benjamintanweihao / yolov3 Goto Github PK

View Code? Open in Web Editor NEW

65.0 3.0 11.0 113.52 MB

YOLOv3 Implementation in TensorFlow 1.1X

Python 100.00%

yolo3 tensorflow keras deep-learning

yolov3's Introduction

🦄 YOLOv3 Implementation in TensorFlow 1.1x + Keras 🦄

How it Looks Like

Watch the demo

Quick Start

On a PC / Mac

Create conda environment depending on whether you have a supported GPU or not:

conda env create -f environment-[c|g]pu.yml
source activate yolov3-[c|g]pu

On a Raspi 3

Install OpenCV 3 with the following instructions.

Then:

pip install tensorflow scikit-learn

Download YOLO Weights

Download weights into the cfg directory:

cd cfg
wget https://pjreddie.com/media/files/yolov3.weights

Demo on Single Image:

python single_image.py

The output is stored on out.png in the root folder.

Demo on Web Cam:

To see it live on your Web Cam:

python webcam.py

Progress

Credits

yolov3's People

Contributors

Stargazers

Watchers

Forkers

collector-m t-zhangshuo michaelzenz wuxiangchao juanlp pclin antonizhubar yyuzhong karthik9001 rushins guo02191

yolov3's Issues

How to train the model?

Shape mismatch error in route layer on specifying input shape in single_image.py

I specified the input shape in single_image.py :
inputs = Input(shape=(416, 416, 3))

This gave me following error:
File "single_image.py", line 11, in
outputs, config = darknet_base(inputs, include_yolo_head=False)
File "/vol1/dbstore/orc_srib/n.hri/Hand_Localization/Experiments/005_yolov3_tflite_pc/YOLOv3/darknet.py", line 67, in darknet_base
x, layers, yolo_layers, ptr = _build_route_layer(x, block, layers, yolo_layers, ptr)
File "/vol1/dbstore/orc_srib/n.hri/Hand_Localization/Experiments/005_yolov3_tflite_pc/YOLOv3/darknet.py", line 224, in _build_route_layer
x = Concatenate(axis=3)(selected_layers)
File "/vol0/tool/ml/cuda9/.venv/py2/tensorflow-1.10.0/lib/python2.7/site-packages/tensorflow/python/keras/engine/base_layer.py", line 728, in call
self.build(input_shapes)
File "/vol0/tool/ml/cuda9/.venv/py2/tensorflow-1.10.0/lib/python2.7/site-packages/tensorflow/python/keras/utils/tf_utils.py", line 148, in wrapper
output_shape = fn(instance, input_shape)
File "/vol0/tool/ml/cuda9/.venv/py2/tensorflow-1.10.0/lib/python2.7/site-packages/tensorflow/python/keras/layers/merge.py", line 392, in build
'Got inputs shapes: %s' % (input_shape))
ValueError: A Concatenate layer requires inputs with matching shapes except for the concat axis. Got inputs shapes: [(None, None, None, 128), (None, 26, 26, 256)]

I can't think of a reason as to why it is able to get shape of one layer only in route layer and not the other.

Error while running tflite_demo.py

Model provided has model identifier 'TE: ', should be 'TFL3'.

Previous errors were :
No module named 'tensorflow.contrib.lite'
ImportError: cannot import name 'interpreter'
ValueError: Could not open '/content/gdrive/My Drive/BEMRR/YOLOv3/model/yolov3.tflite'.

FileNotFoundError: [Errno 2] yolov3-tiny.weights, May work with yolov3.weights

The link https://pjreddie.com/media/files/yolov3.weights provided a file called yolov3.weights not yolov3-tiny.weights . I feel either this setup should work with yolov3.weights or the read me part needs to be edited to have yolov3-tiny.weights downloaded.
https://pjreddie.com/media/files/yolov3-tiny.weights

FileNotFoundError: [Errno 2] No such file or directory: 'C:\Users\NAME\yoloproj\YOLOv3\cfg\yolov3-tiny.weights'

Is it possible to get class scores for all classes ?

Hi, Is it possible to get class scores for all classes ? For example we have a dog object. The model predicts that the object is dog, with score 0.9. Is it possible to get scores for other classes, for example: cat - 0.1 ?
I checked many examples but have not found how to get class scores for all classes ?
Thank you very much.