---
license: apache-2.0
base_model:
  - google/gemma-4-E4B-it
  - google/gemma-4-E2B-it
tags:
  - litert-lm
  - on-device
  - gemma4
---

# Gemma 4 (LiteRT-LM): Memoyad mirror

These are unmodified copies of Gemma 4 LiteRT-LM model files. They are hosted here so the
[Memoyad](https://memoyad.com) app always has a stable download source. The app picks one model
per device. Phones that can run the larger E4B model use it; slower phones use the lighter E2B.

| File | Source | Size (bytes) | SHA-256 |
|---|---|---|---|
| `gemma-4-E4B-it.litertlm` | [litert-community/gemma-4-E4B-it-litert-lm](https://huggingface.co/litert-community/gemma-4-E4B-it-litert-lm) @ `2eee7ac325f20eb8c9ac1d0e972f7c84663062da` | 3659530240 | `0b2a8980ce155fd97673d8e820b4d29d9c7d99b8fa6806f425d969b145bd52e0` |
| `gemma-4-E2B-it.litertlm` | [litert-community/gemma-4-E2B-it-litert-lm](https://huggingface.co/litert-community/gemma-4-E2B-it-litert-lm) @ `b3ca0d2f076785a8f4b2219ddbd2bdb99954eae1` | 2588147712 | `181938105e0eefd105961417e8da75903eacda102c4fce9ce90f50b97139a63c` |

No weights, graphs or tokenizer data were changed. The app pins each download to a commit of this
repository and checks the SHA-256 above. Adding files here never changes what an installed app
downloads.

## License

Licensed under the Apache License, Version 2.0; see [`LICENSE`](https://huggingface.co/tibaf/gemma/blob/main/LICENSE).
The files are redistributed without modification.

Gemma 4 was developed by Google DeepMind. The LiteRT community (Google AI Edge) converted it to
LiteRT-LM. This repository is not affiliated with or endorsed by Google.