Create Notebook Instance

BML provides an interactive code editing and running environment Jupyter Lab with built-in algorithmic frameworks such as TensorFlow, Keras, PyTorch, Caffe, Mxnet, Chainer, CNTK and PaddlePaddle.

Log in to BML Management Console, navigate to select "Model training-> Notebook", click "Create instance", fill in the name of the instance, select the calculation resource package (please see Product Pricing for billing method and package price), select the size of the startup disk, specify the BOS path(optional) to complete the creation (the creation process may take about 1 minute).

Warm Tip: The instance status will continue to incur expenses during "running". Please stop the instance in time when not in use. The instance will be stopped after arrears. If there is a program running, please ensure that the account balance is sufficient. After stopping the instance, all your changes (code, data, installed third-party software package, etc.) will be saved. Click "Run" next time to retrieve it, and then click Jupyter again to continue the previous operation.

Enable Jupyter

Click "Open jupyter" to enter the Jupyter development interface. The platform in the demo folder in the left column provides template code, which is used to introduce how to download data, conduct training, etc., for your learning and reference. Please refer to Appendix for Jupyter operation instructions.