SciSharp / TensorFlow.NET

.NET Standard bindings for Google's TensorFlow for developing, training and deploying Machine Learning models in C# and F#.
https://scisharp.github.io/tensorflow-net-docs
Apache License 2.0
3.25k stars 524 forks source link

Wish for implementation using and storing a -.tflite model #452

Open ghost opened 4 years ago

ghost commented 4 years ago

Tensor Flow has some popular already trained models, that would be nice to use in a dot.net.core. For example the PoseNet Model (https://github.com/tensorflow/tfjs-models/tree/master/posenet). If I want to use the model now I could do the following:

  1. With NPM jou can make use of the models using Type Script
  2. With an internet connection, you could add a js script
  3. Load and run a -.tflite model in C++(https://www.tensorflow.org/lite/guide/inference#load_and_run_a_model_in_c)

Working in c# and .Net.core and Blazor client site works the second way more ore less. But you have to write some .js and you have to rely on an internet connection. It would be really great to just add the model to your source and load a -.tflite model directly into your library. Making the application true Dot.net.

Besides that, I guess ex and importing your models into a .tflite file format would be a great way to share models between software developers that are just working on an implementation off your math they cant alter/break your models etc.

Martijn Carol

Nucs commented 4 years ago

Related issue: https://github.com/SciSharp/TensorFlow.NET/issues/376