README.md 18.9 KB
Newer Older
Francesco Gatti's avatar
README  
Francesco Gatti committed
1
# tkDNN
2
tkDNN is a Deep Neural Network library built with cuDNN and tensorRT primitives, specifically thought to work on NVIDIA Jetson Boards. It has been tested on TK1(branch cudnn2), TX1, TX2, AGX Xavier and several discrete GPU.
Micaela Verucchi's avatar
Micaela Verucchi committed
3
The main goal of this project is to exploit NVIDIA boards as much as possible to obtain the best inference performance. It does not allow training. 
Francesco Gatti's avatar
README  
Francesco Gatti committed
4

Micaela Verucchi's avatar
Micaela Verucchi committed
5
6
7
8

If you use tkDNN in your research, please cite one of the following papers. For use in commercial solutions, write at gattifrancesco@hotmail.it or refer to https://hipert.unimore.it/ .

```
Micaela Verucchi's avatar
Micaela Verucchi committed
9
10
11
Accepted paper @ IRC 2020, will soon been published.
M. Verucchi, L. Bartoli, F. Bagni, F. Gatti, P. Burgio and M. Bertogna, "Real-Time clustering and LiDAR-camera fusion on embedded platforms for self-driving cars",  in proceedings in IEEE Robotic Computing (2020)

Micaela Verucchi's avatar
Micaela Verucchi committed
12
13
14
15
16
Accepted paper @ ETFA 2020, will soon been published.
M. Verucchi, G. Brilli, D. Sapienza, M. Verasani, M. Arena, F. Gatti, A. Capotondi, R. Cavicchioli, M. Bertogna, M. Solieri
"A Systematic Assessment of Embedded Neural Networks for Object Detection", in IEEE International Conference on Emerging Technologies and Factory Automation (2020)
```

Micaela Verucchi's avatar
Micaela Verucchi committed
17
## Index
Francesco Gatti's avatar
Francesco Gatti committed
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
- [tkDNN](#tkdnn)
  - [Index](#index)
  - [Dependencies](#dependencies)
  - [About OpenCV](#about-opencv)
  - [How to compile this repo](#how-to-compile-this-repo)
  - [Workflow](#workflow)
  - [How to export weights](#how-to-export-weights)
    - [1)Export weights from darknet](#1export-weights-from-darknet)
    - [2)Export weights for DLA34 and ResNet101](#2export-weights-for-dla34-and-resnet101)
    - [3)Export weights for CenterNet](#3export-weights-for-centernet)
    - [4)Export weights for MobileNetSSD](#4export-weights-for-mobilenetssd)
  - [Run the demo](#run-the-demo)
    - [FP16 inference](#fp16-inference)
    - [INT8 inference](#int8-inference)
  - [mAP demo](#map-demo)
  - [Existing tests and supported networks](#existing-tests-and-supported-networks)
  - [References](#references)
Micaela Verucchi's avatar
Micaela Verucchi committed
35
36
37
38
39
40




## Dependencies
This branch works on every NVIDIA GPU that supports the dependencies:
41
42
43
* CUDA 10.0
* CUDNN 7.603
* TENSORRT 6.01
44
* OPENCV 3.4
45
* yaml-cpp 0.5.2 (sudo apt install libyaml-cpp-dev)
Francesco Gatti's avatar
README  
Francesco Gatti committed
46

47
48
49
## About OpenCV
To compile and install OpenCV4 with contrib us the script ```install_OpenCV4.sh```. It will download and compile OpenCV in Download folder.
```
Micaela Verucchi's avatar
Micaela Verucchi committed
50
bash scripts/install_OpenCV4.sh
51
52
53
```
When using openCV not compiled with contrib, comment the definition of OPENCV_CUDACONTRIBCONTRIB in include/tkDNN/DetectionNN.h. When commented, the preprocessing of the networks is computed on the CPU, otherwise on the GPU. In the latter case some milliseconds are saved in the end-to-end latency. 

Micaela Verucchi's avatar
Micaela Verucchi committed
54
## How to compile this repo
Francesco Gatti's avatar
Francesco Gatti committed
55
Build with cmake. If using Ubuntu 18.04 a new version of cmake is needed (3.15 or above). 
Francesco Gatti's avatar
README  
Francesco Gatti committed
56
```
Micaela Verucchi's avatar
Micaela Verucchi committed
57
git clone https://github.com/ceccocats/tkDNN
58
cd tkDNN
Francesco Gatti's avatar
README  
Francesco Gatti committed
59
60
mkdir build
cd build
Francesco Gatti's avatar
Francesco Gatti committed
61
cmake .. 
Francesco Gatti's avatar
README  
Francesco Gatti committed
62
63
64
make
```

Micaela Verucchi's avatar
Micaela Verucchi committed
65
66
## Workflow
Steps needed to do inference on tkDNN with a custom neural network. 
Micaela Verucchi's avatar
Micaela Verucchi committed
67
* Build and train a NN model with your favorite framework.
Micaela Verucchi's avatar
Micaela Verucchi committed
68
69
70
71
* Export weights and bias for each layer and save them in a binary file (one for layer).
* Export outputs for each layer and save them in a binary file (one for layer).
* Create a new test and define the network, layer by layer using the weights extracted and the output to check the results. 
* Do inference.
Davide Sapienza's avatar
Davide Sapienza committed
72

Micaela Verucchi's avatar
Micaela Verucchi committed
73
74
## How to export weights

75
Weights are essential for any network to run inference. For each test a folder organized as follow is needed (in the build folder):
Davide Sapienza's avatar
Davide Sapienza committed
76
```
Micaela Verucchi's avatar
Micaela Verucchi committed
77
78
79
    test_nn
        |---- layers/ (folder containing a binary file for each layer with the corresponding wieghts and bias)
        |---- debug/  (folder containing a binary file for each layer with the corresponding outputs)
Davide Sapienza's avatar
Davide Sapienza committed
80
```
Micaela Verucchi's avatar
Micaela Verucchi committed
81
Therefore, once the weights have been exported, the folders layers and debug should be placed in the corresponding test.
Davide Sapienza's avatar
Davide Sapienza committed
82

Micaela Verucchi's avatar
Micaela Verucchi committed
83
### 1)Export weights from darknet
Francesco Gatti's avatar
Francesco Gatti committed
84
To export weights for NNs that are defined in darknet framework, use [this](https://git.hipert.unimore.it/fgatti/darknet.git) fork of darknet and follow these steps to obtain a correct debug and layers folder, ready for tkDNN.
Davide Sapienza's avatar
Davide Sapienza committed
85
86

```
Francesco Gatti's avatar
Francesco Gatti committed
87
git clone https://git.hipert.unimore.it/fgatti/darknet.git
88
cd darknet
Micaela Verucchi's avatar
Micaela Verucchi committed
89
90
91
make
mkdir layers debug
./darknet export <path-to-cfg-file> <path-to-weights> layers
Davide Sapienza's avatar
Davide Sapienza committed
92
```
Micaela Verucchi's avatar
Micaela Verucchi committed
93
N.b. Use compilation with CPU (leave GPU=0 in Makefile) if you also want debug. 
Davide Sapienza's avatar
Davide Sapienza committed
94

Micaela Verucchi's avatar
Micaela Verucchi committed
95
96
### 2)Export weights for DLA34 and ResNet101 
To get weights and outputs needed to run the tests dla34 and resnet101 use the Python script and the Anaconda environment included in the repository.   
Davide Sapienza's avatar
Davide Sapienza committed
97

