labring / FastGPT

FastGPT is a knowledge-based platform built on the LLMs, offers a comprehensive suite of out-of-the-box capabilities such as data processing, RAG retrieval, and visual AI workflow orchestration, letting you easily develop and deploy complex question-answering systems without the need for extensive setup or configuration.
https://fastgpt.in
Other
15.58k stars 4.09k forks source link

embedding的维度数量限制问题 #1850

Open travisgu opened 3 weeks ago

travisgu commented 3 weeks ago

例行检查

你的版本

问题描述, 日志截图

image

看来起embedding的维度不能超过1536,PG Vector的index算法不能超过2000 dimension, 但是通过其他方式进行支持: You can use half-precision indexing to index up to 4,000 dimensions or binary quantization to index up to 64,000 dimensions. Another option is dimensionality reduction.

不知是否有计划在以后的版本支持更高维度的embedding。

复现步骤 使用Ollama上的Qwen:72b模型进行embedding 预期结果

相关截图

c121914yu commented 1 week ago

half 模式后续打算增加,不过得看具体情况。可以考虑先用 milvus