Adding demo for running softmax kernel on Google colab #944
+248
−4
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
meta-cla
bot
added
the
CLA Signed
choijon5
force-pushed
the
softmax_notebook
branch
from
9aab2bf to
3b6a570
Compare
oulgen
approved these changes
Oct 15, 2025
| if [[ "${{ matrix.alias }}" == *"a10g"* ]]; then export HELION_DEV_LOW_VRAM=1; fi | ||
| # -rf: print failed tests | ||
| # --timeout: max allowed time for each test | ||
| pytest -rf --timeout=60 |
There was a problem hiding this comment.
you need to limit this to test folder now
There was a problem hiding this comment.
do you mean I should move this to the test folder?
There was a problem hiding this comment.
pytest -rf --timeout=60 test
notebooks/softmax.ipynb
Outdated
| "metadata": {}, | ||
| "outputs": [], | ||
| "source": [ | ||
| "!pip3 install torch==2.9.0 --index-url https://download.pytorch.org/whl/test/cu126\n", |
There was a problem hiding this comment.
it actually uses 12.4 but PyTorch no longer does 12.4 releases so 12.6 will do :)
User can easily try out Helion with this notebook from our documentation website and GitHub readme. With autotune_effort=quick, it runs quickly and shows good speedup (on T4 - 1.8X speedup).
…ebook with pytest.
…o run notebook on CI so that it can run the notebook cells to setup PyTorch and Helion.
choijon5
force-pushed
the
softmax_notebook
branch
from
3b6a570 to
961c12b
Compare
choijon5
force-pushed
the
softmax_notebook
branch
from
fe44786 to
ea3e3ae
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
To allow users to easily try out Helion, I'm adding a Python notebook to run the softmax kernels. This notebook is linked from our documentation and our GitHub README to run it on Google Colab, which has NVIDIA T4 GPU for free. We achieve good speedup (1.8-1.9X) for this example on T4 and with autotune_effort="quick", it can be run in 5~6 mins on the free version of Colab, instead of 16-17 mins with autotune_effort="full".
Python notebook can be tested with pytest with nbmake. Added this notebook to be run on CI.