TensorFlow.js
E97075
TensorFlow.js is a JavaScript library that enables training and running machine learning models directly in the browser and in Node.js using TensorFlow.
Statements (52)
| Predicate | Object |
|---|---|
| instanceOf |
JavaScript library
→
machine learning framework → open-source software → |
| basedOn | TensorFlow → |
| developer |
Google
→
Google Brain →
surface form:
Google Brain team
|
| feature |
GPU-accelerated computation via WebGL
→
automatic differentiation → define models using high-level APIs → define models using low-level tensor operations → import pre-trained TensorFlow models → model saving and loading → run machine learning models in Node.js → run machine learning models in the browser → support for Core API → support for Layers API → support for pre-trained models → train machine learning models in Node.js → train machine learning models in the browser → |
| license | Apache License 2.0 NERFINISHED → |
| partOf |
TensorFlow
→
surface form:
TensorFlow ecosystem
|
| programmingLanguage |
JavaScript
→
TypeScript programming language →
surface form:
TypeScript
|
| repository | https://github.com/tensorflow/tfjs → |
| supportsDataType |
TypedArray
→
tensor → |
| supportsEnvironment |
Node.js
→
client-side JavaScript → server-side JavaScript → web browser → |
| supportsExecutionBackend |
CPU
→
GPU → Node.js C++ bindings → WebAssembly specification →
surface form:
WebAssembly
WebGL → |
| supportsModelFormat |
Keras HDF5 (via conversion)
→
TensorFlow SavedModel (via conversion) → TensorFlow.js JSON model format → |
| supportsPlatform |
Node.js applications
→
desktop browsers → mobile browsers → |
| supportsTask |
audio recognition
→
image classification → natural language processing → object detection → pose estimation → text classification → transfer learning → |
| useCase |
in-browser machine learning
→
interactive ML-powered web applications → privacy-preserving on-device inference → |
| website | https://www.tensorflow.org/js → |
Referenced by (1)
Full triples — surface form annotated when it differs from this entity's canonical label.