Skip to content

Add accelerator doc #913

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
stevemessick merged 1 commit into from
Jan 29, 2026
Merged

Add accelerator doc #913

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 17 additions & 0 deletions docs/kernels.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -134,6 +134,7 @@ kaggle kernels push -p <FOLDER_PATH> [options]

**Options:**

* `--accelerator <ACCELERATOR_ID>`: ID name of the accelerator to use during the run. E.g. "NvidiaTeslaP100" (aka default GPU), "NvidiaTeslaT4", "TpuV6E8".
* `-p, --path <FOLDER_PATH>`: Path to the folder containing the kernel file (e.g., `.ipynb`, `.Rmd`, `.py`) and the `kernel-metadata.json` file (defaults to the current directory).
* `-t, --timeout <SECONDS>`: Maximum run time in seconds.

Expand All @@ -149,6 +150,22 @@ kaggle kernels push -p tests/kernel

This command uploads your local kernel file and its metadata to Kaggle. If the kernel specified in the metadata exists under your account, it will be updated. Otherwise, a new kernel will be created. After uploading, Kaggle will attempt to run the kernel.

Accelerators available as of Feb 2026:

* NvidiaTeslaP100
* TpuV38
* NvidiaTeslaT4
* NvidiaTeslaT4Highmem
* Tpu1VmV38
* TpuvmV38
Copy link
Contributor

@rosbo rosbo Jan 29, 2026 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Some of these might not be supported anymore. For instance, we only support the 1VM support. Can you try calling kaggle kernels push with different accelerators and see which one succeed and which one fail?

Copy link
Contributor Author

@stevemessick stevemessick Jan 29, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It might be good to have a script to do that, at least until b/479840680 is resolved.

Copy link
Contributor Author

@stevemessick stevemessick Jan 29, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

#914

Copy link

@djherbis djherbis Jan 29, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

A number of these are only available in specific contexts too (ex. specific competitions, specific user groups etc.), some may be unreleased or never be released (testing, admin-only)

Copy link
Contributor Author

@stevemessick stevemessick Jan 29, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Sounds like kaggle k push ... won't solve the problem then. Should I even include a list?

* NvidiaTeslaA100
* NvidiaL4
* TpuV5E8
* NvidiaL4X1
* TpuV6E8
* NvidiaH100
* NvidiaRtxPro6000

## `kaggle kernels pull`

Pulls down the code/notebook and metadata for a kernel.
Expand Down
Loading