Standard DSMLP Jupyter Notebook Server Environments
Educational Technology Services offers two standard notebook servers for the Data Science and Machine Learning Platform: data science and statistical analysis, or machine learning/GPU. Instructors, please let us know in your Instructional Technology Request (CINFO) which you require, and any additional needs. For more detailed information, see An overview of Standard Datahub/DSMLP Containers maintained by UCSD Educational Technology Services.
Data Science Notebook: Popular Python and R data science packages
https://github.com/ucsd-ets/datascience-notebook
SciPy / Machine Learning Notebook: TensorFlow, Pytorch, CUDA
https://github.com/ucsd-ets/scipy-ml-notebook
Standard DSMLP User Container Configuration
- 2 CPU cores
- 4 GB RAM
- 1 GPU (for GPU-enabled courses)
- 100 GB student disk quota
To request increases to the above, respond to your initial DSMLP course ticket, or send an email to dsmlp@ucsd.edu.
Standard DSMLP Course Features
- Assignment distribution/collection/grading: Datahub/DSMLP Assignment Grading Tools
- Test student account for instructor use
- Ability to add TAs to course via Canvas
- Group functionality via Canvas
- Access to GPUs and machine learning packages (e.g., TensorFlow)
- Central hosting and access to instructor-supplied datasets
- Datasets up to 500GB are easily accommodated; larger corpora will be accepted on a space-available basis.
- Datasets can be published system-wide, or restricted to a specific course. Please contact us in advance regarding use of confidential or restricted data.
- Shared course directory by request