diff --git a/docs/source/index.rst b/docs/source/index.rst index bf6eb21..239d981 100644 --- a/docs/source/index.rst +++ b/docs/source/index.rst @@ -10,9 +10,9 @@ TensorFlow Object Detection API tutorial Tensorflow 1.15 has also been released, but seems to be exhibiting `instability issues `_. - A version for TensorFlow 1.9 can be found `here `_. + A version for Tensorflow 1.9 can be found `here `_. - At the time of righting this tutorial, Object Detection model training and evaluation was not migrated to TensorFlow 2.x (see `here `_). From personal tests, it seems that detection using pre-trained models works, however it is not yet possible to train and evaluate models. Once the migration has been completed, a version for TensorFlow 2.x will be produced. + A version for Tensorflow 2.x is in the making and a link will be added here when ready. This is a step-by-step tutorial/guide to setting up and using TensorFlow's Object Detection API to perform, namely, object detection in images/video. diff --git a/docs/source/training.rst b/docs/source/training.rst index 3167297..89ee5b0 100644 --- a/docs/source/training.rst +++ b/docs/source/training.rst @@ -544,11 +544,11 @@ Configuring a Training Pipeline For the purposes of this tutorial we will not be creating a training job from the scratch, but rather we will go through how to reuse one of the pre-trained models provided by TensorFlow. If you would like to train an entirely new model, you can have a look at `TensorFlow's tutorial `_. -The model we shall be using in our examples is the ``ssd_inception_v2_coco`` model, since it provides a relatively good trade-off between performance and speed, however there are a number of other models you can use, all of which are listed in `TensorFlow's detection model zoo `_. More information about the detection performance, as well as reference times of execution, for each of the available pre-trained models can be found `here `_. +The model we shall be using in our examples is the ``ssd_inception_v2_coco`` model, since it provides a relatively good trade-off between performance and speed, however there are a number of other models you can use, all of which are listed in `TensorFlow's detection model zoo `_. More information about the detection performance, as well as reference times of execution, for each of the available pre-trained models can be found `here `_. First of all, we need to get ourselves the sample pipeline configuration file for the specific model we wish to re-train. You can find the specific file for the model of your choice `here `_. In our case, since we shall be using the ``ssd_inception_v2_coco`` model, we shall be downloading the corresponding `ssd_inception_v2_coco.config `_ file. -Apart from the configuration file, we also need to download the latest pre-trained NN for the model we wish to use. This can be done by simply clicking on the name of the desired model in the tables found in `TensorFlow's detection model zoo `_. Clicking on the name of your model should initiate a download for a ``*.tar.gz`` file. +Apart from the configuration file, we also need to download the latest pre-trained NN for the model we wish to use. This can be done by simply clicking on the name of the desired model in the tables found in `TensorFlow's detection model zoo `_. Clicking on the name of your model should initiate a download for a ``*.tar.gz`` file. Once the ``*.tar.gz`` file has been downloaded, open it using a decompression program of your choice (e.g. 7zip, WinZIP, etc.). Next, open the folder that you see when the compressed folder is opened (typically it will have the same name as the compressed folded, without the ``*.tar.gz`` extension), and extract it's contents inside the folder ``training_demo\pre-trained-model``.