From ff7d6a6db277c803ffbd8687bea8ca88fedeec92 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Jukka=20Sepp=C3=A4nen?=
 <40791699+kijai@users.noreply.github.com>
Date: Mon, 14 Oct 2024 10:57:22 +0300
Subject: [PATCH] Update README.md

---
 README.md | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/README.md b/README.md
index 80c06ca..d75fe88 100644
--- a/README.md
+++ b/README.md
@@ -9,7 +9,9 @@ todo
 - optimize memory use
 
 Besides text encoder, which can peak at ~12GB VRAM use, this should run at 9-10GB VRAM when using 1280x768.
-With fp8 and 384p model it can fit under 6GB too.
+With fp8 and 384p model it can fit under 6GB too. Note that these tests were done on 4090, older cards may not support every optimization.
+
+Resolutions outside the model defaults perform poorly.
 
 Model loading has not been optimized at all yet, currently needs everything (choose either of the transformers) from here: