---
license: apache-2.0
language:
- pl
base_model:
- speakleash/Bielik-11B-v2.6-Instruct
library_name: mlx
pipeline_tag: text-generation
tags:
- mlx
- veterinary
- medical
- polish
- reasoning
- clinical-ai
- bielik
- fine-tuned
- sophisticated-merge
- model-soup
- evolutionary
- v3
- apple-silicon
- clinical
- quantized
- q8
inference: false
widget:
- text: Wyjaśnij krótko różnicę między diagnostyką różnicową a rozpoznaniem.
  example_title: Polish instruction prompt
- text: Podsumuj najważniejsze ryzyka w planie wdrożenia.
  example_title: Polish reasoning prompt
---

# svetliq-11b-v3-evolutionary-preview-mlx-q8

`svetliq-11b-v3-evolutionary-preview-mlx-q8` is a Polish veterinary clinical checkpoint in MLX format, derived from `speakleash/Bielik-11B-v2.6-Instruct` and packaged for local Apple Silicon inference.

## Intended use

- Polish veterinary clinical drafting and case reasoning for practitioner review
- Differential diagnosis, triage notes, drug-reference style explanations, and care-plan drafts
- Local Apple Silicon inference where data locality and operator control matter

## Out of scope

- Direct-to-owner veterinary diagnosis or treatment decisions
- Languages other than Polish unless independently evaluated
- Safety-critical decisions without domain expert review
- Claims of benchmark superiority not backed by published evaluation data
- Non-MLX runtime guarantees; this card documents the shipped HF checkpoint, not every possible serving stack

## Training and conversion metadata

| Parameter | Value |
|---|---|
| Repository | `LibraxisAI/svetliq-11b-v3-evolutionary-preview-mlx-q8` |
| Base model | `speakleash/Bielik-11B-v2.6-Instruct` |
| Task | `text-generation` |
| Library | `mlx` |
| Format | MLX / Apple Silicon checkpoint |
| Quantization | Q8 |
| Architecture | LlamaForCausalLM |
| Model files | 3 |
| Config model_type | `llama` |

This card only reports metadata present in the Hugging Face repository, existing card frontmatter, or public config files. Missing benchmark, dataset, or training-run details are left explicit rather than reconstructed.

## Usage

### CLI

```bash
pip install mlx-lm

mlx_lm.generate \
  --model LibraxisAI/svetliq-11b-v3-evolutionary-preview-mlx-q8 \
  --prompt "Opisz krótko objawy odwodnienia u psa i kiedy pilnie skontaktować się z lekarzem weterynarii." \
  --max-tokens 400
```

### Python

```python
from mlx_lm import load, generate

model, tokenizer = load("LibraxisAI/svetliq-11b-v3-evolutionary-preview-mlx-q8")

prompt = "Opisz krótko objawy odwodnienia u psa i kiedy pilnie skontaktować się z lekarzem weterynarii."
response = generate(model, tokenizer, prompt=prompt, max_tokens=400)
print(response)
```

### Multi-turn with the chat template

This checkpoint follows the tokenizer/chat-template contract inherited from `speakleash/Bielik-11B-v2.6-Instruct` when the
template is present in the repository:

```python
from mlx_lm import load, generate

model, tokenizer = load("LibraxisAI/svetliq-11b-v3-evolutionary-preview-mlx-q8")

messages = [
    {"role": "user", "content": "Opisz krótko objawy odwodnienia u psa i kiedy pilnie skontaktować się z lekarzem weterynarii."},
]
prompt = tokenizer.apply_chat_template(messages, add_generation_prompt=True, tokenize=False)
response = generate(model, tokenizer, prompt=prompt, max_tokens=400)
print(response)
```

## Example output

No public sample output is currently declared for this checkpoint. Run the usage example above against your own prompt or audio/image input to inspect behavior.

## Comparison with the base model

| Aspect | Base | This checkpoint |
|---|---|---|
| Lineage | `speakleash/Bielik-11B-v2.6-Instruct` | Polish veterinary-domain checkpoint in MLX format |
| Domain emphasis | General instruction behavior from the base family | Veterinary clinical drafting, Polish case reasoning, and practitioner-facing assistance |
| Published benchmark delta | Not declared in public metadata | Not declared in public metadata |

## Limitations

- No public benchmarks for this checkpoint are declared in the model metadata.
- No public benchmark claims are made by this card unless listed in the frontmatter.
- Validate outputs on your own domain data before relying on this checkpoint.
- Memory use and speed depend heavily on the exact Apple Silicon generation, unified-memory size, and prompt length.
- Veterinary outputs require review by a licensed veterinarian.

## License

`apache-2.0`. Check the upstream/base model license as well when a base model is declared.

## Citation

```bibtex
@misc{libraxisai-svetliq-11b-v3-evolutionary-preview-mlx-q8,
  title = {svetliq-11b-v3-evolutionary-preview-mlx-q8},
  author = {LibraxisAI},
  year = {2026},
  howpublished = {\url{https://huggingface.co/LibraxisAI/svetliq-11b-v3-evolutionary-preview-mlx-q8}},
  note = {MLX checkpoint published by LibraxisAI}
}
```

## Inference tested on

[`LibraxisAI/mlx-batch-server`](https://github.com/LibraxisAI/mlx-batch-server)

## Related

- Base model: [`speakleash/Bielik-11B-v2.6-Instruct`](https://huggingface.co/speakleash/Bielik-11B-v2.6-Instruct)

---

𝚅𝚒𝚋𝚎𝚌𝚛𝚊𝚏𝚝𝚎𝚍. with AI Agents by VetCoders (c)2024-2026 LibraxisAI