zhixiangwei/VLM-150M
收藏官方服务:
资源简介:
VLM-150M是一个经过SFT增强的Qwen2VL模型重新标注的大规模图像-文本数据集,旨在提高文本描述的对齐性和详细程度。
VLM-150M is a large-scale image-text dataset that has been recaptioned using an SFT-enhanced Qwen2VL model to enhance the alignment and detail of textual descriptions.
提供机构:
zhixiangwei


