2. Installation¶
2.1. Install UCTB¶
2.1.1. Step 1: Install TensorFlow¶
You can skip to step 2 if you already installed tensorflow.
You can refer to this page https://www.tensorflow.org/install to install tensorflow, if you have a Nvidia GPU installed on you computer, we highly recommend you to install GPU version of tensorflow.
2.1.2. Step 2: Install UCTB¶
pip install --upgrade UCTB
The following required package will be installed or upgraded with UCTB:
'hmmlearn',
'keras',
'GPUtil',
'numpy',
'pandas',
'python-dateutil',
'scikit-learn',
'scipy',
'statsmodels',
'wget',
'xgboost',
'nni',
'chinesecalendar',
'PyYAML'
2.2. UCTB Docker¶
You can also use UCTB by docker. First pull uctb docker from docker hub.
docker pull dichai/uctb:v0.2.0
And you then can run it.
docker run --runtime=nvidia -it -d dichai/uctb:v0.2.0 /bin/bash