From 1bbebccc0a99068a93820c17338ab408609ef6ac Mon Sep 17 00:00:00 2001
From: Tri Dao <tridpq@gmail.com>
Date: Sat, 9 Jul 2022 23:34:29 -0700
Subject: [PATCH] Edit README to mention bf16 support

---
 README.md | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/README.md b/README.md
index 4d0c780..ba0a6f4 100644
--- a/README.md
+++ b/README.md
@@ -23,8 +23,8 @@ PYTHONPATH=$PWD python benchmarks/benchmark_flash_attention.py
 
 FlashAttention currently supports:
 1. Turing or Ampere GPUs (e.g., A100, RTX 3090, T4, RTX 2080).
-2. fp16.
-3. Head dimensions 16, 32, 64, 128 (bwd requires A100).
+2. fp16 and bf16 (bf16 requires Ampere GPUs).
+3. Head dimensions 16, 32, 64, 128 (head dim 128 backward requires A100).
 
 Our tentative roadmap:
 1. [Jun 2022] Make package pip-installable.