diff options
| author | Laurent Mazare <laurent.mazare@gmail.com> | 2023-08-14 21:30:41 +0100 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2023-08-14 21:30:41 +0100 |
| commit | 90374097dc99b14dfc935318a18c21fc5909291f (patch) | |
| tree | 61a45bfa840f4551a5581a4ab169dce23b13db4b /candle-examples/Cargo.toml | |
| parent | c84883ecf2c240792392353175b634f6ec92a011 (diff) | |
| download | candle-90374097dc99b14dfc935318a18c21fc5909291f.tar.gz candle-90374097dc99b14dfc935318a18c21fc5909291f.tar.bz2 candle-90374097dc99b14dfc935318a18c21fc5909291f.zip | |
Cudnn support (#445)
* Add a cudnn feature to be used for conv2d.
* Allocate the proper workspace.
* Only create a single cudnn handle per cuda device.
* Proper cudnn usage.
* Bugfix.
Diffstat (limited to 'candle-examples/Cargo.toml')
| -rw-r--r-- | candle-examples/Cargo.toml | 1 |
1 files changed, 1 insertions, 0 deletions
diff --git a/candle-examples/Cargo.toml b/candle-examples/Cargo.toml index 54eb0be6..45216a5c 100644 --- a/candle-examples/Cargo.toml +++ b/candle-examples/Cargo.toml @@ -47,6 +47,7 @@ anyhow = { workspace = true } default = [] accelerate = ["dep:accelerate-src", "candle/accelerate", "candle-nn/accelerate", "candle-transformers/accelerate"] cuda = ["candle/cuda", "candle-nn/cuda", "candle-transformers/cuda"] +cudnn = ["candle/cudnn"] flash-attn = ["cuda", "dep:candle-flash-attn"] mkl = ["dep:intel-mkl-src", "candle/mkl", "candle-nn/mkl", "candle-transformers/mkl"] nccl = ["cuda", "cudarc/nccl", "dep:half"] |