-
Notifications
You must be signed in to change notification settings - Fork 4.2k
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
feat(src/layer): add vision_transformer benchmark #3730
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## master #3730 +/- ##
==========================================
+ Coverage 93.33% 93.35% +0.01%
==========================================
Files 672 674 +2
Lines 171577 171980 +403
==========================================
+ Hits 160146 160556 +410
+ Misses 11431 11424 -7 ☔ View full report in Codecov by Sentry. |
…o/ncnn into add-vision-transformer
This pull request introduces 1 alert when merging eb7d4b2 into f14388a - view on LGTM.com new alerts:
|
@nihui test_squeezenet 的错,是要我重造模型? 我也没改涉及 sqz 的东西。 |
@nihui 我没有要改的了。 |
@nihui 好人给 review 一下呗。 |
这套东西,其实就是一个 nn.Linear ,被onnx拆散了 😆 |
@nihui 改好啦,扔掉了 gather; ViT.param 也小了很多。 |
Thanks for your contribution ! |
支持 vision_transformer。
onnx2ncnn.cpp
增加了转换 GELU 和 Crop速度测试 on 8 核 i5 CPU,330MB 大模型。
vision_transformer min = 3236.20 max = 3541.62 avg = 3403.61