-
Notifications
You must be signed in to change notification settings - Fork 33
/
readme_ppyoloe.txt
471 lines (302 loc) · 27.8 KB
/
readme_ppyoloe.txt
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
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
pip install -r requirements.txt -i https://pypi.tuna.tsinghua.edu.cn/simple
----------------------- 转换权重 -----------------------
wget https://paddledet.bj.bcebos.com/models/ppyoloe_crn_s_300e_coco.pdparams
wget https://paddledet.bj.bcebos.com/models/ppyoloe_crn_m_300e_coco.pdparams
wget https://paddledet.bj.bcebos.com/models/ppyoloe_crn_l_300e_coco.pdparams
wget https://paddledet.bj.bcebos.com/models/ppyoloe_crn_x_300e_coco.pdparams
wget https://paddledet.bj.bcebos.com/models/pretrained/CSPResNetb_s_pretrained.pdparams
wget https://paddledet.bj.bcebos.com/models/pretrained/CSPResNetb_m_pretrained.pdparams
wget https://paddledet.bj.bcebos.com/models/pretrained/CSPResNetb_l_pretrained.pdparams
wget https://paddledet.bj.bcebos.com/models/pretrained/CSPResNetb_x_pretrained.pdparams
wget https://paddledet.bj.bcebos.com/models/ppyoloe_plus_crn_s_80e_coco.pdparams
wget https://paddledet.bj.bcebos.com/models/ppyoloe_plus_crn_m_80e_coco.pdparams
wget https://paddledet.bj.bcebos.com/models/ppyoloe_plus_crn_l_80e_coco.pdparams
wget https://paddledet.bj.bcebos.com/models/ppyoloe_plus_crn_x_80e_coco.pdparams
wget https://bj.bcebos.com/v1/paddledet/models/pretrained/ppyoloe_crn_s_obj365_pretrained.pdparams
wget https://bj.bcebos.com/v1/paddledet/models/pretrained/ppyoloe_crn_m_obj365_pretrained.pdparams
wget https://bj.bcebos.com/v1/paddledet/models/pretrained/ppyoloe_crn_l_obj365_pretrained.pdparams
wget https://bj.bcebos.com/v1/paddledet/models/pretrained/ppyoloe_crn_x_obj365_pretrained.pdparams
python tools/convert_weights.py -f exps/ppyoloe/ppyoloe_crn_s_300e_coco.py -c ppyoloe_crn_s_300e_coco.pdparams -oc ppyoloe_crn_s_300e_coco.pth -nc 80
python tools/convert_weights.py -f exps/ppyoloe/ppyoloe_crn_m_300e_coco.py -c ppyoloe_crn_m_300e_coco.pdparams -oc ppyoloe_crn_m_300e_coco.pth -nc 80
python tools/convert_weights.py -f exps/ppyoloe/ppyoloe_crn_l_300e_coco.py -c ppyoloe_crn_l_300e_coco.pdparams -oc ppyoloe_crn_l_300e_coco.pth -nc 80
python tools/convert_weights.py -f exps/ppyoloe/ppyoloe_crn_x_300e_coco.py -c ppyoloe_crn_x_300e_coco.pdparams -oc ppyoloe_crn_x_300e_coco.pth -nc 80
python tools/convert_weights.py -f exps/ppyoloe/ppyoloe_crn_s_300e_coco.py -c CSPResNetb_s_pretrained.pdparams -oc CSPResNetb_s_pretrained.pth -nc 80 --only_backbone True
python tools/convert_weights.py -f exps/ppyoloe/ppyoloe_crn_m_300e_coco.py -c CSPResNetb_m_pretrained.pdparams -oc CSPResNetb_m_pretrained.pth -nc 80 --only_backbone True
python tools/convert_weights.py -f exps/ppyoloe/ppyoloe_crn_l_300e_coco.py -c CSPResNetb_l_pretrained.pdparams -oc CSPResNetb_l_pretrained.pth -nc 80 --only_backbone True
python tools/convert_weights.py -f exps/ppyoloe/ppyoloe_crn_x_300e_coco.py -c CSPResNetb_x_pretrained.pdparams -oc CSPResNetb_x_pretrained.pth -nc 80 --only_backbone True
python tools/convert_weights.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_s_80e_coco.py -c ppyoloe_plus_crn_s_80e_coco.pdparams -oc ppyoloe_plus_crn_s_80e_coco.pth -nc 80
python tools/convert_weights.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_m_80e_coco.py -c ppyoloe_plus_crn_m_80e_coco.pdparams -oc ppyoloe_plus_crn_m_80e_coco.pth -nc 80
python tools/convert_weights.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_l_80e_coco.py -c ppyoloe_plus_crn_l_80e_coco.pdparams -oc ppyoloe_plus_crn_l_80e_coco.pth -nc 80