Micaela Verucchi's avatar
Micaela Verucchi committed
98
Create Anaconda environment and activate it:
Francesco Gatti's avatar
Francesco Gatti committed
99
```
Micaela Verucchi's avatar
Micaela Verucchi committed
100
101
102
conda env create -f file_name.yml
source activate env_name
python <script name>
Francesco Gatti's avatar
Francesco Gatti committed
103
```
Micaela Verucchi's avatar
Micaela Verucchi committed
104
105
### 3)Export weights for CenterNet
To get the weights needed to run Centernet tests use [this](https://github.com/sapienzadavide/CenterNet.git) fork of the original Centernet. 
Francesco Gatti's avatar
Francesco Gatti committed
106
```
Micaela Verucchi's avatar
Micaela Verucchi committed
107
git clone https://github.com/sapienzadavide/CenterNet.git
Francesco Gatti's avatar
Francesco Gatti committed
108
```
Micaela Verucchi's avatar
Micaela Verucchi committed
109
* follow the instruction in the README.md and INSTALL.md
Davide Sapienza's avatar
Davide Sapienza committed
110
111

```
Micaela Verucchi's avatar
Micaela Verucchi committed
112
113
python demo.py --input_res 512 --arch resdcn_101 ctdet --demo /path/to/image/or/folder/or/video/or/webcam --load_model ../models/ctdet_coco_resdcn101.pth --exp_wo --exp_wo_dim 512
python demo.py --input_res 512 --arch dla_34 ctdet --demo /path/to/image/or/folder/or/video/or/webcam --load_model ../models/ctdet_coco_dla_2x.pth --exp_wo --exp_wo_dim 512
Davide Sapienza's avatar
Davide Sapienza committed
114
```
Micaela Verucchi's avatar
Micaela Verucchi committed
115
### 4)Export weights for MobileNetSSD
Micaela Verucchi's avatar
Micaela Verucchi committed
116
To get the weights needed to run Mobilenet tests use [this](https://github.com/mive93/pytorch-ssd) fork of a Pytorch implementation of SSD network. 
Davide Sapienza's avatar
Davide Sapienza committed
117
118

```
Micaela Verucchi's avatar
Micaela Verucchi committed
119
120
121
122
git clone https://github.com/mive93/pytorch-ssd
cd pytorch-ssd
conda env create -f env_mobv2ssd.yml
python run_ssd_live_demo.py mb2-ssd-lite <pth-model-fil> <labels-file>
Davide Sapienza's avatar
Davide Sapienza committed
123
```
Francesco Gatti's avatar
Francesco Gatti committed
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

## Darknet Parser
tkDNN implement and easy parser for darknet cfg files, a network can be converted with *tk::dnn::darknetParser*:
```
// example of parsing yolo4
tk::dnn::Network *net = tk::dnn::darknetParser("yolov4.cfg", "yolov4/layers", "coco.names");
net->print();
```
All models from darknet are now parsed directly from cfg, you still need to export the weights with the descripted tools in the previus section.
<details>
  <summary>Supported layers</summary>
  convolutional
  maxpool
  avgpool
  shortcut
  upsample
  route
  reorg
  region
  yolo
</details>
<details>
  <summary>Supported activations</summary>
  relu
  leaky
  mish
</details>

Micaela Verucchi's avatar
Micaela Verucchi committed
152
## Run the demo
Davide Sapienza's avatar
Davide Sapienza committed
153

Micaela Verucchi's avatar
Micaela Verucchi committed
154
To run the an object detection demo follow these steps (example with yolov3):
Davide Sapienza's avatar
Davide Sapienza committed
155
```
Francesco Gatti's avatar
Francesco Gatti committed
156
rm yolo3_fp32.rt        # be sure to delete(or move) old tensorRT files
Micaela Verucchi's avatar
Micaela Verucchi committed
157
./test_yolo3            # run the yolo test (is slow)
Francesco Gatti's avatar
Francesco Gatti committed
158
./demo yolo3_fp32.rt ../demo/yolo_test.mp4 y
Davide Sapienza's avatar
Davide Sapienza committed
159
```
160
In general the demo program takes 4 parameters:
Davide Sapienza's avatar
Davide Sapienza committed
161
```
162
./demo <network-rt-file> <path-to-video> <kind-of-network> <number-of-classes> <n-batches> <show-flag>
163
```
Micaela Verucchi's avatar
Micaela Verucchi committed
164
165
166
167
where
*  ```<network-rt-file>``` is the rt file generated by a test
*  ```<<path-to-video>``` is the path to a video file or a camera input  
*  ```<kind-of-network>``` is the type of network. Thee types are currently supported: ```y``` (YOLO family), ```c``` (CenterNet family) and ```m``` (MobileNet-SSD family)
168
*  ```<number-of-classes>```is the number of classes the network is trained on
169
170
171
*  ```<n-batches>``` number of batches to use in inference (N.B. you should first export TKDNN_BATCHSIZE to the required n_batches and create again the rt file for the network).
*  ```<show-flag>``` if set to 0 the demo will not show the visualization but save the video into result.mp4 (if n-batches ==1)

