From cb62604d2132bc4f4af22f0a4dc7baaad4a910ce Mon Sep 17 00:00:00 2001 From: pythongosssss <125205205+pythongosssss@users.noreply.github.com> Date: Mon, 25 May 2026 19:35:53 +0100 Subject: [PATCH] feat: map more custom node loaders to asset picker (#12340) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit ## Summary Updates the following node mappings to show the new model picker dialog: - DynamiCrafterModelLoader > checkpoints (https://github.com/kijai/ComfyUI-DynamiCrafterWrapper/blob/d312c62982f1582b7955444d3a22ac948fb9c603/nodes.py#L367) - DynamiCrafterCNLoader > controlnet (https://github.com/kijai/ComfyUI-DynamiCrafterWrapper/blob/d312c62982f1582b7955444d3a22ac948fb9c603/nodes.py#L217) - MelBandRoFormerModelLoader > diffusion_models (https://github.com/kijai/ComfyUI-MelBandRoFormer/blob/92c86854e6654f4aacc97484471af95c98ea16d4/nodes.py#L31) - INPAINT_LoadFooocusInpaint > inpaint (head + patch inputs) (https://github.com/Acly/comfyui-inpaint-nodes/blob/b32f293d3f3ed9f2bc11099b30c8695ff984a593/nodes.py#L129) - AILab_QwenVL_Advanced > LLM/Qwen-VL/* (10 variants) ( "model_name": "Pick the Qwen-VL checkpoint. First run downloads weights into models/LLM/Qwen-VL, so leave disk space.",) ## Changes - **What**: Add custom node mappings ## Review Focus ## Screenshots (if applicable) ┆Issue is synchronized with this [Notion page](https://www.notion.so/PR-12340-feat-map-more-custom-node-loaders-to-asset-picker-3656d73d365081929c0dc9179a9e9c67) by [Unito](https://www.unito.io) --- .../assets/mappings/modelNodeMappings.ts | 22 ++++++++++++++++++- 1 file changed, 21 insertions(+), 1 deletion(-) diff --git a/src/platform/assets/mappings/modelNodeMappings.ts b/src/platform/assets/mappings/modelNodeMappings.ts index e1c3e20992..4eba368b54 100644 --- a/src/platform/assets/mappings/modelNodeMappings.ts +++ b/src/platform/assets/mappings/modelNodeMappings.ts @@ -91,6 +91,16 @@ export const MODEL_NODE_MAPPINGS: ReadonlyArray< 'AILab_QwenVL_PromptEnhancer', 'model_name' ], + ['LLM/Qwen-VL/Qwen2.5-VL-3B-Instruct', 'AILab_QwenVL_Advanced', 'model_name'], + ['LLM/Qwen-VL/Qwen2.5-VL-7B-Instruct', 'AILab_QwenVL_Advanced', 'model_name'], + ['LLM/Qwen-VL/Qwen3-VL-2B-Instruct', 'AILab_QwenVL_Advanced', 'model_name'], + ['LLM/Qwen-VL/Qwen3-VL-2B-Thinking', 'AILab_QwenVL_Advanced', 'model_name'], + ['LLM/Qwen-VL/Qwen3-VL-4B-Instruct', 'AILab_QwenVL_Advanced', 'model_name'], + ['LLM/Qwen-VL/Qwen3-VL-4B-Thinking', 'AILab_QwenVL_Advanced', 'model_name'], + ['LLM/Qwen-VL/Qwen3-VL-8B-Instruct', 'AILab_QwenVL_Advanced', 'model_name'], + ['LLM/Qwen-VL/Qwen3-VL-8B-Thinking', 'AILab_QwenVL_Advanced', 'model_name'], + ['LLM/Qwen-VL/Qwen3-VL-32B-Instruct', 'AILab_QwenVL_Advanced', 'model_name'], + ['LLM/Qwen-VL/Qwen3-VL-32B-Thinking', 'AILab_QwenVL_Advanced', 'model_name'], ['LLM/checkpoints', 'LoadChatGLM3', 'chatglm3_checkpoint'], // ---- Qwen3 TTS (ComfyUI-FunBox) ---- @@ -148,6 +158,8 @@ export const MODEL_NODE_MAPPINGS: ReadonlyArray< 'DownloadAndLoadDynamiCrafterCNModel', 'model' ], + ['checkpoints', 'DynamiCrafterModelLoader', 'ckpt_name'], + ['controlnet', 'DynamiCrafterCNLoader', 'ckpt_name'], // ---- LayerStyle (ComfyUI_LayerStyle_Advance) ---- ['BEN', 'LS_LoadBenModel', 'model'], @@ -157,6 +169,8 @@ export const MODEL_NODE_MAPPINGS: ReadonlyArray< // ---- Inpaint (comfyui-inpaint-nodes) ---- ['inpaint', 'INPAINT_LoadInpaintModel', 'model_name'], + ['inpaint', 'INPAINT_LoadFooocusInpaint', 'head'], + ['inpaint', 'INPAINT_LoadFooocusInpaint', 'patch'], // ---- LayerDiffuse (comfyui-layerdiffuse) ---- ['layer_model', 'LayeredDiffusionApply', 'config'], @@ -212,9 +226,15 @@ export const MODEL_NODE_MAPPINGS: ReadonlyArray< ['ultralytics/bbox', 'UltralyticsDetectorProvider', 'model_name'], ['ultralytics/segm', 'UltralyticsDetectorProvider', 'model_name'], + // ---- Mel-Band RoFormer audio separation (ComfyUI-MelBandRoFormer) ---- + ['diffusion_models', 'MelBandRoFormerModelLoader', 'model_name'], + // ---- ComfyUI core geometry estimation (MoGe) ---- ['geometry_estimation', 'LoadMoGeModel', 'model_name'], // ---- ComfyUI core optical flow (RAFT) ---- - ['optical_flow', 'OpticalFlowLoader', 'model_name'] + ['optical_flow', 'OpticalFlowLoader', 'model_name'], + + // ---- WanVideo (ComfyUI-WanVideoWrapper) ---- + ['loras', 'WanVideoLoraSelect', 'lora'] ] as const satisfies ReadonlyArray