python tools/convert_weights.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_x_80e_coco.py -c ppyoloe_plus_crn_x_80e_coco.pdparams -oc ppyoloe_plus_crn_x_80e_coco.pth -nc 80
python tools/convert_weights.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_s_80e_coco.py -c ppyoloe_crn_s_obj365_pretrained.pdparams -oc ppyoloe_crn_s_obj365_pretrained.pth -nc 365
python tools/convert_weights.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_m_80e_coco.py -c ppyoloe_crn_m_obj365_pretrained.pdparams -oc ppyoloe_crn_m_obj365_pretrained.pth -nc 365
python tools/convert_weights.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_l_80e_coco.py -c ppyoloe_crn_l_obj365_pretrained.pdparams -oc ppyoloe_crn_l_obj365_pretrained.pth -nc 365
python tools/convert_weights.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_x_80e_coco.py -c ppyoloe_crn_x_obj365_pretrained.pdparams -oc ppyoloe_crn_x_obj365_pretrained.pth -nc 365
----------------------- 预测 -----------------------
python tools/demo.py image -f exps/ppyoloe/ppyoloe_crn_s_300e_coco.py -c ppyoloe_crn_s_300e_coco.pth --path assets/000000000019.jpg --conf 0.15 --tsize 640 --save_result --device gpu
python tools/demo.py image -f exps/ppyoloe/ppyoloe_crn_m_300e_coco.py -c ppyoloe_crn_m_300e_coco.pth --path assets/000000000019.jpg --conf 0.15 --tsize 640 --save_result --device gpu
python tools/demo.py image -f exps/ppyoloe/ppyoloe_crn_l_300e_coco.py -c ppyoloe_crn_l_300e_coco.pth --path assets/000000000019.jpg --conf 0.15 --tsize 640 --save_result --device gpu
python tools/demo.py image -f exps/ppyoloe/ppyoloe_crn_x_300e_coco.py -c ppyoloe_crn_x_300e_coco.pth --path assets/000000000019.jpg --conf 0.15 --tsize 640 --save_result --device gpu
python tools/demo.py image -f exps/ppyoloe/ppyoloe_crn_l_voc2012.py -c 6.pth --path assets/000000000019.jpg --conf 0.15 --tsize 640 --save_result --device gpu
python tools/demo.py image -f exps/ppyoloe_plus/ppyoloe_plus_crn_s_80e_coco.py -c ppyoloe_plus_crn_s_80e_coco.pth --path assets/000000000019.jpg --conf 0.15 --tsize 640 --save_result --device gpu
----------------------- 迁移学习，带上-c（--ckpt）参数读取预训练模型。 -----------------------
后台启动：
nohup xxx > ppyolo.log 2>&1 &
复现paddle版ppyoloe_s迁移学习（不冻结骨干网络）:（可以加--fp16， -eb表示验证时的批大小）
python tools/train.py -f exps/ppyoloe/ppyoloe_crn_s_voc2012.py -d 1 -b 4 -eb 2 -c ppyoloe_crn_s_300e_coco.pth --fp16
python tools/eval.py -f exps/ppyoloe/ppyoloe_crn_s_voc2012.py -d 1 -b 4 -c 16.pth --conf 0.01 --tsize 640
python tools/demo.py image -f exps/ppyoloe/ppyoloe_crn_s_voc2012.py -c 16.pth --path assets/000000000019.jpg --conf 0.15 --tsize 640 --save_result --device gpu
2机2卡训练：(发现一个隐藏知识点：获得损失（训练）、推理 都要放在模型的forward()中进行，否则DDP会计算错误结果。)
export CUDA_VISIBLE_DEVICES=0
python tools/train.py -f exps/ppyoloe/ppyoloe_crn_s_voc2012.py --dist-url tcp://192.168.0.106:12312 --num_machines 2 --machine_rank 0 -b 4 -eb 2 -c ppyoloe_crn_s_300e_coco.pth --fp16
1机2卡训练：(发现一个隐藏知识点：获得损失（训练）、推理 都要放在模型的forward()中进行，否则DDP会计算错误结果。)
export CUDA_VISIBLE_DEVICES=0,1
nohup python tools/train.py -f exps/ppyoloe/ppyoloe_crn_s_voc2012.py -d 2 -b 24 -eb 24 -w 4 -ew 4 -lrs 0.7 -c ppyoloe_crn_s_300e_coco.pth > ppyoloe_s.log 2>&1 &
tail -n 20 ppyoloe_s.log
实测ppyoloe_s的AP(0.50:0.95)可以到达0.48+、AP(0.50)可以到达0.68+、AP(small)可以到达0.15+。
- - - - - - - - - - - - - - - - - - - - - -
训练 ppyoloe_xs
python tools/train.py -f exps/ppyoloe/ppyoloe_crn_xs_voc2012.py -d 1 -b 4 -eb 2 -c ppyoloe_crn_s_300e_coco.pth
