Skip to content
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

Jump to bottom

add searchsorted op #7949

Merged
merged 69 commits into from
Apr 29, 2022
Merged

add searchsorted op #7949

merged 69 commits into from
Apr 29, 2022

Conversation

yoonlee888
Copy link
Contributor

@yoonlee888 yoonlee888 commented Apr 1, 2022 •
edited
Loading

背景：NERF网络需要用到这个算子
算子描述：
参考pytorch的实现https://pytorch.org/docs/stable/generated/torch.searchsorted.html?highlight=searchsorted#torch.searchsorted
接口与pytorch 1.10 版本实现完全对齐。

@hengzi hengzi self-requested a review April 1, 2022 07:43
hengzi
hengzi reviewed Apr 1, 2022
View reviewed changes
Copy link
Contributor

@hengzi hengzi left a comment

Choose a reason for hiding this comment

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

python/oneflow/framework/docstr/ 添加api文档

@hengzi hengzi requested a review from oneflow-ci-bot April 2, 2022 08:03
@github-actions
Copy link
Contributor

github-actions bot commented Apr 2, 2022

Code got formatted by CI. Please request CI again if you still want to have this PR merged. If the PR is from a forked repo, please download the patch files from the GitHub Actions web page and apply them locally.

@hengzi hengzi removed the request for review from oneflow-ci-bot April 2, 2022 08:10
@yoonlee888 yoonlee888 requested review from oneflow-ci-bot and Flowingsun007 and removed request for oneflow-ci-bot April 28, 2022 01:10
@github-actions
Copy link
Contributor

View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/7949/

@github-actions
Copy link
Contributor

Speed stats: 
GPU Name: NVIDIA GeForce GTX 1080 

❌ OneFlow resnet50 time: 129.1ms (= 12911.3ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 143.3ms (= 14328.5ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.11 (= 143.3ms / 129.1ms)

OneFlow resnet50 time: 83.5ms (= 8354.7ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 85.3ms (= 8531.9ms / 100, input_shape=[8, 3, 224, 224])
❌ Relative speed: 1.02 (= 85.3ms / 83.5ms)

OneFlow resnet50 time: 53.5ms (= 10702.0ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 53.1ms (= 10610.9ms / 200, input_shape=[4, 3, 224, 224])
❌ Relative speed: 0.99 (= 53.1ms / 53.5ms)

OneFlow resnet50 time: 44.4ms (= 8884.0ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 40.7ms (= 8146.6ms / 200, input_shape=[2, 3, 224, 224])
❌ Relative speed: 0.92 (= 40.7ms / 44.4ms)

OneFlow resnet50 time: 39.1ms (= 7819.1ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 40.8ms (= 8151.5ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.04 (= 40.8ms / 39.1ms)

OneFlow swin dataloader time: 0.255s (= 51.018s / 200, num_workers=1)
PyTorch swin dataloader time: 0.152s (= 30.332s / 200, num_workers=1)
Relative speed: 0.595 (= 0.152s / 0.255s)

OneFlow swin dataloader time: 0.069s (= 13.755s / 200, num_workers=4)
PyTorch swin dataloader time: 0.043s (= 8.518s / 200, num_workers=4)
Relative speed: 0.619 (= 0.043s / 0.069s)

OneFlow swin dataloader time: 0.039s (= 7.701s / 200, num_workers=8)
PyTorch swin dataloader time: 0.022s (= 4.452s / 200, num_workers=8)
Relative speed: 0.578 (= 0.022s / 0.039s)

❌ OneFlow resnet50 time: 145.2ms (= 14517.6ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 168.0ms (= 16800.3ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.16 (= 168.0ms / 145.2ms)

OneFlow resnet50 time: 93.1ms (= 9308.9ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 109.7ms (= 10967.7ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.18 (= 109.7ms / 93.1ms)

OneFlow resnet50 time: 77.7ms (= 15539.9ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 88.7ms (= 17735.6ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
❌ Relative speed: 1.14 (= 88.7ms / 77.7ms)

OneFlow resnet50 time: 63.7ms (= 12745.9ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 74.5ms (= 14900.1ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
❌ Relative speed: 1.17 (= 74.5ms / 63.7ms)

OneFlow resnet50 time: 56.0ms (= 11209.2ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 68.7ms (= 13735.6ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.23 (= 68.7ms / 56.0ms)

