diff --git a/nodes_sampler.py b/nodes_sampler.py
index e724a19..a44073d 100644
--- a/nodes_sampler.py
+++ b/nodes_sampler.py
@@ -1090,6 +1090,7 @@ class WanVideoSampler:
             lynx_ref_latent = lynx_embeds.get("ref_latent", None)
             lynx_ref_latent_uncond = lynx_embeds.get("ref_latent_uncond", None)
             lynx_ref_text_embed = lynx_embeds.get("ref_text_embed", None)
+            lynx_ref_text_embed = dict_to_device(lynx_ref_text_embed, device)
             lynx_cfg_scale = lynx_embeds.get("cfg_scale", 1.0)
             if not isinstance(lynx_cfg_scale, list):
                 lynx_cfg_scale = [lynx_cfg_scale] * (steps + 1)
@@ -1111,7 +1112,7 @@ class WanVideoSampler:
                 lynx_ref_buffer = transformer(
                     [lynx_ref_input.to(device, dtype)],
                     torch.tensor([0], device=device),
-                    [emb.to(device) for emb in lynx_ref_text_embed["prompt_embeds"]],
+                    lynx_ref_text_embed["prompt_embeds"],
                     seq_len=math.ceil((lynx_ref_latent.shape[2] * lynx_ref_latent.shape[3]) / 4 * lynx_ref_latent.shape[1]),
                     lynx_embeds=lynx_embeds
                 )
@@ -1125,7 +1126,7 @@ class WanVideoSampler:
                     lynx_ref_buffer_uncond = transformer(
                         [lynx_ref_input_uncond.to(device, dtype)],
                         torch.tensor([0], device=device),
-                        [emb.to(device) for emb in lynx_ref_text_embed["prompt_embeds"]],
+                        lynx_ref_text_embed["prompt_embeds"],
                         seq_len=math.ceil((lynx_ref_latent.shape[2] * lynx_ref_latent.shape[3]) / 4 * lynx_ref_latent.shape[1]),
                         lynx_embeds=lynx_embeds,
                         is_uncond=True