实测ppyoloe_xs的AP(0.50:0.95)可以到达0.xx+、AP(0.50)可以到达0.xx+、AP(small)可以到达0.xx+。
- - - - - - - - - - - - - - - - - - - - - -
复现paddle版ppyoloe_l迁移学习（冻结了骨干网络）:（可以加--fp16， -eb表示验证时的批大小）
python tools/train.py -f exps/ppyoloe/ppyoloe_crn_l_voc2012.py -d 1 -b 8 -eb 2 -c ppyoloe_crn_l_300e_coco.pth --fp16
python tools/eval.py -f exps/ppyoloe/ppyoloe_crn_l_voc2012.py -d 1 -b 4 -c 16.pth --conf 0.01 --tsize 640
python tools/demo.py image -f exps/ppyoloe/ppyoloe_crn_l_voc2012.py -c 16.pth --path assets/000000000019.jpg --conf 0.15 --tsize 640 --save_result --device gpu
1机2卡训练：(发现一个隐藏知识点：获得损失（训练）、推理 都要放在模型的forward()中进行，否则DDP会计算错误结果。)
export CUDA_VISIBLE_DEVICES=0,1
nohup python tools/train.py -f exps/ppyoloe/ppyoloe_crn_l_voc2012.py -d 2 -b 8 -eb 2 -c ppyoloe_crn_l_300e_coco.pth --fp16 > ppyoloe_l.log 2>&1 &
tail -n 20 ppyoloe_l.log
实测ppyoloe_l的AP(0.50:0.95)可以到达0.66+、AP(0.50)可以到达0.85+、AP(small)可以到达0.28+。
- - - - - - - - - - - - - - - - - - - - - -
感兴趣的同学可以加载 CSPResNetb_l_pretrained.pth 从头训练ppyoloe_l（voc2012数据集）：
python tools/train.py -f exps/ppyoloe/ppyoloe_crn_l_voc2012.py -d 1 -b 8 -eb 2 -c CSPResNetb_l_pretrained.pth --fp16
训练完16个epoch后，实测ppyoloe_l的AP(0.50:0.95)可以到达0.05+、AP(0.50)可以到达0.16+、AP(small)可以到达0.00+。
- - - - - - - - - - - - - - - - - - - - - -
ppyoloe_plus_s迁移学习（不冻结骨干网络）:（可以加--fp16， -eb表示验证时的批大小）
python tools/eval.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_s_voc2012.py -d 1 -b 4 -c 16.pth --conf 0.01 --tsize 640
python tools/demo.py image -f exps/ppyoloe_plus/ppyoloe_plus_crn_s_voc2012.py -c 16.pth --path assets/000000000019.jpg --conf 0.15 --tsize 640 --save_result --device gpu
1机2卡训练：(发现一个隐藏知识点：获得损失（训练）、推理 都要放在模型的forward()中进行，否则DDP会计算错误结果。)
export CUDA_VISIBLE_DEVICES=0,1
nohup python tools/train.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_s_voc2012.py -d 2 -b 16 -eb 8 -w 4 -ew 4 -lrs 1.0 -c ppyoloe_crn_s_obj365_pretrained.pth --fp16 > ppyoloe_plus_s_from_obj365.log 2>&1 &
tail -n 20 ppyoloe_plus_s_from_obj365.log
实测 ppyoloe_plus_s_from_obj365 的AP(0.50:0.95)可以到达0.59+、AP(0.50)可以到达0.78+、AP(small)可以到达0.20+。
日志见 train_ppyolo_in_voc2012/mmdet_ppyoloe_plus_s_from_obj365_2gpu.txt
- - - - - - -
读 COCO 预训练模型（实际上由obj365 fine-tune 得到）进行fine-tune：
export CUDA_VISIBLE_DEVICES=0,1
nohup python tools/train.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_s_voc2012.py -d 2 -b 16 -eb 8 -w 4 -ew 4 -lrs 1.0 -c ppyoloe_plus_crn_s_80e_coco.pth > ppyoloe_plus_s_from_coco.log 2>&1 &
实测 ppyoloe_plus_s_from_coco 的AP(0.50:0.95)可以到达0.62+、AP(0.50)可以到达0.81+、AP(small)可以到达0.22+。
日志见(这是在完全空闲的服务器，2卡4090测速) train_ppyolo_in_voc2012/mmdet_ppyoloe_plus_s_from_obj365(to_coco)_2gpu.txt
tensorboard --logdir=./PPYOLOEPlus_outputs/
- - - - - - -
做 ppyoloe_s 的对比实验（读 ppyoloe 的模型）：
export CUDA_VISIBLE_DEVICES=0,1
nohup python tools/train.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_s_voc2012.py -d 2 -b 16 -eb 8 -w 4 -ew 4 -lrs 1.0 -c ppyoloe_crn_s_300e_coco.pth --fp16 > ppyoloe_s_from_coco.log 2>&1 &
实测 ppyoloe_s_from_coco 的AP(0.50:0.95)可以到达0.61+、AP(0.50)可以到达0.80+、AP(small)可以到达0.23+。
日志见 train_ppyolo_in_voc2012/mmdet_ppyoloe_plus_s_from_coco_2gpu.txt
- - - - - - - - - - - - - - - - - - - - - -
pcp:
export CUDA_VISIBLE_DEVICES=0,1
nohup python tools/train.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_s_pcp.py -d 2 -b 24 -eb 8 -c ppyoloe_crn_s_obj365_pretrained.pth > ppyoloe_plus_s_from_obj365.log 2>&1 &
实测 ppyoloe_plus_s_from_obj365 的AP(0.50:0.95)可以到达0.xx+、AP(0.50)可以到达0.xx+、AP(small)可以到达0.xx+。
日志见 train_ppyolo_in_voc2012/xxxxxxxxxx.txt
- - - - - - -
