colossalai.nn.layer.parallel_2d

colossalai.nn.layer.parallel_2d.split_tensor_2d(input_, dim=0)[source]

Splits 2D tensor in specified dimension across cols.

Parameters

  • input (torch.tensor) – Input tensor.

  • dim (int) – Specified dimension in which to split.

Returns

The tensor has been split.

Return type

torch.tensor

colossalai.nn.layer.parallel_2d.reduce_by_batch_2d(input_, reduce_mean=False)[source]

All-reduce the input from the model parallel region.

Parameters

  • input (torch.tensor) – input matrix.

  • reduce_mean (bool, optional) – If set to True, it will divide the output by column parallel size, default to False.

class colossalai.nn.layer.parallel_2d.Linear2D(in_features, out_features, bias=True, dtype=None, skip_bias_add=False, weight_initializer=<function kaiming_uniform_.<locals>.initializer>, bias_initializer=<function xavier_uniform_.<locals>.initializer>)[source]

Linear layer for 2D parallelism

Parameters

  • in_features (int) – size of each input sample.

  • out_features (int) – size of each output sample.

  • bias (bool, optional) – If set to False, the layer will not learn an additive bias, defaults to True.

  • dtype (torch.dtype, optional) – The dtype of parameters, defaults to None.

  • skip_bias_add (bool, optional) – If set to True, it will skip bias add for linear layer, which is preserved for kernel fusion, defaults to False.

  • weight_initializer (typing.Callable, optional) – The initializer of weight, defaults to kaiming uniform initializer.

  • bias_initializer (typing.Callable, optional) – The initializer of bias, defaults to xavier uniform initializer.

More details about initializer please refer to init.

class colossalai.nn.layer.parallel_2d.LayerNorm2D(normalized_shape, eps=1e-05, dtype=None)[source]

Layer Normalization for 2D parallelism.

Parameters

  • normalized_shape (int) – input shape from an expected input of size. \([* \times \text{normalized_shape}[0] \times \text{normalized_shape}[1] \times \ldots \times \text{normalized_shape}[-1]]\) If a single integer is used, it is treated as a singleton list, and this module will normalize over the last dimension which is expected to be of that specific size.

  • eps (float, optional) – a value added to the denominator for numerical stability, defaults to 1e-05.

  • dtype (torch.dtype, optional) – The dtype of parameters, defaults to None.

class colossalai.nn.layer.parallel_2d.Classifier2D(in_features, num_classes, weight=None, bias=True, dtype=None, weight_initializer=<function kaiming_uniform_.<locals>.initializer>, bias_initializer=<function xavier_uniform_.<locals>.initializer>)[source]

Classifier for 2D parallelism.

Parameters

  • in_features (int) – size of each input sample.

  • num_classes (int) – number of classes.

  • weight (torch.nn.Parameter, optional) – weight of the classifier, defaults to None.

  • bias (bool, optional) – If set to False, the layer will not learn an additive bias, defaults to True.

  • dtype (torch.dtype, optional) – The dtype of parameters, defaults to None.

  • weight_initializer (typing.Callable, optional) – The initializer of weight, defaults to kaiming uniform initializer.

  • bias_initializer (typing.Callable, optional) – The initializer of bias, defaults to xavier uniform initializer.

More details about initializer please refer to init.

class colossalai.nn.layer.parallel_2d.PatchEmbedding2D(img_size, patch_size, in_chans, embed_size, flatten=True, dtype=None, weight_initializer=<function kaiming_uniform_.<locals>.initializer>, bias_initializer=<function xavier_uniform_.<locals>.initializer>, position_embed_initializer=<function zeros_.<locals>.initializer>)[source]

2D Image to Patch Embedding.

Parameters

  • img_size (int) – image size.

  • patch_size (int) – patch size.

  • in_chans (int) – number of channels of input image.

  • embed_size (int) – size of embedding.

  • dtype (torch.dtype, optional) – The dtype of parameters, defaults to None.

  • flatten (bool, optional) – whether to flatten output tensor, defaults to True.

  • weight_initializer (typing.Callable, optional) – The initializer of weight, defaults to kaiming uniform initializer.

