A new custom node for ComfyUI, named MOSS-Image Vision Text Encode, has been developed. This node allows reference images to be processed through the Qwen3.5-9B vision tower, with its embeddings integrated into the text conditioning. This enables images to influence the generation process as part of the prompt, rather than solely through VAE latents. AI
IMPACT Enables more nuanced image-to-text conditioning in generative models.
RANK_REASON This is a custom node for an existing tool (ComfyUI) that integrates a specific model (Qwen3.5-9B) for image conditioning.
- ComfyUI
- lodestones/MOSS-Image-SFT-8-steps-dmd
- MOSS-Image
- Qwen3.5-9B
- ThetaCursed/ComfyUI-MOSSImageTextEncoder
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →