docs: publish professional LumynaX model card

README.md

@@ -15,124 +15,96 @@ tags:
 - abteex-ai-labs
 - local-first
 ---
+
 # LumynaX Infused SmolLM2 360M Instruct GGUF
 
-LumynaX Infused SmolLM2 360M Instruct GGUF is a Hugging Face-ready LumynaX GGUF release package. It wraps a
-prebuilt GGUF artifact for local `llama.cpp`, `llama-cpp-python`, Ollama, LM
-Studio, Jan, or compatible GGUF runtimes.
+<!-- lumynax-public-release-card:v4 -->
+
+<p align="center">
+  <img src="docs/lumynax-release-overview.svg" alt="LumynaX Infused SmolLM2 360M Instruct GGUF 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>
 
-## Provenance
+![LumynaX: infusion release](https://img.shields.io/badge/LumynaX-infusion%20release-16a34a) ![access: public and non-gated](https://img.shields.io/badge/access-public%20and%20non--gated-0ea5e9) ![runtime: llama cpp](https://img.shields.io/badge/runtime-llama%20cpp-2563eb) ![format: GGUF](https://img.shields.io/badge/format-GGUF-7c3aed) ![audit: pass](https://img.shields.io/badge/audit-pass-ea580c) ![docs: v4](https://img.shields.io/badge/docs-v4-111827)
 
-- upstream base model: `HuggingFaceTB/SmolLM2-360M-Instruct`
-- source GGUF repo: `HuggingFaceTB/SmolLM2-360M-Instruct-GGUF`
-- packaged GGUF file: `smollm2-360m-instruct-q8_0.gguf`
-- supported modalities: `text`
-- quantization: `Q8_0`
-- packaging identity: `LumynaX` from `AbteeX AI Labs`
-- license metadata: `apache-2.0`
-- weight claim: this release packages the referenced GGUF artifact and does not claim a private weight merge
+## Executive Summary
 
-## Quick Start
+This repository is a complete LumynaX release package for `AbteeXAILab/lumynax-infused-smollm2-360m-gguf`. 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 records this as a LumynaX packaging and inference-chain layer around the listed upstream artifact; it does not claim a private LumynaX weight merge.
+
+## Quickstart
 
 ```bash
+hf download AbteeXAILab/lumynax-infused-smollm2-360m-gguf --local-dir lumynax-infused-smollm2-360m-gguf
+cd lumynax-infused-smollm2-360m-gguf
 pip install -r requirements.txt
 python quickstart.py --interactive
-python quickstart.py --prompt "Say hello in one short sentence."
 ```
 
-To use the bundled llama.cpp fallback directly:
+Direct llama.cpp smoke command:
 
 ```bash
-python quickstart.py --llama-cli C:\path\to\llama-cli.exe --prompt "Say hello."
+llama-cli -m "smollm2-360m-instruct-q8_0.gguf" -p "Who are you? Answer as LumynaX in two sentences." -n 160
 ```
 
-
-## Ollama
+Ollama path:
 
 ```bash
-cd ollama
-powershell -NoProfile -ExecutionPolicy Bypass -File ./create_ollama_model.ps1
+ollama create lumynax-infused-smollm2-360m-gguf -f ollama/Modelfile
 ollama run lumynax-infused-smollm2-360m-gguf
 ```
 
-## Included Files
-
-- `smollm2-360m-instruct-q8_0.gguf`: primary GGUF model artifact
-- `quickstart.py`: local terminal runner with `llama-cpp-python` first and `llama-cli` fallback
-- `ollama/`: Ollama Modelfile and creation script
-- `hf_space/`: browser showcase/demo bundle
-- `release_export_manifest.json`: package metadata and runtime defaults
-- `checksums.sha256`: release integrity manifest
-
-## Publisher
-
-- organization / lab: `AbteeX AI Labs`
-- website: `https://abteex.com`
-- Hugging Face owner account: `AbteeXAILab`
-- recommended model repo: `AbteeXAILab/lumynax-infused-smollm2-360m-gguf`
-- recommended local model name: `lumynax-infused-smollm2-360m-gguf`
-
-<!-- lumynax-public-release-block:v3 -->
-
-## Complete LumynaX Release Card
-
-![LumynaX: infused](https://img.shields.io/badge/LumynaX-infused-22c55e) ![runtime: llama cpp](https://img.shields.io/badge/runtime-llama%20cpp-1f6feb) ![format: GGUF](https://img.shields.io/badge/format-GGUF-7c3aed) ![package: prebuilt gguf release](https://img.shields.io/badge/package-prebuilt%20gguf%20release-0f766e) ![docs: v3](https://img.shields.io/badge/docs-v3-f97316)
-
-This section is the expanded public-facing package card for this LumynaX release. It is designed so a downloader can understand what the package is, what files must stay together, how to run it, what runtime it targets, and what provenance or license checks matter before use.
-
-## Release At A Glance
+## Model Profile
 
 | Field | Value |
 | --- | --- |
-| HF repo | `AbteeXAILab/lumynax-infused-smollm2-360m-gguf` |
-| Public status | `public` |
-| Release type | `Local text generation` |
-| Delivery | `standalone_prebuilt_gguf_release` |
-| Package state | `prebuilt_gguf_release` |
-| Runtime backend | `llama_cpp` |
+| Release | `LumynaX Infused SmolLM2 360M Instruct GGUF` |
+| Repository | `AbteeXAILab/lumynax-infused-smollm2-360m-gguf` |
+| Mode | `Local-first text generation package` |
+| Runtime | `llama_cpp` |
 | Prompt format | `chatml` |
 | Modalities | `text` |
 | Primary artifact | `smollm2-360m-instruct-q8_0.gguf` |
-| Total detected weight size | `368.50 MB` |
+| Detected weight size | `368.50 MB` |
+| Package state | `prebuilt_gguf_release` |
+| Delivery | `standalone_prebuilt_gguf_release` |
 | Upstream/base | `HuggingFaceTB/SmolLM2-360M-Instruct` |
+| Upstream kind | `official_base_weights` |
 | Source GGUF | `HuggingFaceTB/SmolLM2-360M-Instruct-GGUF` |
 | Quantization | `Q8_0` |
 | License metadata | `apache-2.0` |
-| Last refreshed | `2026-05-11` |
-
-## Visual Release Map
+| Refreshed | `2026-05-11` |
+| Upstream license link | [license](https://huggingface.co/HuggingFaceTB/SmolLM2-360M-Instruct-GGUF) |
 
-![LumynaX release map](docs/lumynax-release-map.svg)
+## Runtime Path
 
-## What "LumynaX-Infused" Means Here
+<p align="center">
+  <img src="docs/lumynax-runtime-flow.svg" alt="LumynaX Infused SmolLM2 360M Instruct GGUF runtime flow" width="100%" />
+</p>
 
-This package belongs to the LumynaX model and inference-chain family from AbteeX AI Labs. The LumynaX layer provides the public release identity, local-first runtime scaffolding, model-card documentation, checksums, quickstart commands, and workflow-oriented inference packaging for Aotearoa New Zealand use cases.
+## Capability Profile
 
-Manifest indicates no private LumynaX weight merge was applied; this is a public release/package layer around the listed upstream artifact.
-
-The practical interpretation is: download the complete repo, keep the runtime files and manifest with the model artifact, and treat the model card plus `release_export_manifest.json` as the source of truth for provenance.
-
-## What This Package Is Best For
-
-Use this for fast smoke tests, demos, packaging validation, and low-resource local runs.
-
-## Download And Run
-
-```bash
-git lfs install
-git clone https://huggingface.co/AbteeXAILab/lumynax-infused-smollm2-360m-gguf
-cd lumynax-infused-smollm2-360m-gguf
-pip install -r requirements.txt
-python quickstart.py --interactive
-```
-
-Direct llama.cpp-style smoke prompt:
-
-```bash
-python quickstart.py --prompt "Who are you? Answer as LumynaX in one sentence."
-```
+| Field | Value |
+| --- | --- |
+| Primary fit | Use this for fast smoke tests, demos, packaging validation, and low-resource local runs. |
+| 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. |
 
-## Required Runtime Components
+## Runtime Files
 
 | Component | Status | Path |
 | --- | --- | --- |
@@ -144,56 +116,90 @@ python quickstart.py --prompt "Who are you? Answer as LumynaX in one sentence."
 | 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
 
-| File | Size |
+| Artifact | Size |
 | --- | ---: |
 | `smollm2-360m-instruct-q8_0.gguf` | 368.50 MB |
 
-## Integrity And Validation
+## 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:
 
-- `checksums.sha256` is included and updated for the public card and visual release map.
-- `release_export_manifest.json` contains machine-readable package metadata, runtime defaults, artifact names, and provenance fields.
-- `quickstart.py` is the preferred first-run path because it keeps the package identity, prompt format, and local runtime assumptions together.
-- Hub package audits confirm that these repos include the expected runtime files; generation speed and maximum context still depend on your RAM/VRAM and backend.
+```text
+You are LumynaX operating from the LumynaX Infused SmolLM2 360M Instruct GGUF package identity. Be helpful, clear, and honest about provenance.
+```
 
-## Hardware Guidance
+## Validation Status
 
-Small package. Usually appropriate for CPU smoke tests and lightweight demos.
+| Field | Value |
+| --- | --- |
+| Runtime audit | `pass` |
+| Public access audit | `public and non-gated` |
+| Anonymous metadata access | `True` |
+| Anonymous file listing | `True` |
+| Quickstart syntax | `pass` |
+| Manifest references | `pass` |
+| Checksum references | `pass` |
 
-## Prompting And Identity
+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.
 
-Ask identity/provenance questions directly:
+## Integrity Checks
 
-```text
-Who are you?
-What model package is this?
-What is the upstream model and license?
-What files do I need to keep together to run this locally?
+After download, compare the model artifact against `checksums.sha256`.
+
+```bash
+sha256sum "smollm2-360m-instruct-q8_0.gguf"
+cat checksums.sha256
+```
+
+On Windows PowerShell:
+
+```powershell
+Get-FileHash -Algorithm SHA256 "smollm2-360m-instruct-q8_0.gguf"
+Get-Content checksums.sha256
 ```
 
-The LumynaX identity should be presented at runtime, while provenance remains explicit. Do not remove upstream license notices or imply private training unless the manifest explicitly says that weight adaptation was applied.
+## Provenance And License
+
+- Publisher: AbteeX AI Labs.
+- Family: LumynaX model and inference-chain release family.
+- Upstream/base: `HuggingFaceTB/SmolLM2-360M-Instruct`.
+- Source GGUF: `HuggingFaceTB/SmolLM2-360M-Instruct-GGUF`.
+- License metadata: `apache-2.0`.
+- License link: https://huggingface.co/HuggingFaceTB/SmolLM2-360M-Instruct-GGUF.
 
-## License And Use Notes
+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.
 
-- Check `LICENSE.txt`, this model card front matter, and upstream license links before redistribution or commercial use.
-- Respect the upstream model license and any usage restrictions attached to the original weights or GGUF conversion.
-- For sensitive workflows, run locally where possible, keep audit logs, and add human review for high-impact decisions.
+## Limitations And Responsible Use
 
-## Support Files For Automation
+- 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 should prefer these files:
+## Automation Notes
 
-- `release_export_manifest.json` for structured metadata.
-- `checksums.sha256` for file integrity.
-- `quickstart.py` for local smoke tests.
-- `requirements.txt` for Python dependencies.
-- `ollama/Modelfile` where an Ollama path is included.
+Automation should read these files before launching:
 
-## Related Demo
+- `release_export_manifest.json`
+- `checksums.sha256`
+- `quickstart.py`
+- `requirements.txt`
+- `ollama/Modelfile` when present
 
-A lightweight public LumynaX demo is available at:
+## Related LumynaX Demo
 
-- `https://huggingface.co/spaces/AbteeXAILab/lumynax-live-demo`
+Try the public browser demo:
 
+- https://huggingface.co/spaces/AbteeXAILab/lumynax-live-demo

checksums.sha256

@@ -9,10 +9,12 @@ c46f2b285b76fa0a180e3945a11fe035f318a73168b565973771474fb1b20494  merged_model/P
 5e05007a23fd50cf67d323643af03c20ecf8aacd8728ebb159cfc9163a98e21c  ollama/create_ollama_model.ps1
 84f079674b6370ff673f3308c9759d17e88c94b276d31e470b826357fcf4cd6c  ollama/Modelfile
 1be0702ab536c4e402638faa5fcaffc368a245bf42930f4ece357299eee1e3cd  quickstart.py
-b502512a1b8d215813c8e3e9818d01ab01d72a6513099b99674dbc96b12e7948  README.md
+ea78dfcc83e878e13f520f3c88902b3b8d7b5813d869f71c3c70c59e1cfe1cc3  README.md
 006cd08485755e0e71386b4a41a1401e8b72b737b7479670fd091976225398b5  release_export_manifest.json
 2a7ca962dd79646b8470b45ec926ade0a4eb01ebdd93452e6990d8997666e378  requirements.txt
 48ab3034d0dd401fbc721eb1df3217902fee7dab9078992d66431f09b7750201  smollm2-360m-instruct-q8_0.gguf
 40211b0cb75f9ef6f3205da77372128ca8abe3dd159fbfcf42fee6b6f70845bc  UPLOAD_TO_HF.md
 2dfede0e6610c473959c963b292fcec325452acba33fd1bba21110e04933df53  VERSION.txt
 c6a7ec3c652a97f205d9cd8f1ddc6c8cfd3f2b61274bf7175dc92f8ab09d8ecc  docs/lumynax-release-map.svg
+54d44454133fa9e35a089675bc741ffbc86987745c67de3e9789165cff665382  docs/lumynax-release-overview.svg
+b13dba2284b76a058560386943506d9926563454971089067397d57f05cb9ad6  docs/lumynax-runtime-flow.svg