读 COCO 预训练模型（实际上由obj365 fine-tune 得到）进行fine-tune：
export CUDA_VISIBLE_DEVICES=0,1
nohup python tools/train.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_s_pcp.py -d 2 -b 24 -eb 8 -c ppyoloe_plus_crn_s_80e_coco.pth > ppyoloe_plus_s_from_coco.log 2>&1 &
实测 ppyoloe_plus_s_from_coco 的AP(0.50:0.95)可以到达0.xx+、AP(0.50)可以到达0.xx+、AP(small)可以到达0.xx+。
日志见 train_ppyolo_in_voc2012/xxxxxxxxxx.txt
- - - - - - -
做 ppyoloe_s 的对比实验（读 ppyoloe 的模型）：
export CUDA_VISIBLE_DEVICES=0,1
nohup python tools/train.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_s_pcp.py -d 2 -b 24 -eb 8 -c ppyoloe_crn_s_300e_coco.pth > ppyoloe_s_from_coco.log 2>&1 &
实测 ppyoloe_s_from_coco 的AP(0.50:0.95)可以到达0.xx+、AP(0.50)可以到达0.xx+、AP(small)可以到达0.xx+。
日志见 train_ppyolo_in_voc2012/xxxxxxxxxx.txt
----------------------- 知识蒸馏 -----------------------
蒸馏损失有 F.binary_cross_entropy, 所以不使用 fp16, 除非 F.binary_cross_entropy 前转成float32 且 使用 with torch.cuda.amp.autocast(enabled=False):
1.先训练老师模型:
读 COCO 预训练模型（实际上由obj365 fine-tune 得到）进行fine-tune：
export CUDA_VISIBLE_DEVICES=0,1
nohup python tools/train.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_l_voc2012.py -d 2 -b 16 -eb 8 -w 4 -ew 4 -lrs 1.0 -c ppyoloe_plus_crn_l_80e_coco.pth > ppyoloe_plus_l_from_coco.log 2>&1 &
实测 ppyoloe_plus_l_from_coco 的AP(0.50:0.95)可以到达0.715+、AP(0.50)可以到达0.881+、AP(small)可以到达0.314+。
日志见 train_ppyolo_in_voc2012/ppyoloe_plus_l_from_obj365(to_coco)_2gpu.txt
2.蒸馏:
读 COCO 预训练模型（实际上由obj365 fine-tune 得到）进行fine-tune：
export CUDA_VISIBLE_DEVICES=0,1
nohup python tools/train.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_s_voc2012.py -d 2 -b 16 -eb 8 -w 4 -ew 4 -lrs 1.0 -c ppyoloe_plus_crn_s_80e_coco.pth -sf exps/slim/distill/ppyoloe_plus_crn_l_voc2012_l2s.py -sc PPYOLOEPlus_outputs/ppyoloe_plus_crn_l_voc2012/16.pth > ppyoloe_plus_s_from_coco.log 2>&1 &
恢复训练:
nohup python tools/train.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_s_voc2012.py -d 2 -b 16 -eb 8 -w 4 -ew 4 -lrs 1.0 -c 7.pth --resume -sf exps/slim/distill/ppyoloe_plus_crn_l_voc2012_l2s.py -sc PPYOLOEPlus_outputs/ppyoloe_plus_crn_l_voc2012/16.pth > ppyoloe_plus_s_from_coco.log 2>&1 &
(单卡调试)
python tools/train.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_s_voc2012.py -d 1 -b 4 -eb 4 -w 1 -ew 1 -lrs 1.0 -c ppyoloe_plus_crn_s_80e_coco.pth -sf exps/slim/distill/ppyoloe_plus_crn_l_voc2012_l2s.py -sc PPYOLOEPlus_outputs/ppyoloe_plus_crn_l_voc2012/16.pth
实测 ppyoloe_plus_s_from_coco 的AP(0.50:0.95)可以到达0.xxx+、AP(0.50)可以到达0.xxx+、AP(small)可以到达0.xxx+。
日志见 train_ppyolo_in_voc2012/ppyoloe_plus_s_distill_l2s.txt
仔细检查：
单卡和多卡时，PPdetModelEMA 的 state_dict 不保存老师的权重，训练时ema也不更新老师的权重
单卡和多卡训练时时，老师的权重不被训练，bn的均值和方差也不会变化
evaluate_and_save_model()
aaaaaaaa.py 对比查看 老师权重的变化
----------------------- 恢复训练（加上参数--resume） -----------------------
----------------------- 评估 -----------------------
python tools/eval.py -f exps/ppyoloe/ppyoloe_crn_s_300e_coco.py -d 1 -b 8 -w 4 -c ppyoloe_crn_s_300e_coco.pth --conf 0.01 --tsize 640
Average forward time: 21.56 ms, Average NMS time: 0.00 ms, Average inference time: 21.56 ms
Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.423
Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.593
Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.458
Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.237
Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.464
Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.582
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.336
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.546
