From fd0c5b0e3eb71a74efcbe5fbd31ea77ffaae7d70 Mon Sep 17 00:00:00 2001
From: Disty0 <disty@disty.xyz>
Date: Wed, 11 Jun 2025 15:12:22 +0300
Subject: [PATCH] Update changelog

---
 CHANGELOG.md | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index d8e022c8e..7dcd87c17 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -9,14 +9,18 @@
   - Increase the medvram mode threshold from 8GB to 12GB  
   - Set CPU backend to use FP32 by default  
   - Relax Python version checks for Zluda  
-  - Don't override user set gfx version with ROCm  
   - Make VAE options not require model reload  
 
 - **Torch**  
   - Set default to `torch==2.7.1`  
-  - Support ROCm 6.4 with `---use-nightly`  
   - Force upgrade pip when installing Torch  
 
+- **ROCm**  
+  - Support ROCm 6.4 with `---use-nightly`  
+  - Don't override user set gfx version  
+  - Don't override gfx version with RX 9000  
+  - Fix flash-atten repo  
+
 - **SDNQ Quantization**  
   - Add group size support for convolutional layers  
   - Add quantized matmul support for for convolutional layers  
@@ -31,7 +35,6 @@
   - PixArt Sigma Small and Large loading  
   - TAESD previews with PixArt  
   - VAE Tiling with non-default tile sizes  
-  - flash-atten repo with ROCm  
 
 ## Update for 2025-06-02