tensorflow-yolov4-tflite

YOLOv4, YOLOv4-tiny Implemented in Tensorflow 2.0. Convert YOLO v4, YOLOv3, YOLO tiny .weights to .pb, .tflite and trt format for tensorflow, tensorflow lite, tensorRT.

Download yolov4.weights file: https://drive.google.com/open?id=1cewMfusmPjYWbrnuJRuKhPMwRe_b9PaT

Prerequisites

• Tensorflow 2.3.0rc0

Performance

Demo

# Convert darknet weights to tensorflow
yolov4
python savemodel.py --weights ./data/yolov4.weights --output ./checkpoints/yolov4-416 --inputsize 416 --model yolov4
yolov4-tiny
python savemodel.py --weights ./data/yolov4-tiny.weights --output ./checkpoints/yolov4-tiny-416 --inputsize 416 --model yolov4 --tiny
Run demo tensorflow
python detect.py --weights ./checkpoints/yolov4-416 --size 416 --model yolov4 --image ./data/kite.jpg
python detect.py --weights ./checkpoints/yolov4-tiny-416 --size 416 --model yolov4 --image ./data/kite.jpg --tiny

If you want to run yolov3 or yolov3-tiny change `--model yolov3` in command

Output

Yolov4 original weight

Yolov4 tflite int8

Convert to tflite

# Save tf model for tflite converting
python savemodel.py --weights ./data/yolov4.weights --output ./checkpoints/yolov4-416 --inputsize 416 --model yolov4 --framework tflite
yolov4
python convert_tflite.py --weights ./checkpoints/yolov4-416 --output ./checkpoints/yolov4-416.tflite
yolov4 quantize float16
python converttflite.py --weights ./checkpoints/yolov4-416 --output ./checkpoints/yolov4-416-fp16.tflite --quantizemode float16
yolov4 quantize int8
python converttflite.py --weights ./checkpoints/yolov4-416 --output ./checkpoints/yolov4-416-int8.tflite --quantizemode int8 --dataset ./coco_dataset/coco/val207.txt
Run demo tflite model
python detect.py --weights ./checkpoints/yolov4-416.tflite --size 416 --model yolov4 --image ./data/kite.jpg --framework tflite

Yolov4 and Yolov4-tiny int8 quantization have some issues. I will try to fix that. You can try Yolov3 and Yolov3-tiny int8 quantization

Convert to TensorRT

# yolov3
python savemodel.py --weights ./data/yolov3.weights --output ./checkpoints/yolov3.tf --inputsize 416 --model yolov3
python converttrt.py --weights ./checkpoints/yolov3.tf --quantizemode float16 --output ./checkpoints/yolov3-trt-fp16-416
yolov3-tiny
python savemodel.py --weights ./data/yolov3-tiny.weights --output ./checkpoints/yolov3-tiny.tf --inputsize 416 --tiny
python converttrt.py --weights ./checkpoints/yolov3-tiny.tf --quantizemode float16 --output ./checkpoints/yolov3-tiny-trt-fp16-416
yolov4
python savemodel.py --weights ./data/yolov4.weights --output ./checkpoints/yolov4.tf --inputsize 416 --model yolov4
python converttrt.py --weights ./checkpoints/yolov4.tf --quantizemode float16 --output ./checkpoints/yolov4-trt-fp16-416

Evaluate on COCO 2017 Dataset

# run script in /script/getcocodataset_2017.sh to download COCO 2017 Dataset
preprocess coco dataset
cd data
mkdir dataset
cd ..
cd scripts
python cococonvert.py --input ./coco/annotations/instancesval2017.json --output val2017.pkl
python cocoannotation.py --cocopath ./coco 
cd ..
evaluate yolov4 model
python evaluate.py --weights ./data/yolov4.weights
cd mAP/extra
python remove_space.py
cd ..
python main.py --output resultsyolov4tf

mAP50 on COCO 2017 Dataset

| Detection | 512x512 | 416x416 | 320x320 | |-------------|---------|---------|---------| | YoloV3 | 55.43 | 52.32 | | | YoloV4 | 61.96 | 57.33 | |

Benchmark

python benchmarks.py --size 416 --model yolov4 --weights ./data/yolov4.weights

TensorRT performance

| YoloV4 416 images/s | FP32 | FP16 | INT8 | |---------------------|----------|----------|----------| | Batch size 1 | 55 | 116 | | | Batch size 8 | 70 | 152 | |

Tesla P100

| Detection | 512x512 | 416x416 | 320x320 | |-------------|---------|---------|---------| | YoloV3 FPS | 40.6 | 49.4 | 61.3 | | YoloV4 FPS | 33.4 | 41.7 | 50.0 |

Tesla K80

| Detection | 512x512 | 416x416 | 320x320 | |-------------|---------|---------|---------| | YoloV3 FPS | 10.8 | 12.9 | 17.6 | | YoloV4 FPS | 9.6 | 11.7 | 16.0 |

Tesla T4

| Detection | 512x512 | 416x416 | 320x320 | |-------------|---------|---------|---------| | YoloV3 FPS | 27.6 | 32.3 | 45.1 | | YoloV4 FPS | 24.0 | 30.3 | 40.1 |

Tesla P4

| Detection | 512x512 | 416x416 | 320x320 | |-------------|---------|---------|---------| | YoloV3 FPS | 20.2 | 24.2 | 31.2 | | YoloV4 FPS | 16.2 | 20.2 | 26.5 |

Macbook Pro 15 (2.3GHz i7)

| Detection | 512x512 | 416x416 | 320x320 | |-------------|---------|---------|---------| | YoloV3 FPS | | | | | YoloV4 FPS | | | |

Traning your own model

# Prepare your dataset
If you want to train from scratch:
In config.py set FISRTSTAGEEPOCHS=0 
Run script:
python train.py
Transfer learning: 
python train.py --weights ./data/yolov4.weights

The training performance is not fully reproduced yet, so I recommended to use Alex's Darknet to train your own data, then convert the .weights to tensorflow or tflite.

TODO

• [x] Convert YOLOv4 to TensorRT
• [x] YOLOv4 tflite on android
• [ ] YOLOv4 tflite on ios
• [x] Training code
• [x] Update scale xy
• [ ] ciou
• [ ] Mosaic data augmentation
• [x] Mish activation
• [x] yolov4 tflite version
• [x] yolov4 in8 tflite version for mobile

References

* YOLOv4: Optimal Speed and Accuracy of Object Detection YOLOv4. * darknet My project is inspired by these previous fantastic YOLOv3 implementations: * Yolov3 tensorflow * Yolov3 tf2