From 87a01773be23b996e38217a6a574c2de08ac560f Mon Sep 17 00:00:00 2001
From: leejet <leejet714@gmail.com>
Date: Fri, 24 Jul 2026 22:37:26 +0800
Subject: [PATCH] fix: add frame dimension for Hunyuan IMG2VID encoding (#1816)

---
 src/stable-diffusion.cpp | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/stable-diffusion.cpp b/src/stable-diffusion.cpp
index 860c0679..bb5c3e09 100644
--- a/src/stable-diffusion.cpp
+++ b/src/stable-diffusion.cpp
@@ -5789,7 +5789,7 @@ static std::optional<ImageGenerationLatents> prepare_video_generation_latents(sd
         auto encode_condition_frame = [&](const sd::Tensor<float>& image,
                                           int64_t latent_frame,
                                           const char* name) -> bool {
-            auto encoded = sd_ctx->sd->encode_first_stage(image);
+            auto encoded = sd_ctx->sd->encode_first_stage(image.unsqueeze(2));
             if (encoded.empty()) {
                 LOG_ERROR("failed to encode Hunyuan Video %s conditioning frame", name);
                 return false;