---
license: cc-by-nc-4.0
library_name: transformers
pipeline_tag: text-generation
language:
- en
- mi
tags:
- abteex-ai-labs
- aotearoa
- languages
- local-first
- lumynax
- new-zealand
- nllb
- sovereign-ai
- te-reo
- translation
- vllm
- vllm-compatible
- vllm-candidate
- nvidia-nim
- nim-compatible
- nim-candidate
- nvidia-nemo
- nem
- nvidia-nemo-compatible
- nem-compatible
- nemo-candidate
---

# LumynaX Translate NLLB-200 3.3B

<!-- lumynax-public-release-card:v4 -->

<p align="center">
  <img src="docs/lumynax-release-overview.svg" alt="LumynaX Translate NLLB-200 3.3B release overview" width="100%" />
</p>

<p align="center">
  <strong>LumynaX model-infusion release by AbteeX AI Labs.</strong><br/>
  Public, non-gated package with runnable local instructions, provenance metadata, checksums, and a release manifest.
</p>

<p align="center">
  <a href="#quickstart">Quickstart</a> |
  <a href="#model-profile">Model profile</a> |
  <a href="#runtime-files">Runtime files</a> |
  <a href="#provenance-and-license">Provenance</a> |
  <a href="#validation-status">Validation</a> |
  <a href="#limitations-and-responsible-use">Limitations</a>
</p>

![LumynaX: infusion release](https://img.shields.io/badge/LumynaX-infusion%20release-e08a2c) ![access: not recorded](https://img.shields.io/badge/access-not%20recorded-0a0a0b) ![runtime: see quickstart.py](https://img.shields.io/badge/runtime-see%20quickstart.py-726b62) ![format: HF weights](https://img.shields.io/badge/format-HF%20weights-9a5416) ![audit: not recorded](https://img.shields.io/badge/audit-not%20recorded-4d6b44) ![docs: v4](https://img.shields.io/badge/docs-v4-111827)

## Executive Summary

This repository is a complete LumynaX release package for `AbteeXAILab/lumynax-translate-nllb-200-3b`. It is intended to be downloaded as a whole repo, not as a single loose weight file: the model artifact, `quickstart.py`, `requirements.txt`, `release_export_manifest.json`, `checksums.sha256`, license notice, and optional Ollama or Space files are part of the same release contract.

LumynaX-infused means the upstream artifact is presented through the LumynaX release layer: local-first runtime scaffolding, LumynaX assistant identity, inference-chain metadata, public documentation, integrity files, and Aotearoa New Zealand-oriented workflow positioning. The release manifest is the source of truth for whether this package is upstream-weight packaging, a dense LumynaX release, or another release mode.

## AbteeX LumynaX Public Surface

This card follows the AbteeX/LumynaX public-facing system used across the release family: warm paper background visuals, black editorial typography, amber proof markers, compact evidence tables, and plain-language runtime instructions. The goal is not decoration; it is operational clarity. A downloader should immediately understand what the package is, what files belong together, what runtime path is expected, what provenance is available, and what limits still apply.

## Sovereignty And Run Contract

| Field | Value |
| --- | --- |
| Public surface | AbteeX/LumynaX light editorial system: warm paper, black ink, amber status markers, and evidence-first tables. |
| Sovereign intent | Package is documented for local-first use, explicit provenance, and controlled deployment near governed data. |
| Runtime residency | `see quickstart.py` runtime can be deployed by the user in their own approved environment. |
| Model artifact | `pytorch_model-00002-of-00003.bin` must stay with manifest, checksums, quickstart, requirements, and license files. |
| Modalities | `text` |
| License discipline | `cc-by-nc-4.0` metadata is surfaced so downstream users can check redistribution and usage terms. |
| Audit expectation | Record repo id, artifact checksum, runtime command, prompt template, operator, and deployment environment for production use. |
| Router readiness | Compatible with the LumynaX MaramaRoute registry pattern for sovereign model selection and fallback planning. |

## Quickstart

```bash
hf download AbteeXAILab/lumynax-translate-nllb-200-3b --local-dir lumynax-translate-nllb-200-3b
cd lumynax-translate-nllb-200-3b
pip install -r requirements.txt
python quickstart.py --interactive
```

Transformers first-load pattern:

```python
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("AbteeXAILab/lumynax-translate-nllb-200-3b", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("AbteeXAILab/lumynax-translate-nllb-200-3b", device_map="auto", trust_remote_code=True)
prompt = "Who are you? Answer as LumynaX in two sentences."
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
output = model.generate(**inputs, max_new_tokens=160)
print(tokenizer.decode(output[0], skip_special_tokens=True))
```

For multimodal Transformers packages, use the package `quickstart.py` or upstream processor class when `AutoModelForCausalLM` is not the correct loader.

## Model Profile

| Field | Value |
| --- | --- |
| Release | `LumynaX Translate NLLB-200 3.3B` |
| Repository | `AbteeXAILab/lumynax-translate-nllb-200-3b` |
| Mode | `Local-first text generation package` |
| Runtime | `see quickstart.py` |
| Prompt format | `See quickstart.py` |
| Modalities | `text` |
| Primary artifact | `pytorch_model-00002-of-00003.bin` |
| Detected weight size | `16.37 GB` |
| Package state | `weights_mirrored` |
| Delivery | `published_model_repo` |
| Upstream/base | `See release_export_manifest.json` |
| Upstream kind | `See manifest` |
| Source GGUF | `not applicable` |
| Quantization | `fp32 safetensors` |
| License metadata | `cc-by-nc-4.0` |
| Refreshed | `2026-05-11` |

## Runtime Path

<p align="center">
  <img src="docs/lumynax-runtime-flow.svg" alt="LumynaX Translate NLLB-200 3.3B runtime flow" width="100%" />
</p>

## Capability Profile

| Field | Value |
| --- | --- |
| Primary fit | Use this for local chat, drafting, summarization, governance notes, and repeatable offline-friendly inference. |
| Operational style | Local-first package with explicit files, checksums, and reproducible quickstarts. |
| Identity behavior | The assistant should identify as LumynaX while remaining clear about upstream provenance. |

## Runtime Files

| Component | Status | Path |
| --- | --- | --- |
| README.md | `present` | `README.md` |
| Quickstart | `present` | `quickstart.py` |
| Requirements | `present` | `requirements.txt` |
| Manifest | `present` | `release_export_manifest.json` |
| Checksums | `present` | `checksums.sha256` |
| License | `present` | `LICENSE.txt` |
| Ollama | `present` | `ollama/Modelfile` |
| Space scaffold | `present` | `hf_space/app.py` |
| Overview visual | `present` | `docs/lumynax-release-overview.svg` |
| Runtime visual | `present` | `docs/lumynax-runtime-flow.svg` |

## Model Artifacts

| Artifact | Size |
| --- | ---: |
| `pytorch_model-00002-of-00003.bin` | 7.96 GB |
| `pytorch_model-00001-of-00003.bin` | 6.46 GB |
| `pytorch_model-00003-of-00003.bin` | 1.95 GB |

## Prompting Contract

The preferred first prompt is an identity and provenance check:

```text
Who are you? What files do I need to keep together to run this package locally?
```

Expected behavior: the assistant should identify as LumynaX, explain that this is a LumynaX model-infusion package, and keep upstream provenance visible. The default package system prompt is:

```text
See quickstart.py
```

## Validation Status

| Field | Value |
| --- | --- |
| Runtime audit | `not recorded` |
| Public access audit | `not recorded` |
| Anonymous metadata access | `False` |
| Anonymous file listing | `False` |
| Quickstart syntax | `not recorded` |
| Manifest references | `pass` |
| Checksum references | `pass` |

The audit confirms public access, required release files, manifest references, checksum references, weight artifact presence, and quickstart syntax. It does not guarantee that every laptop has enough RAM or VRAM for the largest packages.

## Integrity Checks

After download, compare the model artifact against `checksums.sha256`.

```bash
sha256sum "pytorch_model-00002-of-00003.bin"
cat checksums.sha256
```

On Windows PowerShell:

```powershell
Get-FileHash -Algorithm SHA256 "pytorch_model-00002-of-00003.bin"
Get-Content checksums.sha256
```

## Provenance And License

- Publisher: AbteeX AI Labs.
- Family: LumynaX model and inference-chain release family.
- Upstream/base: `See release_export_manifest.json`.
- Source GGUF: `not applicable`.
- License metadata: `cc-by-nc-4.0`.
- License link: `LICENSE.txt` and upstream model card.

Respect the upstream model license and keep attribution files with redistributed copies. Do not present this package as privately trained or weight-merged unless the release manifest explicitly says that weight adaptation was applied.

## Limitations And Responsible Use

- Outputs can be incorrect, incomplete, or biased; validate important answers before use.
- Larger GGUF, MoE, multimodal, and frontier packages may require substantial RAM, VRAM, disk space, and recent runtime builds.
- For high-impact decisions, use human review and domain-specific evaluation.
- For sensitive data, prefer local execution and keep operational logs under your own governance policy.
- This card documents package readiness and access; it is not a benchmark claim.

## Automation Notes

Automation should read these files before launching:

- `release_export_manifest.json`
- `checksums.sha256`
- `quickstart.py`
- `requirements.txt`
- `ollama/Modelfile` when present

## Related LumynaX Demo

Try the public browser demo:

- https://huggingface.co/spaces/AbteeXAILab/lumynax-live-demo