---
title: Gemma-3-Glitter-12B-8bit
canonical_url: "https://www.modelscope.cn/models/mlx-community/Gemma-3-Glitter-12B-8bit"
md_url: "https://www.modelscope.cn/models/mlx-community/Gemma-3-Glitter-12B-8bit.md"
repository: mlx-community/Gemma-3-Glitter-12B-8bit
last_updated: 2025-10-09
pipeline_tag: image-text-to-text
tasks:
  - image-text-to-text
model_type:
  - gemma3
architectures:
  - Gemma3ForConditionalGeneration
base_model:
  - allura-org/Gemma-3-Glitter-12B
base_model_relation: quantized
parameters: 3.3B
tensor_type:
  - U32
  - BF16
library_name:
  - mlx
  - safetensors
  - pytorch
frameworks:
  - pytorch
inference_backends:
  - "deploy_task vlm/text/emb"
  - "sglang 0.5.2"
  - "vllm 0.9.2"
downloads: 68
stars: 0
tags:
  - mergekit
  - merge
  - mlx
  - mlx-my-repo
---

# Gemma-3-Glitter-12B-8bit

> Gemma-3-Glitter-12B-8bit - mlx-community 在 ModelScope 开源的模型。mlx-community/Gemma-3-Glitter-12B-8bit

mlx-community/Gemma-3-Glitter-12B-8bit 是 ModelScope 魔搭社区上的 3.3B 参数image-text-to-text模型，基于 allura-org/Gemma-3-Glitter-12B 构建，可用 deploy_task vlm/text/emb、sglang 0.5.2、vllm 0.9.2 部署。

- **Repository**: mlx-community/Gemma-3-Glitter-12B-8bit
- **Tasks**: image-text-to-text
- **Parameters**: 3.3B
- **Base model**: allura-org/Gemma-3-Glitter-12B
- **Inference backends**: deploy_task vlm/text/emb, sglang 0.5.2, vllm 0.9.2
- **Tags**: mergekit, merge, mlx, mlx-my-repo
- **Downloads**: 68
- **Stars**: 0
- **Last updated**: 2025-10-09

Source: https://www.modelscope.cn/models/mlx-community/Gemma-3-Glitter-12B-8bit

---

# mlx-community/Gemma-3-Glitter-12B-8bit

The Model [mlx-community/Gemma-3-Glitter-12B-8bit](https://huggingface.co/mlx-community/Gemma-3-Glitter-12B-8bit) was converted to MLX format from [allura-org/Gemma-3-Glitter-12B](https://huggingface.co/allura-org/Gemma-3-Glitter-12B) using mlx-lm version **0.26.4**.

## Use with mlx

```bash
pip install mlx-lm
```

```python
from mlx_lm import load, generate

model, tokenizer = load("mlx-community/Gemma-3-Glitter-12B-8bit")

prompt="hello"

if hasattr(tokenizer, "apply_chat_template") and tokenizer.chat_template is not None:
    messages = [{"role": "user", "content": prompt}]
    prompt = tokenizer.apply_chat_template(
        messages, tokenize=False, add_generation_prompt=True
    )

response = generate(model, tokenizer, prompt=prompt, verbose=True)
```