From b9198d076b43a3769494fb0ee2682096fdf37ded Mon Sep 17 00:00:00 2001 From: markus583 Date: Sun, 22 Sep 2024 15:22:08 +0200 Subject: [PATCH] remove exact timings. --- README.md | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/README.md b/README.md index a0f04d69..914c56a5 100644 --- a/README.md +++ b/README.md @@ -60,13 +60,13 @@ sat = SaT("sat-3l-sm", onnx_providers=["CUDAExecutionProvider"]) # PyTorch GPU >>> model = SaT("sat-3l-sm") >>> model.half().to("cuda") ->>> %timeit list(model.split(texts)) -138 ms ± 8.41 ms per loop (mean ± std. dev. of 7 runs, 1 loop each) +>>> list(model.split(texts)) +# quite fast already, but... # onnxruntime GPU >>> model = SaT("sat-3l-sm", ort_providers=["CUDAExecutionProvider"]) >>> %timeit list(model.split(texts)) -198 ms ± 1.36 ms per loop (mean ± std. dev. of 7 runs, 1 loop each) +# ...this should be ~50% faster! ``` If you wish to use LoRA in combination with an ONNX model: