Nodes/ComfyUI Llava-OneVision/OneVision Caption Folder
ComfyUI Node

OneVision Caption Folder

A ComfyUI node in LLaVA-OneVision with 10 inputs and 1 output.

By kijai·Created 2 years ago·Updated 7 months ago· 102
OneVision Caption Folder
  • llava_model
  • STRING
folder_path
promptYou are AI captioning tool, you caption images in very elaborate detail without referring to the image as 'the image', the results should be useful for image model training purposes. You focus on the composition, style and action any possible subject is performing. You don't make assumptions or try to tell a story. You also describe the background of the image separately. Caption this image:
max_tokens512
keep_model_loadedtrue
temperature0.20
seed1
max_image_size1024
prefix
suffix
CategoryLLaVA-OneVision

Inputs (10)

NameTypeDefaultDescription
llava_modelLLAVAMODEL
folder_pathSTRING
promptSTRINGYou are AI captioning tool, you caption images in very elaborate detail without referring to the image as 'the image', the results should be useful for image model training purposes. You focus on the composition, style and action any possible subject is performing. You don't make assumptions or try to tell a story. You also describe the background of the image separately. Caption this image:
max_tokensINT5121–8192
keep_model_loadedBOOLEANtrue
temperatureFLOAT0.200–1
seedINT11–18446744073709550000
max_image_sizeINT1024256–8192
prefixSTRING
suffixSTRING

Outputs (1)

NameTypeDescription
STRINGSTRING