---
license: mit
---
The corresponding Image/Audio/Video/Music files can be found in:

**Pre-Train Datasets:**

**CC3M:** liuhaotian/LLaVA-CC3M-Pretrain-595K

**LAION:** laion/LAION-Audio-300M

**OpenVid:** nkp37/OpenVid-1M

**Fine-Tune Datasets:**

**Image:** liuhaotian/LLaVA-Instruct-150K

**Audio:** gijs/tacos-captioning & gijs/clothoaqa

**Video:** lmms-lab/LLaVA-Video-178K

**Music:** TwinkStart/Nsynth & CLAPv2/MusicCaps & jfforero/MUSGEN-EmoMusic & DynamicSuperb/ChordClassification_AcousticGuitarAndPiano

**Test Datasets:**

**Image:** https://cocodataset.org/ & https://cs.stanford.edu/people/dorarad/gqa/ & https://ocr-vqa.github.io/ & https://textvqa.org/

Other modality test files are included in the corresponding **Fine-Tune Datasets**.