embedding model bge-m3 how to set return_sparse=True ?

[colinTmx]

Note that the issue tracker is NOT the place for general support.

[qinxuye]

createing_embdding should be able to pass extra params, you can try it using Xinference Python client.

[colinTmx]

createing_embdding should be able to pass extra params, you can try it using Xinference Python client.

encode() got an unexpected keyword argument 'return_sparse'. what's the keyword?

[github-actions[bot]]

This issue is stale because it has been open for 7 days with no activity.

[colinTmx]

Hope it can be supported as soon as possible.

[stolennumber]

Hope it can be supported as soon as possible.

[pengjunfeng11]

createing_embdding should be able to pass extra params, you can try it using Xinference Python client.

我查看了源代码，发现这个问题的根本是由于使用snetence_transformer进行embedding，而s_t的编码都是稠密向量（仅以bge-m3为例，该模型在官方贴中提到了这一点[https://huggingface.co/BAAI/bge-m3]），所以无法解决这个问题。目前看来解决方法之一，是更改默认的框架，为用户提供出了sentence transformer以外的其他框架。问问官方后续有这步工作吗？如果有的话那么就不必重复造轮子？

默认只使用sentence transformer

return_sparse参数最终会放入forward方法，导致报错。

[github-actions[bot]]

This issue is stale because it has been open for 7 days with no activity.