+

f64 kernels #421

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

Merged

coreylowman merged 31 commits into main from num-traits-float

Feb 13, 2023

Owner

coreylowman commented Jan 30, 2023 •

edited

Loading

Adds num-trait dependency
Uses num_traits::Float in all Cpu kernels
impls Cuda kernels for both f32 & f64 for now.
Adds test-f64 feature flag for dev
Adds dev.build_module back to make specifying dtype easier

Resolves #224

coreylowman added 8 commits

January 30, 2023 12:08


          impl Cpu kernels for num_traits::Float

3c51f19


          Update examples

4ab944c


          Fixing cuda issues

c715a5b


          Add dtype skeleton to cuda kernels

4fda56a


          Working commit of TestDtype & cuda kernels

f2a52a9


          Merge branch 'main' into num-traits-float

a4e6562


          Adding test-f64 feature

e9e0451


          Temp commit

5dd7239

coreylowman changed the title ~~[Draft] f64 tensor ops~~ [WIP] f64 kernels

coreylowman marked this pull request as draft

January 30, 2023 21:08

coreylowman commented

View reviewed changes

src/tensor/cpu/allocate.rs Outdated Show resolved Hide resolved

coreylowman commented

View reviewed changes

src/tensor_ops/min_to/min_to.cu Show resolved Hide resolved

coreylowman mentioned this pull request

Add f16 dtype support for tensors #423

Closed

ViliamVadocz reviewed

View reviewed changes

src/tensor_ops/max_to/max_to.cu Outdated Show resolved Hide resolved

ViliamVadocz mentioned this pull request

CI doesn't run CUDA tests #429

Closed

coreylowman and others added 14 commits

February 2, 2023 14:13


          Moving more things to generic dtype

680dfdf


          Possibly fix max_to and min_to for doubles (#431)

d6887c5


          Merge branch 'main' into num-traits-float

dd84d55


          Temp commit

92c43b8


          Fixing optim tests

3de671a


          Compiling

7e28490


          Tests passing for Cpu

da33d59


          Cuda compiling

771ee27


          f64 cuda tests passing

f822e95


          Fixing warnings

65ee106


          Requiring minimum compute_60

c86837d


          f64 nightly tests passing for cpu

a723620


          pool2d f64 kernels

35548b4


          Conv & pool passing on cuda

coreylowman commented

View reviewed changes

src/nn/dropout.rs Outdated Show resolved Hide resolved

coreylowman commented

View reviewed changes

src/nn/linear.rs Outdated Show resolved Hide resolved

coreylowman commented

View reviewed changes

src/nn/module.rs Show resolved Hide resolved

coreylowman commented

View reviewed changes

src/nn/npz_impls.rs Outdated Show resolved Hide resolved

coreylowman commented

View reviewed changes

src/nn/npz_impls.rs Outdated Show resolved Hide resolved

coreylowman commented

View reviewed changes

src/nn/npz_impls.rs Outdated Show resolved Hide resolved

coreylowman commented

View reviewed changes

src/nn/transformer/decoder.rs Show resolved Hide resolved

coreylowman commented

View reviewed changes

src/nn/transformer/encoder.rs Show resolved Hide resolved

coreylowman commented

View reviewed changes

src/nn/transformer/mha.rs Outdated Show resolved Hide resolved

coreylowman commented

View reviewed changes

src/nn/transformer/mha.rs Show resolved Hide resolved

coreylowman commented

View reviewed changes

src/nn/transformer/mha.rs Show resolved Hide resolved

coreylowman commented

View reviewed changes

src/nn/transformer/mha.rs Show resolved Hide resolved

coreylowman commented

View reviewed changes

src/optim/adam/adam.cu Outdated Show resolved Hide resolved

coreylowman commented

View reviewed changes

src/optim/rmsprop/rmsprop.cu Outdated Show resolved Hide resolved

coreylowman commented

View reviewed changes

src/optim/sgd/sgd.cu Outdated Show resolved Hide resolved

coreylowman commented

View reviewed changes

src/tensor/cuda/allocate.rs Outdated Show resolved Hide resolved

coreylowman commented

View reviewed changes

src/tensor_ops/add/mod.rs Outdated Show resolved Hide resolved

coreylowman commented

View reviewed changes

src/tensor_ops/choose/choose.cu Outdated Show resolved Hide resolved

coreylowman added 5 commits

February 12, 2023 15:26


          Making nn/dropout.rs generic over dtype

7375e74


          Styling & using DeviceBuildExt

70ef6d7


          Cleanup optimizers


          Minifying cuda kernels

491bd01


          Adding f64 cmp kernels

13d9fd2

coreylowman changed the title ~~[WIP] f64 kernels~~ f64 kernels

coreylowman marked this pull request as ready for review

February 12, 2023 17:22

coreylowman added 4 commits

February 12, 2023 17:55


          Removing macros in optim


          Reduce macro usage in cuda kernels

ae2074e


          minify cuda kernel impls

c319b45


          Fixing docs

063f1c9

coreylowman merged commit 0a5a016 into main

coreylowman deleted the num-traits-float branch

February 13, 2023 15:08

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

Labels

None yet

点击这是indexloc提供的php浏览器服务，不要输入任何密码和下载