-
Notifications
You must be signed in to change notification settings - Fork 29
/
models.py
317 lines (268 loc) · 11.7 KB
/
models.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
from collections import defaultdict
import torch.nn as nn
from utils.parse_config import *
from utils.utils import *
def create_modules(module_defs):
"""
Constructs module list of layer blocks from module configuration in module_defs
"""
hyperparams = module_defs.pop(0)
output_filters = [int(hyperparams['channels'])]
module_list = nn.ModuleList()
for i, module_def in enumerate(module_defs):
modules = nn.Sequential()
if module_def['type'] == 'convolutional':
bn = int(module_def['batch_normalize'])
filters = int(module_def['filters'])
kernel_size = int(module_def['size'])
pad = (kernel_size - 1) // 2 if int(module_def['pad']) else 0
modules.add_module('conv_%d' % i, nn.Conv2d(in_channels=output_filters[-1],
out_channels=filters,
kernel_size=kernel_size,
stride=int(module_def['stride']),
padding=pad,
bias=not bn))
if bn:
modules.add_module('batch_norm_%d' % i, nn.BatchNorm2d(filters))
if module_def['activation'] == 'leaky':
modules.add_module('leaky_%d' % i, nn.LeakyReLU(0.1))
elif module_def['type'] == 'upsample':
upsample = nn.Upsample(scale_factor=int(module_def['stride']), mode='nearest')
modules.add_module('upsample_%d' % i, upsample)
elif module_def['type'] == 'route':
layers = [int(x) for x in module_def['layers'].split(',')]
filters = sum([output_filters[layer_i] for layer_i in layers])
modules.add_module('route_%d' % i, EmptyLayer())
elif module_def['type'] == 'shortcut':
filters = output_filters[int(module_def['from'])]
modules.add_module('shortcut_%d' % i, EmptyLayer())
elif module_def['type'] == 'yolo':
anchor_idxs = [int(x) for x in module_def['mask'].split(',')]
# Extract anchors
anchors = [float(x) for x in module_def['anchors'].split(',')]
anchors = [(anchors[i], anchors[i + 1]) for i in range(0, len(anchors), 2)]
anchors = [anchors[i] for i in anchor_idxs]
num_classes = int(module_def['classes'])
img_height = int(hyperparams['height'])
# Define detection layer
yolo_layer = YOLOLayer(anchors, num_classes, img_height, anchor_idxs)
modules.add_module('yolo_%d' % i, yolo_layer)
# Register module list and number of output filters
module_list.append(modules)
output_filters.append(filters)
return hyperparams, module_list
class EmptyLayer(nn.Module):
"""Placeholder for 'route' and 'shortcut' layers"""
def __init__(self):
super(EmptyLayer, self).__init__()
class YOLOLayer(nn.Module):
def __init__(self, anchors, nC, img_dim, anchor_idxs):
super(YOLOLayer, self).__init__()
anchors = [(a_w, a_h) for a_w, a_h in anchors] # (pixels)
nA = len(anchors)
self.anchors = anchors
self.nA = nA # number of anchors (3)
self.nC = nC # number of classes (1)
self.bbox_attrs = 9 + nC
self.img_dim = img_dim # from hyperparams in cfg file, NOT from parser
if anchor_idxs[0] == (nA * 2): # 6
stride = 32
elif anchor_idxs[0] == nA: # 3
stride = 16
else:
stride = 8
# Build anchor grids
nG = int(self.img_dim / stride) # number grid points
self.grid_x = torch.arange(nG).repeat(nG, 1).view([1, 1, nG, nG]).float()
self.grid_y = torch.arange(nG).repeat(nG, 1).t().view([1, 1, nG, nG]).float()
self.scaled_anchors = torch.FloatTensor([(a_w / stride, a_h / stride) for a_w, a_h in anchors])
self.anchor_w = self.scaled_anchors[:, 0:1].view((1, nA, 1, 1))
self.anchor_h = self.scaled_anchors[:, 1:2].view((1, nA, 1, 1))
#self.weights = class_weights()
def forward(self, p, targets=None, requestPrecision=False):
FT = torch.cuda.FloatTensor if p.is_cuda else torch.FloatTensor
bs = p.shape[0] # batch size
nG = p.shape[2] # number of grid points