Davide Sapienza's avatar
Davide Sapienza committed
172
N.b. By default it is used FP32 inference
Micaela Verucchi's avatar
Micaela Verucchi committed
173
174

![demo](https://user-images.githubusercontent.com/11562617/72547657-540e7800-388d-11ea-83c6-49dfea2a0607.gif)
175

Davide Sapienza's avatar
Davide Sapienza committed
176
177
178
179
180
### FP16 inference

To run the an object detection demo with FP16 inference follow these steps (example with yolov3):
```
export TKDNN_MODE=FP16  # set the half floating point optimization
Francesco Gatti's avatar
Francesco Gatti committed
181
rm yolo3_fp16.rt        # be sure to delete(or move) old tensorRT files
Davide Sapienza's avatar
Davide Sapienza committed
182
./test_yolo3            # run the yolo test (is slow)
Francesco Gatti's avatar
Francesco Gatti committed
183
./demo yolo3_fp16.rt ../demo/yolo_test.mp4 y
Davide Sapienza's avatar
Davide Sapienza committed
184
185
186
187
188
```
N.b. Using FP16 inference will lead to some errors in the results (first or second decimal). 

### INT8 inference

Micaela Verucchi's avatar
Micaela Verucchi committed
189
190
191
192
193
194
To run the an object detection demo with INT8 inference three environment variables need to be set:
  * ```export TKDNN_MODE=INT8```: set the 8-bit integer optimization
  * ```export TKDNN_CALIB_IMG_PATH=/path/to/calibration/image_list.txt``` : image_list.txt has in each line the absolute path to a calibration image
  * ```export TKDNN_CALIB_LABEL_PATH=/path/to/calibration/label_list.txt```: label_list.txt has in each line the absolute path to a calibration label
  
You should provide image_list.txt and label_list.txt, using training images. However, if you want to quickly test the INT8 inference you can run (from this repo root folder)
Davide Sapienza's avatar
Davide Sapienza committed
195
```
Micaela Verucchi's avatar
Micaela Verucchi committed
196
197
198
bash scripts/download_validation.sh COCO
```
to automatically download COCO2017 validation (inside demo folder) and create those needed file. Use BDD insted of COCO to download BDD validation. 
Davide Sapienza's avatar
Davide Sapienza committed
199

Micaela Verucchi's avatar
Micaela Verucchi committed
200
201
202
203
204
Then a complete example using yolo3 and COCO dataset would be:
```
export TKDNN_MODE=INT8
export TKDNN_CALIB_LABEL_PATH=../demo/COCO_val2017/all_labels.txt
export TKDNN_CALIB_IMG_PATH=../demo/COCO_val2017/all_images.txt
Francesco Gatti's avatar
Francesco Gatti committed
205
rm yolo3_int8.rt        # be sure to delete(or move) old tensorRT files
Davide Sapienza's avatar
Davide Sapienza committed
206
./test_yolo3            # run the yolo test (is slow)
Francesco Gatti's avatar
Francesco Gatti committed
207
./demo yolo3_int8.rt ../demo/yolo_test.mp4 y
Davide Sapienza's avatar
Davide Sapienza committed
208
```
Micaela Verucchi's avatar
Micaela Verucchi committed
209
210
211
212
213
N.B. 
 * Using INT8 inference will lead to some errors in the results. 
 * The test will be slower: this is due to the INT8 calibration, which may take some time to complete. 
 * INT8 calibration requires TensorRT version greater than or equal to 6.0
 * Only 100 images are used to create the calibration table by default (set in the code).
Davide Sapienza's avatar
Davide Sapienza committed
214

215
216
217
### BatchSize bigger than 1
```
export TKDNN_BATCHSIZE=2
Francesco Gatti's avatar
Francesco Gatti committed
218
219
220
221
222
223
# build tensorRT files
```
This will create a TensorRT file with the desidered **max** batch size.
The test will still run with a batch of 1, but the created tensorRT can manage the desidered batch size.

### Test batch Inference
Francesco Gatti's avatar
Francesco Gatti committed
224
This will test the network with random input and check if the output of each batch is the same.
Francesco Gatti's avatar
Francesco Gatti committed
225
226
227
228
229
230
231
232
233
```
./test_rtinference <network-rt-file> <number-of-batches>
# <number-of-batches> should be less or equal to the max batch size of the <network-rt-file>

