---
language:
  - en
  - zh
  - ko
  - ja
license: apache-2.0
base_model: Jackrong/Gemopus-4-E4B-it
library_name: mlx
tags:
  - mlx
  - gemma
  - gemma4
  - edge-ai
  - instruction-tuned
  - reasoning
  - privacy
  - human-preference-alignment
pipeline_tag: image-text-to-text
---
# 🦆 zecanard/Gemopus-4-E4B-it-MLX-8bit-int8-affine

[This model](https://huggingface.co/zecanard/Gemopus-4-E4B-it-MLX-8bit-int8-affine) was converted to MLX from [`Jackrong/Gemopus-4-E4B-it`](https://huggingface.co/Jackrong/Gemopus-4-E4B-it) using `mlx-vlm` version **0.6.3**.
Please refer to the [original model card](https://huggingface.co/Jackrong/Gemopus-4-E4B-it) for more details.

## 🌟 Quality

Quantized vision language model with an effective **9.446 bits per weight**.

`mlx_vlm.convert --quantize --q-group-size 32 --q-bits 8 --q-mode affine`

## 🛠️ Customizations

This quant includes a bugfix related to tools calling. It is aware of the current date, and also enables thinking (if available). You may disable this behavior by deleting the following line from the chat template, or changing `true` to `false`:

`{%- set enable_thinking = true %}`

You may need to adjust your environment’s **Reasoning Section Parsing** to recognize `<|channel>thought` as the **Start String**, and `<channel|>` as the **End String**.

## 🖥️ Use with `mlx`

```bash
pip install -U mlx-vlm
```

```bash
mlx_vlm.generate --model zecanard/Gemopus-4-E4B-it-MLX-8bit-int8-affine --max-tokens 100 --temperature 0 --prompt "Describe this image." --image <path_to_image>
```