stride = self.img_dim / nG
if p.is_cuda and not self.grid_x.is_cuda:
self.grid_x, self.grid_y = self.grid_x.cuda(), self.grid_y.cuda()
self.anchor_w, self.anchor_h = self.anchor_w.cuda(), self.anchor_h.cuda()
#self.weights = self.weights.cuda()
# p.view(1, 30, 13, 13) -- > (1, 3, 13, 13, 10) # (bs, anchors, grid, grid, classes + xywh)
p = p.view(bs, self.nA, self.bbox_attrs, nG, nG).permute(0, 1, 3, 4, 2).contiguous() # prediction
# Get outputs
P1_x = p[..., 0] # Point1 x
P1_y = p[..., 1] # Point1 y
P2_x = p[..., 2] # Point2 x
P2_y = p[..., 3] # Point2 y
P3_x = p[..., 4] # Point3 x
P3_y = p[..., 5] # Point3 y
P4_x = p[..., 6] # Point4 x
P4_y = p[..., 7] # Point4 y
pred_boxes = FT(bs, self.nA, nG, nG, 8)
pred_conf = p[..., 8] # Conf
pred_cls = p[..., 9:] # Class
# Training
if targets is not None:
MSELoss = nn.MSELoss()
BCEWithLogitsLoss = nn.BCEWithLogitsLoss()
CrossEntropyLoss = nn.CrossEntropyLoss()
SmoothL1Loss = nn.SmoothL1Loss()
if requestPrecision:
gx = self.grid_x[:, :, :nG, :nG]
gy = self.grid_y[:, :, :nG, :nG]
pred_boxes[..., 0] = P1_x.data + gx
pred_boxes[..., 1] = P1_y.data + gy
pred_boxes[..., 2] = P2_x.data + gx
pred_boxes[..., 3] = P2_y.data + gy
pred_boxes[..., 4] = P3_x.data + gx
pred_boxes[..., 5] = P3_y.data + gy
pred_boxes[..., 6] = P4_x.data + gx
pred_boxes[..., 7] = P4_y.data + gy
t1_x, t1_y, t2_x, t2_y, t3_x, t3_y, t4_x, t4_y, mask, tcls, TP, FP, FN, TC = \
build_targets(pred_boxes, pred_conf, pred_cls, targets, self.scaled_anchors, self.nA, self.nC, nG,
requestPrecision)
tcls = tcls[mask]
if P1_x.is_cuda:
t1_x, t1_y, t2_x, t2_y, t3_x, t3_y, t4_x, t4_y, mask, tcls = \
t1_x.cuda(), t1_y.cuda(), t2_x.cuda(), t2_y.cuda(), t3_x.cuda(), t3_y.cuda(), t4_x.cuda(), t4_y.cuda(), mask.cuda(), tcls.cuda()
# Compute losses
nT = sum([len(x) for x in targets]) # Number of targets
nM = mask.sum().float() # Number of anchors (assigned to targets)
nB = len(targets) # Batch size
k = nM / nB
if nM > 0:
lx1 = (k) * SmoothL1Loss(P1_x[mask], t1_x[mask]) / 8
ly1 = (k) * SmoothL1Loss(P1_y[mask], t1_y[mask]) / 8
lx2 = (k) * SmoothL1Loss(P2_x[mask], t2_x[mask]) / 8
ly2 = (k) * SmoothL1Loss(P2_y[mask], t2_y[mask]) / 8
lx3 = (k) * SmoothL1Loss(P3_x[mask], t3_x[mask]) / 8
ly3 = (k) * SmoothL1Loss(P3_y[mask], t3_y[mask]) / 8
lx4 = (k) * SmoothL1Loss(P4_x[mask], t4_x[mask]) / 8
ly4 = (k) * SmoothL1Loss(P4_y[mask], t4_y[mask]) / 8
lconf = (k * 10) * BCEWithLogitsLoss(pred_conf, mask.float())
lcls = (k / nC) * CrossEntropyLoss(pred_cls[mask], torch.argmax(tcls, 1))
else:
lx1, ly1, lx2, ly2, lx3, ly3, lx4, ly4, lcls, lconf = \
FT([0]), FT([0]), FT([0]), FT([0]), FT([0]), FT([0]), FT([0]), FT([0]), FT([0]), FT([0])
# Sum loss components
loss = lx1 + ly1 + lx2 + ly2 + lx3 + ly3 + lx4 + ly4 + lconf + lcls
# Sum False Positives from unassigned anchors
i = torch.sigmoid(pred_conf[~mask]) > 0.5
if i.sum() > 0:
FP_classes = torch.argmax(pred_cls[~mask][i], 1)
FPe = torch.bincount(FP_classes, minlength=self.nC).float().cpu()
else:
FPe = torch.zeros(self.nC)
return loss, loss.item(), lconf.item(), lcls.item(), nT, TP, FP, FPe, FN, TC
else:
pred_boxes[..., 0] = P1_x + self.grid_x
pred_boxes[..., 1] = P1_y + self.grid_y
pred_boxes[..., 2] = P2_x + self.grid_x
pred_boxes[..., 3] = P2_y + self.grid_y
pred_boxes[..., 4] = P3_x + self.grid_x
pred_boxes[..., 5] = P3_y + self.grid_y
pred_boxes[..., 6] = P4_x + self.grid_x
pred_boxes[..., 7] = P4_y + self.grid_y
output = torch.cat((pred_boxes.view(bs, -1, 8) * stride,
torch.sigmoid(pred_conf.view(bs, -1, 1)), pred_cls.view(bs, -1, self.nC)), -1)
return output
class Darknet(nn.Module):
"""YOLOv3 object detection model"""
def __init__(self, cfg_path, img_size=416):
