Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Add support for more flexible object detection model selection (quantized, faster-rcnn, etc...)

See original GitHub issue

I stumbled upon this project from a LinkedIn post, and it peeked my interest as I’ve implemented a TensorFlow component in Home Assistant, a popular open source home automation platform.

One of the requests I get frequently is the ability to train and use custom object labels. Up to this point, I’ve told people that the effort required to do so is too great for most users.

On to the feature request… I noticed here that you are defaulting the model to an SSD MobileNet model. How large is the effort to allow users to choose the base model to train? I, for example, am using the faster_rcnn_inception_v2_coco model since the SSD model is not accurate enough for the distance and angle of the cameras I have mounted.

Issue Analytics

State:
Created 4 years ago
Comments:7 (4 by maintainers)

Top GitHub Comments

1reaction

bourdakos1commented, Apr 15, 2019

Another thing to note, is that the core ml/tflite conversions won’t work anymore, they are only compatible with ssd at the moment

0reactions

husnejahancommented, May 25, 2019

hi, I am getting this error"(Using settings from config.yaml) error Invalid Cloud Object Storage credentials." how to change or start process again

Top Results From Across the Web

Benchmarking Object Detection Deep Learning Models in ...

This work benchmarks deep learning object detection models in embedded devices. Furthermore, some hardware selection guidelines are included, together with ...

Object Detection with Deep Learning: A Review - arXiv

Our review begins with a brief introduction on the history of deep learning and its representative tool, namely Convolutional Neural Network. (CNN). Then...

Unable to train a quantization aware Faster RCNN + ... - GitHub

When I try to fine tune a Faster RCNN + Resnet50 object detector I get a runtime error. Stack trace: Traceback (most recent...

Deep Learning for Generic Object Detection: A Survey

4a) is to assess the presence of objects from a given set of object classes in an image; i.e., assigning one or more...

R-CNN, Fast R-CNN, Faster R-CNN, YOLO — Object ...

Another integral part of computer vision is object detection. Object detection aids in pose estimation, vehicle detection, surveillance etc.