Skip to content

Latest commit

 

History

History
87 lines (70 loc) · 2.65 KB

README.md

File metadata and controls

87 lines (70 loc) · 2.65 KB

uai-sdk

What is UCloud AI SDK

  1. This SDK include pack tool to pack local AI inference program into UCloud web service image which could run on UCloud UAI Inference Platform.
  2. This SDK include whole APIs which could interactive with UCloud UAI Inference Platform.
  3. This SDK include pack tool to help pack local AI trainning program into UCloud Docker image which could run on UCloud UAI Train Platform.
  4. This SDK include whole APIs which could interactive with UCloud UAI Train Platforms

Supported AI Platform

UAI Inference

CHECK ONUAI Inference Base Docker Images

1. GPU

  • Tensorflow (1.6.0 tested)
  • Tensorflow (1.7.0 tested)
  • Caffe (1.0.0 tested)

2. CPU

  • Tensorflow (1.1.0 tested)
  • Tensorflow (1.2.0 tested)
  • Tensorflow (1.3.0 tested)
  • Tensorflow (1.4.0 tested)
  • Tensorflow (1.6.0 tested)
  • Tensorflow (1.7.0 tested)
  • Tensorflow (1.8.0, cuda9 tested)
  • Tensorflow (1.9.0, cuda9 tested)
  • Tensorflow (1.10.0, cuda9 tested)
  • Tensorflow (1.11.0, cuda9 tested)
  • Tensorflow (1.12.0, cuda9 tested)
  • Tensorflow (2.0.0a, cuda10 tested)
  • MXNet (0.9.5 tested)
  • MXNet (1.0.0 tested)
  • Keras (1.2.0 tested)
  • Keras (2.0.8 tested)
  • Caffe (1.0.0 tested)
  • Intel-Caffe
  • XGBoost (0.7 tested, python interface only)

UAI Train

CHECK ONUAI Train Base Docker Images

  • Tensorflow (0.11.0 tested)
  • Tensorflow(1.1.0 tested)
  • Tensorflow(1.2.0 tested)
  • Tensorflow(1.3.0 tested)
  • Tensorflow(1.4.0 tested)
  • Tensorflow(1.5.0, cuda9 tested)
  • Tensorflow(1.6.0, cuda9 tested)
  • Tensorflow (1.7.0, cuda9 tested)
  • Tensorflow (1.8.0, cuda9 tested)
  • Tensorflow (1.9.0, cuda9 tested)
  • Tensorflow (1.10.0, cuda9 tested)
  • Tensorflow (1.11.0, cuda9 tested)
  • Tensorflow (1.12.0, cuda9 tested)
  • Tensorflow (2.0.0a, cuda10 tested)
  • MXNet (0.9.5 tested)
  • MXNet (1.0.0, cuda9 tested)
  • Keras (2.0.8 tested)
  • Keras (2.1.6, tf-1.8 testes)
  • Caffe (1.0.0 tested)
  • Caffe2 (Detection)
  • PyTorch (0.2.0)
  • PyTorch (0.4)

UAI Train Supporting Distributed Training

  • Tensorflow Distributed Training (examples include slim/cifar/wide-deep)
  • MXNet Distributed Training

How to install

  1. Install your deep learning python package, such as Tensorflow, MXNet, Keras, Caffe (tested version preferred)

  2. Install uai-sdk

     $ sudo python setup.py install
    
  3. Install all other lib your inference code required

How to interactive with UCloud UAI Platform

UAI Inference Docs

https://docs.ucloud.cn/ai/uai-inference/index

UAI Train Docs

https://docs.ucloud.cn/ai/uai-train/use