pointpillar相关的其它文章链接如下：
上一篇文章，我们梳理了数据增强和数据处理，并且得到了相应的pillar数据。下面我们继续讲pointpillar中的网络结构。
整体网络结构如下：
1. VFE
功能：这部分是简化版的pointnet网络，将经过数据增强和数据处理过后的pillar(N,4)数据，经过BN层、Relu激活层和max pool层得到(C, H, W)数据。
在VFE之前的data_dict的数据如下所示：
'''
batch_dict:
points:(N,5) --> (batch_index,x,y,z,r) batch_index代表了该点云数据在当前batch中的index
frame_id:(batch_size,) -->帧ID-->我们存放的是npy的绝对地址，batch_size个地址
gt_boxes:(batch_size,N,8)--> (x,y,z,dx,dy,dz,ry,class)，
use_lead_xyz:(batch_size,) --> (1,1,1,1)，batch_size个1
voxels:(M,32,4) --> (x,y,z,r)
voxel_coords:(M,4) --> (batch_index,z,y,x) batch_index代表了该点云数据在当前batch中的index
voxel_num_points:(M,):每个voxel内的点云
batch_size：batch_size大小
'''
随后经过VFE之后，就可以把原始的点云结构
（
N
∗
4
）
（N*4）
（N∗4）变换成了
(
D
，
P
，
N
)
(D，P，N)
(D，P，N)，其中 D代表了每个点云的特征维度，也就是每个点云10个特征(论文中只有9维)，P代表了所有非空的立方柱体，N代表了每个pillar中最多会有多少个点。具体操作以及说明如下：
- D ( x , y , z , x c , r , y c , z c , x p , y p , z p ) D(x,y,z, x_c ,r, y_c , z_c , x_p ,y_p,z_p) D(x,y,z,xc,r,yc,zc,xp,yp,zp)：xyz表示点云的真实坐标，下标c代表了每个点云到该点所对应pillar中所有点平均值的偏移量，下标p表示该点距离所在pillar中心点的偏移量。
- P：代表了所有非空的立方柱体，yaml配置中有最大值MAX_NUMBER_OF_VOXELS。
- N：代表了每个pillar中最多会有多少个点，实际操作取32。
得到 ( D ， P ， N ) (D，P，N) (D，P，N)的张量后，接下来这里使用了一个简化版的pointnet网络对点云的数据进行特征提取（即将这些点通过MLP升维，然后跟着BN层和Relu激活层），得到一个 ( C ， P ， N ) (C，P，N) (C，P，N)形状的张量，之后再使用max pooling操作提取每个pillar中最能代表该pillar的点。那么输出会变成 ( C ， P ， N ) − > ( C ， P ) − > ( C , H , W ) (C，P，N)->(C，P)->(C, H, W) (C，P，N)−>(C，P)−>(C,H,W)
这部分代码在：pcdet/models/backbones_3d/vfe/pillar_vfe.py，具体的注释代码如下：
import torch
import torch.nn as nn
import torch.nn.functional as F
from .vfe_template import VFETemplate
class PFNLayer(nn.Module):
def __init__(self,
in_channels,
out_channels,
use_norm=True,
last_layer=False):
super().__init__()
self.last_vfe = last_layer
self.use_norm = use_norm
if not self.last_vfe:
out_channels = out_channels // 2
# x的维度由（M, 32, 10）升维成了（M, 32, 64）,max pool之后32才去掉
if self.use_norm:
self.linear = nn.Linear(in_channels, out_channels, bias=False)
self.norm = nn.BatchNorm1d(out_channels, eps=1e-3, momentum=0.01)
else:
self.linear = nn.Linear(in_channels, out_channels, bias=True)
self.part = 50000
def forward(self, inputs):
if inputs.shape[0] > self.part:
# nn.Linear performs randomly when batch size is too large
num_parts = inputs.shape[0] // self.part
part_linear_out = [self.linear(inputs[num_part*self.part:(num_part+1)*self.part])
for num_part in range(num_parts+1)]
x = torch.cat(part_linear_out, dim=0)
else:
x = self.linear(inputs)
torch.backends.cudnn.enabled = False
#permute变换维度，(M, 64, 32) --> (M, 32, 64)
# 这里之所以变换维度，是因为BatchNorm1d在通道维度上进行,对于图像来说默认模式为[N,C,H*W],通道在第二个维度上
x = self.norm(x.permute(0, 2, 1)).permute(0, 2, 1) if self.use_norm else x
torch.backends.cudnn.enabled = True
x = F.relu(x)