@github-actions
Copy link
Contributor

View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/7949/

@github-actions
Copy link
Contributor

CI failed when running job: cuda-module. PR label automerge has been removed

@github-actions
Copy link
Contributor

Speed stats: 
GPU Name: GeForce GTX 1080 

✔️ OneFlow resnet50 time: 128.5ms (= 12848.5ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 142.6ms (= 14259.7ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.11 (= 142.6ms / 128.5ms)

OneFlow resnet50 time: 80.3ms (= 8033.5ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 87.1ms (= 8713.0ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.08 (= 87.1ms / 80.3ms)

OneFlow resnet50 time: 49.8ms (= 9958.4ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 61.6ms (= 12321.0ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.24 (= 61.6ms / 49.8ms)

OneFlow resnet50 time: 39.5ms (= 7898.8ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 49.5ms (= 9905.1ms / 200, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 1.25 (= 49.5ms / 39.5ms)

OneFlow resnet50 time: 36.9ms (= 7384.0ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 40.1ms (= 8020.0ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 1.09 (= 40.1ms / 36.9ms)

OneFlow swin dataloader time: 0.260s (= 52.063s / 200, num_workers=1)
PyTorch swin dataloader time: 0.151s (= 30.136s / 200, num_workers=1)
Relative speed: 0.579 (= 0.151s / 0.260s)

OneFlow swin dataloader time: 0.068s (= 13.626s / 200, num_workers=4)
PyTorch swin dataloader time: 0.043s (= 8.593s / 200, num_workers=4)
Relative speed: 0.631 (= 0.043s / 0.068s)

OneFlow swin dataloader time: 0.036s (= 7.154s / 200, num_workers=8)
PyTorch swin dataloader time: 0.021s (= 4.269s / 200, num_workers=8)
Relative speed: 0.597 (= 0.021s / 0.036s)

✔️ OneFlow resnet50 time: 135.9ms (= 13590.4ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 162.9ms (= 16290.9ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.20 (= 162.9ms / 135.9ms)

OneFlow resnet50 time: 88.0ms (= 8801.7ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 110.8ms (= 11084.3ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.26 (= 110.8ms / 88.0ms)

OneFlow resnet50 time: 64.0ms (= 12804.2ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 87.8ms (= 17563.2ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.37 (= 87.8ms / 64.0ms)

OneFlow resnet50 time: 50.3ms (= 10051.4ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 77.2ms (= 15448.0ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.54 (= 77.2ms / 50.3ms)

OneFlow resnet50 time: 49.0ms (= 9790.8ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 77.7ms (= 15538.5ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.59 (= 77.7ms / 49.0ms)

@github-actions
Copy link
Contributor

CI failed when running job: cuda-benchmark. PR label automerge has been removed

@yoonlee888 yoonlee888 requested review from oneflow-ci-bot and removed request for oneflow-ci-bot April 29, 2022 02:34
@github-actions
Copy link
Contributor

View latest API docs preview at: https://staging.oneflow.info/docs/Oneflow-Inc/oneflow/pr/7949/

@mergify mergify bot merged commit 395da40 into master Apr 29, 2022
@mergify mergify bot deleted the searchsorted branch April 29, 2022 03:57
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@BBuf BBuf BBuf approved these changes

@hengzi hengzi hengzi approved these changes

@hjchen2 hjchen2 Awaiting requested review from hjchen2 hjchen2 is a code owner

@jackalcooper jackalcooper Awaiting requested review from jackalcooper jackalcooper is a code owner

@daquexian daquexian Awaiting requested review from daquexian

@liujuncheng liujuncheng Awaiting requested review from liujuncheng liujuncheng is a code owner

@guo-ran guo-ran Awaiting requested review from guo-ran

@MARD1NO MARD1NO Awaiting requested review from MARD1NO

@doombeaker doombeaker Awaiting requested review from doombeaker doombeaker is a code owner

@Flowingsun007 Flowingsun007 Awaiting requested review from Flowingsun007

@oneflow-ci-bot oneflow-ci-bot Awaiting requested review from oneflow-ci-bot

Assignees
No one assigned
Labels
automerge enhancement op
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
5 participants
@yoonlee888 @hengzi @Flowingsun007 @BBuf @oneflow-ci-bot