From 69c83a60dbbcc837c985d34ee5aa23b606d6d156 Mon Sep 17 00:00:00 2001
From: kijai <40791699+kijai@users.noreply.github.com>
Date: Wed, 4 Mar 2026 01:30:40 +0200
Subject: [PATCH] Small fixes

---
 comfy/supported_models.py    |  3 +++
 comfy_extras/nodes_rtdetr.py | 18 ++++++++++--------
 2 files changed, 13 insertions(+), 8 deletions(-)

diff --git a/comfy/supported_models.py b/comfy/supported_models.py
index 9bc628bc2..6606d27e4 100644
--- a/comfy/supported_models.py
+++ b/comfy/supported_models.py
@@ -1746,6 +1746,9 @@ class RT_DETR_v4(supported_models_base.BASE):
         out = model_base.RT_DETR_v4(self, device=device)
         return out
 
+    def clip_target(self, state_dict={}):
+        return None
+
 models = [LotusD, Stable_Zero123, SD15_instructpix2pix, SD15, SD20, SD21UnclipL, SD21UnclipH, SDXL_instructpix2pix, SDXLRefiner, SDXL, SSD1B, KOALA_700M, KOALA_1B, Segmind_Vega, SD_X4Upscaler, Stable_Cascade_C, Stable_Cascade_B, SV3D_u, SV3D_p, SD3, StableAudio, AuraFlow, PixArtAlpha, PixArtSigma, HunyuanDiT, HunyuanDiT1, FluxInpaint, Flux, LongCatImage, FluxSchnell, GenmoMochi, LTXV, LTXAV, HunyuanVideo15_SR_Distilled, HunyuanVideo15, HunyuanImage21Refiner, HunyuanImage21, HunyuanVideoSkyreelsI2V, HunyuanVideoI2V, HunyuanVideo, CosmosT2V, CosmosI2V, CosmosT2IPredict2, CosmosI2VPredict2, ZImagePixelSpace, ZImage, Lumina2, WAN22_T2V, WAN21_T2V, WAN21_I2V, WAN21_FunControl2V, WAN21_Vace, WAN21_Camera, WAN22_Camera, WAN22_S2V, WAN21_HuMo, WAN22_Animate, WAN21_FlowRVS, WAN21_SCAIL, Hunyuan3Dv2mini, Hunyuan3Dv2, Hunyuan3Dv2_1, HiDream, Chroma, ChromaRadiance, ACEStep, ACEStep15, Omnigen2, QwenImage, Flux2, Kandinsky5Image, Kandinsky5, Anima, RT_DETR_v4]
 
 models += [SVD_img2vid]
diff --git a/comfy_extras/nodes_rtdetr.py b/comfy_extras/nodes_rtdetr.py
index 60c3c9b92..61307e268 100644
--- a/comfy_extras/nodes_rtdetr.py
+++ b/comfy_extras/nodes_rtdetr.py
@@ -82,16 +82,18 @@ class DrawBBoxes(io.ComfyNode):
 
     @classmethod
     def execute(cls, bboxes, image=None) -> io.NodeOutput:
-        # Normalise bboxes to a list-of-lists (one list of detections per image).
-        # It may arrive as: a bare dict, a flat list of dicts, or a list of lists.
+        # Normalise to list[list[dict]], then fit to batch size B.
         B = image.shape[0] if image is not None else 1
         if isinstance(bboxes, dict):
-            bboxes = [[bboxes]] * B
-        elif not isinstance(bboxes, list) or len(bboxes) == 0:
-            bboxes = [[]] * B
-        elif not isinstance(bboxes[0], list):
-            # flat list of dicts: same detections for every image
-            bboxes = [bboxes] * B
+            bboxes = [[bboxes]]
+        elif not isinstance(bboxes, list) or not bboxes:
+            bboxes = [[]]
+        elif isinstance(bboxes[0], dict):
+            bboxes = [bboxes]  # flat list → same detections for every image
+
+        if len(bboxes) == 1:
+            bboxes = bboxes * B
+        bboxes = (bboxes + [[]] * B)[:B]
 
         if image is None:
             B = len(bboxes)