---
license: apache-2.0
language:
- pt
base_model:
- amalia-llm/AMALIA-9B-0626-SFT
base_model_relation: quantized
pipeline_tag: text-generation
library_name: mlx
tags:
- mlx
- quantized
- portuguese
- european-portuguese
- amalia
---

# AMALIA-9B-0626-SFT — MLX 8-bit

8-bit [MLX](https://github.com/ml-explore/mlx) quantization of [amalia-llm/AMALIA-9B-0626-SFT](https://huggingface.co/amalia-llm/AMALIA-9B-0626-SFT) (~9.7 GB), for Apple Silicon Macs via [mlx-lm](https://github.com/ml-explore/mlx-lm).

AMALIA is an open-source language model targeting European Portuguese, developed by a consortium of Portuguese universities and research centres and funded by the Government of Portugal. This is the SFT (supervised fine-tuned, instruction-following) variant. See the [original model card](https://huggingface.co/amalia-llm/AMALIA-9B-0626-SFT) and the [technical report](https://arxiv.org/abs/2603.26511) for full details.

Converted with `mlx_lm.convert` (8-bit, group size 64).

## Usage

```bash
pip install mlx-lm
mlx_lm.generate --model layerx-labs/AMALIA-9B-0626-SFT-MLX-8bit --prompt "Explica o que é o Arquivo.pt."

# or chat
mlx_lm.chat --model layerx-labs/AMALIA-9B-0626-SFT-MLX-8bit
```

```python
from mlx_lm import load, generate
model, tokenizer = load("layerx-labs/AMALIA-9B-0626-SFT-MLX-8bit")
prompt = tokenizer.apply_chat_template([{"role": "user", "content": "Explica o que é o Arquivo.pt."}], add_generation_prompt=True, tokenize=False)
print(generate(model, tokenizer, prompt=prompt, max_tokens=256))
```

## License

Apache 2.0, same as the original model. All credit for the model goes to the [AMALIA project](https://amaliallm.pt/).