embedding_bag(input, weight, offsets=None, max_norm=None, norm_type=2, scale_grad_by_freq=False, mode='mean', sparse=False, per_sample_weights=None, include_last_offset=False, padding_idx=None)¶
Computes sums, means or maxes of bags of embeddings, without instantiating the intermediate embeddings.
torch.nn.EmbeddingBagfor more details.
This operation may produce nondeterministic gradients when given tensors on a CUDA device. See Reproducibility for more information.
input (LongTensor) – Tensor containing bags of indices into the embedding matrix
weight (Tensor) – The embedding matrix with number of rows equal to the maximum possible index + 1, and number of columns equal to the embedding size
offsets (LongTensor, optional) – Only used when
offsetsdetermines the starting index position of each bag (sequence) in
max_norm (float, optional) – If given, each embedding vector with norm larger than
max_normis renormalized to have norm
max_norm. Note: this will modify
norm_type (float, optional) – The
p-norm to compute for the
scale_grad_by_freq (boolean, optional) – if given, this will scale gradients by the inverse of frequency of the words in the mini-batch. Default
False. Note: this option is not supported when
mode (string, optional) –
"max". Specifies the way to reduce the bag. Default:
sparse (bool, optional) – if
True, gradient w.r.t.
weightwill be a sparse tensor. See Notes under
torch.nn.Embeddingfor more details regarding sparse gradients. Note: this option is not supported when
per_sample_weights (Tensor, optional) – a tensor of float / double weights, or None to indicate all weights should be taken to be 1. If specified,
per_sample_weightsmust have exactly the same shape as input and is treated as having the same
offsets, if those are not None.
include_last_offset (bool, optional) – if
True, the size of offsets is equal to the number of bags + 1. The last element is the size of the input, or the ending index position of the last bag (sequence).
padding_idx (int, optional) – If specified, the entries at
padding_idxdo not contribute to the gradient; therefore, the embedding vector at
padding_idxis not updated during training, i.e. it remains as a fixed “pad”. Note that the embedding vector at
padding_idxis excluded from the reduction.
inputis 2D of shape (B, N), it will be treated as
Bbags (sequences) each of fixed length
N, and this will return
Bvalues aggregated in a way depending on the
offsetsis ignored and required to be
Nonein this case.
inputis 1D of shape (N), it will be treated as a concatenation of multiple bags (sequences).
offsetsis required to be a 1D tensor containing the starting index positions of each bag in
input. Therefore, for
offsetsof shape (B),
inputwill be viewed as having
Bbags. Empty bags (i.e., having 0-length) will have returned vectors filled by zeros.
weight(Tensor): the learnable weights of the module of shape (num_embeddings, embedding_dim)
per_sample_weights(Tensor, optional). Has the same shape as
output: aggregated embedding values of shape (B, embedding_dim)
>>> # an Embedding module containing 10 tensors of size 3 >>> embedding_matrix = torch.rand(10, 3) >>> # a batch of 2 samples of 4 indices each >>> input = torch.tensor([1,2,4,5,4,3,2,9]) >>> offsets = torch.tensor([0,4]) >>> F.embedding_bag(input, embedding_matrix, offsets) tensor([[ 0.3397, 0.3552, 0.5545], [ 0.5893, 0.4386, 0.5882]]) >>> # example with padding_idx >>> embedding_matrix = torch.rand(10, 3) >>> input = torch.tensor([2, 2, 2, 2, 4, 3, 2, 9]) >>> offsets = torch.tensor([0,4]) >>> F.embedding_bag(input, embedding_matrix, offsets, padding_idx=2, mode='sum') tensor([[ 0.0000, 0.0000, 0.0000], [-0.7082, 3.2145, -2.6251]])