CaptionForge

JLC_CaptionForge

The main capstone/orchestration node. It consumes Pass A raw caption records, runs the distillation and validation stages, and exports final captions. The VLM-validated natural paragraph is the authoritative long caption. Formatting stages should not blindly rewrite that natural caption.

Captioning/CaptionForge · 9 inputs · 40 parameters · 5 outputs

JLC_CaptionForge_Pipeline_Planner

The central planning node for normal runs. It coordinates: - input image path or direct image passthrough - recursive folder traversal - filename glob filtering - output directory - run name - overwrite behavior - Pass A witness run counts - seed schedules - sampling schedules - max image size - max token budget - LoRA trigger word - user caption anchor - distiller settings - validator settings - final export settings - derived JSONL/TXT/config paths

Captioning/CaptionForge · 7 inputs · 31 parameters · 3 outputs

JLC_CaptionForgeExtraOptions

Build a CAPTIONFORGE_EXTRA_OPTIONS payload for caption-template controls.

Captioning/CaptionForge · 2 outputs

JLC_CaptionForgeJoy

Canonical JoyCaption raw-caption witness CapNode for CaptionForge.

Captioning/CaptionForge/Captioning Nodes · 5 inputs · 18 parameters · 5 outputs

JLC_CaptionForgeOllamaCaption

Ollama-backed VLM Pass A witness. This node delegates image-caption generation to a local Ollama server rather than loading Hugging Face/PyTorch weights inside ComfyUI. It can use configured Ollama VLM tags such as: text gemma4:26b qwen3.6:35B-A3B huihuiai/gemma-4-abliterated:26b Its purpose is to provide access to other raw-caption witness alternatives. It's function is parallel to the Joy Caption and Qwen Caption nodes, and should not be confused with the later VLM validator/capstone role.

Captioning/CaptionForge/Caption Nodes · 7 inputs · 18 parameters · 5 outputs

JLC_CaptionForgeQwen

Canonical Qwen raw-caption witness CapNode for CaptionForge.

Captioning/CaptionForge/Captioning Nodes · 6 inputs · 17 parameters · 5 outputs