vis4d.op.detect3d.bevformer.temporal_self_attention

An attention module used in BEVFormer based on Deformable-Detr.

Classes

TemporalSelfAttention([embed_dims, ...])

Temperal Self Attention.
class TemporalSelfAttention(embed_dims=256, num_heads=8, num_levels=4, num_points=4, num_bev_queue=2, im2col_step=64, dropout=0.1, batch_first=True)[source]

Temperal Self Attention.

Init.

Parameters:

  • embed_dims (int) – The embedding dimension of Attention. Default: 256.

  • num_heads (int) – Parallel attention heads. Default: 64.

  • num_levels (int) – The number of feature map used in Attention. Default: 4.

  • num_points (int) – The number of sampling points for each query in each head. Default: 4.

  • num_bev_queue (int) – In this version, we only use one history BEV and one currenct BEV. The length of BEV queue is 2.

  • im2col_step (int) – The step used in image_to_column. Default: 64.

  • dropout (float) – A Dropout layer on inp_identity. Default: 0.1.

  • batch_first (bool) – Key, Query and Value are shape of (batch, n, embed_dim) or (n, batch, embed_dim). Default to True.

forward(query, reference_points, value, spatial_shapes, level_start_index, key_padding_mask=None, identity=None, query_pos=None)[source]

Forward Function of MultiScaleDeformAttention.

Parameters:

  • query (Tensor) – Query of Transformer with shape (num_query, bs, embed_dims).

  • reference_points (Tensor) – The normalized reference points with shape (bs, num_query, num_levels, 2), all elements is range in [0, 1], top-left (0,0), bottom-right (1, 1), including padding area. or (N, Length_{query}, num_levels, 4), add additional two dimensions is (w, h) to form reference boxes.

  • value (Tensor) – The value tensor with shape (num_key, bs, embed_dims).

  • spatial_shapes (Tensor) – Spatial shape of features in different levels. With shape (num_levels, 2), last dimension represents (h, w).

  • level_start_index (Tensor) – The start index of each level. A tensor has shape (num_levels, ) and can be represented as [0, h_0*w_0, h_0*w_0+h_1*w_1, …].

  • key_padding_mask (Tensor) – ByteTensor for value, with shape [bs, num_key].

  • identity (Tensor) – The tensor used for addition, with the same shape as query. Default None. If None, query will be used.

  • query_pos (Tensor, optional) – The positional encoding for query. Default: None.

Returns:

forwarded results with shape [num_query, bs, embed_dims].

Return type:

Tensor

init_weights()[source]

Default initialization for Parameters of Module.

Return type:

None