---
title: README
emoji: 🌖
colorFrom: yellow
colorTo: indigo
sdk: static
pinned: false
---

# LiteRT Community

[LiteRT](https://ai.google.dev/edge/litert) is Google's on-device framework for high-performance ML & GenAI deployment on edge platforms. It is the improved successor to TensorFlow Lite. On this community page, you can find ready-to-run LiteRT models for a wide range of ML/AI tasks. 

Within this ecosystem, [LiteRT-LM](https://ai.google.dev/edge/litert-lm) specializes in cutting edge GenAI. Recognizing that LLMs now function as complex pipelines of related models rather than single standalone models, LiteRT-LM leverages LiteRT to deliver an optimized solution for running LLMs on-device. 

Both LiteRT and LiteRT-LM can run on a variety of devices including Android, iOS, Windows, macOS, Linux, IoT and Web allow easy deployment and scaling across a diverse device landscape. 

# Trying it Live

<div align="center">

| [<svg xmlns="http://www.w3.org/2000/svg" height="72px" viewBox="0 -960 960 960" width="72px" fill="currentColor"><path d="M40-240q9-107 65.5-197T256-580l-74-128q-6-9-3-19t13-15q8-5 18-2t16 12l74 128q86-36 180-36t180 36l74-128q6-9 16-12t18 2q10 5 13 15t-3 19l-74 128q94 53 150.5 143T920-240H40Zm275.5-124.5Q330-379 330-400t-14.5-35.5Q301-450 280-450t-35.5 14.5Q230-421 230-400t14.5 35.5Q259-350 280-350t35.5-14.5Zm400 0Q730-379 730-400t-14.5-35.5Q701-450 680-450t-35.5 14.5Q630-421 630-400t14.5 35.5Q659-350 680-350t35.5-14.5Z"/></svg>](https://play.google.com/store/apps/details?id=com.google.ai.edge.gallery&pli=1) | [<svg xmlns="http://www.w3.org/2000/svg" height="84px" viewBox="0 -960 960 960" width="84px" fill="currentColor"><path d="M160-615v-60h60v60h-60Zm0 335v-275h60v275h-60Zm292 0H347q-24.75 0-42.37-17.63Q287-315.25 287-340v-280q0-24.75 17.63-42.38Q322.25-680 347-680h105q24.75 0 42.38 17.62Q512-644.75 512-620v280q0 24.75-17.62 42.37Q476.75-280 452-280Zm-105-60h105v-280H347v280Zm228 60v-60h165v-114H635q-24.75 0-42.37-17.63Q575-489.25 575-514v-106q0-24.75 17.63-42.38Q610.25-680 635-680h165v60H635v106h105q24.75 0 42.38 17.62Q800-478.75 800-454v114q0 24.75-17.62 42.37Q764.75-280 740-280H575Z"/></svg>](https://apps.apple.com/us/app/google-ai-edge-gallery/id6749645337) | [<svg xmlns="http://www.w3.org/2000/svg" height="72px" viewBox="0 -960 960 960" width="72px" fill="currentColor"><path d="M320-120v-40l80-80H160q-33 0-56.5-23.5T80-320v-440q0-33 23.5-56.5T160-840h640q33 0 56.5 23.5T880-760v440q0 33-23.5 56.5T800-240H560l80 80v40H320ZM160-440h640v-320H160v320Zm0 0v-320 320Z"/></svg>](https://ai.google.dev/edge/litert-lm/cli) | [<svg xmlns="http://www.w3.org/2000/svg" height="72px" viewBox="0 -960 960 960" width="72px" fill="currentColor"><path d="M160-120q-33 0-56.5-23.5T80-200v-560q0-33 23.5-56.5T160-840h560q33 0 56.5 23.5T800-760v80h80v80h-80v80h80v80h-80v80h80v80h-80v80q0 33-23.5 56.5T720-120H160Zm0-80h560v-560H160v560Zm80-80h200v-160H240v160Zm240-280h160v-120H480v120Zm-240 80h200v-200H240v200Zm240 200h160v-240H480v240ZM160-760v560-560Z"/></svg>](https://ai.google.dev/edge/litert-lm/cli) | [<svg xmlns="http://www.w3.org/2000/svg" height="72px" viewBox="0 -960 960 960" width="72px" fill="currentColor"><path d="M838-79 710-207v103h-60v-206h206v60H752l128 128-42 43Zm-358-1q-83 0-156-31.5T197-197q-54-54-85.5-126.36T80-478q0-83.49 31.5-156.93Q143-708.36 197-762.68 251-817 324-848.5 397-880 480-880t156 31.5q73 31.5 127 85.82 54 54.32 85.5 127.75Q880-561.49 880-478q0 23-2 44.5t-7 43.5h-63q6-21.67 9-43.33 3-21.67 3-44.47 0-22.8-2.95-45.6-2.94-22.8-8.83-45.6H648q2 23 4 45.5t2 45q0 22.5-1.25 44.5T649-390h-61q3-22 4.5-44t1.5-44q0-22.75-1.5-45.5T588-569H373.42q-3.42 23-4.92 45.5t-1.5 45q0 22.5 1.5 44.5t4.5 44h197v60H384q14 53 34 104t62 86q23 0 45-2.5t45-7.5v60q-23 5-45 7.5T480-80ZM151.78-390H312q-2.5-22-3.75-44T307-478q0-22.75 1-45.5t3-45.5H151.71q-5.85 22.8-8.78 45.6-2.93 22.8-2.93 45.6t2.95 44.47q2.94 21.66 8.83 43.33ZM172-629h149.59q11.41-48 28.91-93.5T395-810q-71 24-129.5 69.5T172-629Zm222 478q-26-41-43.5-86T323-330H172q33 67 91 114t131 65Zm-10-478h193q-13-54-36-104t-61-89q-38 40-61 89.5T384-629Zm255.34 0H788q-35-66-93-112t-129-68q27 41 44.5 86.5t28.84 93.5Z"/></svg>](https://huggingface.co/spaces/tylermullen/Gemma4) |
| :---: | :---: | :---: | :---: | :---: |
| [Android](https://play.google.com/store/apps/details?id=com.google.ai.edge.gallery&pli=1) | [iOS](https://apps.apple.com/us/app/google-ai-edge-gallery/id6749645337) | [Desktop](https://ai.google.dev/edge/litert-lm/cli) | [IoT](https://ai.google.dev/edge/litert-lm/cli) | [Web](https://huggingface.co/spaces/tylermullen/Gemma4) |

