catchzeng / object-detection-api

Make it easy to train and deploy Object Detection(SSD) and Image Segmentation(Mask R-CNN) Model Using TensorFlow Object Detection API.

object-detection-api

Make it easy to train and deploy Object Detection(SSD) and Image Segmentation(Mask R-CNN) Model Using TensorFlow Object Detection API.

Requirements

• conda
• protoc

Use the following command to check if the installation is successful.

$ conda --version
conda 4.9.2
$ protoc --version
libprotoc 3.17.1

Installation

Conda

$ conda create -n  od python=3.8.5 && conda activate od && make install
......
----------------------------------------------------------------------
Ran 24 tests in 21.869s

OK (skipped=1)

Install directly (such as in colab)

$ make install
......
----------------------------------------------------------------------
Ran 24 tests in 21.869s

OK (skipped=1)

For details, please refer to colab demo.

Note: If the installation fails, you can refer to the detailed steps in the official document.

Note: The update of cudaDNN and toolkit may not be as fast as TensorFlow. Therefore, if your machine has a GPU, after the installation is complete, you need to downgrade TensorFlow to the version supported by cudaDNN and toolkit in order to support GPU training. Take 2.8.0 as an example:

$ pip install --upgrade tf-models-official==2.8.0
$ pip install --upgrade tensorflow==2.8.0

Usage

Train

Object Detection

The easiest way to Train a Custom Object Detection Model Using TensorFlow Object Detection API

Image Segmentation

The easiest way to Train a Custom Image Segmentation Model Using TensorFlow Object Detection API Mask R-CNN

Deploy

Deploy image segmentation (Mask R-CNN) model service with TensorFlow Serving & Flask

object-detection-api's Issues

training error

File "/usr/local/lib/python3.6/dist-packages/tensorflow/python/eager/execute.py", line 60, in quick_execute
    inputs, attrs, num_outputs)
tensorflow.python.framework.errors_impl.InvalidArgumentError: 2 root error(s) found.
  (0) Invalid argument:  indices[0] = 0 is not in [0, 0)
	 [[node GatherV2_8 (defined at /local/lib/python3.6/dist-packages/object_detection/utils/shape_utils.py:246) ]]
	 [[MultiDeviceIteratorGetNextFromShard]]
	 [[RemoteCall]]
	 [[while/body/_1/IteratorGetNext]]
	 [[Identity_11/_5982]]
  (1) Invalid argument:  indices[0] = 0 is not in [0, 0)
	 [[node GatherV2_8 (defined at /local/lib/python3.6/dist-packages/object_detection/utils/shape_utils.py:246) ]]
	 [[MultiDeviceIteratorGetNextFromShard]]
	 [[RemoteCall]]
	 [[while/body/_1/IteratorGetNext]]

got this error during training, training didn't start

Unable to train with colab

2 root error(s) found.
(0) UNIMPLEMENTED: DNN library is not found.
[[{{node functional_1/conv1_conv/Conv2D}}]]
[[StatefulPartitionedCall/SecondStagePostprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Reshape_5/_126]]
(1) UNIMPLEMENTED: DNN library is not found.
[[{{node functional_1/conv1_conv/Conv2D}}]]
0 successful operations.

Pls upload the requirements.txt , Instead of installing from make file

using lighter backbone

is it possible to train with lighter models like resenet18, resnet50?