Open-Qwen2VL-Data

Introduction This repository contains the data for Open-Qwen2VL: Compute-Efficient Pre-Training of Fully-Open Multimodal LLMs on Academic Resources. Project page: Code: Dataset ccs ebdataset: CC3M-CC12M-SBU filtered by CLIP, we directly download the webdataset based on the released of curated subset of BLIP-1 datacomp medium dfn webdataset: DataComp-Medium-128M filtered by DFN, we just… See the full description on the dataset page:

Source: Hugging Face Hub (weizhiwang/Open-Qwen2VL-Data). Metadata imported from the dataset’s Hub tags.