Skip to content

[RELEASE] version == 0.5.0#479

Merged
vadiklyutiy merged 3 commits intomainfrom
vadim/ver-promotion
Dec 20, 2024
Merged

[RELEASE] version == 0.5.0#479
vadiklyutiy merged 3 commits intomainfrom
vadim/ver-promotion

Conversation

@vadiklyutiy
Copy link
Copy Markdown
Collaborator

version == 0.5.0

@vadiklyutiy vadiklyutiy merged commit a0fff01 into main Dec 20, 2024
@vadiklyutiy vadiklyutiy deleted the vadim/ver-promotion branch December 20, 2024 17:00
vadiklyutiy added a commit that referenced this pull request Dec 26, 2024
This PR introduce `bench_task.py` scripts. It is similar to another
scripts in `tests/benchmarks/bench_*.py`.

As input it uses `task.pickle` file (dump of task) that is always saved
in our cache.

Using 
`python tests/benchmarks/bench_task.py --task <path to .pickle file>`

I found that it is a very convenient way for both debugging and
performance analysis. Compared to the full model run it is much faster.
Compared to `bench_op.py` they shouldn't do extra work for finding task
shape and modifying a run to use this shape.
AndreSlavescu pushed a commit to AndreSlavescu/hidet that referenced this pull request May 31, 2025
This PR introduce `bench_task.py` scripts. It is similar to another
scripts in `tests/benchmarks/bench_*.py`.

As input it uses `task.pickle` file (dump of task) that is always saved
in our cache.

Using 
`python tests/benchmarks/bench_task.py --task <path to .pickle file>`

I found that it is a very convenient way for both debugging and
performance analysis. Compared to the full model run it is much faster.
Compared to `bench_op.py` they shouldn't do extra work for finding task
shape and modifying a run to use this shape.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant