Note

You are reading the documentation for MMSelfSup 0.x, which will soon be deprecated by the end of 2022. We recommend you upgrade to MMSelfSup 1.0.0rc versions to enjoy fruitful new features and better performance brought by OpenMMLab 2.0. Check out the changelog, code and documentation of MMSelfSup 1.0.0rc for more details.

Source code for mmselfsup.models.utils.position_embedding

# Copyright (c) OpenMMLab. All rights reserved.
import torch


def build_2d_sincos_position_embedding(patches_resolution,
                                       embed_dims,
                                       temperature=10000.,
                                       cls_token=False):
    """The function is to build position embedding for model to obtain the
    position information of the image patches."""

    if isinstance(patches_resolution, int):
        patches_resolution = (patches_resolution, patches_resolution)

    h, w = patches_resolution
    grid_w = torch.arange(w, dtype=torch.float32)
    grid_h = torch.arange(h, dtype=torch.float32)
    grid_w, grid_h = torch.meshgrid(grid_w, grid_h)
    assert embed_dims % 4 == 0, \
        'Embed dimension must be divisible by 4.'
    pos_dim = embed_dims // 4

    omega = torch.arange(pos_dim, dtype=torch.float32) / pos_dim
    omega = 1. / (temperature**omega)
    out_w = torch.einsum('m,d->md', [grid_w.flatten(), omega])
    out_h = torch.einsum('m,d->md', [grid_h.flatten(), omega])

    pos_emb = torch.cat(
        [
            torch.sin(out_w),
            torch.cos(out_w),
            torch.sin(out_h),
            torch.cos(out_h)
        ],
        dim=1,
    )[None, :, :]

    if cls_token:
        cls_token_pe = torch.zeros([1, 1, embed_dims], dtype=torch.float32)
        pos_emb = torch.cat([cls_token_pe, pos_emb], dim=1)

    return pos_emb