yolov3-tf2
This repo provides a clean implementation of YoloV3 in TensorFlow 2.0 using all the best practices.
TensorFlow 2.0
yolov3
with pre-trained Weights
yolov3-tiny
with pre-trained Weights
Inference example
Transfer learning example
Eager mode training with tf.GradientTape
Graph mode training with model.fit
Functional model with tf.keras.layers
Input pipeline using tf.data
Tensorflow Serving
Vectorized transformations
GPU accelerated
Fully integrated with absl-py
from abseil.io
Clean implementation
Following the best practices
MIT License
# Tensorflow CPUconda env create -f conda-cpu.yml conda activate yolov3-tf2-cpu# Tensorflow GPUconda env create -f conda-gpu.yml conda activate yolov3-tf2-gpu
pip install -r requirements.txt
# Ubuntu 18.04sudo apt-add-repository -r ppa:graphics-drivers/ppa sudo apt install nvidia-driver-430# Windows/Otherhttps://www.nvidia.com/Download/index.aspx
# yolov3wget https://pjreddie.com/media/files/yolov3.weights -O data/yolov3.weights python convert.py# yolov3-tinywget https://pjreddie.com/media/files/yolov3-tiny.weights -O data/yolov3-tiny.weights python convert.py --weights ./data/yolov3-tiny.weights --output ./checkpoints/yolov3-tiny.tf --tiny
# yolov3python detect.py --image ./data/meme.jpg# yolov3-tinypython detect.py --weights ./checkpoints/yolov3-tiny.tf --tiny --image ./data/street.jpg# webcampython detect_video.py --video 0# video filepython detect_video.py --video path_to_file.mp4 --weights ./checkpoints/yolov3-tiny.tf --tiny# video file with outputpython detect_video.py --video path_to_file.mp4 --output ./output.avi
You need to generate tfrecord following the TensorFlow Object Detection API. For example you can use Microsoft VOTT to generate such dataset. You can also use this script to create the pascal voc dataset.
python train.py --batch_size 8 --dataset ~/Data/voc2012.tfrecord --val_dataset ~/Data/voc2012_val.tfrecord --epochs 100 --mode eager_tf --transfer fine_tune python train.py --batch_size 8 --dataset ~/Data/voc2012.tfrecord --val_dataset ~/Data/voc2012_val.tfrecord --epochs 100 --mode fit --transfer none python train.py --batch_size 8 --dataset ~/Data/voc2012.tfrecord --val_dataset ~/Data/voc2012_val.tfrecord --epochs 100 --mode fit --transfer no_output python train.py --batch_size 8 --dataset ~/Data/voc2012.tfrecord --val_dataset ~/Data/voc2012_val.tfrecord --epochs 10 --mode eager_fit --transfer fine_tune --weights ./checkpoints/yolov3-tiny.tf --tiny
You can export the model to tf serving
python export_tfserving.py --output serving/yolov3/1/ # verify tfserving graph saved_model_cli show --dir serving/yolov3/1/ --tag_set serve --signature_def serving_default
The inputs are preprocessed images (see dataset.transform_iamges
)
outputs are
yolo_nms_0: bounding boxes yolo_nms_1: scores yolo_nms_2: classes yolo_nms_3: numbers of valid detections
Numbers are obtained with rough calculations from detect_video.py
Detection | 416x416 | 320x320 | 608x608 |
---|---|---|---|
YoloV3 | 1000ms | 500ms | 1546ms |
YoloV3-Tiny | 100ms | 58ms | 208ms |
Detection | 416x416 | 320x320 | 608x608 |
---|---|---|---|
YoloV3 | 74ms | 57ms | 129ms |
YoloV3-Tiny | 18ms | 15ms | 28ms |
Detection | 416x416 | 320x320 | 608x608 |
---|---|---|---|
YoloV3 | 66ms | 50ms | 123ms |
YoloV3-Tiny | 15ms | 10ms | 24ms |
Detection | 416x416 |
---|---|
YoloV3 predict_on_batch | 29-32ms |
YoloV3 predict_on_batch + TensorRT | 22-28ms |
Darknet version of YoloV3 at 416x416 takes 29ms on Titan X. Considering Titan X has about double the benchmark of Tesla M60, Performance-wise this implementation is pretty comparable.
Great addition for existing TensorFlow experts. Not very easy to use without some intermediate understanding of TensorFlow graphs. It is annoying when you accidentally use incompatible features like tensor.shape[0] or some sort of python control flow that works fine in eager mode, but totally breaks down when you try to compile the model to graph.
When calling model(x) directly, we are executing the graph in eager mode. Formodel.predict
, tf actually compiles the graph on the first run and then
execute in graph mode. So if you are only running the model once, model(x)
is
faster since there is no compilation needed. Otherwise, model.predict
or
using exported SavedModel graph is much faster (by 2x). For non real-time usage,model.predict_on_batch
is even faster as tested by @AnaRhisT94)
Extremely useful for debugging purpose, you can set breakpoints anywhere.
You can compile all the keras fitting functionalities with gradient tape using therun_eagerly
argument in model.compile. From my limited testing, all training methods
including GradientTape, keras.fit, eager or not yeilds similar performance. But graph
mode is still preferred since it's a tiny bit more efficient.
@tf.function is very cool. It's like an in-between version of eager and graph. You can step through the function by disabling tf.function and then gain performance when you enable it in production. Important note, you should not pass any non-tensor parameter to @tf.function, it will cause re-compilation on every call. I am not sure whats the best way other than using globals.
Absolutely amazing. If you don't know already, absl.py is officially used by internal projects at Google. It standardizes application interface for Python and many other languages. After using it within Google, I was so excited to hear abseil going open source. It includes many decades of best practices learned from creating large size scalable applications. I literally have nothing bad to say about it, strongly recommend absl.py to everybody.
very hard with pure functional API because the layer ordering is different in tf.keras and darknet. The clean solution here is creating sub-models in keras. Keras is not able to save nested model in h5 format properly, TF Checkpoint is recommended since its offically supported by TensorFlow.
It doesn't work very well for transfer learning. There are many articles and github issues all over the internet. I used a simple hack to make it work nicer on transfer learning with small batches.
I know it's very confusion but the output is tuple of shape
( [N, 13, 13, 3, 6], [N, 26, 26, 3, 6], [N, 52, 52, 3, 6] )
where N is the number of labels in batch and the last dimension "6" represents[x, y, w, h, obj, class]
of the bounding boxes.
the default threshold is 0.5 for both IOU and score, you can adjust them
according to your need by setting --yolo_iou_threshold
and--yolo_score_threshold
flags
convert.py: --output: path to output (default: './checkpoints/yolov3.tf') --[no]tiny: yolov3 or yolov3-tiny (default: 'false') --weights: path to weights file (default: './data/yolov3.weights') --num_classes: number of classes in the model (default: '80') (an integer) detect.py: --classes: path to classes file (default: './data/coco.names') --image: path to input image (default: './data/girl.png') --output: path to output image (default: './output.jpg') --[no]tiny: yolov3 or yolov3-tiny (default: 'false') --weights: path to weights file (default: './checkpoints/yolov3.tf') --num_classes: number of classes in the model (default: '80') (an integer) detect_video.py: --classes: path to classes file (default: './data/coco.names') --video: path to input video (use 0 for cam) (default: './data/video.mp4') --output: path to output video (remember to set right codec for given format. e.g. XVID for .avi) (default: None) --output_format: codec used in VideoWriter when saving video to file (default: 'XVID) --[no]tiny: yolov3 or yolov3-tiny (default: 'false') --weights: path to weights file (default: './checkpoints/yolov3.tf') --num_classes: number of classes in the model (default: '80') (an integer)train.py: --batch_size: batch size (default: '8') (an integer) --classes: path to classes file (default: './data/coco.names') --dataset: path to dataset (default: '') --epochs: number of epochs (default: '2') (an integer) --learning_rate: learning rate (default: '0.001') (a number) --mode: <fit|eager_fit|eager_tf>: fit: model.fit, eager_fit: model.fit(run_eagerly=True), eager_tf: custom GradientTape (default: 'fit') --num_classes: number of classes in the model (default: '80') (an integer) --size: image size (default: '416') (an integer) --[no]tiny: yolov3 or yolov3-tiny (default: 'false') --transfer: <none|darknet|no_output|frozen|fine_tune>: none: Training from scratch, darknet: Transfer darknet, no_output: Transfer all but output, frozen: Transfer and freeze all, fine_tune: Transfer all and freeze darknet only (default: 'none') --val_dataset: path to validation dataset (default: '') --weights: path to weights file (default: './checkpoints/yolov3.tf')
Updated to Tensorflow to v2.0.0 Release
It is pretty much impossible to implement this from the yolov3 paper alone. I had to reference the official (very hard to understand) and many un-official (many minor errors) repos to piece together the complete picture.
official yolov3 implementation
explinations of parameters
models
loss functions
data transformations
loss functions
models
batch normalization fix
上一篇:keras-yolo2
下一篇:yololib
还没有评论,说两句吧!
热门资源
Keras-ResNeXt
Keras ResNeXt Implementation of ResNeXt models...
seetafaceJNI
项目介绍 基于中科院seetaface2进行封装的JAVA...
spark-corenlp
This package wraps Stanford CoreNLP annotators ...
capsnet-with-caps...
CapsNet with capsule-wise convolution Project ...
inferno-boilerplate
This is a very basic boilerplate example for pe...
智能在线
400-630-6780
聆听.建议反馈
E-mail: support@tusaishared.com