Nodes/ComfyUI-ArchAi3d-Qwen/ArchAi3D_Qwen_Encoder
ComfyUI Node

ArchAi3D_Qwen_Encoder

A ComfyUI node in ArchAi3d/Qwen with 18 inputs and 3 outputs.

By amir84ferdos·Created 10 months ago·Updated 4 months ago· 65
ArchAi3D_Qwen_Encoder
  • clip
  • vae
  • image1_vl
  • image2_vl
  • image3_vl
  • image1_latent
  • image2_latent
  • image3_latent
  • conditioning
  • latent
  • formatted_prompt
prompt
system_prompt
image1_labelImage 1
image2_labelImage 2
image3_labelImage 3
conditioning_strength1.00
image1_latent_strength1.00
image2_latent_strength1.00
image3_latent_strength1.00
debug_modefalse
CategoryArchAi3d/Qwen

Inputs (18)

NameTypeDefaultDescription
clipCLIPQwen-VL CLIP model for encoding text and vision tokens
promptSTRINGText prompt (vision tokens inserted automatically in ChatML format)
system_promptSTRINGOptional system prompt (wrapped in ChatML <|im_start|>system block)
image1_labelSTRINGImage 1Custom label for Image 1 (e.g., 'Image 1 (target)', 'Image 1 (room)')
image2_labelSTRINGImage 2Custom label for Image 2 (e.g., 'Image 2 (style ref)', 'Image 2 (material)')
image3_labelSTRINGImage 3Custom label for Image 3 (e.g., 'Image 3 (color ref)', 'Image 3 (lighting)')
conditioning_strengthFLOAT1.000–2Global strength for text+vision embeddings (1.0=normal, <1.0=weaker, >1.0=stronger)
image1_latent_strengthFLOAT1.000–2Image1 latent strength (1.0=normal, <1.0=weaker, >1.0=stronger)
image2_latent_strengthFLOAT1.000–2Image2 latent strength (1.0=normal, <1.0=weaker, >1.0=stronger)
image3_latent_strengthFLOAT1.000–2Image3 latent strength (1.0=normal, <1.0=weaker, >1.0=stronger)
debug_modeBOOLEANfalseEnable console logging (shows strengths, shapes, and formatted prompt)
vaeoptVAEVAE for encoding reference latents (required if using latent images)
image1_vloptIMAGEImage 1 for vision encoder (RGB only, expects correct size)
image2_vloptIMAGEImage 2 for vision encoder (RGB only, expects correct size)
image3_vloptIMAGEImage 3 for vision encoder (RGB only, expects correct size)
image1_latentoptIMAGEImage 1 for reference latent (RGB only, expects correct size)
image2_latentoptIMAGEImage 2 for reference latent (RGB only, expects correct size)
image3_latentoptIMAGEImage 3 for reference latent (RGB only, expects correct size)

Outputs (3)

NameTypeDescription
conditioningCONDITIONINGText+vision embeddings with reference latents metadata attached
latentLATENTImage1 latent in standard format (for VAEDecode or other latent nodes)
formatted_promptSTRINGFinal ChatML-formatted prompt with vision tokens (for debugging)