MNeMoNiCuZ-ComfyUI-mnemic-nodes/web/docs/MNeMiC_LoadTextImagePairsList.md

🖼️+📝 Load Text-Image Pairs (List)

The list version of 🖼️+📝 Load Text-Image Pair (Single): returns the whole dataset as parallel lists, rotated so the seed picks the starting point.

Inputs

• seed — Where the lists start. Set its control to increment to rotate.
• input_path: Folder holding the images and caption files. Relative paths start in ComfyUI's input folder.
• force_reload — Re-read from disk instead of using the cache. Turn it on after changing files on disk, then off again.
• image_input / text_input (optional) — Data already in the workflow. When both are connected they take priority over the folder.
• limit_count (optional) — Cap on how many pairs to return. 0 returns all of them.
• text_format_extension (optional) — Caption file extension without the dot. Default txt.

Outputs

All in the same order, one entry per pair:

• image_list — The images, as one batch.
• string_list — The captions.
• image_path_list — Full paths.
• image_filename_list — Filenames without extensions.
• total_count — How many pairs the dataset holds.

Notes

The preview shows only the first selected image and its text, and updates when input_path, seed or caption extension changes. Missing captions leave the text preview blank. Images without captions remain in the output with empty text. The preview uses the cached batch when available. Connected inputs show only the first output image and text after execution.

The folder is read once and cached per path; force_reload clears it. Images whose dimensions differ from the first are resized and centre-cropped so the batch is uniform. A single text input paired with several images is applied to all of them; a list of texts is truncated to match the shorter of the two.