EmbeddingGemma 2 Text 270M (LiteRT-LM)

EmbeddingGemma 2 Text 270M (LiteRT-LM)

Turns text into vectors for search, clustering and question answering on phones and laptops. A 270M text-only EmbeddingGemma 2 build packaged for LiteRT-LM, covering over 100 languages and code.

Made by litert-community on Hugging Face. These notes are Mindket's, written from its model page and model card; the model's own files and license are on Hugging Face.

About

What it is: A text-only EmbeddingGemma 2 model with 270M parameters, packaged for the LiteRT-LM runtime. It turns text into vectors that you can compare, so apps can search, group and match documents on the device itself. It handles over 100 languages and programming code.

Who it suits: App developers who want text search or classification to run on a phone, a laptop or other small hardware, without sending text to a server. Anyone who stores many vectors and wants them smaller will also find it useful.

Industry | Who uses it | How they use it | Example
Mobile apps | App developers | Embed notes or messages on the device for search | Searching saved notes on a phone
Software | Product teams | Cluster or classify documents in an app | Sorting support tickets by topic
Knowledge tools | Builders of question answering | Retrieve passages that match a question | Finding the right page of a manual
Research | Embedding researchers | Test small multilingual embedders | Comparing results at different vector sizes

You can skip it if your task needs images or audio, since this variant handles text only.

Where to use it:

• Android and iOS, through Google's AI Edge Gallery app
• macOS, and a web demo for embedding search
• MediaPipe tasks such as semantic image search and decision making
• Windows, macOS or Linux with Python, to fetch the files

Use case | Who | What they get back
Document search | App developers | Documents ranked by closeness to a query
Classification and clustering | Product teams | Text grouped or labeled by meaning
Question answering | Builders | Passages that match a question

How to set it up:

• Python, to install the Hugging Face client
• A LiteRT-LM runtime in your app or on your device
• Room for a model of 270M parameters

1. Install the Hugging Face client.
2. Download the model files to your computer.
3. Load them with LiteRT-LM in your app, following Google's overview guide.

pip install -U huggingface_hub
hf download litert-community/embeddinggemma-2-text-270m-litert-lm

How to use it: Feed text to the model through LiteRT-LM and compare the vectors it returns to find similar texts. You can try it first in Google's AI Edge Gallery app or the web demo. The source shows no code, so follow the LiteRT-LM guide for the exact calls.

Tool | What it does
LiteRT-LM | Runs the model on the device
AI Edge Gallery | App for trying the model on Android and iOS
MediaPipe | Offers ready tasks that use the embedder, such as semantic retrieval

Example: You embed a folder of short notes on a phone, then embed the question "where did I save the router password". The model returns vectors, and the notes closest to the question come back first. Nothing leaves the device.

For newcomers and experts: If you are new, an embedding is a list of numbers that captures what a text means, so similar texts get similar lists. Try the AI Edge Gallery demo before writing any code.

If you know the field: the model combines a 130M transformer with a 140M embedder. It uses Matryoshka Representation Learning, so you can cut vectors down to 128 dimensions and store far less.

Text
LiteRT-LM model
Vector
Compare vectors
Search or group results

Good to know: Smaller vectors cost less to store, and the model supports sizes down to 128 dimensions.

• Parameters: 270M in total
• Languages: over 100, plus programming code
• Scope: text only
• License: Apache 2.0
• Cost to run: none named
• Data it reaches: model files downloaded to your computer, plus the text you choose to embed
• Not for: images or audio, since this variant is text only

On Hugging Face, 15 November 2026

• Library: litert-lm
• Downloads in the last 30 days: 16,202
• Likes: 14
• License: apache-2.0
• Based on: google/embeddinggemma-2
• Tags: litert-lm

Links

• Model: https://huggingface.co/litert-community/embeddinggemma-2-text-270m-litert-lm
• Files: https://huggingface.co/litert-community/embeddinggemma-2-text-270m-litert-lm/tree/main

Get the model

Download its files with Hugging Face's command line:

pip install -U huggingface_hub
hf download litert-community/embeddinggemma-2-text-270m-litert-lm

Then follow the usage in its model card. Nothing in this kit runs by itself.