---
title: Apertus-SEA-LION-v4-8B-IT
canonical_url: "https://www.modelscope.cn/models/mlx-community/Apertus-SEA-LION-v4-8B-IT"
md_url: "https://www.modelscope.cn/models/mlx-community/Apertus-SEA-LION-v4-8B-IT.md"
repository: mlx-community/Apertus-SEA-LION-v4-8B-IT
last_updated: 2026-02-19
license: apache-2.0
pipeline_tag: text-generation
tasks:
  - text-generation
model_type:
  - apertus
architectures:
  - ApertusForCausalLM
base_model:
  - aisingapore/Apertus-SEA-LION-v4-8B-IT
base_model_relation: finetune
parameters: 5.8B
tensor_type:
  - BF16
library_name:
  - mlx
  - safetensors
  - pytorch
frameworks:
  - pytorch
language:
  - en
  - zh
  - vi
  - id
  - th
  - fil
  - ta
  - ms
  - my
downloads: 12
stars: 0
tags:
  - mlx
---

# Apertus-SEA-LION-v4-8B-IT

> Apertus-SEA-LION-v4-8B-IT - mlx-community 在 ModelScope 开源的模型。mlx-community/Apertus-SEA-LION-v4-8B-IT

mlx-community/Apertus-SEA-LION-v4-8B-IT 是 ModelScope 魔搭社区上的 5.8B 参数text-generation模型，采用 apache-2.0 许可，基于 aisingapore/Apertus-SEA-LION-v4-8B-IT 构建。

- **Repository**: mlx-community/Apertus-SEA-LION-v4-8B-IT
- **License**: apache-2.0
- **Tasks**: text-generation
- **Parameters**: 5.8B
- **Base model**: aisingapore/Apertus-SEA-LION-v4-8B-IT
- **Tags**: mlx
- **Downloads**: 12
- **Stars**: 0
- **Last updated**: 2026-02-19

Source: https://www.modelscope.cn/models/mlx-community/Apertus-SEA-LION-v4-8B-IT

---

# mlx-community/Apertus-SEA-LION-v4-8B-IT

This model [mlx-community/Apertus-SEA-LION-v4-8B-IT](https://huggingface.co/mlx-community/Apertus-SEA-LION-v4-8B-IT) was
converted to MLX format from [aisingapore/Apertus-SEA-LION-v4-8B-IT](https://huggingface.co/aisingapore/Apertus-SEA-LION-v4-8B-IT)
using mlx-lm version **0.30.6**.

## Use with mlx

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

```python
from mlx_lm import load, generate

model, tokenizer = load("mlx-community/Apertus-SEA-LION-v4-8B-IT")

prompt = "hello"

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

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