mirror of
https://github.com/open-mmlab/mmocr.git
synced 2025-06-03 21:54:47 +08:00
2d743cfa19
* [Feature] Add RepeatAugSampler * initial commit * spts inference done * merge repeat_aug (bug in multi-node?) * fix inference * train done * rm readme * Revert "merge repeat_aug (bug in multi-node?)" This reverts commit 393506a97cbe6d75ad1f28611ea10eba6b8fa4b3. * Revert "[Feature] Add RepeatAugSampler" This reverts commit 2089b02b4844157670033766f257b5d1bca452ce. * remove utils * readme & conversion script * update readme * fix * optimize * rename cfg & del compose * fix * fix
56 lines
1.9 KiB
Python
56 lines
1.9 KiB
Python
import math
|
|
|
|
import torch
|
|
from torch import Tensor, nn
|
|
|
|
|
|
class PositionEmbeddingSine(nn.Module):
|
|
"""This is a more standard version of the position embedding, very similar
|
|
to the one used by the Attention is all you need paper, generalized to work
|
|
on images.
|
|
|
|
Adapted from https://github.com/shannanyinxiang/SPTS.
|
|
"""
|
|
|
|
def __init__(self,
|
|
num_pos_feats=64,
|
|
temperature=10000,
|
|
normalize=True,
|
|
scale=None):
|
|
super().__init__()
|
|
self.num_pos_feats = num_pos_feats
|
|
self.temperature = temperature
|
|
self.normalize = normalize
|
|
if scale is not None and normalize is False:
|
|
raise ValueError('normalize should be True if scale is passed')
|
|
if scale is None:
|
|
scale = 2 * math.pi
|
|
self.scale = scale
|
|
|
|
def forward(self, mask: Tensor):
|
|
assert mask is not None
|
|
not_mask = ~mask
|
|
y_embed = not_mask.cumsum(1, dtype=torch.float32)
|
|
x_embed = not_mask.cumsum(2, dtype=torch.float32)
|
|
if self.normalize:
|
|
eps = 1e-6
|
|
y_embed = y_embed / (y_embed[:, -1:, :] + eps) * self.scale
|
|
x_embed = x_embed / (x_embed[:, :, -1:] + eps) * self.scale
|
|
|
|
dim_t = torch.arange(
|
|
self.num_pos_feats, dtype=torch.float32, device=mask.device)
|
|
dim_t = self.temperature**(2 *
|
|
torch.div(dim_t, 2, rounding_mode='floor') /
|
|
self.num_pos_feats)
|
|
|
|
pos_x = x_embed[:, :, :, None] / dim_t
|
|
pos_y = y_embed[:, :, :, None] / dim_t
|
|
pos_x = torch.stack(
|
|
(pos_x[:, :, :, 0::2].sin(), pos_x[:, :, :, 1::2].cos()),
|
|
dim=4).flatten(3)
|
|
pos_y = torch.stack(
|
|
(pos_y[:, :, :, 0::2].sin(), pos_y[:, :, :, 1::2].cos()),
|
|
dim=4).flatten(3)
|
|
pos = torch.cat((pos_y, pos_x), dim=3).permute(0, 3, 1, 2)
|
|
return pos
|