This commit is contained in:
Glenn Jocher 2019-08-11 15:17:40 +02:00
parent 05914799fa
commit 636c1cff7a
1 changed files with 24 additions and 24 deletions

View File

@ -4,7 +4,7 @@ from pathlib import Path
import torch.nn.functional as F import torch.nn.functional as F
ONNX_EXPORT = False ONNX_EXPORT = True
def create_modules(module_defs): def create_modules(module_defs):
@ -34,7 +34,7 @@ def create_modules(module_defs):
modules.add_module('BatchNorm2d', nn.BatchNorm2d(filters, momentum=0.1)) modules.add_module('BatchNorm2d', nn.BatchNorm2d(filters, momentum=0.1))
if mdef['activation'] == 'leaky': # TODO: activation study https://github.com/ultralytics/yolov3/issues/441 if mdef['activation'] == 'leaky': # TODO: activation study https://github.com/ultralytics/yolov3/issues/441
modules.add_module('activation', nn.LeakyReLU(0.1, inplace=True)) modules.add_module('activation', nn.LeakyReLU(0.1, inplace=True))
# modules.add_module('activation', nn.PReLU(num_parameters=1)) # modules.add_module('activation', nn.PReLU(num_parameters=1, init=0.10))
# modules.add_module('activation', Swish()) # modules.add_module('activation', Swish())
elif mdef['type'] == 'maxpool': elif mdef['type'] == 'maxpool':
@ -105,7 +105,7 @@ class YOLOLayer(nn.Module):
stride = [32, 16, 8][yolo_index] # stride of this layer stride = [32, 16, 8][yolo_index] # stride of this layer
nx = int(img_size[1] / stride) # number x grid points nx = int(img_size[1] / stride) # number x grid points
ny = int(img_size[0] / stride) # number y grid points ny = int(img_size[0] / stride) # number y grid points
create_grids(self, max(img_size), (nx, ny)) create_grids(self, img_size, (nx, ny))
def forward(self, p, img_size, var=None): def forward(self, p, img_size, var=None):
if ONNX_EXPORT: if ONNX_EXPORT:
@ -127,33 +127,33 @@ class YOLOLayer(nn.Module):
grid_xy = self.grid_xy.repeat((1, self.na, 1, 1, 1)).view((1, -1, 2)) grid_xy = self.grid_xy.repeat((1, self.na, 1, 1, 1)).view((1, -1, 2))
anchor_wh = self.anchor_wh.repeat((1, 1, self.nx, self.ny, 1)).view((1, -1, 2)) / ngu anchor_wh = self.anchor_wh.repeat((1, 1, self.nx, self.ny, 1)).view((1, -1, 2)) / ngu
# p = p.view(-1, 5 + self.nc) p = p.view(-1, 5 + self.nc)
# xy = torch.sigmoid(p[..., 0:2]) + grid_xy[0] # x, y xy = torch.sigmoid(p[..., 0:2]) + grid_xy[0] # x, y
# wh = torch.exp(p[..., 2:4]) * anchor_wh[0] # width, height wh = torch.exp(p[..., 2:4]) * anchor_wh[0] # width, height
# p_conf = torch.sigmoid(p[:, 4:5]) # Conf p_conf = torch.sigmoid(p[:, 4:5]) # Conf
# p_cls = F.softmax(p[:, 5:85], 1) * p_conf # SSD-like conf p_cls = F.softmax(p[:, 5:85], 1) * p_conf # SSD-like conf
# return torch.cat((xy / ngu[0], wh, p_conf, p_cls), 1).t() return torch.cat((xy / ngu[0], wh, p_conf, p_cls), 1).t()
p = p.view(1, -1, 5 + self.nc) # p = p.view(1, -1, 5 + self.nc)
xy = torch.sigmoid(p[..., 0:2]) + grid_xy # x, y # xy = torch.sigmoid(p[..., 0:2]) + grid_xy # x, y