# example
export TKDNN_BATCHSIZE=4           # set max batch size
rm yolo3_fp32.rt                   # be sure to delete(or move) old tensorRT files
./test_yolo3                       # build RT file
./test_rtinference yolo3_fp32.rt 4 # test with a batch size of 4
234
235
```

236
## mAP demo
Davide Sapienza's avatar
Davide Sapienza committed
237

238
239
To compute mAP, precision, recall and f1score, run the map_demo.

240
241
A validation set is needed. 
To download COCO_val2017 (80 classes) run (form the root folder): 
xavier's avatar
xavier committed
242
```
243
bash scripts/download_validation.sh COCO
xavier's avatar
xavier committed
244
```
245
246
247
248
249
To download Berkeley_val (10 classes) run (form the root folder): 
```
bash scripts/download_validation.sh BDD
```

xavier's avatar
xavier committed
250
To compute the map, the following parameters are needed:
251
```
Micaela Verucchi's avatar
Micaela Verucchi committed
252
./map_demo <network rt> <network type [y|c|m]> <labels file path> <config file path>
253
254
```
where 
Micaela Verucchi's avatar
Micaela Verucchi committed
255
* ```<network rt>```: rt file of a chosen network on which compute the mAP.
Micaela Verucchi's avatar
Micaela Verucchi committed
256
* ```<network type [y|c|m]>```: type of network. Right now only y(yolo), c(centernet) and m(mobilenet) are allowed
Micaela Verucchi's avatar
Micaela Verucchi committed
257
* ```<labels file path>```: path to a text file containing all the paths of the ground-truth labels. It is important that all the labels of the ground-truth are in a folder called 'labels'. In the folder containing the folder 'labels' there should be also a folder 'images', containing all the ground-truth images having the same same as the labels. To better understand, if there is a label path/to/labels/000001.txt there should be a corresponding image path/to/images/000001.jpg. 
Micaela Verucchi's avatar
Micaela Verucchi committed
258
* ```<config file path>```: path to a yaml file with the parameters needed for the mAP computation, similar to demo/config.yaml
259
260
261
262

Example:

