Gemma 4 E2B (LiteRT-LM mirror)

This repository is a redistribution mirror of litert-community/gemma-4-E2B-it-litert-lm, hosted to provide a stable, controlled URL for downstream apps that ship on-device inference.

The model weights, tokenizer and chat template are distributed unchanged from the upstream repository. No retraining, fine-tuning or quantisation has been performed in this mirror.

Files

gemma-4-E2B-it.litertlm — the LiteRT-LM packaged model file (≈2.6 GB).

The SHA-256 of the .litertlm file matches the upstream object:

ab7838cdfc8f77e54d8ca45eadceb20452d9f01e4bfade03e5dce27911b27e42

Apps that download this file should verify the hash before loading.

License

Released under the same license declared by the upstream repository: Apache License 2.0 — see LICENSE.

The underlying Gemma model weights are originally produced by Google and were released under the Gemma Terms of Use and the Gemma Prohibited Use Policy. Users of this mirror are bound by those policies in addition to the Apache 2.0 license. See the NOTICE file for full attribution and links.

Attribution

Original model authors: Google (Gemma team)
LiteRT-LM packaging: the litert-community organisation on Hugging Face
Mirror maintainer: this repository owner

This mirror exists for redistribution stability only; all credit for the underlying research and packaging belongs to Google and the LiteRT-LM community. Please cite the upstream sources in any work that uses this model.

Downloads last month: 39

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for multimodel/gemma-4-E2B-it-litert-lm

Base model

google/gemma-3-4b-pt

Finetuned

google/gemma-3-4b-it

Finetuned

(772)

this model