cache_dir – Where to save pretrained model. Default to None

• kwargs – Additional arguments for the tokenizer if not pretrained.

Returns:

• DistilBert Tokenizer (DistilBertTokenizer):

  DistilBert Tokenizer.

• DistilBert Model (DistilBertForSequenceClassification):

  DistilBert model for sentence classification.

Return type:

2-element tuple of DistilBert Tokenizer, DistilBert Model

References

https://huggingface.co/docs/transformers/main/en/model_doc/distilbert

Examples

>>> from tint.models import DistilBert

>>> tokenizer, model = DistilBert("distilbert-base-uncased")

class tint.models.MLP(units: list, bias: Union[list, bool] = True, dropout: Union[list, float] = 0.0, norm: Optional[Union[list, str]] = None, activations: Union[list, str] = 'relu', activation_final: Optional[str] = None)

Base MLP class.

The following batch norms are available:

• BatchNorm1d: 'batch_norm_1d'

and the following activations are available:

• CELU: 'celu'

• ELU: 'elu'

• LeakyReLU: 'leaky_relu'

• LogSoftmax: 'log_softmax'

• ReLU: 'relu'

• ReLU6: 'relu6'

• Sigmoid: 'sigmoid'

• Softmax: 'softmax'

• Softplus: 'softplus'

• SoftSign: 'softsign'

• Tanh: 'tanh'

• Tanhshrink: 'tanhshrink'

For more insights into specific arguments of the MLP, please refer to Linear pytorch documentation.

Parameters:

• units (list) – A list of units, which creates the layers. Default to None

• bias (list, bool) – Whether to add bias to each layer. Default to True

• dropout (list, float) – Dropout rates. Default to 0.0

• norm (list, str) – Normalisation layers. Either a list or a string. Default to None

• activations (list, str) – Activation functions. Either a list or a string. Default to 'relu'

• activation_final (str) – Final activation. Default to None

References

https://pytorch.org/docs/stable/generated/torch.nn.Linear.html#torch.nn.Linear

Examples

>>> import torch.nn as nn
>>> from tint.models import MLP

>>> mlp = MLP(units=[5, 10, 1])  # Simple fc with relu activations.
>>> mlp = MLP(units=[5, 10, 1], dropout=.1)  # Adding dropout.
>>> mlp = MLP(units=[5, 10, 1], activations="elu")  # Elu activations.

forward(x: Tensor) → Tensor

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class tint.models.Net(layers: Union[List[Module], Module], loss: Union[str, Callable] = 'mse', optim: str = 'adam', lr: float = 0.001, lr_scheduler: Optional[Union[dict, str]] = None, lr_scheduler_args: Optional[dict] = None, l2: float = 0.0)

Base Net class.

This provides a wrapper around any Pytorch model into the Pytorch Lightning framework.

Net adds a loss and an optimizer to the model. The following losses are available:

• MAE: 'l1'

• MSE: 'mse'

• NLL: 'nll'

• CrossEntropy: 'cross_entropy'

• CrossEntropy with soft labels: 'soft_cross_entropy'

• BCE with logits: 'bce_with_logits'

The following optimizer are available:

• SGD: 'sgd'

• Adam: 'adam'

It is also possible to pass a custom learning rate to the Net, as well as a learning rate scheduler. Both SGD and Adam also support l2 regularisation.

Parameters:

• layers (list, nn.Module) – The base layers. Can be either a Pytorch module or a list of Pytorch modules.

• loss (str, callable) – Which loss to use. Default to 'mse'

• optim (str) – Which optimizer to use. Default to 'adam'

• lr (float) – Learning rate. Default to 1e-3

• lr_scheduler (dict, str) – Learning rate scheduler. Either a dict (custom scheduler) or a string. Default to None

• lr_scheduler_args (dict) – Additional args for the scheduler. Default to None

• l2 (float) – L2 regularisation. Default to 0.0

References

https://pytorch-lightning.readthedocs.io/en/latest/common/lightning_module.html

Examples

>>> import torch.nn as nn
>>> from tint.models import MLP, Net

>>> mlp = MLP(units=[5, 10, 1])  # Simple fc with relu activations.
>>> net = Net([mlp])  # Wrap the mlp into a PyTorch Lightning Net

configure_optimizers()

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple.

