raynardj/amz-product-title-emb

该数据集是使用Gemini嵌入模型text-embedding-005对亚马逊产品标题数据集进行嵌入处理的结果。每个样本包含产品标题文本（content字段）和对应的768维浮点数嵌入向量（embeddings字段，采用float16精度以减小存储需求）。数据集专为训练或机器学习任务设计，支持流式加载以实现快速训练启动。数据集规模较大，包含约1487万个样本，总大小约26GB。This dataset contains embeddings generated using the Gemini embedding model text-embedding-005 on an Amazon product title dataset. Each example includes the product title text (content field) and its corresponding 768-dimensional embedding vector (embeddings field, stored in float16 precision to reduce storage). The dataset is designed for training or machine learning tasks and supports streaming for fast training startup. It is large-scale, with approximately 14.87 million examples and a total size of about 26GB.