From 6c4843567d56b76774527c9ad9deb2b37405d28b Mon Sep 17 00:00:00 2001
From: Rattus <rattus128@gmail.com>
Date: Mon, 16 Mar 2026 21:57:26 +1000
Subject: [PATCH] wan: encode frames 2x2.

Reduce VRAM usage greatly by encoding frames 2 at a time rather than
4.
---
 comfy/ldm/wan/vae.py | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/comfy/ldm/wan/vae.py b/comfy/ldm/wan/vae.py
index 65296afa8..a4a567130 100644
--- a/comfy/ldm/wan/vae.py
+++ b/comfy/ldm/wan/vae.py
@@ -458,11 +458,12 @@ class WanVAE(nn.Module):
         conv_idx = [0]
         ## cache
         t = x.shape[2]
-        iter_ = 1 + (t - 1) // 4
+        t = 1 + ((t - 1) // 4) * 4
+        iter_ = 1 + (t - 1) // 2
         feat_map = None
         if iter_ > 1:
             feat_map = [None] * count_cache_layers(self.encoder)
-        ## 对encode输入的x，按时间拆分为1、4、4、4....
+        ## 对encode输入的x，按时间拆分为1、2、2、2....(总帧数先按4N+1向下取整)
         for i in range(iter_):
             conv_idx = [0]
             if i == 0:
@@ -472,7 +473,7 @@ class WanVAE(nn.Module):
                     feat_idx=conv_idx)
             else:
                 out_ = self.encoder(
-                    x[:, :, 1 + 4 * (i - 1):1 + 4 * i, :, :],
+                    x[:, :, 1 + 2 * (i - 1):1 + 2 * i, :, :],
                     feat_cache=feat_map,
                     feat_idx=conv_idx,
                     final=(i == (iter_ - 1)))