-
Notifications
You must be signed in to change notification settings - Fork 131
fix: CUDA実行プロバイダーにおける畳み込みアルゴリズム検索をEXHAUSTIVE (0)からDEFAULT (2)に変更 #1164
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.
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
fix: CUDA実行プロバイダーにおける畳み込みアルゴリズム検索をEXHAUSTIVE (0)からDEFAULT (2)に変更 #1164
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Pull Request Overview
This PR changes the CUDA execution provider's convolution algorithm search from EXHAUSTIVE (0) to DEFAULT (2) to significantly reduce warmup time when using CUDA execution provider. This reverts a change that occurred during migration from nbigaouette/onnxruntime-rs to pykeio/ort.
- Changed
CUDAExecutionProviderconfiguration to useDefaultinstead ofExhaustivealgorithm search - Applied the change consistently across three different usage locations in the codebase
- Addresses performance regression introduced during library migration
Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.
|
ちょうど24時間くらい前にcargo-codspeed v4.0.0がリリースされたことによってCIが落ちた。 @Sanzentyo 対処としてはこちらで別PRを作ろうと思うので、何もしなくてOKです。 |
…_algorithm_search`を設定しない
Hiroshiba
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM!!
データがたくさんあるから自信を持って設定を選べますね!ありがたい!!
内容
CUDAExecutionProviderのcudnn_conv_algo_searchをデフォルトのEXHAUSTIVE (0)からDEFAULT(2)へ変更するDEFAULT (2)からEXHAUSTIVE (0)に変わっていたものを元に戻すを参照のこと
関連 Issue
Fixes: #1163
Closes: #1162