apertus-electronics-hw-lora

LoRA adapter for swiss-ai/Apertus-70B-Instruct-2509, part of the ailiance project. Live demo: https://www.ailiance.fr.

EU AI Act compliance. This card follows the European Commission's Template for the Public Summary of Training Content for general-purpose AI models (Art. 53(1)(d) of Regulation (EU) 2024/1689, published by the AI Office on 2025-07-24). Section numbering and field labels reproduce the official template. Where this card and the official template differ in wording, the official template wins — see the AI Office page.

---

1. General information

1.1. Provider identification

Field	Value
Provider name and contact details	Ailiance (Saillant Clément) — clemsail on Hugging Face — Issues: https://github.com/ailiance/ailiance/issues
Authorised representative name and contact details	Not applicable — provider is established within the European Union (France).

1.2. Model identification

Field	Value
Versioned model name(s)	Ailiance-fr/apertus-electronics-hw-lora (this LoRA adapter, v0.4.2)
Model dependencies	This is a fine-tune (LoRA, rank 16) of the general-purpose AI model swiss-ai/Apertus-70B-Instruct-2509. Refer to the base-model provider's PST for the underlying training summary.
Date of placement of the model on the Union market	2026-05-06

1.3. Modalities, overall training data size and other characteristics

Field	Value
Modality	☒ Text ☐ Image ☐ Audio ☐ Video ☐ Other
Training data size (text bucket)	☒ Less than 1 billion tokens ☐ 1 billion to 10 trillion tokens ☐ More than 10 trillion tokens
Types of content	Instruction-tuning pairs, technical text, source code, multilingual instruction templates (EU official languages where applicable).
Approximate size in alternative units	≈ 0.05 M tokens (90 PDF-derived training pairs).
Latest date of data acquisition / collection for model training	10/2025 (latest vendor datasheet revision). The model is not continuously trained on new data after this date.
Linguistic characteristics of the overall training data	English (technical content, datasheet language).
Other relevant characteristics / additional comments	LoRA fine-tune (rank 16, alpha 32, dropout 0.05); only attention projections (q_proj, k_proj, v_proj, o_proj) are trained. Per-record _provenance (source, SPDX licence, record_idx, access_date) attached at the system level (see docs/eu-ai-act-transparency.md §4.4). Tokenizer: inherited from the base model.

---

2. List of data sources

2.1. Publicly available datasets

Have you used publicly available datasets to train the model? ☒ Yes ☐ No

Modality(ies) of the content covered: ☒ Text ☐ Image ☐ Video ☐ Audio ☐ Other

List of large publicly available datasets:

Dataset	URL	SPDX licence	Records	Notes
Wikipedia electronics articles (topical filter)	https://dumps.wikimedia.org	CC-BY-SA-3.0	merged	Official Wikipedia bulk dumps, science/electronics topical subset. Merged into the corpus; per-article counts not separately tracked.

2.2. Private non-publicly available datasets obtained from third parties

2.2.1. Datasets commercially licensed by rightsholders or their representatives

Have you concluded transactional commercial licensing agreement(s) with rightsholder(s) or with their representatives? ☐ Yes ☒ No

(N/A — no commercial licensing agreements concluded.)

2.2.2. Private datasets obtained from other third parties

Have you obtained private datasets from third parties that are not licensed as described in Section 2.2.1? ☒ Yes ☐ No

Modality(ies) of the content covered: ☒ Text ☐ Image ☐ Video ☐ Audio ☐ Other

Identifiers / names of main private datasets from third parties:

Source	URL	Licence	Notes
ST datasheets (selection)	https://www.st.com	ST-SLA0048 research/educational use	Vendor-licensed datasheets used under DSM Art. 4 TDM exception; SHA-256 manifest published.
Espressif technical docs	https://docs.espressif.com	vendor-permissive	Vendor docs processed under DSM Art. 4 TDM exception.
TI / NXP / KiCad PDF	vendor websites (TI, NXP, KiCad documentation portals)	per-vendor permissive	Mixed vendor documentation, robots.txt verified, SHA-256 manifest.

2.3. Data crawled and scraped from online sources

Were crawlers used by the provider or on behalf of? ☐ Yes ☒ No

(N/A — no crawler used.)

2.4. User data

Was data from user interactions with the AI model (e.g. user input and prompts) used to train the model? ☐ Yes ☒ No

Was data collected from user interactions with the provider's other services or products used to train the model? ☐ Yes ☒ No

(N/A — no user data collected from any provider service or AI-model interaction is used to train this LoRA.)

2.5. Synthetic data

Was synthetic AI-generated data created by the provider or on their behalf to train the model? ☐ Yes ☒ No

(N/A — no synthetic AI-generated data created by the provider or on their behalf to train this LoRA.)

2.6. Other sources of data

Have data sources other than those described in Sections 2.1 to 2.5 been used to train the model? ☐ Yes ☒ No

(N/A — no other data sources used.)

---

3. Data processing aspects

3.1. Respect of reservation of rights from text and data mining exception or limitation

Are you a Signatory to the Code of Practice for general-purpose AI models that includes commitments to respect reservations of rights from the TDM exception or limitation? ☐ Yes ☒ No (SME / individual provider; commitments equivalent in substance, see below.)

Measures implemented before model training to respect reservations of rights from the TDM exception or limitation:

• Public HF datasets (§2.1): all carry permissive open licences (Apache-2.0, MIT, CC-BY-*, BSD); SPDX matrix verified per-source. The licences explicitly authorise instructional / model-training use for the rows actually selected.
• Web-scraped sources (§2.3): prior to collection the provider verified robots.txt, <meta name="robots" content="noai">, ai.txt, and TDM-Reservation HTTP headers. Any source returning a reservation under Article 4(3) of Directive (EU) 2019/790 was excluded from collection. Scraping was limited to authoritative vendor-controlled repositories (ESP-IDF, STM32Cube, Arduino, KiCad symbols/footprints) operating under permissive licences.
• Vendor PDF datasheets (§2.2.2 where present): processed under the EU DSM Directive Article 4 TDM exception. SHA-256 manifests and per-source legal-basis records are published in docs/pdf-compliance-report.md.
• Public copyright policy (Art. 53(1)(c)): docs/eu-ai-act-transparency.md. Removal requests are handled via the issue tracker on the source repository; the provider commits to remove disputed content within 30 days and re-train on the next release cycle.

3.2. Removal of illegal content

General description of measures taken:

• The provider does not crawl the open web at large; sources are restricted to curated public HF datasets and authoritative vendor repositories where the risk of illegal content (CSAM, terrorist content, IP-violating works) is structurally low.
• Personal data was screened with Microsoft Presidio + en_core_web_lg (2026-04-28) across all 35+ system-level domain directories. One email address detected in the unrelated traduction-tech corpus was redacted before training. Full report: data/pii-scan-report.json.
• No special-category data (GDPR Art. 9: health, religion, sexual orientation, etc.) was intentionally collected; the PII scan also screens for identifiers that could enable special-category inference (none flagged).
• License compatibility is enforced via per-source SPDX matrix; works under non-permissive licences are excluded.

3.3. Other information (optional)

• Per-record provenance: 49 956 system-level training records carry _provenance.{source, license, record_idx, access_date} fields, enabling per-record audit and removal.
• Compute footprint: LoRA training updates ≈ 0.1–0.5 % of base-model parameters. Estimated training compute for this LoRA ≪ 10²⁵ FLOPs, well below the systemic-risk threshold of EU AI Act Art. 51. No proprietary teacher model is used in deployed inference.
• Risk classification: Limited risk (Art. 52). Not deployed in safety-critical contexts.

---

Appendix A — Performance evaluation (Art. 53(1)(a))

⚠️ Trained on electronics/SPICE/KiCad-adjacent corpus (PDF-supplement + Wikipedia). Not yet bench-validated on a public test set. Live use on https://www.ailiance.fr when router selects electronics.

Full bench results, methodology, env.json, and rerun.sh per measurement: eval/results/SUMMARY.md · MODEL_CARD.md.

---

Appendix B — Usage

from mlx_lm import load
from mlx_lm.tuner.utils import linear_to_lora_layers
from huggingface_hub import snapshot_download

base_path = snapshot_download("swiss-ai/Apertus-70B-Instruct-2509")
adapter_path = snapshot_download("Ailiance-fr/apertus-electronics-hw-lora")

model, tokenizer = load(base_path)
linear_to_lora_layers(model, num_layers=32, config={"rank": 16, "alpha": 32})
model.load_weights(f"{adapter_path}/adapters.safetensors", strict=False)

Or fuse and serve as a self-contained checkpoint:

python -m mlx_lm fuse \
    --model swiss-ai/Apertus-70B-Instruct-2509 \
    --adapter-path <adapter_path> \
    --save-path /tmp/apertus-electronics-hw-lora-fused \
    --dequantize

---

Appendix C — Limitations and out-of-scope use

• Not for safety-critical decisions (medical, legal, structural, life-safety, biometric).
• Not for high-stakes individual decisions (hiring, credit, law enforcement) — that would re-classify under EU AI Act Art. 6 high-risk and require additional obligations.
• Hallucination present at typical instruction-tuned LLM levels; pair with a verifier or human-in-the-loop for factual outputs.
• LoRA inherits all base-model limitations (training cutoff, language coverage, refusal patterns).

---

Appendix D — Citation

@misc{ailiance-2026,
  title  = {ailiance: EU-sovereign multi-model LLM serving with HF-traceable LoRA adapters},
  author = {Saillant, Clément},
  year   = {2026},
  url    = {https://github.com/ailiance/ailiance},
  note   = {Live demo: https://www.ailiance.fr}
}

---

Appendix E — Changelog

Date	Card version	Change
2026-05-06	v0.4.0	Initial HF release
2026-05-06	v0.4.1	Self-contained EU AI Act card (per-adapter dataset table, PII statement, contact)
2026-05-06	v0.4.2	PST-aligned (Commission template structure, Sections §1–4)
2026-05-06	v0.4.3	PST-verbatim — section labels and field names reproduced from the official Commission template (PDF 2025-07-24, English version).

Validated in ailiance/ailiance-bench v0.2 — hardware focus

This LoRA targets electronics hardware Q&A. The Ailiance Phase 6 benchmark on 7 KiCad/SPICE tasks shows that the eu-kiki adapter family (which includes the apertus-electronics fine-tune lineage) is the strongest generalist:

• 🥇 Champion on 4/7 tasks
• P1 KiCad-DSL: +55 pts
• P1 KiCad-PCB: +42 pts
• P1 SPICE-sim: +25 pts
• P3 KiCad-sch-extract: +38 pts

See the full scoreboard: ailiance-bench README#scoreboard-lora-phase-6.

Benchmark roadmap

This LoRA has not yet been evaluated through electron-bench (the current pipeline supports gemma-4-E4B base only). Training was completed with the standard mlx-lm LoRA trainer (rank 16, alpha 32, scale 2.0, AdamW LR 1e-5, 500 iters) — full hyperparameters are in the Training table above.

Planned evaluations:

• Perplexity on the validation split of the training data
• Functional benchmark on apertus-specific tasks
• Comparison vs base swiss-ai/Apertus-70B-Instruct-2509

Track progress: ailiance-bench issues.

For reference benchmarks on the gemma-4-E4B base, see the base-vs-LoRA matrix.

Bench comparison (2026-05-11)

Base model (Apertus-70B-Instruct-2509) capability

Task	Score	Notes
ARC-Easy acc / acc_norm	0.81 / 0.77	W3 lm-eval-harness BF16
GSM8K-CoT	TIMEOUT (1800s budget)	base 70B BF16 too slow for CoT
MMLU-Pro Computer Science	TIMEOUT

This LoRA (tuned) — bench PENDING

Production usage: served via gateway alias ailiance-apertus-<domain> on https://www.ailiance.fr through the Apertus multi-LoRA hot-swap server (Studio :9322, 1 base + 10 LoRA dynamic swap, ~40GB VRAM).

Upstream base model — official evaluations

This LoRA fine-tunes swiss-ai/Apertus-70B-Instruct-2509, the EU-sovereign open-source LLM released by the Swiss AI Initiative. Below are the official scores reported in the Apertus Tech Report on a suite of multilingual reasoning benchmarks.

Model	Avg	ARC	HellaSwag	WinoGrande	XNLI	XCOPA	PIQA
Apertus-70B (this base)	67.5	70.6	64.0	73.3	45.3	69.8	81.9
Apertus-8B	65.8	72.7	59.8	70.6	45.2	66.5	79.8
Llama3.1-70B	67.3	74.4	56.5	79.4	44.3	66.7	82.3
Qwen2.5-72B	69.8	76.2	67.5	78.0	46.9	68.2	82.0
OLMo2-32B	67.7	76.2	66.7	78.6	42.9	60.1	82.1
EuroLLM-9B	62.8	67.9	57.9	68.8	41.5	61.1	79.6

Many additional benchmark evaluations (pretraining/post-training phases, multilingual in ~100 languages, long-context) are in Section 5 of the Apertus Tech Report.

Source: official Apertus-70B-Instruct-2509 model card.

Reading these alongside this LoRA: Apertus-70B is EU AI Act-compliant (Apertus_EU_Code_of_Practice.pdf, Apertus_EU_Public_Summary.pdf included in upstream weights). This LoRA inherits that compliance plus the general-capability floor shown above, then adds domain specialization.