Morteza89/gemma-4-26b-a4b-it-int4-ov

Gemma 4 26B A4B IT OpenVINO INT4

This repository contains an unofficial OpenVINO™ IR conversion of google/gemma-4-26B-A4B-it with INT4 weight compression.

It is intended for local OpenVINO inference using Optimum Intel, and supports vision-language generation with an image and text prompt.

---

Validate locally

From the parent folder:

cd ~/ov_models

python ov_gemma_4_26B_A4B_it_INT4/examples/run_gemma4_it_vlm_openvino.py \
  --model-dir ov_gemma_4_26B_A4B_it_INT4 \
  --device GPU \
  --max-new-tokens 128 \
  --prompt "Describe this image in one paragraph."

## Original model

- Original model: `google/gemma-4-26B-A4B-it`
- Model family: Gemma 4
- Converted format: OpenVINO IR
- Weight format: INT4
- Task: image-text-to-text / vision-language generation

This is an unofficial converted repository. Please refer to the original model card for full license terms, usage restrictions, intended use, safety notes, and citation guidance.

## Conversion command

This model was converted with Optimum Intel using OpenVINO export:

```bash
optimum-cli export openvino \
  --model google/gemma-4-26B-A4B-it \
  --weight-format int4 \
  --trust-remote-code \
  ov_gemma_4_26B_A4B_it_INT4