Returns:

Any of these 6 options.

• Single optimizer.

• List or Tuple of optimizers.

• Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).

• Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.

• Tuple of dictionaries as described above, with an optional "frequency" key.

• None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

# The ReduceLROnPlateau scheduler requires a monitor
def configure_optimizers(self):
    optimizer = Adam(...)
    return {
        "optimizer": optimizer,
        "lr_scheduler": {
            "scheduler": ReduceLROnPlateau(optimizer, ...),
            "monitor": "metric_to_track",
            "frequency": "indicates how often the metric is updated"
            # If "monitor" references validation metrics, then "frequency" should be set to a
            # multiple of "trainer.check_val_every_n_epoch".
        },
    }


# In the case of two optimizers, only one using the ReduceLROnPlateau scheduler
def configure_optimizers(self):
    optimizer1 = Adam(...)
    optimizer2 = SGD(...)
    scheduler1 = ReduceLROnPlateau(optimizer1, ...)
    scheduler2 = LambdaLR(optimizer2, ...)
    return (
        {
            "optimizer": optimizer1,
            "lr_scheduler": {
                "scheduler": scheduler1,
                "monitor": "metric_to_track",
            },
        },
        {"optimizer": optimizer2, "lr_scheduler": scheduler2},
    )

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

The frequency value specified in a dict along with the optimizer key is an int corresponding to the number of sequential batches optimized with the specific optimizer. It should be given to none or to all of the optimizers. There is a difference between passing multiple optimizers in a list, and passing multiple optimizers in dictionaries with a frequency of 1:

• In the former case, all optimizers will operate on the given batch in each optimization step.

• In the latter, only one optimizer will operate on the given batch at every step.

This is different from the frequency value specified in the lr_scheduler_config mentioned above.

def configure_optimizers(self):
    optimizer_one = torch.optim.SGD(self.model.parameters(), lr=0.01)
    optimizer_two = torch.optim.SGD(self.model.parameters(), lr=0.01)
    return [
        {"optimizer": optimizer_one, "frequency": 5},
        {"optimizer": optimizer_two, "frequency": 10},
    ]

In this example, the first optimizer will be used for the first 5 steps, the second optimizer for the next 10 steps and that cycle will continue. If an LR scheduler is specified for an optimizer using the lr_scheduler key in the above dict, the scheduler will only be updated when its optimizer is being used.

Examples:

# most cases. no learning rate scheduler
def configure_optimizers(self):
    return Adam(self.parameters(), lr=1e-3)

# multiple optimizer case (e.g.: GAN)
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    return gen_opt, dis_opt

# example with learning rate schedulers
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    dis_sch = CosineAnnealing(dis_opt, T_max=10)
    return [gen_opt, dis_opt], [dis_sch]

# example with step-based learning rate schedulers
# each optimizer has its own scheduler
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    gen_sch = {
        'scheduler': ExponentialLR(gen_opt, 0.99),
        'interval': 'step'  # called after each training step
    }
    dis_sch = CosineAnnealing(dis_opt, T_max=10) # called every epoch
    return [gen_opt, dis_opt], [gen_sch, dis_sch]

# example with optimizer frequencies
# see training procedure in `Improved Training of Wasserstein GANs`, Algorithm 1
# https://arxiv.org/abs/1704.00028
def configure_optimizers(self):
    gen_opt = Adam(self.model_gen.parameters(), lr=0.01)
    dis_opt = Adam(self.model_dis.parameters(), lr=0.02)
    n_critic = 5
    return (
        {'optimizer': dis_opt, 'frequency': n_critic},
        {'optimizer': gen_opt, 'frequency': 1}
    )

Note

Some things to know:

• Lightning calls .backward() and .step() on each optimizer as needed.

