diff options
| author | Nicolas Patry <patry.nicolas@protonmail.com> | 2023-12-20 09:54:19 +0100 |
|---|---|---|
| committer | Nicolas Patry <patry.nicolas@protonmail.com> | 2023-12-20 09:54:19 +0100 |
| commit | 9b5e4843a63180a2803b1e836b4ca90f14281d03 (patch) | |
| tree | 32dc8468d9201cf079d437f96ad5cdadb2dddb42 /Cargo.toml | |
| parent | 03641293eeb1dd0ff3d5a93e85c7f9eb289704e4 (diff) | |
| download | candle-9b5e4843a63180a2803b1e836b4ca90f14281d03.tar.gz candle-9b5e4843a63180a2803b1e836b4ca90f14281d03.tar.bz2 candle-9b5e4843a63180a2803b1e836b4ca90f14281d03.zip | |
Optimizing decode matmul (Phi at 28tok/s on M3).
Adding some benchmark in order to help checking out matmul performance.
Diffstat (limited to 'Cargo.toml')
| -rw-r--r-- | Cargo.toml | 1 |
1 files changed, 1 insertions, 0 deletions
@@ -32,6 +32,7 @@ accelerate-src = { version = "0.3.2" } anyhow = { version = "1", features = ["backtrace"] } byteorder = "1.4.3" clap = { version = "4.2.4", features = ["derive"] } +criterion = { version = "0.5.1", default-features=false } cudarc = { version = "0.9.14", features = ["f16"] } gemm = { version = "0.16.6", features = ["wasm-simd128-enable"] } hf-hub = "0.3.0" |