[GH-ISSUE #4778] Support for jina-embeddings-v2-base-zh #28769

Closed
opened 2026-04-22 07:17:38 -05:00 by GiteaMirror · 8 comments
Owner

Originally created by @wwjCMP on GitHub (Jun 2, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/4778

https://huggingface.co/jinaai/jina-embeddings-v2-base-zh

Originally created by @wwjCMP on GitHub (Jun 2, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/4778 https://huggingface.co/jinaai/jina-embeddings-v2-base-zh
GiteaMirror added the model label 2026-04-22 07:17:38 -05:00
Author
Owner

@thinkverse commented on GitHub (Jun 2, 2024):

I am pinging @JoanFM for this issue.

<!-- gh-comment-id:2143822753 --> @thinkverse commented on GitHub (Jun 2, 2024): I am pinging @JoanFM for this issue.
Author
Owner

@JoanFM commented on GitHub (Jun 2, 2024):

Hey @wwjCMP

the only problem for supporting this feature is to get the right preprocessing and tokenization right.

I am not sure if there is currently supprt for another chinese model from which I can take inspiration.

<!-- gh-comment-id:2143881133 --> @JoanFM commented on GitHub (Jun 2, 2024): Hey @wwjCMP the only problem for supporting this feature is to get the right preprocessing and tokenization right. I am not sure if there is currently supprt for another chinese model from which I can take inspiration.
Author
Owner

@wwjCMP commented on GitHub (Jun 2, 2024):

Hey @wwjCMP

the only problem for supporting this feature is to get the right preprocessing and tokenization right.

I am not sure if there is currently supprt for another chinese model from which I can take inspiration.

I can't quite grasp what you mean. But, here is a Chinese embedding model that I am currently using.

https://huggingface.co/aspire/acge_text_embedding
mofanke/acge_text_embedding:latest

<!-- gh-comment-id:2143885714 --> @wwjCMP commented on GitHub (Jun 2, 2024): > Hey @wwjCMP > > the only problem for supporting this feature is to get the right preprocessing and tokenization right. > > I am not sure if there is currently supprt for another chinese model from which I can take inspiration. I can't quite grasp what you mean. But, here is a Chinese embedding model that I am currently using. https://huggingface.co/aspire/acge_text_embedding mofanke/acge_text_embedding:latest
Author
Owner

@JoanFM commented on GitHub (Jun 2, 2024):

Hey @wwjCMP
the only problem for supporting this feature is to get the right preprocessing and tokenization right.
I am not sure if there is currently supprt for another chinese model from which I can take inspiration.

I can't quite grasp what you mean. But, here is a Chinese embedding model that I am currently using.

https://huggingface.co/aspire/acge_text_embedding mofanke/acge_text_embedding:latest

I mean, is this model available through llama.cpp?

<!-- gh-comment-id:2143957078 --> @JoanFM commented on GitHub (Jun 2, 2024): > > Hey @wwjCMP > > the only problem for supporting this feature is to get the right preprocessing and tokenization right. > > I am not sure if there is currently supprt for another chinese model from which I can take inspiration. > > I can't quite grasp what you mean. But, here is a Chinese embedding model that I am currently using. > > https://huggingface.co/aspire/acge_text_embedding mofanke/acge_text_embedding:latest I mean, is this model available through llama.cpp?
Author
Owner

@JoanFM commented on GitHub (Jun 6, 2024):

hey @wwjCMP ,

I open a PR in llama.cpp to support this model https://github.com/ggerganov/llama.cpp/pull/7795

<!-- gh-comment-id:2151724227 --> @JoanFM commented on GitHub (Jun 6, 2024): hey @wwjCMP , I open a PR in `llama.cpp` to support this model https://github.com/ggerganov/llama.cpp/pull/7795
Author
Owner

@EntropyYue commented on GitHub (Oct 30, 2024):

Already supported, you can download it here EntropyYue/jina-embeddings-v2-base-zh.

<!-- gh-comment-id:2447787574 --> @EntropyYue commented on GitHub (Oct 30, 2024): Already supported, you can download it here [EntropyYue/jina-embeddings-v2-base-zh](https://ollama.com/EntropyYue/jina-embeddings-v2-base-zh).
Author
Owner

@JoanFM commented on GitHub (Nov 5, 2024):

But how did you do @EntropyYue?

How is it supported in llama.cpp? Are you sure the tokenization and the result is equivalent to the usage of HF?

<!-- gh-comment-id:2457623105 --> @JoanFM commented on GitHub (Nov 5, 2024): But how did you do @EntropyYue? How is it supported in `llama.cpp`? Are you sure the tokenization and the result is equivalent to the usage of `HF`?
Author
Owner

@EntropyYue commented on GitHub (Nov 5, 2024):

But how did you do @EntropyYue?

How is it supported in llama.cpp? Are you sure the tokenization and the result is equivalent to the usage of HF?

It was successfully converted in the hf-to-gguf script. And it seems to have no problem using it either. may be it has inssue like your pr, I'm not sure

<!-- gh-comment-id:2457930320 --> @EntropyYue commented on GitHub (Nov 5, 2024): > But how did you do @EntropyYue? > > How is it supported in `llama.cpp`? Are you sure the tokenization and the result is equivalent to the usage of `HF`? It was successfully converted in the `hf-to-gguf` script. And it seems to have no problem using it either. may be it has inssue like your pr, I'm not sure
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#28769