```
xavier's avatar
xavier committed
263
cd build
Davide Sapienza's avatar
Davide Sapienza committed
264
./map_demo dla34_cnet_FP32.rt c ../demo/COCO_val2017/all_labels.txt ../demo/config.yaml
xavier's avatar
xavier committed
265
```
Micaela Verucchi's avatar
Micaela Verucchi committed
266

267
268
This demo also creates a json file named ```net_name_COCO_res.json``` containing all the detections computed. The detections are in COCO format, the correct format to subit the results to [CodaLab COCO detection challenge](https://competitions.codalab.org/competitions/20794#participate).

Micaela Verucchi's avatar
Micaela Verucchi committed
269
## Existing tests and supported networks
Micaela Verucchi's avatar
Micaela Verucchi committed
270
271
272

| Test Name         | Network                                       | Dataset                                                       | N Classes | Input size    | Weights                                                                   |
| :---------------- | :-------------------------------------------- | :-----------------------------------------------------------: | :-------: | :-----------: | :------------------------------------------------------------------------ |
Micaela Verucchi's avatar
Micaela Verucchi committed
273
| yolo              | YOLO v2<sup>1</sup>                           | [COCO 2014](http://cocodataset.org/)                          | 80        | 608x608       | [weights](https://cloud.hipert.unimore.it/s/nf4PJ3k8bxBETwL/download)                                                                   |
Micaela Verucchi's avatar
Micaela Verucchi committed
274
275
276
| yolo_224          | YOLO v2<sup>1</sup>                           | [COCO 2014](http://cocodataset.org/)                          | 80        | 224x224       | weights                                                                   |
| yolo_berkeley     | YOLO v2<sup>1</sup>                           | [BDD100K  ](https://bair.berkeley.edu/blog/2018/05/30/bdd/)   | 10        | 416x736       | weights                                                                   |
| yolo_relu         | YOLO v2 (with ReLU, not Leaky)<sup>1</sup>    | [COCO 2014](http://cocodataset.org/)                          | 80        | 416x416       | weights                                                                   |
Micaela Verucchi's avatar
Micaela Verucchi committed
277
| yolo_tiny         | YOLO v2 tiny<sup>1</sup>                      | [COCO 2014](http://cocodataset.org/)                          | 80        | 416x416       | [weights](https://cloud.hipert.unimore.it/s/m3orfJr8pGrN5mQ/download)                                                                   |
Micaela Verucchi's avatar
Micaela Verucchi committed
278
| yolo_voc          | YOLO v2<sup>1</sup>                           | [VOC      ](http://host.robots.ox.ac.uk/pascal/VOC/)          | 21        | 416x416       | [weights](https://cloud.hipert.unimore.it/s/DJC5Fi2pEjfNDP9/download)                                                                   |
Micaela Verucchi's avatar
Micaela Verucchi committed
279
| yolo3             | YOLO v3<sup>2</sup>                           | [COCO 2014](http://cocodataset.org/)                          | 80        | 416x416       | [weights](https://cloud.hipert.unimore.it/s/jPXmHyptpLoNdNR/download)     |
280
| yolo3_512   | YOLO v3<sup>2</sup>                                 | [COCO 2017](http://cocodataset.org/)                          | 80        | 512x512       | [weights](https://cloud.hipert.unimore.it/s/RGecMeGLD4cXEWL/download)     |
Micaela Verucchi's avatar
Micaela Verucchi committed
281
| yolo3_berkeley    | YOLO v3<sup>2</sup>                           | [BDD100K  ](https://bair.berkeley.edu/blog/2018/05/30/bdd/)   | 10        | 320x544       | [weights](https://cloud.hipert.unimore.it/s/o5cHa4AjTKS64oD/download)                                                                   |
Micaela Verucchi's avatar
Micaela Verucchi committed
282
283
| yolo3_coco4       | YOLO v3<sup>2</sup>                           | [COCO 2014](http://cocodataset.org/)                          | 4         | 416x416       | [weights](https://cloud.hipert.unimore.it/s/o27NDzSAartbyc4/download)                                                                   |