# 完成pointnet的最大池化操作，找出每个pillar中最能代表该pillar的点
x_max = torch.max(x, dim=1, keepdim=True)[0]
if self.last_vfe:
return x_max
else:
x_repeat = x_max.repeat(1, inputs.shape[1], 1)
x_concatenated = torch.cat([x, x_repeat], dim=2)
return x_concatenated
class PillarVFE(VFETemplate):
def __init__(self, model_cfg, num_point_features, voxel_size, point_cloud_range, **kwargs):
super().__init__(model_cfg=model_cfg)
self.use_norm = self.model_cfg.USE_NORM
self.with_distance = self.model_cfg.WITH_DISTANCE
self.use_absolute_xyz = self.model_cfg.USE_ABSLOTE_XYZ
# num_point_features:10
num_point_features += 6 if self.use_absolute_xyz else 3
if self.with_distance:
num_point_features += 1
#[64]
self.num_filters = self.model_cfg.NUM_FILTERS
assert len(self.num_filters) > 0
# num_filters: [10, 64]
num_filters = [num_point_features] + list(self.num_filters)
pfn_layers = []
#len(num_filters) - 1 == 1
for i in range(len(num_filters) - 1):
in_filters = num_filters[i] # 10
out_filters = num_filters[i + 1] # 64
pfn_layers.append(
PFNLayer(in_filters, out_filters, self.use_norm, last_layer=(i >= len(num_filters) - 2))
)
#收集PFN层，在forward中执行
self.pfn_layers = nn.ModuleList(pfn_layers)
self.voxel_x = voxel_size[0]
self.voxel_y = voxel_size[1]
self.voxel_z = voxel_size[2]
self.x_offset = self.voxel_x / 2 + point_cloud_range[0]
self.y_offset = self.voxel_y / 2 + point_cloud_range[1]
self.z_offset = self.voxel_z / 2 + point_cloud_range[2]
def get_output_feature_dim(self):
return self.num_filters[-1]
def get_paddings_indicator(self, actual_num, max_num, axis=0):
'''
指出一个pillar中哪些是真实数据,哪些是填充的0数据
'''
actual_num = torch.unsqueeze(actual_num, axis + 1)
max_num_shape = [1] * len(actual_num.shape)
max_num_shape[axis + 1] = -1
max_num = torch.arange(max_num, dtype=torch.int, device=actual_num.device).view(max_num_shape)
paddings_indicator = actual_num.int() > max_num
return paddings_indicator
def forward(self, batch_dict, **kwargs):
'''
batch_dict:
points:(N,5) --> (batch_index,x,y,z,r) batch_index代表了该点云数据在当前batch中的index
frame_id:(batch_size,) -->帧ID-->我们存放的是npy的绝对地址，batch_size个地址
gt_boxes:(batch_size,N,8)--> (x,y,z,dx,dy,dz,ry,class)，
use_lead_xyz:(batch_size,) --> (1,1,1,1)，batch_size个1
voxels:(M,32,4) --> (x,y,z,r)
voxel_coords:(M,4) --> (batch_index,z,y,x) batch_index代表了该点云数据在当前batch中的index
voxel_num_points:(M,):每个voxel内的点云
batch_size:4：batch_size大小
'''
voxel_features, voxel_num_points, coords = batch_dict['voxels'], batch_dict['voxel_num_points'], batch_dict['voxel_coords']
#求每个pillar中所有点云的平均值,设置keepdim=True的，则保留原来的维度信息
points_mean = voxel_features[:, :, :3].sum(dim=1, keepdim=True) / voxel_num_points.type_as(voxel_features).view(-1, 1, 1)
#每个点云数据减去该点对应pillar的平均值，得到差值 xc,yc,zc
