From 2ba0dd8bd3725d8ec6173261ab0016f9f336eeb8 Mon Sep 17 00:00:00 2001
From: Han Qi <hanq@google.com>
Date: Tue, 30 Jul 2024 22:24:28 +0000
Subject: [PATCH] Add newest llama-3 benchmarks

---
 benchmarks/summary.md | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/benchmarks/summary.md b/benchmarks/summary.md
index 41b011de..e4487cd0 100644
--- a/benchmarks/summary.md
+++ b/benchmarks/summary.md
@@ -7,6 +7,9 @@ Date | Device  | dtype | batch size | cache length |max input length |max output
 ----| ------- | ------ |---------- | -------------|-----------------|------------------|----------------------
 2024-04-24 | TPU v5e-8 | bfloat16 | 128 | 2048 | 1024 | 1024 | 8249 
 2024-04-24 | TPU v5e-8 | int8 | 256 | 2048 | 1024 | 1024 | 10873
+2024-07-29 | TPU v5e-8 | int8 | 256 | 2048 | 1024 | 1024 | 8471.54
+
+**NOTE:(2024-07-29)** Looks like we have a regression in the past 3 month. We are working in fixing it.
 
 
 ## Gemma - 7B