</div>

Not sure where to start? We recommend first trying our models in our Google AI Edge Gallery app on [Android](https://play.google.com/store/apps/details?id=com.google.ai.edge.gallery&pli=1) and [iOS](https://apps.apple.com/us/app/google-ai-edge-gallery/id6749645337).

## 🌟 Community's Picks of the Week
* **JEV Laya Series** (`System One`): [laya-LiteRT](https://huggingface.co/litert-community/laya-LiteRT) / [Laya-English-LiteRT](https://huggingface.co/litert-community/Laya-English-LiteRT) / [Laya-Multilingual-LiteRT](https://huggingface.co/litert-community/Laya-Multilingual-LiteRT)
  A lightweight, efficient text classification suite optimized for LiteRT, featuring base, English-specific, and multilingual variants for on-device NLP.
* **[PaddleOCR-VL-1.6](https://huggingface.co/litert-community/PaddleOCR-VL-1.6)** (`Image-Text-to-Text`) High-accuracy vision-language model tailored for robust on-device document understanding and OCR tasks.
* **[decider-2b-vision-LiteRT](https://huggingface.co/litert-community/decider-2b-vision-LiteRT)** (`Image-Text-to-Text`) A compact 2B multimodal vision model converted for real-time visual reasoning and image understanding at the edge.
* **[Spark-X2.5-4B](https://huggingface.co/litert-community/Spark-X2.5-4B)** (`Text Generation`) The most popular edge LLM this week, offering strong instruction-following capabilities within a 4B footprint.
* **[Audio8-TTS-Preview-0.6b](https://huggingface.co/litert-community/Audio8-TTS-Preview-0.6b)** (`Text-to-Speech`) A tiny 0.6B parameter TTS model enabling ultra-fast, low-latency voice synthesis on resource-constrained devices.
* **[Nemotron-3-Diarization-LiteRT](https://huggingface.co/litert-community/Nemotron-3-Diarization-LiteRT)** (`Voice Activity Detection`) Optimized edge diarization model for precise multi-speaker segmentation and voice tracking.

# Community Contributions

Are we missing your favorite model? You can convert and run [PyTorch](https://github.com/google-ai-edge/litert-torch), [TensorFlow](https://ai.google.dev/edge/litert/models/convert_tf), or [JAX](https://ai.google.dev/edge/litert/models/convert_jax) models to the classic TFLite format using the LiteRT conversion and optimization tools. Or for LLMs, you can use the [LiteRT Torch Generative API](https://github.com/google-ai-edge/ai-edge-torch/tree/main/ai_edge_torch/generative). When your model is ready, join the LiteRT community org and upload the model here for others to try!

## Submission Guidelines

* **Model Conversion:** Your model must be converted using the LiteRT-Torch or Generative API in `.tflite` or `.litertlm` formats.
* **Location:** All contributions must be made in the [community contribution folder](https://huggingface.co/collections/litert-community/community-contribution) (e.g., `LiteRT-Community/community-contribution/mobile-bert`).
* **Licensing:** If your model has a license, it must be either an Apache 2.0 or MIT open-source license. Models with no license are also accepted.
* **Data Provenance and Privacy:** If applicable, you must include a summary of the model's training data in your submission. Additionally, you must provide confirmation that all Personally Identifiable Information (PII) has been scrubbed from the data.

## Policies & Support

* **Content Policy:** All uploaded models must fully comply with the [Hugging Face content guidelines](https://huggingface.co/collections/litert-community/community-contribution).
* **Community Moderation:** The Google AI Edge team will not actively moderate community contributions to this community, except for reactive removal in clear cases of spam or abuse.
* **Conversion Requests:** In the event that you cannot successfully convert a `.tflite` or `.litertlm` model, you may submit a PR requesting conversion support from the LiteRT team.