f_cluster = voxel_features[:, :, :3] - points_mean
# 创建每个点云到该pillar的坐标中心点偏移量空数据 xp,yp,zp
f_center = torch.zeros_like(voxel_features[:, :, :3])
'''
coords是每个网格点的坐标，即[432, 496, 1]，需要乘以每个pillar的长宽得到点云数据中实际的长宽（单位米）
同时为了获得每个pillar的中心点坐标，还需要加上每个pillar长宽的一半得到中心点坐标
每个点的x、y、z减去对应pillar的坐标中心点，得到每个点到该点中心点的偏移量
'''
f_center[:, :, 0] = voxel_features[:, :, 0] - (coords[:, 3].to(voxel_features.dtype).unsqueeze(1) * self.voxel_x + self.x_offset)
f_center[:, :, 1] = voxel_features[:, :, 1] - (coords[:, 2].to(voxel_features.dtype).unsqueeze(1) * self.voxel_y + self.y_offset)
f_center[:, :, 2] = voxel_features[:, :, 2] - (coords[:, 1].to(voxel_features.dtype).unsqueeze(1) * self.voxel_z + self.z_offset)
#配置中使用了绝对坐标，直接组合即可。
if self.use_absolute_xyz:
features = [voxel_features, f_cluster, f_center] #10个特征，直接组合
else:
features = [voxel_features[..., 3:], f_cluster, f_center]
#距离信息，False
if self.with_distance:
points_dist = torch.norm(voxel_features[:, :, :3], 2, 2, keepdim=True)
features.append(points_dist)
features = torch.cat(features, dim=-1)
voxel_count = features.shape[1]
#mask中指明了每个pillar中哪些是需要被保留的数据
mask = self.get_paddings_indicator(voxel_num_points, voxel_count, axis=0)
mask = torch.unsqueeze(mask, -1).type_as(voxel_features)
#由0填充的数据，在计算出现xc,yc,zc和xp,yp,zp时会有值
#features中去掉0值信息。
features *= mask
#执行上面收集的PFN层，每个pillar抽象出64维特征
for pfn in self.pfn_layers:
features = pfn(features)
features = features.squeeze()
batch_dict['pillar_features'] = features
return batch_dict
2. MAP_TO_BEV
功能：将得到的pillar数据，投影至二维坐标中。
在经过简化版的pointnet网络提取出每个pillar的特征信息后，就需要将每个的pillar数据重新放回原来的坐标中，也就是二维坐标，组成 伪图像 数据。
对应到论文中就是stacked pillars，将生成的pillar按照坐标索引还原到原空间中。
这部分代码在：pcdet/models/backbones_2d/map_to_bev/pointpillar_scatter.py，具体的注释代码如下：
import torch
import torch.nn as nn
class PointPillarScatter(nn.Module):
def __init__(self, model_cfg, grid_size, **kwargs):
super().__init__()
self.model_cfg = model_cfg
self.num_bev_features = self.model_cfg.NUM_BEV_FEATURES #64
self.nx, self.ny, self.nz = grid_size # [432,496,1]
assert self.nz == 1
def forward(self, batch_dict, **kwargs):
'''
batch_dict['pillar_features']-->为VFE得到的数据(M, 64)
voxel_coords:(M,4) --> (batch_index,z,y,x) batch_index代表了该点云数据在当前batch中的index
'''
pillar_features, coords = batch_dict['pillar_features'], batch_dict['voxel_coords']
batch_spatial_features = []
# 根据batch_index，获取batch_size大小
batch_size = coords[:, 0].max().int().item() + 1
for batch_idx in range(batch_size):
# 创建一个空间坐标所有用来接受pillar中的数据
# spatial_feature 维度 (64,214272)
spatial_feature = torch.zeros(
self.num_bev_features,
self.nz * self.nx * self.ny,
dtype=pillar_features.dtype,
device=pillar_features.device)
batch_mask = coords[:, 0] == batch_idx #返回mask，[True, False...]
