dimtoneff-ComfyUI-VL-Nodes

Files

T

History

dimtoneff 31ac10c7ba 1. Added a new TextSave_VL node (inspired from "was" pack):

- Added skip_overwrite parameter (BOOLEAN, default False)
   - When enabled, prevents overwriting existing files
   - Useful when purging bad captions and re-running workflows to save only missing descriptions

2. LoadImagesFromDirBatch_VL and LoadImagesFromDirList_VL nodes:
   - Added traverse_subdirs parameter (BOOLEAN, default False)
   - When enabled, recursively scans all subdirectories
   - Returns full absolute paths instead of just filenames
   - When disabled, maintains original behavior (current directory only)
   - Useful for organizing images in folder structures

These enhancements provide better control over file handling in batch image captioning workflows while maintaining full backward compatibility.

2025-08-29 02:38:11 +02:00

Batch_images_caption_workflow.jpg

1. Added a new TextSave_VL node (inspired from "was" pack):

2025-08-29 02:38:11 +02:00

Batch_images_caption_workflow.json

1. Added a new TextSave_VL node (inspired from "was" pack):

2025-08-29 02:38:11 +02:00

internVL3_5_workflow.jpg

- added OpenGVLab/InternVL3_5-1B/38B models support for GGUF and transformers (GGUF may or may not work for you, depends on the llama.cpp versions and python bindings)

2025-08-27 13:32:09 +02:00

internVL3_5_workflow.json

- added OpenGVLab/InternVL3_5-1B/38B models support for GGUF and transformers (GGUF may or may not work for you, depends on the llama.cpp versions and python bindings)