This project mainly include three parts.
1.Provides training methods for multiple mainstream object detection datasets(coco2017, coco2014, BDD100k, Visdrone, Hand)
2.Provides a mainstream model compression algorithm including pruning, quantization, and knowledge distillation.
3.Provides multiple backbone for yolov3 including Darknet-YOLOv3，Tiny-YOLOv3，Mobilenetv3-YOLOv3
Source using Pytorch implementation to ultralytics/yolov3 for yolov3 source code. Pruning method based on BN layer by coldlarry/YOLOv3-complete-pruning, thank you to both of you.
January 4, 2020. Provides download links and training methods to the Visdrone dataset.
January 19, 2020. Dior, Bdd100k and Visdrone training will be provided, as well as the converted weights file.
March 1, 2020. Provides Mobilenetv3 backbone.
April 7, 2020. Implement two models based on Mobilenetv3: Yolov3-Mobilenet, and Yolov3tin-Mobilene-small, provide pre-training weights, extend the normal pruning methods to the two Mobilenet-based models.
April 27, 2020. Update mobilenetv3 pre-training weights, add a layer pruning method, methods from the tanluren/yolov3-channel-and-layer-pruning/yolov3, Thanks for sharing.
May 22, 2020 Updated some new optimizations from ultralytics/yolov3, update cfg file and weights of YOLOv4.
May 22, 2020, the 8-bit fixed-point quantization method was updated and some bugs were fixed.
July 12, 2020, the problem of map returning to 0 after pruning and of yolov3-mobilenet was fixed. See issue#41 for more details.
July 14, 2020. Mobilenet supports a bn fusion method and two methods of extreme pruning that based on shortcut.
September 30, 2020, the BN_Fold training method was updated to reduce the precision loss caused by BN fusion, and the POW (2) quantization method targeted at FPGA was updated. See the quantization section for details.
Our project based on ultralytics/yolov3, see ultralytics/yolov3 for details. Here is a brief explanation:
numpytorch >= 1.1.0opencv-pythontqdm
| Function | |
|---|---|
| Multi-Backbone training | √ |
| Multi-Datasets | √ |
| Pruning | √ |
| Quantization | √ |
| Knowledge Distillation | √ |
python3 train.py --data ... --cfg ... For training model command, the -pt command is required when using coco pre-training model.
python3 test.py --data ... --cfg ... Test command
python3 detect.py --data ... --cfg ... --source ... Detection command, the default address of source is data/samples, the output result is saved in the output file, and the detection resource can be pictures, videos.
This project provides preprocessed datasets for the YOLOv3, configuration files (.cfg), dataset index files (.data), dataset category files (.names), and anchor box sizes (including 9 boxes for YOLOv3 and 6 boxes for tiny- YOLOv3) that are reclustered using the K-means algorithm. mAP
| Dataset | YOLOv3-640 | YOLOv4-640 | YOLOv3-mobilenet-640 |
|---|---|---|---|
| Dior | 0.749 | ||
| bdd100k | 0.543 | ||
| visdrone | 0.311 | 0.383 | 0.348 |
Download the following address, download and unzip the folder to copy to the data directory can be used.
-
Extract code：hjln
-
Extract code：rhqx
-
Extract code：k8ms
Training command
python3 train.py --data data/coco2017.data --batch-size ... --weights weights/yolov3-608.weights -pt --cfg cfg/yolov3/yolov3.cfg --img-size ... --epochs ...-
Extract code：vnuq
-
Extract code：l8wz
Training command
python3 train.py --data data/dior.data --batch-size ... --weights weights/yolov3-608.weights -pt --cfg cfg/yolov3/yolov3-onDIOR.cfg --img-size ... --epochs ...-
Extract code：8duw
-
Extract code：xeqo
Training command
python3 train.py --data data/bdd100k.data --batch-size ... --weights weights/yolov3-608.weights -pt --cfg cfg/yolov3/yolov3-bdd100k.cfg --img-size ... --epochs ...-
Extract code：dy4c
-
Extract code：87lf
-
Extract code：xblu
Extract code：fb6y
Training command
python train.py --data data/visdrone.data --batch-size ... --weights weights/yolov3-608.weights -pt --cfg cfg/yolov3/yolov3-visdrone.cfg --img-size ... --epochs ...-
Extract code：3du4
Training command
python train.py --data data/oxfordhand.data --batch-size ... --weights weights/yolov3-608.weights -pt --cfg cfg/yolov3/yolov3-visdrone.cfg --img-size ... --epochs ...The DIRO dataset is one of the largest, most diverse, and publicly available object detection datasets in the Earth observation community. Among them, the number of instances of ships and vehicles is high, which achieves a good balance between small instances and large ones. The images were collected from Google Earth.
Bdd100 is a large, diverse data set of driving videos containing 100,000 videos. Each video was about 40 seconds long, and the researchers marked bounding boxes for all 100,000 key frames of objects that often appeared on the road. The data set covers different weather conditions, including sunny, cloudy and rainy days, and different times of day and night.