wh = torch.exp(p[..., 2:4]) * anchor_wh # width, height # wh = torch.exp(p[..., 2:4]) * anchor_wh # width, height
p_conf = torch.sigmoid(p[..., 4:5]) # Conf # p_conf = torch.sigmoid(p[..., 4:5]) # Conf
p_cls = p[..., 5:5 + self.nc] # p_cls = p[..., 5:5 + self.nc]
# Broadcasting only supported on first dimension in CoreML. See onnx-coreml/_operators.py # # Broadcasting only supported on first dimension in CoreML. See onnx-coreml/_operators.py
# p_cls = F.softmax(p_cls, 2) * p_conf # SSD-like conf # # p_cls = F.softmax(p_cls, 2) * p_conf # SSD-like conf
p_cls = torch.exp(p_cls).permute((2, 1, 0)) # p_cls = torch.exp(p_cls).permute((2, 1, 0))
p_cls = p_cls / p_cls.sum(0).unsqueeze(0) * p_conf.permute((2, 1, 0)) # F.softmax() equivalent # p_cls = p_cls / p_cls.sum(0).unsqueeze(0) * p_conf.permute((2, 1, 0)) # F.softmax() equivalent
p_cls = p_cls.permute(2, 1, 0) # p_cls = p_cls.permute(2, 1, 0)
return torch.cat((xy / ngu, wh, p_conf, p_cls), 2).squeeze().t() # return torch.cat((xy / ngu, wh, p_conf, p_cls), 2).squeeze().t()
else: # inference else: # inference
# s = 1.5 # scale_xy (pxy = pxy * s - (s - 1) / 2) # s = 1.5 # scale_xy (pxy = pxy * s - (s - 1) / 2)
io = p.clone() # inference output io = p.clone() # inference output
io[..., 0:2] = torch.sigmoid(io[..., 0:2]) + self.grid_xy # xy io[..., 0:2] = torch.sigmoid(io[..., 0:2]) + self.grid_xy # xy
io[..., 2:4] = torch.exp(io[..., 2:4]) * self.anchor_wh # wh yolo method io[..., 2:4] = torch.exp(io[..., 2:4]) * self.anchor_wh # wh yolo method
# io[..., 2:4] = ((torch.sigmoid(io[..., 2:4]) * 2) ** 3) * self.anchor_wh # wh power method
io[..., :4] *= self.stride io[..., :4] *= self.stride
# io[..., 2:4] = ((torch.sigmoid(io[..., 2:4]) * 2) ** 3) * self.anchor_wh # wh power method
io[..., 4:] = torch.sigmoid(io[..., 4:]) # p_conf, p_cls io[..., 4:] = torch.sigmoid(io[..., 4:]) # p_conf, p_cls
# io[..., 5:] = F.softmax(io[..., 5:], dim=4) # p_cls # io[..., 5:] = F.softmax(io[..., 5:], dim=4) # p_cls
@ -181,7 +181,7 @@ class Darknet(nn.Module):
self.seen = np.array([0], dtype=np.int64) # (int64) number of images seen during training self.seen = np.array([0], dtype=np.int64) # (int64) number of images seen during training
def forward(self, x, var=None): def forward(self, x, var=None):
img_size = max(x.shape[-2:]) img_size = x.shape[-2:]
layer_outputs = [] layer_outputs = []
output = [] output = []
@ -241,8 +241,8 @@ def get_yolo_layers(model):
def create_grids(self, img_size=416, ng=(13, 13), device='cpu', type=torch.float32): def create_grids(self, img_size=416, ng=(13, 13), device='cpu', type=torch.float32):
nx, ny = ng # x and y grid size nx, ny = ng # x and y grid size
self.img_size = img_size self.img_size = max(img_size)
self.stride = img_size / max(ng) self.stride = self.img_size / max(ng)
# build xy offsets # build xy offsets
yv, xv = torch.meshgrid([torch.arange(ny), torch.arange(nx)]) yv, xv = torch.meshgrid([torch.arange(ny), torch.arange(nx)])