Skip to main content

Welcome

Efficiency and precision are at your fingertips with our meticulously crafted software stack, meticulously curated by NVIDIA and orchestrated by CGC.

Explore our extensive catalog of applications designed to streamline your workflow, empowering you to:

  • Data Preparation: Seamlessly prepare your datasets for AI tasks.
  • AI Model Development and Training: Facilitate the creation and training of cutting-edge AI models.
  • Model Deployment into Production: Simplify the process of deploying your models into production environments.

For a comprehensive solution, delve into our bundled end-to-end applications. Within this suite, you'll discover:

  • RAG: Explore and enhance retrieval-augmented generation capabilities for Llama 13B model.

Should you find any gaps in our app offerings that could enhance your workflow or benefit your team, please don't hesitate to contact us. We're committed to continually improving our suite of applications to meet your evolving needs.

  • Custom images - Execute any application using your own custom image. Private images can be retrieved from any repository that you have access to.
  • Jobs - providing custom Docker images along with a startup command, for application to run once. We believe this feature allows for greater flexibility and customization for our customers.

Flexible resource customization​

Among multiple parameters that are passed during resource/job/db creation, user can specify special CGC key-value relationships, that overwrites specific mechanics with user input.

  • image_pull_secret_name: by default it does the same thing as --repository-secret TEXT. That is lower level key, that is used for that CGC argument under the hood.
  • app_token: by default it is random system generated APP_TOKEN, that is used for all our templates for authorization. Using that key is preferred, when you want to overwrite template password (unless you do not want to display correct password within cgc compute list -d)
  • custom_command: by default in the CGC CLI, it is piped value. That key overwrites command for the container. Usage of custom_command always merges arguments into the command - as both overwriting command and arguments at the same time is not possible.
  • custom_image: by default it does the same thing as --image TEXT. That is lower level key, that is used for that CGC argument under the hood.