The VisDrone2019 dataset was collected by AISKYEYE team at the Machine Learning and Data Mining Laboratory at Tianjin University, China. Benchmark data set contains 288 video clips, and consists of 261908 frames and 10209 frames a static image, by all sorts of installed on the unmanned aerial vehicle (uav) camera capture, covers a wide range of aspects, including location (thousands of kilometers apart from China in 14 different cities), environment (city and country), object (pedestrians, vehicles, bicycles, etc.) and density (sparse and crowded scenario). This data set was collected using a variety of uav platforms (i.e., uAvs with different models) in a variety of situations and under various weather and light conditions. These frames are manually marked with more than 2.6 million border frames, which are often targets of interest, such as pedestrians, cars, bicycles and tricycles. Some important attributes are also provided, including scene visibility, object categories, and occlusion, to improve data utilization.
Based on mobilenetv3, two kinds of network structures are designed
| Structure | backbone | Postprocessing | Parameters | GFLOPS | mAP0.5 | mAP0.5:0.95 | speed(inference/NMS/total) | FPS |
|---|---|---|---|---|---|---|---|---|
| YOLOv3 | 38.74M | 20.39M | 59.13M | 117.3 | 0.580 | 0.340 | 12.3/1.7/14.0 ms | 71.4fps |
| YOLOv3tiny | 6.00M | 2.45M | 8.45M | 9.9 | 0.347 | 0.168 | 3.5/1.8/5.3 ms | 188.7fps |
| YOLOv3-mobilenetv3 | 2.84M | 20.25M | 23.09M | 32.2 | 0.547 | 0.346 | 7.9/1.8/9.7 ms | 103.1fps |
| YOLOv3tiny-mobilenetv3-small | 0.92M | 2.00M | 2.92M | 2.9 | 0.379 | 0.214 | 5.2/1.9/7.1 ms | 140.8fps |
| YOLOv4 | - | - | 61.35M | 107.1 | 0.650 | 0.438 | 13.5/1.8/15.3 ms | 65.4fps |
| YOLOv4-tiny | - | - | 5.78M | 12.3 | 0.435 | 0.225 | 4.1/1.7/5.8 ms | 172.4fps |
-
YOLOv3,YOLOv3tiny and YOLOv4 were trained and tested on Coco2014, and Yolov3-Mobilenetv3 and YOLOv3tiny Mobilenetv3-Small were trained and tested on Coco2017.
-
The reasoning speed is tested on GTX2080ti*4, and the image size is 608.
-
The training test set should match the training set, as mismatch will cause the problem of map virtual height. Because reference issue
1.YOLOv3
python3 train.py --data data/... --batch-size ... -pt --weights weights/yolov3-608.weights --cfg cfg/yolov3/yolov3.cfg --img_size ...Weights Download
-
Extract code：k8ms
2.YOLOv3tiny
python3 train.py --data data/... --batch-size ... -pt --weights weights/yolov3tiny.weights --cfg cfg/yolov3tiny/yolov3-tiny.cfg --img_size ...-
Extract code：udfe
3.YOLOv3tiny-mobilenet-small
python3 train.py --data data/... --batch-size ... -pt --weights weights/yolov3tiny-mobilenet-small.weights --cfg cfg/yolov3tiny-mobilenet-small/yolov3tiny-mobilenet-small-coco.cfg --img_size ...-
Extract code：pxz4
4.YOLOv3-mobilenet
python3 train.py --data data/... --batch-size ... -pt --weights weights/yolov3-mobilenet.weights --cfg cfg/yolov3-mobilenet/yolov3-mobilenet-coco.cfg --img_size ...-
Extract code：3vm8
5.YOLOv4
python3 train.py --data data/... --batch-size ... -pt --weights weights/yolov4.weights --cfg cfg/yolov4/yolov4.cfg --img_size ...-
Extract code：njdg
| method | advantage | disadvantage |
|---|---|---|
| Normal pruning | Not prune for shortcut layer. It has a considerable and stable compression rate that requires no fine tuning. | The compression rate is not extreme. |
| Shortcut pruning | Very high compression rate. | need fine-tuning. |
| Silmming | Shortcut fusion method was used to improve the precision of shear planting. | Best way for shortcut pruning |
| Regular pruning | Designed for hardware deployment, the number of filters after pruning is a multiple of 2, no fine-tuning, support tiny-yolov3 and Mobilenet series. | Part of the compression ratio is sacrificed for regularization. |
| layer pruning | ResBlock is used as the basic unit for purning, which is conducive to hardware deployment. | It can only be cut backbone, with a limited prunning rate. |
| layer-channel pruning | First, use channel pruning and then use layer pruning, and pruning rate was very high. | Accuracy may be affected. |
1.Training
python3 train.py --data ... -pt --batch-size ... --weights ... --cfg ...2.Sparse training
-srSparse training，--sSpecifies the sparsity factor，--pruneSpecify the sparsity type.
--prune 0 is the sparsity of normal pruning and regular pruning.
--prune 1 is the sparsity of shortcut pruning.
--prune 2 is the sparsity of layer pruning.
command：
python3 train.py --data ... -pt --batch-size 32 --weights ... --cfg ... -sr --s 0.001 --prune 0 3.Pruning
- normal pruning
python3 normal_prune.py --cfg ... --data ... --weights ... --percent ...- regular pruning
python3 regular_prune.py --cfg ... --data ... --weights ... --percent ...- shortcut pruning
python3 shortcut_prune.py --cfg ... --data ... --weights ... --percent ...- silmming
python3 slim_prune.py --cfg ... --data ... --weights ... --percent ...- layer pruning
python3 layer_prune.py --cfg ... --data ... --weights ... --shortcut ...- layer-channel pruning
python3 layer_channel_prune.py --cfg ... --data ... --weights ... --shortcut ... --percent ...It is important to note that the cfg and weights variables in OPT need to be pointed to the cfg and weights files generated by step 2.
In addition, you can get more compression by increasing the percent value in the code. (If the sparsity is not enough and the percent value is too high, the program will report an error.)
1.normal pruning oxfordhand，img_size = 608，test on GTX2080Ti*4
| model | parameter before pruning | mAP before pruning | inference time before pruning | percent | parameter after pruning | mAP after pruning | inference time after pruning |
|---|---|---|---|---|---|---|---|
| yolov3(without fine tuning) | 58.67M | 0.806 | 0.1139s | 0.8 | 10.32M | 0.802 | 0.0844s |
| yolov3-mobilenet(fine tuning) | 22.75M | 0.812 | 0.0345s | 0.97 | 2.72M | 0.795 | 0.0211s |
| yolov3tiny(fine tuning) | 8.27M | 0.708 | 0.0144s | 0.5 | 1.13M | 0.641 | 0.0116s |
2.regular pruning oxfordhand，img_size = 608，test ong GTX2080Ti*4
| model | parameter before pruning | mAP before pruning | inference time before pruning | percent | parameter after pruning | mAP after pruning | inference time after pruning |
|---|---|---|---|---|---|---|---|
| yolov3(without fine tuning) | 58.67M | 0.806 | 0.1139s | 0.8 | 12.15M | 0.805 | 0.0874s |
| yolov3-mobilenet(fine tuning) | 22.75M | 0.812 | 0.0345s | 0.97 | 2.75M | 0.803 | 0.0208s |
| yolov3tiny(fine tuning) | 8.27M | 0.708 | 0.0144s | 0.5 | 1.82M | 0.703 | 0.0122s |
--quantized 2 Dorefa quantization method
python train.py --data ... --batch-size ... --weights ... --cfg ... --img-size ... --epochs ... --quantized 2--quantized 1 Google quantization method
python train.py --data ... --batch-size ... --weights ... --cfg ... --img-size ... --epochs ... --quantized 3--BN_Flod using BN Flod training, --FPGA Pow(2) quantization for FPGA.
oxfordhand, yolov3, 640image-size
| method | mAP |
|---|---|
| Baseline | 0.847 |
| Google8bit | 0.851 |
| Google8bit + BN Flod | 0.851 |
| Google8bit + BN Flod + FPGA | 0.852 |
| Google4bit + BN Flod + FPGA | 0.842 |
The distillation method is based on the basic distillation method proposed by Hinton in 2015, and has been partially improved in combination with the detection network.
Distilling the Knowledge in a Neural Network paper
command : --t_cfg --t_weights --KDstr
--t_cfg cfg file of teacher model
--t_weights weights file of teacher model
--KDstr KD strategy
`--KDstr 1` KLloss can be obtained directly from the output of Tencher network and the output of Student network and added to the overall loss.
`--KDstr 2` To distinguish between Boxloss and ClassLoss, the student does not learn directly from the teacher. L2 distance is calculated respectively for Student, teacher and GT. When Student is greater than teacher, an additional loss is added for Student and GT.
`--KDstr 3` To distinguish between Boxloss and ClassLoss, the student learns directly from the teacher.
`--KDstr 4` KDloss is divided into three categories, boxLoss, ClassLoss and FeatureLoss.
`--KDstr 5` On the basis of KDstr 4, the fine-grain-mask is added into the feature
example:
python train.py --data ... --batch-size ... --weights ... --cfg ... --img-size ... --epochs ... --t_cfg ... --t_weights ...Usually, the pre-compression model is used as the teacher model, and the post-compression model is used as the student model for distillation training to improve the mAP of student network.
oxfordhand，yolov3tiny as teacher model，normal pruning yolov3tiny as student model
| teacher model | mAP of teacher model | student model | directly fine tuning | KDstr 1 | KDstr 2 | KDstr 3 | KDstr 4(L1) | KDstr 5(L1) |
|---|---|---|---|---|---|---|---|---|
| yolov3tiny608 | 0.708 | normal pruning yolov3tiny608 | 0.658 | 0.666 | 0.661 | 0.672 | 0.673 | 0.674 |