this_coords = coords[batch_mask, :] #获取当前的batch_idx的数
#计算pillar的索引，该点之前所有行的点总和加上该点所在的列即可
indices = this_coords[:, 1] + this_coords[:, 2] * self.nx + this_coords[:, 3]
indices = indices.type(torch.long) # 转换数据类型
pillars = pillar_features[batch_mask, :]
pillars = pillars.t()
# 在索引位置填充pillars
spatial_feature[:, indices] = pillars
# 将空间特征加入list,每个元素为(64, 214272)
batch_spatial_features.append(spatial_feature)
# 在第0个维度将所有的数据堆叠在一起
batch_spatial_features = torch.stack(batch_spatial_features, 0)
# reshape回原空间(伪图像) （4, 64, 214272）--> (4, 64, 496, 432)
batch_spatial_features = batch_spatial_features.view(batch_size, self.num_bev_features * self.nz, self.ny, self.nx)
batch_dict['spatial_features'] = batch_spatial_features
#返回数据
return batch_dict
3. BACKBONE_2D
功能：骨干网络，提取特征
经过上面的映射操作，将原来的pillar提取最大的数值后放回到相应的坐标后，就可以得到类似于图像的数据了；只有在有pillar非空的坐标处有提取的点云数据，其余地方都是0数据，所以得到的一个（batch_size，64, 432, 496）的张量还是很稀疏的。
BACKBONE_2D的输入特征维度（batch_size，64, 432, 496），输出的特征维度为[batch_size, 384, 248, 216]。
需要说明的是，主干网络构建了下采样和上采样网络，分别为加入到了blocks和deblocks中，上采样和下采样的具体操作可查看下列代码和注释。
这部分代码在：pcdet/models/backbones_2d/base_bev_backbone.py，具体的注释代码如下：
import numpy as np
import torch
import torch.nn as nn
class BaseBEVBackbone(nn.Module):
# input_channels = 64
def __init__(self, model_cfg, input_channels):
super().__init__()
self.model_cfg = model_cfg
# 层参数
if self.model_cfg.get('LAYER_NUMS', None) is not None:
assert len(self.model_cfg.LAYER_NUMS) == len(self.model_cfg.LAYER_STRIDES) == len(self.model_cfg.NUM_FILTERS)
layer_nums = self.model_cfg.LAYER_NUMS #[3, 5, 5]
layer_strides = self.model_cfg.LAYER_STRIDES #[2, 2, 2]
num_filters = self.model_cfg.NUM_FILTERS # [64, 128, 256]
else:
layer_nums = layer_strides = num_filters = []
# 上采样参数
if self.model_cfg.get('UPSAMPLE_STRIDES', None) is not None:
assert len(self.model_cfg.UPSAMPLE_STRIDES) == len(self.model_cfg.NUM_UPSAMPLE_FILTERS)
num_upsample_filters = self.model_cfg.NUM_UPSAMPLE_FILTERS #[128, 128, 128]
upsample_strides = self.model_cfg.UPSAMPLE_STRIDES # [1, 2, 4]
else:
upsample_strides = num_upsample_filters = []
num_levels = len(layer_nums) #3层
c_in_list = [input_channels, *num_filters[:-1]] # [64, 64, 128]，(*表示取值，在列表中使用，num_filters取除了最后一个元素)
self.blocks = nn.ModuleList()
self.deblocks = nn.ModuleList()
# 开始处理3层网络
# 通道数分别为：# (64,64)-->(64,128)-->(128,256)
for idx in range(num_levels):
#需要说明的是，经过这里的层，feature map变小, (h - kernel + 2p )/s + 1
cur_layers = [
nn.ZeroPad2d(1),
#layer_strides：[2, 2, 2]
nn.Conv2d(
c_in_list[idx], num_filters[idx], kernel_size=3,
stride=layer_strides[idx], padding=0, bias=False
),
nn.BatchNorm2d(num_filters[idx], eps=1e-3, momentum=0.01),
nn.ReLU()
]
# 根据layer_nums堆叠卷积层，网络层分别为[3, 5, 5]，输入输出的通道数不变
#需要说明的是，经过这里的层，feature map大小不变
for k in range(layer_nums[idx]):
cur_layers.extend([
nn.Conv2d(num_filters[idx], num_filters[idx], kernel_size=3, padding=1, bias=False),
nn.BatchNorm2d(num_filters[idx], eps=1e-3, momentum=0.01),
nn.ReLU()
])
#添加下采样层
self.blocks.append(nn.Sequential(*cur_layers))
#上采样，upsample_strides：[1, 2, 4]
if len(upsample_strides) > 0:
stride = upsample_strides[idx]
if stride >= 1:
# ConvTranspose2d：逆卷积/转置卷积，包含两步先对原 tensor 进行 上采样，然后再进行一次常规的卷积操作
#上采样，每两行(列)中插入（stride-1）个零行（列）
#卷积操作：以输入的kernel_size, 步长始终为1（和输入的stride的值无关）。