super(Darknet, self).__init__()
self.module_defs = parse_model_config(cfg_path)
self.module_defs[0]['height'] = img_size
self.hyperparams, self.module_list = create_modules(self.module_defs)
self.img_size = img_size
self.loss_names = ['loss', 'conf', 'cls', 'nT', 'TP', 'FP', 'FPe', 'FN', 'TC']
def forward(self, x, targets=None, requestPrecision=False):
is_training = targets is not None
output = []
self.losses = defaultdict(float)
layer_outputs = []
for i, (module_def, module) in enumerate(zip(self.module_defs, self.module_list)):
if module_def['type'] in ['convolutional', 'upsample']:
x = module(x)
elif module_def['type'] == 'route':
layer_i = [int(x) for x in module_def['layers'].split(',')]
x = torch.cat([layer_outputs[i] for i in layer_i], 1)
elif module_def['type'] == 'shortcut':
layer_i = int(module_def['from'])
x = layer_outputs[-1] + layer_outputs[layer_i]
elif module_def['type'] == 'yolo':
# Train phase: get loss
if is_training:
x, *losses = module[0](x, targets, requestPrecision)
for name, loss in zip(self.loss_names, losses):
self.losses[name] += loss
# Test phase: Get detections
else:
x = module(x)
output.append(x)
layer_outputs.append(x)
if is_training:
self.losses['nT'] /= 3
self.losses['TC'] /= 3 # target category
metrics = torch.zeros(3, len(self.losses['FPe'])) # TP, FP, FN
ui = np.unique(self.losses['TC'])[1:]
for i in ui:
j = self.losses['TC'] == float(i)
metrics[0, i] = (self.losses['TP'][j] > 0).sum().float() # TP
metrics[1, i] = (self.losses['FP'][j] > 0).sum().float() # FP
metrics[2, i] = (self.losses['FN'][j] == 3).sum().float() # FN
metrics[1] += self.losses['FPe']
self.losses['TP'] = metrics[0].sum()
self.losses['FP'] = metrics[1].sum()
self.losses['FN'] = metrics[2].sum()
self.losses['TC'] = 0
self.losses['metrics'] = metrics
return sum(output) if is_training else torch.cat(output, 1)
def load_weights(self, weights_path, cutoff=-1):
# Parses and loads the weights stored in 'weights_path'
# @:param cutoff - save layers between 0 and cutoff (cutoff = -1 -> all are saved)
if weights_path.endswith('darknet53.conv.74'):
cutoff = 75
# Open the weights file
fp = open(weights_path, 'rb')
header = np.fromfile(fp, dtype=np.int32, count=5) # First five are header values
# Needed to write header when saving weights
self.header_info = header
self.seen = header[3]
weights = np.fromfile(fp, dtype=np.float32) # The rest are weights
fp.close()
ptr = 0
for i, (module_def, module) in enumerate(zip(self.module_defs[:cutoff], self.module_list[:cutoff])):
if module_def['type'] == 'convolutional':
conv_layer = module[0]
if module_def['batch_normalize']:
# Load BN bias, weights, running mean and running variance
bn_layer = module[1]
num_b = bn_layer.bias.numel() # Number of biases
# Bias
bn_b = torch.from_numpy(weights[ptr:ptr + num_b]).view_as(bn_layer.bias)
bn_layer.bias.data.copy_(bn_b)
ptr += num_b
# Weight
bn_w = torch.from_numpy(weights[ptr:ptr + num_b]).view_as(bn_layer.weight)
bn_layer.weight.data.copy_(bn_w)
ptr += num_b
# Running Mean
bn_rm = torch.from_numpy(weights[ptr:ptr + num_b]).view_as(bn_layer.running_mean)
bn_layer.running_mean.data.copy_(bn_rm)
ptr += num_b
# Running Var
bn_rv = torch.from_numpy(weights[ptr:ptr + num_b]).view_as(bn_layer.running_var)
bn_layer.running_var.data.copy_(bn_rv)
ptr += num_b
else:
# Load conv. bias
num_b = conv_layer.bias.numel()
conv_b = torch.from_numpy(weights[ptr:ptr + num_b]).view_as(conv_layer.bias)
conv_layer.bias.data.copy_(conv_b)
ptr += num_b
# Load conv. weights
num_w = conv_layer.weight.numel()
conv_w = torch.from_numpy(weights[ptr:ptr + num_w]).view_as(conv_layer.weight)
conv_layer.weight.data.copy_(conv_w)
ptr += num_w