  • bias_initializer (typing.Callable, optional) – The initializer of bias, defaults to xavier uniform initializer.

  • position_embed_initializer (typing.Callable, optional) – The initializer of position embedding, defaults to zeros initializer.

More details about initializer please refer to init.

class colossalai.nn.layer.parallel_2d.Embedding2D(num_embeddings, embedding_dim, padding_idx=None, dtype=None, weight_initializer=<function normal_.<locals>.initializer>, *args, **kwargs)[source]

Embedding for 2D parallelism.

Parameters

  • num_embeddings (int) – number of embeddings.

  • embedding_dim (int) – dimension of embedding.

  • padding_idx (int, optional) – If specified, the entries at padding_idx do not contribute to the gradient; therefore, the embedding vector at padding_idx is not updated during training, i.e. it remains as a fixed “pad”, defaults to None.

  • dtype (torch.dtype, optional) – The dtype of parameters, defaults to None.

  • weight_initializer (typing.Callable, optional) – he initializer of weight, defaults to normal initializer.

The args and kwargs used in :class:torch.nn.functional.embedding should contain:

max_norm (float, optional): If given, each embedding vector with norm larger than max_norm is
            renormalized to have norm max_norm. Note: this will modify weight in-place.
norm_type (float, optional): The p of the p-norm to compute for the max_norm option. Default 2.
scale_grad_by_freq (bool, optional): If given, this will scale gradients by the inverse
            of frequency of the words in the mini-batch. Default False.
sparse (bool, optional): If True, gradient w.r.t. weight will be a sparse tensor. Default False.

More details about args and kwargs could be found in Embedding.

More details about initializer please refer to init

class colossalai.nn.layer.parallel_2d.VocabParallelEmbedding2D(num_embeddings, embedding_dim, padding_idx=None, dtype=None, weight_initializer=<function normal_.<locals>.initializer>, *args, **kwargs)[source]

Embedding parallelized in the vocabulary dimension.

Parameters

  • num_embeddings (int) – number of embeddings.

  • embedding_dim (int) – dimension of embedding.

  • padding_idx (int, optional) – If specified, the entries at padding_idx do not contribute to the gradient; therefore, the embedding vector at padding_idx is not updated during training, i.e. it remains as a fixed “pad”, defaults to None.

  • dtype (torch.dtype, optional) – The dtype of parameters, defaults to None.

  • weight_initializer (typing.Callable, optional) – he initializer of weight, defaults to normal initializer.

The args and kwargs used in :class:torch.nn.functional.embedding should contain:

max_norm (float, optional): If given, each embedding vector with norm larger than max_norm is
            renormalized to have norm max_norm. Note: this will modify weight in-place.
norm_type (float, optional): The p of the p-norm to compute for the max_norm option. Default 2.
scale_grad_by_freq (bool, optional): If given, this will scale gradients by the inverse
            of frequency of the words in the mini-batch. Default False.
sparse (bool, optional): If True, gradient w.r.t. weight will be a sparse tensor. Default False.

More details about args and kwargs could be found in Embedding.

More details about initializer please refer to init.

class colossalai.nn.layer.parallel_2d.VocabParallelClassifier2D(in_features, num_classes, weight=None, bias=True, dtype=None, weight_initializer=<function kaiming_uniform_.<locals>.initializer>, bias_initializer=<function xavier_uniform_.<locals>.initializer>)[source]

Vocab parallel classifier layer for 2D parallelism.

Parameters

  • in_features (int) – size of each input sample.

  • num_classes (int) – number of classes.

  • weight (torch.nn.Parameter, optional) – weight of the classifier, defaults to None.

  • bias (bool, optional) – If set to False, the layer will not learn an additive bias, defaults to True.

  • dtype (torch.dtype, optional) – The dtype of parameters, defaults to None.

  • weight_initializer (typing.Callable, optional) – The initializer of weight, defaults to kaiming uniform initializer.

  • bias_initializer (typing.Callable, optional) – The initializer of bias, defaults to xavier uniform initializer.

More details about initializer please refer to init.