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.583
Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.356
Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.640
Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.756
python tools/eval.py -f exps/ppyoloe/ppyoloe_crn_s_300e_coco.py -d 1 -b 8 -w 4 -c ppyoloe_crn_s_300e_coco.pth --conf 0.01 --tsize 320
Average forward time: 14.12 ms, Average NMS time: 0.00 ms, Average inference time: 14.12 ms
Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.341
Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.491
Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.362
Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.117
Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.378
Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.571
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.288
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.450
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.477
Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.174
Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.549
Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.744
python tools/eval.py -f exps/ppyoloe/ppyoloe_crn_m_300e_coco.py -d 1 -b 8 -c ppyoloe_crn_m_300e_coco.pth --conf 0.01 --tsize 640
Average forward time: 35.90 ms, Average NMS time: 0.00 ms, Average inference time: 35.90 ms
Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.482
Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.654
Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.523
Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.294
Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.527
Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.645
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.365
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.597
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.635
Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.427
Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.687
Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.803
python tools/eval.py -f exps/ppyoloe/ppyoloe_crn_l_300e_coco.py -d 1 -b 4 -c ppyoloe_crn_l_300e_coco.pth --conf 0.01 --tsize 640
Average forward time: 49.45 ms, Average NMS time: 0.00 ms, Average inference time: 49.45 ms
Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.505
Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.681
Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.547
Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.335
Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.552
Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.671
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.377
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.617
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.657
Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.461
Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.709
Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.819
python tools/eval.py -f exps/ppyoloe/ppyoloe_crn_x_300e_coco.py -d 1 -b 2 -c ppyoloe_crn_x_300e_coco.pth --conf 0.01 --tsize 640
Average forward time: 79.69 ms, Average NMS time: 0.00 ms, Average inference time: 79.70 ms
Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.515
Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.690
Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.555
Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.338
Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.562
Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.682
