GetInputTensors is part of the API

Feat/get input tensors (#172)

* feat: flag for console output

* feat: GetInputTensors

* fix: test and bug fix

* feat: bump version

* feat: more tests

Update dependencies versions

Bump version (#171)

* feat: test are run in //

* feat: bump module versions

* WARN: test should be able to run in //

go module release

This is a pre-release according to issue #168

Tiny YOLO v2 is running

This is a pre-release to distribute a small cli able to run the Tiny YOLO v2 model for object detection.

To run this utility, you need to download the ONNX model from the zoo.

To run this utility:

✗  ./tiny_yolov2-darwin -h
Usage of ./tiny_yolov2-darwin:
  -h    help
  -img string
        path of an input jpeg image (use - for stdin)
  -model string
        path to the model file (default "model.onnx")
  -output string
        path of an output png file (use - for stdout)
  -s    silent mode (useful if output is -)
This application is configured via the environment. The following environment
variables can be used:

KEY                          TYPE     DEFAULT    REQUIRED    DESCRIPTION
YOLO_CONFIDENCE_THRESHOLD    Float    0.30       true
YOLO_PROBA_THRESHOLD         Float    0.90       true

This utility is provided for fun without further support.

Enjoy!

Emotion fer+ and draft of BCHW utilities

About

This is a pre-release able to run the emotion fer+ model.
On top of that, it holds the premises of a new package to manipulate images and BCHW tensors (see internal/x).

Binary attachment

The attachment is a simple tool that expects the emotion fer + model as input (model.onnx) and a picture encoded in gray/png and of size 64x64. (See the Makefile for options of ImageMagick to convert the picture to a suitable format).

You need to grab the model.onnx file from here

MacOS with brew

On MacOS you can install imagesnap and imagemagick and run macos_fun.sh.

Otherwise, simply download the Makefile and type make install to install everything and make analyze1.jpg to create a picture and analyze it.

Ability to run mnist against the master of Gorgonia

feat: mnist example (onnx version 1.3) can successfully run with Gorg…

…onia

* chore: ignore binary

* feat: create the skeleton of a utility to run a model from the zoo

* chore: prepare the implementation of the auto-padding

* feat: the utility is working

* feat: add the same_upper auto-paddding

* chore: new informations

* chore: add some doc

This is a release I made for my blog

please cf https://blog.owulveryck.info/2019/04/03/from-a-project-to-a-product-the-state-of-onnx-go.html#one-more-thing

v0.1-mnist-cli

To use it:
curl https://www.cntk.ai/OnnxModels/mnist/opset_7/mnist.tar.gz | tar -C /tmp -xzvf -
./mnist-reader.darwin -model /tmp/mnist/model.onnx

then point your browser to http://localhost:8100

To build it:

$ git checkout v0.1-mnist-cli
$ cd example/gorgonia/mnist-reader
$ curl https://www.cntk.ai/OnnxModels/mnist/opset_7/mnist.tar.gz | tar -C /tmp -xzf -
$ go run main.go -model /tmp/mnist/model.onnx