From 8fcfa1dd2d20c3d55ba20a8b715d14e086b5368b Mon Sep 17 00:00:00 2001
From: Zhongjie Duan <35051019+Artiprocher@users.noreply.github.com>
Date: Tue, 26 Aug 2025 17:22:25 +0800
Subject: [PATCH] bugfix

---
 diffsynth/models/qwen_image_dit.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/diffsynth/models/qwen_image_dit.py b/diffsynth/models/qwen_image_dit.py
index e0d493c..13cb5ca 100644
--- a/diffsynth/models/qwen_image_dit.py
+++ b/diffsynth/models/qwen_image_dit.py
@@ -94,7 +94,7 @@ class QwenEmbedRope(nn.Module):
 
     def _expand_pos_freqs_if_needed(self, video_fhw, txt_seq_lens):
         if isinstance(video_fhw, list):
-            video_fhw = video_fhw[0]
+            video_fhw = tuple(max([i[j] for i in video_fhw]) for j in range(3))
         _, height, width = video_fhw
         if self.scale_rope:
             max_vid_index = max(height // 2, width // 2)