# num_upsample_filters：输出通道[128, 128, 128]
self.deblocks.append(nn.Sequential(
nn.ConvTranspose2d(
num_filters[idx], num_upsample_filters[idx],
upsample_strides[idx],
stride=upsample_strides[idx], bias=False
),
nn.BatchNorm2d(num_upsample_filters[idx], eps=1e-3, momentum=0.01),
nn.ReLU()
))
else:
stride = np.round(1 / stride).astype(np.int)
self.deblocks.append(nn.Sequential(
nn.Conv2d(
num_filters[idx], num_upsample_filters[idx],
stride,
stride=stride, bias=False
),
nn.BatchNorm2d(num_upsample_filters[idx], eps=1e-3, momentum=0.01),
nn.ReLU()
))
c_in = sum(num_upsample_filters) # 128 + 128 + 128 = 384
#下面不执行：3 > 3 False
if len(upsample_strides) > num_levels:
self.deblocks.append(nn.Sequential(
nn.ConvTranspose2d(c_in, c_in, upsample_strides[-1], stride=upsample_strides[-1], bias=False),
nn.BatchNorm2d(c_in, eps=1e-3, momentum=0.01),
nn.ReLU(),
))
#输出特征384
self.num_bev_features = c_in
def forward(self, data_dict):
"""
Args:
data_dict:
spatial_features:(batch_size, 64, 496, 432)
Returns:
"""
spatial_features = data_dict['spatial_features']
ups = []
ret_dict = {}
x = spatial_features #输入维度：[batch_size, 64, 496, 432]
for i in range(len(self.blocks)):
#下采样
x = self.blocks[i](x)
#下采样之后，x的shape分别为：torch.Size([batch_size, 64, 248, 216])，torch.Size([batch_size, 128, 124, 108])，torch.Size([batch_size, 256, 62, 54])
#spatial_features的shape一直为：torch.Size([batch_size, 64, 496, 432])
stride = int(spatial_features.shape[2] / x.shape[2]) #三次分别为2，4，8
ret_dict['spatial_features_%dx' % stride] = x
#上采样
#上采样不影响x的值，上采样后的值在ups中，ups中的元素维度都是：torch.Size([batch_size, 128, 248, 216])
if len(self.deblocks) > 0:
ups.append(self.deblocks[i](x))
else:
ups.append(x)
if len(ups) > 1:
x = torch.cat(ups, dim=1)
elif len(ups) == 1:
x = ups[0]
if len(self.deblocks) > len(self.blocks):
x = self.deblocks[-1](x)
#保存结果
#输出维度：[batch_size, 384, 248, 216]，特征更多，尺度减小为原先的一半
data_dict['spatial_features_2d'] = x
return data_dict
4. DENSE_HEAD
功能：
一共有三个类别的先验框，每个先验框都有 两个方向，分别是BEV视角下的0度和90度，每个类别的先验框只有一种尺度信息；分别是车 [3.9, 1.6, 1.56]、人[0.8, 0.6, 1.73]、自行车[1.76, 0.6, 1.73]（单位：米）。其中Car的先验框如下所示：
{
'class_name': 'Car', #类别
'anchor_sizes': [[3.9, 1.6, 1.56]], #先验框的尺寸
'anchor_rotations': [0, 1.57], #两种角度
'anchor_bottom_heights': [-1.78], #先验框最低点高度
'align_center': False,
'feature_map_stride': 2,
'matched_threshold': 0.6,#iou，正样本阈值
'unmatched_threshold': 0.45#iou，负样本阈值，两个阈值中间的不计算损失
},
在anchor匹配GT的过程中，使用的是 2D IOU 匹配方式，直接从生成的特征图也就是BEV视角进行匹配，没有考虑高度的信息。
每个anchor都需要预测7个参数，分别是 ( x , y , z , w , l , h , θ , c l s ) (x, y, z, w, l, h, θ, cls) (x,y,z,w,l,h,θ,cls)。
- x , y , z x, y, z x,y,z预测一个anchor的中心坐标在点云中的位置；
- w ， l ， h w，l，h w，l，h分别预测了一个anchor的长宽高数据;
- θ θ θ预测了box的旋转角度。 同时，因为在角度预测时候很难区分两个完全相反的box，所以PiontPillars的检测头中还添加了对一个anchor的方向预测，这里使用了一个
基于softmax的方向分类box的两个朝向信息。- cls为预测的类别。
这部分代码在：pcdet/models/dense_heads/anchor_head_single.py，具体的注释代码如下：
import numpy as np
import torch.nn as nn
from .anchor_head_template import AnchorHeadTemplate
class AnchorHeadSingle(AnchorHeadTemplate):
'''
Args:
model_cfg: AnchorHeadSingle的配置
input_channels: 384 输入通道数
num_class: 3
class_names: ['Car','Pedestrian','Cyclist']
grid_size: (432, 496, 1)
point_cloud_range: (0, -39.68, -3, 69.12, 39.68, 1)
predict_boxes_when_training: False
'''
def __init__(self, model_cfg, input_channels, num_class, class_names, grid_size, point_cloud_range,
predict_boxes_when_training=True, **kwargs):
super().__init__(
model_cfg=model_cfg, num_class=num_class, class_names=class_names, grid_size=grid_size, point_cloud_range=point_cloud_range,
predict_boxes_when_training=predict_boxes_when_training
)
#在父类中调用generate_anchors中生成anchors和num_anchors_per_location
# 每个点会生成不同类别的2个先验框(anchor)，也就是说num_anchors_per_location：[2, 2, 2,]-》3类，每类2个anchor
#所以每个点生成6个先验框(anchor)
self.num_anchors_per_location = sum(self.num_anchors_per_location)
#类别， 1x1 卷积：conv_cls: Conv2d(384, 18, kernel_size=(1, 1), stride=(1, 1))
#每个点6个anchor，每个anchor预测3个类别，所以输出的类别为6*3
self.conv_cls = nn.Conv2d(
input_channels, self.num_anchors_per_location * self.num_class,
kernel_size=1
)
#box，1x1 卷积：conv_box: Conv2d(384, 42, kernel_size=(1, 1), stride=(1, 1))
#每个点6个anchor，每个anchor预测7个值（x, y, z, w, l, h, θ），所以输出的值为6*7
self.conv_box = nn.Conv2d(
input_channels, self.num_anchors_per_location * self.box_coder.code_size, #self.box_coder.code_size默认为7
kernel_size=1
)
# 是否使用方向分类，1x1 卷积：conv_dir: Conv2d(384, 12, kernel_size=(1, 1), stride=(1, 1))
#每个点6个anchor，每个anchor预测2个方向(正负)，所以输出的值为6*2
if self.model_cfg.get('USE_DIRECTION_CLASSIFIER', None) is not None:
self.conv_dir_cls = nn.Conv2d(
input_channels,
self.num_anchors_per_location * self.model_cfg.NUM_DIR_BINS,
kernel_size=1
)
else:
self.conv_dir_cls = None
self.init_weights()
#初始化参数
def init_weights(self):
pi = 0.01
nn.init.constant_(self.conv_cls.bias, -np.log((1 - pi) / pi))
nn.init.normal_(self.conv_box.weight, mean=0, std=0.001)
def forward(self, data_dict):
# spatial_features_2d 维度 ：（batch_size, 384, 248, 216）
spatial_features_2d = data_dict['spatial_features_2d']
#cls_preds的维度为：torch.Size([batch_size, 18, 248, 216])
#每个点6个anchor，每个anchor预测3个类别，所以输出的类别为6*3
cls_preds = self.conv_cls(spatial_features_2d)
#box_preds的维度为：torch.Size([batch_size, 42, 248, 216])
#每个点6个anchor，每个anchor预测7个值（x, y, z, w, l, h, θ），所以输出的值为6*7
box_preds = self.conv_box(spatial_features_2d)
#调整顺序
cls_preds = cls_preds.permute(0, 2, 3, 1).contiguous() # [N, H, W, C]
box_preds = box_preds.permute(0, 2, 3, 1).contiguous() # [N, H, W, C]
self.forward_ret_dict['cls_preds'] = cls_preds
self.forward_ret_dict['box_preds'] = box_preds
#方向预测，也就是正负预测
if self.conv_dir_cls is not None:
#dir_cls_preds的维度为：torch.Size([batch_size, 12, 248, 216])
#每个点6个anchor，每个anchor预测2个方向(正负)，所以输出的值为6*2
dir_cls_preds = self.conv_dir_cls(spatial_features_2d)
#调整顺序
dir_cls_preds = dir_cls_preds.permute(0, 2, 3, 1).contiguous()
self.forward_ret_dict['dir_cls_preds'] = dir_cls_preds
else:
dir_cls_preds = None
#如果是在训练模式的时候，需要对每个先验框分配GT来计算loss
if self.training:
targets_dict = self.assign_targets(
gt_boxes=data_dict['gt_boxes']
)
#分配的
self.forward_ret_dict.update(targets_dict)
#非训练模式，则直接生成进行box的预测
if not self.training or self.predict_boxes_when_training:
batch_cls_preds, batch_box_preds = self.generate_predicted_boxes(
batch_size=data_dict['batch_size'],
cls_preds=cls_preds, box_preds=box_preds, dir_cls_preds=dir_cls_preds
)
data_dict['batch_cls_preds'] = batch_cls_preds
data_dict['batch_box_preds'] = batch_box_preds
data_dict['cls_preds_normalized'] = False
return data_dict
3万+
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
