The Microsoft Cognitive Toolkit - Cognitive Toolkit - CNTK
Dive into the world of deep learning with The Microsoft Cognitive Toolkit - CNTK! 🧠💻 Train neural networks with ease using parallelization across multiple GPUs, support for popular model types, and easy framework interoperability with ONNX. #AI #deeplearning #CNTK
- The Microsoft Cognitive Toolkit (CNTK) is an open-source toolkit for commercial-grade distributed deep learning.
- CNTK describes neural networks as a series of computational steps via a directed graph.
- It allows easy realization and combination of popular model types like feed-forward DNNs, CNNs, RNNs/LSTMs.
- CNTK implements SGD learning with automatic differentiation and parallelization across multiple GPUs and servers.
- The latest release of CNTK is 2.7, supporting 64-bit Linux or Windows operating systems.
- CNTK can be included in Python, C#, and C++ programs or used as a standalone machine-learning tool through BrainScript.
- It supports ONNX format for framework interoperability and shared optimization.
- ONNX allows moving models between frameworks like CNTK, Caffe2, MXNet, PyTorch.
- The 1-bit SGD in CNTK no longer requires a separate license and is available under the GitHub license.
- Installation options include pre-compiled binary packages or compiling from the GitHub source provided.