• If learning rate scheduler is specified in configure_optimizers() with key "interval" (default “epoch”) in the scheduler configuration, Lightning will call the scheduler’s .step() method automatically in case of automatic optimization.

• If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizers.

• If you use multiple optimizers, training_step() will have an additional optimizer_idx parameter.

• If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.

• If you use multiple optimizers, gradients will be calculated only for the parameters of current optimizer at each training step.

• If you need to control how often those optimizers step or override the default .step() schedule, override the optimizer_step() hook.

forward(x: Tensor) → Tensor

Same as torch.nn.Module.forward().

Parameters:

• *args – Whatever you decide to pass into the forward method.

• **kwargs – Keyword arguments are also possible.

Returns:

Your model’s output

predict_step(batch, batch_idx, dataloader_idx=0)

Step function called during predict(). By default, it calls forward(). Override to add any processing logic.

The predict_step() is used to scale inference on multi-devices.

To prevent an OOM error, it is possible to use BasePredictionWriter callback to write the predictions to disk or database after each batch or on epoch end.

The BasePredictionWriter should be used while using a spawn based accelerator. This happens for Trainer(strategy="ddp_spawn") or training on 8 TPU cores with Trainer(accelerator="tpu", devices=8) as predictions won’t be returned.

Example

class MyModel(LightningModule):

    def predict_step(self, batch, batch_idx, dataloader_idx=0):
        return self(batch)

dm = ...
model = MyModel()
trainer = Trainer(accelerator="gpu", devices=2)
predictions = trainer.predict(model, dm)

Parameters:

• batch – Current batch.

• batch_idx – Index of current batch.

• dataloader_idx – Index of the current dataloader.

Returns:

Predicted output

test_step(batch, batch_idx)

Operates on a single batch of data from the test set. In this step you’d normally generate examples or calculate anything of interest such as accuracy.

# the pseudocode for these calls
test_outs = []
for test_batch in test_data:
    out = test_step(test_batch)
    test_outs.append(out)
test_epoch_end(test_outs)

Parameters:

• batch – The output of your DataLoader.

• batch_idx – The index of this batch.

• dataloader_id – The index of the dataloader that produced this batch. (only if multiple test dataloaders used).

Returns:

Any of.

• Any object or value

• None - Testing will skip to the next batch

# if you have one test dataloader:
def test_step(self, batch, batch_idx):
    ...


# if you have multiple test dataloaders:
def test_step(self, batch, batch_idx, dataloader_idx=0):
    ...

Examples:

# CASE 1: A single test dataset
def test_step(self, batch, batch_idx):
    x, y = batch

    # implement your own
    out = self(x)
    loss = self.loss(out, y)

    # log 6 example images
    # or generated text... or whatever
    sample_imgs = x[:6]
    grid = torchvision.utils.make_grid(sample_imgs)
    self.logger.experiment.add_image('example_images', grid, 0)

    # calculate acc
    labels_hat = torch.argmax(out, dim=1)
    test_acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)

    # log the outputs!
    self.log_dict({'test_loss': loss, 'test_acc': test_acc})

If you pass in multiple test dataloaders, test_step() will have an additional argument. We recommend setting the default value of 0 so that you can quickly switch between single and multiple dataloaders.

# CASE 2: multiple test dataloaders
def test_step(self, batch, batch_idx, dataloader_idx=0):
    # dataloader_idx tells you which dataset this is.
    ...

Note

If you don’t need to test you don’t need to implement this method.

Note

When the test_step() is called, the model has been put in eval mode and PyTorch gradients have been disabled. At the end of the test epoch, the model goes back to training mode and gradients are enabled.

training_step(batch, batch_idx)

Here you compute and return the training loss and some additional metrics for e.g. the progress bar or logger.

