https://github.com/oracle-quickstart/oci-gpu-jupyter
https://github.com/oracle-quickstart/oci-gpu-jupyter
oracle-led
Last synced: 10 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/oracle-quickstart/oci-gpu-jupyter
- Owner: oracle-quickstart
- License: upl-1.0
- Created: 2020-09-27T01:37:52.000Z (about 5 years ago)
- Default Branch: master
- Last Pushed: 2021-06-28T18:47:50.000Z (over 4 years ago)
- Last Synced: 2024-12-29T20:16:01.049Z (12 months ago)
- Topics: oracle-led
- Language: HCL
- Homepage:
- Size: 43 KB
- Stars: 1
- Watchers: 1
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- Contributing: CONTRIBUTING.md
- License: LICENSE
- Security: Security.md
Awesome Lists containing this project
README
## Use Case
Create a sandboxed machine learning environment that has Tensorflow, SciKit
and other libraries preconfigured.
## Create ML environment based on Image in Marketplace
The VM is created based on the .
All the steps described in the marketplace entry are automated. In addition, a functional network is created along with the option of multiple disks that are backed up via policy.
### Using this example
* Copy terraform.template.tfvars to terraform.tfvars
* Update terraform.tfvars with credential information for your tenancy
* Modify variables.tf as needed
* Provision the infrastructure:
```
$ terraform init
$ terraform plan
$ terraform apply
```
### Connect to Jupyter Notebook
* This automation uses ssh redirection for security. This provides for secure access to the notebooks. This approach is not recommended for a team environment. A sample ssh redirect command is below. The output from terraform apply will display a custom command for your environment.
```
ssh -i -L 8081:localhost:8080 opc@
```
* Use your local browser to connect to the Jupyter
```
http://localhost:8081
http://localhost:8081/notebooks/Examples/deeplearning_bootcamp-master
```
* Enter the password as configured in variables.tf
Known Issue:
If name is changed (app_description, app_tag or app_environment),
the resources are re-created properly, but jupyter notebooks are not
provisioned.