From 148650d157b8a6fa993ab01c352529a80db4e718 Mon Sep 17 00:00:00 2001 From: ANANDHU S <71482562+anandhu-eng@users.noreply.github.com> Date: Sat, 19 Oct 2024 15:50:17 +0530 Subject: [PATCH] enabled amd implementation for llama2 --- docs/benchmarks/language/llama2-70b.md | 7 +++++++ main.py | 4 ++++ 2 files changed, 11 insertions(+) diff --git a/docs/benchmarks/language/llama2-70b.md b/docs/benchmarks/language/llama2-70b.md index e68693716..40c62cf71 100644 --- a/docs/benchmarks/language/llama2-70b.md +++ b/docs/benchmarks/language/llama2-70b.md @@ -25,3 +25,10 @@ hide: {{ mlperf_inference_implementation_readme (4, "llama2-70b-99", "neuralmagic") }} {{ mlperf_inference_implementation_readme (4, "llama2-70b-99.9", "neuralmagic") }} + +=== "AMD" + ## AMD MLPerf Implementation + +{{ mlperf_inference_implementation_readme (4, "llama2-70b-99", "amd") }} + +{{ mlperf_inference_implementation_readme (4, "llama2-70b-99.9", "amd") }} diff --git a/main.py b/main.py index 38e6e1650..a4f09d559 100755 --- a/main.py +++ b/main.py @@ -43,6 +43,10 @@ def mlperf_inference_implementation_readme(spaces, model, implementation, *, imp devices = [ "CUDA" ] frameworks = [ "TensorRT" ] + elif implementation == "amd": + devices = [ "ROCm" ] + frameworks = [ "pytorch" ] + elif implementation == "neuralmagic": devices = [ "CUDA" ] frameworks = [ "pytorch" ]