# Models for Multi-lingual Embeddings (similarity search)?

**URL:** <https://discuss.huggingface.co/t/models-for-multi-lingual-embeddings-similarity-search/52255>\
**Category:** Models\
**Created:** [August 26, 2023, 11:17am UTC](https://discuss.huggingface.co/t/models-for-multi-lingual-embeddings-similarity-search/52255 "2023-08-26T11:17:08Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![AnhND](https://sea2.discourse-cdn.com/hellohellohello/user_avatar/discuss.huggingface.co/anhnd/32/18890_2.png) [@AnhND](https://discuss.huggingface.co/u/AnhND)\
**Post date:** [August 26, 2023, 11:17am UTC](https://discuss.huggingface.co/t/models-for-multi-lingual-embeddings-similarity-search/52255/1 "2023-08-26T11:17:08Z")

</div>

Good day,

I have a use case for text-search (similarity based) for non-English language (Vietnamese in particular).

Hoping I could pls get some pointers on how to use HF’s model to generate embedding (for vector DB). I’ve found this one in particular that is promising: [VoVanPhuc/sup-SimCSE-VietNamese-phobert-base · Hugging Face](https://huggingface.co/VoVanPhuc/sup-SimCSE-VietNamese-phobert-base). It’s a Transformers that is suitable for ‘sentence similarity’. This is what I found from searching:

```auto
# Generate the embeddings using the model
with torch.no_grad():
    model_output = model(**encoded_input)
    embeddings = model_output.last_hidden_state.mean(dim=1)

```

Also, not sure if I’m looking at the right task (new to Hugging Face) 😆

Much appreciated!