Parameters:

• batch (Tensor | (Tensor, …) | [Tensor, …]) – The output of your DataLoader. A tensor, tuple or list.

• batch_idx (int) – Integer displaying index of this batch

• optimizer_idx (int) – When using multiple optimizers, this argument will also be present.

• hiddens (Any) – Passed in if truncated_bptt_steps > 0.

Returns:

Any of.

• Tensor - The loss tensor

• dict - A dictionary. Can include any keys, but must include the key 'loss'

• None - Training will skip to the next batch. This is only for automatic optimization.

  This is not supported for multi-GPU, TPU, IPU, or DeepSpeed.

In this step you’d normally do the forward pass and calculate the loss for a batch. You can also do fancier things like multiple forward passes or something model specific.

Example:

def training_step(self, batch, batch_idx):
    x, y, z = batch
    out = self.encoder(x)
    loss = self.loss(out, x)
    return loss

If you define multiple optimizers, this step will be called with an additional optimizer_idx parameter.

# Multiple optimizers (e.g.: GANs)
def training_step(self, batch, batch_idx, optimizer_idx):
    if optimizer_idx == 0:
        # do training_step with encoder
        ...
    if optimizer_idx == 1:
        # do training_step with decoder
        ...

If you add truncated back propagation through time you will also get an additional argument with the hidden states of the previous step.

# Truncated back-propagation through time
def training_step(self, batch, batch_idx, hiddens):
    # hiddens are the hidden states from the previous truncated backprop step
    out, hiddens = self.lstm(data, hiddens)
    loss = ...
    return {"loss": loss, "hiddens": hiddens}

Note

The loss value shown in the progress bar is smoothed (averaged) over the last values, so it differs from the actual loss returned in train/validation step.

Note

When accumulate_grad_batches > 1, the loss returned here will be automatically normalized by accumulate_grad_batches internally.

validation_step(batch, batch_idx)

Operates on a single batch of data from the validation set. In this step you’d might generate examples or calculate anything of interest like accuracy.

# the pseudocode for these calls
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    val_outs.append(out)
validation_epoch_end(val_outs)

Parameters:

• batch – The output of your DataLoader.

• batch_idx – The index of this batch.

• dataloader_idx – The index of the dataloader that produced this batch. (only if multiple val dataloaders used)

Returns:

• Any object or value

• None - Validation will skip to the next batch

# pseudocode of order
val_outs = []
for val_batch in val_data:
    out = validation_step(val_batch)
    if defined("validation_step_end"):
        out = validation_step_end(out)
    val_outs.append(out)
val_outs = validation_epoch_end(val_outs)

# if you have one val dataloader:
def validation_step(self, batch, batch_idx):
    ...


# if you have multiple val dataloaders:
def validation_step(self, batch, batch_idx, dataloader_idx=0):
    ...

Examples:

# CASE 1: A single validation dataset
def validation_step(self, batch, batch_idx):
    x, y = batch

    # implement your own
    out = self(x)
    loss = self.loss(out, y)

    # log 6 example images
    # or generated text... or whatever
    sample_imgs = x[:6]
    grid = torchvision.utils.make_grid(sample_imgs)
    self.logger.experiment.add_image('example_images', grid, 0)

    # calculate acc
    labels_hat = torch.argmax(out, dim=1)
    val_acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)

    # log the outputs!
    self.log_dict({'val_loss': loss, 'val_acc': val_acc})

If you pass in multiple val dataloaders, validation_step() will have an additional argument. We recommend setting the default value of 0 so that you can quickly switch between single and multiple dataloaders.

# CASE 2: multiple validation dataloaders
def validation_step(self, batch, batch_idx, dataloader_idx=0):
    # dataloader_idx tells you which dataset this is.
    ...

Note

If you don’t need to validate you don’t need to implement this method.

