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

Dev add contiguous view ops #7503

Merged
merged 60 commits into from
Mar 1, 2022
Merged

Dev add contiguous view ops #7503

merged 60 commits into from
Mar 1, 2022

Conversation

Flowingsun007
Copy link
Contributor

@Flowingsun007 Flowingsun007 commented Feb 15, 2022 •
edited
Loading

此pr为搬运#6854 中的基于view机制实现的部分view op，包括：

  • view op
  • unsqueeze op
  • squeeze op

（这部分view op的引入不会改变tensor的contiguous属性）

@Flowingsun007 Flowingsun007 marked this pull request as ready for review February 15, 2022 07:58
@Flowingsun007 Flowingsun007 enabled auto-merge (squash) February 15, 2022 08:49
BBuf
BBuf reviewed Feb 21, 2022
View reviewed changes
oneflow/core/framework/tensor_methods.cpp Outdated Show resolved Hide resolved
oneflow/core/framework/tensor_methods.cpp Outdated Show resolved Hide resolved
oneflow/core/framework/tensor_methods.cpp Outdated Show resolved Hide resolved
oneflow/core/framework/tensor_methods.cpp Outdated Show resolved Hide resolved
oneflow/core/functional/impl/array_functor.cpp Outdated Show resolved Hide resolved
@github-actions
Copy link
Contributor

github-actions bot commented Mar 1, 2022

Speed stats: 
GPU Name: GeForce GTX 1080 

✔️ OneFlow resnet50 time: 128.6ms (= 12857.6ms / 100, input_shape=[16, 3, 224, 224])
PyTorch resnet50 time: 145.6ms (= 14557.9ms / 100, input_shape=[16, 3, 224, 224])
✔️ Relative speed: 1.13 (= 145.6ms / 128.6ms)

❌ OneFlow resnet50 time: 79.3ms (= 7934.9ms / 100, input_shape=[8, 3, 224, 224])
PyTorch resnet50 time: 87.3ms (= 8727.9ms / 100, input_shape=[8, 3, 224, 224])
✔️ Relative speed: 1.10 (= 87.3ms / 79.3ms)

OneFlow resnet50 time: 53.3ms (= 10653.8ms / 200, input_shape=[4, 3, 224, 224])
PyTorch resnet50 time: 58.9ms (= 11775.3ms / 200, input_shape=[4, 3, 224, 224])
✔️ Relative speed: 1.11 (= 58.9ms / 53.3ms)

OneFlow resnet50 time: 43.1ms (= 8627.7ms / 200, input_shape=[2, 3, 224, 224])
PyTorch resnet50 time: 46.7ms (= 9343.6ms / 200, input_shape=[2, 3, 224, 224])
✔️ Relative speed: 1.08 (= 46.7ms / 43.1ms)

OneFlow resnet50 time: 39.4ms (= 7879.4ms / 200, input_shape=[1, 3, 224, 224])
PyTorch resnet50 time: 38.1ms (= 7619.6ms / 200, input_shape=[1, 3, 224, 224])
✔️ Relative speed: 0.97 (= 38.1ms / 39.4ms)

✔️ OneFlow resnet50 time: 142.4ms (= 14241.3ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 159.6ms (= 15960.3ms / 100, input_shape=[16, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.12 (= 159.6ms / 142.4ms)

OneFlow resnet50 time: 89.2ms (= 8923.2ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 104.7ms (= 10473.0ms / 100, input_shape=[8, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.17 (= 104.7ms / 89.2ms)

OneFlow resnet50 time: 61.4ms (= 12287.9ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 77.1ms (= 15418.1ms / 200, input_shape=[4, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.25 (= 77.1ms / 61.4ms)

OneFlow resnet50 time: 53.5ms (= 10700.8ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 65.7ms (= 13145.5ms / 200, input_shape=[2, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.23 (= 65.7ms / 53.5ms)

OneFlow resnet50 time: 48.1ms (= 9615.5ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
PyTorch resnet50 time: 60.6ms (= 12120.7ms / 200, input_shape=[1, 3, 224, 224], ddp, world size=2)
✔️ Relative speed: 1.26 (= 60.6ms / 48.1ms)

@github-actions
Copy link
Contributor

github-actions bot commented Mar 1, 2022

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

@github-actions github-actions bot removed the automerge label Mar 1, 2022
@oneflow-ci-bot oneflow-ci-bot removed their request for review March 1, 2022 07:31
@oneflow-ci-bot oneflow-ci-bot removed their request for review March 1, 2022 07:55
@oneflow-ci-bot oneflow-ci-bot removed their request for review March 1, 2022 08:17
@Flowingsun007 Flowingsun007 requested review from oneflow-ci-bot and removed request for oneflow-ci-bot March 1, 2022 12:47
@oneflow-ci-bot oneflow-ci-bot removed their request for review March 1, 2022 13:16
@oneflow-ci-bot oneflow-ci-bot requested review from oneflow-ci-bot and removed request for oneflow-ci-bot March 1, 2022 13:16
@oneflow-ci-bot oneflow-ci-bot requested review from oneflow-ci-bot and removed request for oneflow-ci-bot March 1, 2022 14:33
@oneflow-ci-bot oneflow-ci-bot self-requested a review March 1, 2022 16:26
@Flowingsun007 Flowingsun007 merged commit bc2a158 into master Mar 1, 2022
@Flowingsun007 Flowingsun007 deleted the dev_add_contiguous_view_ops branch March 1, 2022 19:35
@Flowingsun007 Flowingsun007 mentioned this pull request Mar 11, 2022
Closed
marigoold pushed a commit that referenced this pull request Mar 15, 2022
@Flowingsun007 @marigoold
Dev add contiguous view ops (#7503)
225c074 
* view op

* narrow op

* squeeze unsqueeze op

* revert narrow
wyg1997 pushed a commit that referenced this pull request Mar 17, 2022
@Flowingsun007 @wyg1997
Dev add contiguous view ops (#7503)
55599f7 
* view op

* narrow op

* squeeze unsqueeze op

* revert narrow
lixinqi
lixinqi reviewed Jul 22, 2022
View reviewed changes
oneflow/core/functional/impl/common.cpp
return target_stride;
}

Maybe<Shape> InferShape(const std::shared_ptr<one::Tensor>& x, const Shape& shape) {
Copy link
Contributor

Choose a reason for hiding this comment

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

这个函数是谁写的呢？函数名称没有信息量。另外它不需要x吧，只需要x->shape()

Copy link
Contributor Author

Choose a reason for hiding this comment

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

好的，我后续提个pr refine一下

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@wyg1997 wyg1997 wyg1997 left review comments

@hjchen2 hjchen2 hjchen2 left review comments

@lixinqi lixinqi lixinqi left review comments

@BBuf BBuf BBuf approved these changes

@daquexian daquexian daquexian approved these changes

@chengtbf chengtbf Awaiting requested review from chengtbf

@jackalcooper jackalcooper Awaiting requested review from jackalcooper

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

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

Successfully merging this pull request may close these issues.
7 participants
@Flowingsun007 @lixinqi @daquexian @hjchen2 @BBuf @wyg1997 @oneflow-ci-bot