Tīrtha · embedding model
EmbeddingGemma 2 (text)
Turns text into a list of numbers (a vector) that captures its meaning, so you can search, group and retrieve text by meaning instead of exact words. Send up to 64 texts in one call, up to 100,000 characters in total. Each text must fit the model's length limit, so very long texts, code or non-English text may need to be shorter.
Open weights · Apache-2.0
by Google · Apache-2.0 · open weights
About this model
- Made for
- search and retrieval (RAG), classification, clustering, semantic similarity
- Languages
- 100+ languages
- Vector size
- 768 numbers per text
- Model size
- 744M parameters
- Licence
- Apache-2.0
How to use it
- Start each query with
task: search result | query:. - Write each document as
title: {title} | text: {content}, withnoneas the title when there is none. - Compare vectors with cosine similarity.
- Split long text into chunks; each model has its own length limit.
How to call it
POST https://api.tirtha.ai/v1/embeddings
{"model": "tirtha/embeddinggemma-2",
"input": ["first text", "second text"]}
Start using EmbeddingGemma 2 (text) in 3 steps
Read the docs- Get an API keyFree to create. It works for every model in the catalogue.Get an API key
- Add creditsPay as you go. Credits work on every model.Add credits
- Point your client at TīrthaAny OpenAI-compatible client works. Use your key and this base URL.base_url: "https://api.tirtha.ai/v1" api_key: "YOUR_TIRTHA_KEY"