# Copyright (c) Meta Platforms, Inc. and affiliates. # SPDX-License-Identifier: LicenseRef-SAM-License # From SAM 3D Body (https://github.com/facebookresearch/sam-3d-body), modified for FAB; licence: third_party_licenses/SAM-License.txt import torch import torch.nn as nn def drop_path( x: torch.Tensor, drop_prob: float = 0.0, training: bool = False ) -> torch.Tensor: """Drop paths (Stochastic Depth) per sample (when applied in main path of residual blocks). We follow the implementation https://github.com/rwightman/pytorch-image-models/blob/a2727c1bf78ba0d7b5727f5f95e37fb7f8866b1f/timm/models/layers/drop.py # noqa: E501 """ if not training: return x keep_prob = 1 - drop_prob # handle tensors with different dimensions, not just 4D tensors. shape = (x.shape[0],) + (1,) * (x.ndim - 1) random_tensor = keep_prob + torch.rand(shape, dtype=x.dtype, device=x.device) output = x.div(keep_prob) * random_tensor.floor() return output class DropPath(nn.Module): """Drop paths (Stochastic Depth) per sample (when applied in main path of residual blocks). We follow the implementation https://github.com/rwightman/pytorch-image-models/blob/a2727c1bf78ba0d7b5727f5f95e37fb7f8866b1f/timm/models/layers/drop.py # noqa: E501 Args: drop_prob (float): Probability of the path to be zeroed. Default: 0.1 """ def __init__(self, drop_prob: float = 0.1): super().__init__() self.drop_prob = drop_prob def forward(self, x: torch.Tensor) -> torch.Tensor: return drop_path(x, self.drop_prob, self.training)