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.383
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.628
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.669
Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.484
Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.718
Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.823
python tools/eval.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_s_80e_coco.py -d 1 -b 8 -c ppyoloe_plus_crn_s_80e_coco.pth --conf 0.01 --tsize 640
Average forward time: 15.57 ms, Average NMS time: 0.00 ms, Average inference time: 15.57 ms
Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.431
Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.599
Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.474
Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.256
Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.469
Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.583
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.342
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.567
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.625
Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.424
Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.688
Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.777
python tools/eval.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_m_80e_coco.py -d 1 -b 8 -c ppyoloe_plus_crn_m_80e_coco.pth --conf 0.01 --tsize 640
Average forward time: 20.78 ms, Average NMS time: 0.00 ms, Average inference time: 20.78 ms
Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.492
Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.663
Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.539
Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.306
Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.534
Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.656
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.373
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.614
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.673
Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.491
Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.728
Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.823
python tools/eval.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_l_80e_coco.py -d 1 -b 8 -c ppyoloe_plus_crn_l_80e_coco.pth --conf 0.01 --tsize 640
Average forward time: 22.93 ms, Average NMS time: 0.00 ms, Average inference time: 22.93 ms
Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.523
Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.694
Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.573
Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.344
Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.571
Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.685
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.385
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.641
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.699
Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.521
Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.755
Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.846
python tools/eval.py -f exps/ppyoloe_plus/ppyoloe_plus_crn_x_80e_coco.py -d 1 -b 8 -c ppyoloe_plus_crn_x_80e_coco.pth --conf 0.01 --tsize 640
Average forward time: 27.18 ms, Average NMS time: 0.00 ms, Average inference time: 27.18 ms
Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.542
Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.712
Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.594
Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.372
Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.587
Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.699
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.394
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.657
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.711
Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.542
Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.765
Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.851
----------------------- 复现COCO上的精度 -----------------------
(5)ppyoloe_crn_s_300e_coco
(
PaddleDetection-release-2.4 的总批大小是 256=8*32
PaddleDetection-release-2.6 的总批大小是 64=8*8
另外，加上--fp16实测会导致loss出现nan，所以不加，不使用混合精度训练。
)
export CUDA_VISIBLE_DEVICES=0,1,2,3,4,5,6,7
nohup python tools/train.py -f exps/ppyoloe/ppyoloe_crn_s_300e_coco.py -d 8 -b 64 -eb 64 -w 4 -ew 4 -c CSPResNetb_s_pretrained.pth > ppyoloe_crn_s_300e_coco_8gpu.log 2>&1 &
训练日志见 train_coco/ppyoloe_s_8gpu.txt
实测训300 epochs后，最高mAP为42.10，基本上能达到转换的官方权重( Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.423)
python tools/eval.py -f exps/ppyoloe/ppyoloe_crn_s_300e_coco.py -d 8 -b 64 -c 300.pth --conf 0.01 --tsize 640
只有双卡的时候：
export CUDA_VISIBLE_DEVICES=0,1
nohup python tools/train.py -f exps/ppyoloe/ppyoloe_crn_s_300e_coco.py -d 2 -b 16 -eb 16 -w 4 -ew 4 -c CSPResNetb_s_pretrained.pth > ppyoloe_crn_s_300e_coco_2gpu.log 2>&1 &
----------------------- 导出为ncnn -----------------------
python tools/demo.py ncnn -f exps/ppyoloe/ppyoloe_crn_s_300e_coco.py -c ppyoloe_crn_s_300e_coco.pth --ncnn_output_path ppyoloe_crn_s_300e_coco
python tools/demo.py ncnn -f exps/ppyoloe/ppyoloe_crn_m_300e_coco.py -c ppyoloe_crn_m_300e_coco.pth --ncnn_output_path ppyoloe_crn_m_300e_coco
python tools/demo.py ncnn -f exps/ppyoloe/ppyoloe_crn_l_300e_coco.py -c ppyoloe_crn_l_300e_coco.pth --ncnn_output_path ppyoloe_crn_l_300e_coco
python tools/demo.py ncnn -f exps/ppyoloe/ppyoloe_crn_x_300e_coco.py -c ppyoloe_crn_x_300e_coco.pth --ncnn_output_path ppyoloe_crn_x_300e_coco
python tools/demo.py ncnn -f exps/ppyoloe/ppyoloe_crn_l_voc2012.py -c 6.pth --ncnn_output_path ppyoloe_crn_l_voc2012_epoch_6
cd build/examples
./test2_06_ppyoloe_ncnn ../../my_tests/000000000019.jpg ppyoloe_crn_s_300e_coco.param ppyoloe_crn_s_300e_coco.bin
cd build/examples
./test2_06_ppyoloe_ncnn ../../my_tests/000000000019.jpg ppyoloe_crn_m_300e_coco.param ppyoloe_crn_m_300e_coco.bin
cd build/examples
./test2_06_ppyoloe_ncnn ../../my_tests/000000000019.jpg ppyoloe_crn_l_300e_coco.param ppyoloe_crn_l_300e_coco.bin
cd build/examples
./test2_06_ppyoloe_ncnn ../../my_tests/000000000019.jpg ppyoloe_crn_x_300e_coco.param ppyoloe_crn_x_300e_coco.bin
cd build/examples
./test2_06_ppyoloe_ncnn ../../my_tests/000000000019.jpg ppyoloe_crn_l_voc2012_epoch_6.param ppyoloe_crn_l_voc2012_epoch_6.bin
./test2_06_ppyoloe_ncnn ../../my_tests/000000013659.jpg ppyoloe_crn_s_300e_coco.param ppyoloe_crn_s_300e_coco.bin
./test2_06_ppyoloe_ncnn ../../my_tests/000000013659.jpg ppyoloe_crn_m_300e_coco.param ppyoloe_crn_m_300e_coco.bin
./test2_06_ppyoloe_ncnn ../../my_tests/000000013659.jpg ppyoloe_crn_l_300e_coco.param ppyoloe_crn_l_300e_coco.bin
./test2_06_ppyoloe_ncnn ../../my_tests/000000013659.jpg ppyoloe_crn_x_300e_coco.param ppyoloe_crn_x_300e_coco.bin
./test2_06_ppyoloe_ncnn ../../my_tests/000000013659.jpg ppyoloe_crn_l_voc2012_epoch_6.param ppyoloe_crn_l_voc2012_epoch_6.bin