| yolo3_flir        | YOLO v3<sup>2</sup>                           | [FREE FLIR](https://www.flir.com/oem/adas/adas-dataset-form/) | 3         | 320x544       | [weights](https://cloud.hipert.unimore.it/s/62DECncmF6bMMiH/download)                                                                   |
Micaela Verucchi's avatar
Micaela Verucchi committed
284
| yolo3_tiny        | YOLO v3 tiny<sup>2</sup>                      | [COCO 2014](http://cocodataset.org/)                          | 80        | 416x416       | [weights](https://cloud.hipert.unimore.it/s/LMcSHtWaLeps8yN/download)     |
285
| yolo3_tiny512     | YOLO v3 tiny<sup>2</sup>                      | [COCO 2017](http://cocodataset.org/)                          | 80        | 512x512       | [weights](https://cloud.hipert.unimore.it/s/8Zt6bHwHADqP4JC/download)     |
Micaela Verucchi's avatar
Micaela Verucchi committed
286
| dla34             | Deep Leayer Aggreagtion (DLA) 34<sup>3</sup>  | [COCO 2014](http://cocodataset.org/)                          | 80        | 224x224       | weights                                                                   |
Micaela Verucchi's avatar
Micaela Verucchi committed
287
| dla34_cnet        | Centernet (DLA34 backend)<sup>4</sup>         | [COCO 2017](http://cocodataset.org/)                          | 80        | 512x512       | [weights](https://cloud.hipert.unimore.it/s/KRZBbCQsKAtQwpZ/download)     |
Micaela Verucchi's avatar
Micaela Verucchi committed
288
| mobilenetv2ssd    | Mobilnet v2 SSD Lite<sup>5</sup>              | [VOC      ](http://host.robots.ox.ac.uk/pascal/VOC/)          | 21        | 300x300       | [weights](https://cloud.hipert.unimore.it/s/x4ZfxBKN23zAJQp/download)     |
289
| mobilenetv2ssd512 | Mobilnet v2 SSD Lite<sup>5</sup>              | [COCO 2017](http://cocodataset.org/)                          | 81        | 512x512       | [weights](https://cloud.hipert.unimore.it/s/pdCw2dYyHMJrcEM/download)     |
Micaela Verucchi's avatar
Micaela Verucchi committed
290
| resnet101         | Resnet 101<sup>6</sup>                        | [COCO 2014](http://cocodataset.org/)                          | 80        | 224x224       | weights                                                                   |
Micaela Verucchi's avatar
Micaela Verucchi committed
291
| resnet101_cnet    | Centernet (Resnet101 backend)<sup>4</sup>     | [COCO 2017](http://cocodataset.org/)                          | 80        | 512x512       | [weights](https://cloud.hipert.unimore.it/s/5BTjHMWBcJk8g3i/download)     |
Micaela Verucchi's avatar
Micaela Verucchi committed
292
| csresnext50-panet-spp    | Cross Stage Partial Network <sup>7</sup>     | [COCO 2014](http://cocodataset.org/)                          | 80        | 416x416       | [weights](https://cloud.hipert.unimore.it/s/Kcs4xBozwY4wFx8/download)     |
Micaela Verucchi's avatar
Micaela Verucchi committed
293
| yolo4             | Yolov4 <sup>8</sup>                           | [COCO 2017](http://cocodataset.org/)                          | 80        | 416x416       | [weights](https://cloud.hipert.unimore.it/s/d97CFzYqCPCp5Hg/download)     |
Micaela Verucchi's avatar
Micaela Verucchi committed
294
295
296
297
298
299
300
301
302


## References

1. Redmon, Joseph, and Ali Farhadi. "YOLO9000: better, faster, stronger." Proceedings of the IEEE conference on computer vision and pattern recognition. 2017.
2. Redmon, Joseph, and Ali Farhadi. "Yolov3: An incremental improvement." arXiv preprint arXiv:1804.02767 (2018).
3. Yu, Fisher, et al. "Deep layer aggregation." Proceedings of the IEEE conference on computer vision and pattern recognition. 2018.
4. Zhou, Xingyi, Dequan Wang, and Philipp Krähenbühl. "Objects as points." arXiv preprint arXiv:1904.07850 (2019).
5. Sandler, Mark, et al. "Mobilenetv2: Inverted residuals and linear bottlenecks." Proceedings of the IEEE conference on computer vision and pattern recognition. 2018.
Micaela Verucchi's avatar
Micaela Verucchi committed
303
6. He, Kaiming, et al. "Deep residual learning for image recognition." Proceedings of the IEEE conference on computer vision and pattern recognition. 2016.
Micaela Verucchi's avatar
Micaela Verucchi committed
304
7. Wang, Chien-Yao, et al. "CSPNet: A New Backbone that can Enhance Learning Capability of CNN." arXiv preprint arXiv:1911.11929 (2019).
Micaela Verucchi's avatar
Micaela Verucchi committed
305
8. Bochkovskiy, Alexey, Chien-Yao Wang, and Hong-Yuan Mark Liao. "YOLOv4: Optimal Speed and Accuracy of Object Detection." arXiv preprint arXiv:2004.10934 (2020).