Note

When the validation_step() is called, the model has been put in eval mode and PyTorch gradients have been disabled. At the end of validation, the model goes back to training mode and gradients are enabled.

class tint.models.RNN(input_size: int, rnn: Union[RNNBase, str] = 'rnn', hidden_size: int = 32, num_layers: int = 1, bias: bool = True, dropout: float = 0.0, bidirectional: bool = False, many_to_one: bool = False)

A base recurrent model class.

The following RNN are supported:

• RNN: 'rnn'

• LSTM: 'lstm'

• GRU: 'gru'

Parameters:

• input_size (int) – Input size of the model.

• rnn (nn.RNNBase, str) – Which type of RNN to use. Default to 'rnn'

• hidden_size (int) – The number of features in the hidden state h. Default to 32

• num_layers (int) – Number of recurrent layers. Default to 1

• bias (bool) – Whether to use bias. Default to True

• dropout (float) – Dropout rates. Default to 0.0

• bidirectional (bool) – If True, becomes a bidirectional RNN. Default to False

• many_to_one (bool) – Whether to reduce the temporal dimension. Default to False

References

https://pytorch.org/docs/stable/nn.html#recurrent-layers

Examples

>>> from tint.models import RNN

>>> rnn = RNN(10, hidden_size=32)
>>> gru = RNN(10, rnn="gru", bidirectional=True)
>>> lstm = RNN(10, rnn="lstm", many_to_one=True)

forward(x: Tensor) → Tensor

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

tint.models.Roberta(pretrained_model_name_or_path: Optional[str] = None, config=None, vocab_file=None, cache_dir=None, **kwargs)

Get Roberta model for sentence classification, either as a pre-trained model or from scratch.

Parameters:

• pretrained_model_name_or_path – Path of the pre-trained model. If None, return an untrained Roberta model. Default to None

• config – Config of the Roberta. Required when not loading a pre-trained model, otherwise unused. Default to None

• vocab_file – Path to a vocab file for the tokenizer. Default to None

• cache_dir – Where to save pretrained model. Default to None

• kwargs – Additional arguments for the tokenizer if not pretrained.

Returns:

• Roberta Tokenizer (RobertaTokenizer):

  Roberta Tokenizer.

• Roberta Model (RobertaForSequenceClassification):

  Roberta model for sentence classification.

Return type:

2-element tuple of Roberta Tokenizer, Roberta Model

References

https://huggingface.co/docs/transformers/main/en/model_doc/roberta

Examples

>>> from tint.models import Roberta

 >>> tokenizer, model = Roberta("roberta-base")

class tint.models.TransformerEncoder(d_model: int, nhead: int = 1, dim_feedforward: int = 32, num_layers: int = 1, dropout: float = 0.0, activation: str = 'relu', layer_norm_eps: float = 1e-05, norm_first: bool = False, enable_nested_tensor: bool = False, many_to_one: bool = False)

A base transformer encoder model class.

Parameters:

• d_model (int) – Input size of the model.

• nhead (int) – Number of heads. Default to 1

• dim_feedforward (int) – Dimension of the feedforward network model. Default to 32

• num_layers (int) – Number of layers. Default to 1

• dropout (float) – Dropout rates. Default to 0.0

• activation (str) – Activation function. Default to 'relu'

• layer_norm_eps (float) – Eps value in layer normalization components. Default to 1e-5

• norm_first (bool) – If True, layer norm is done prior to attention and feedforward operations, respectively. Default to False

• enable_nested_tensor (bool) – If True, input will automatically convert to nested tensor. Default to False

• many_to_one (bool) – Whether to reduce the temporal dimension. Default to False

References

https://pytorch.org/docs/stable/nn.html#transformer-layers

Examples

>>> from tint.models import TransformerEncoder

>>> transformer = TransformerEncoder(10)
>>> transformer = TransformerEncoder(10, nhead=2, dropout=0.1)

forward(x: Tensor) → Tensor

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

property src_mask

Generate a square mask for the sequence. The masked positions are filled with float(‘-inf’). Unmasked positions are filled with float(0.0).

Returns:

A mask.

Return type:

th.Tensor