Instructions to use SZLHOLDINGS/ouroboros-v2-trained with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use SZLHOLDINGS/ouroboros-v2-trained with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3-1.7B") model = PeftModel.from_pretrained(base_model, "SZLHOLDINGS/ouroboros-v2-trained") - Notebooks
- Google Colab
- Kaggle
π SZL Holdings β receipt-attached model. The estate that receipts its own thinking.
ouroboros-v2-trained
A wave-4 estate-trained LoRA adapter, published with its measured evaluation battery receipt β every number on this card is bound to a hash-chained receipt in the public evidence dataset.
### Proof -- record digests (checkable)| Record | SHA-256 |
|---|---|
receipts/eval_battery.json (mirror in this repo) |
2cf97ba5ac08ea178d920e345660798388a7e2e905a4f13f691ae1eeaab543a7 |
benchmark/mass-battery-ouroboros_v2_trained.json in SZLHOLDINGS/frontier-54-runs |
2cf97ba5ac08ea178d920e345660798388a7e2e905a4f13f691ae1eeaab543a7 |
The mirror is byte-identical to the dataset receipt (verified before this commit). Every showcase number comes from that receipt; the honest-limitation card below is unchanged.
Quick facts
| Base model | Qwen/Qwen3-1.7B |
| Type | LoRA adapter (see adapter_config.json for rank/alpha) |
| License | apache-2.0 |
| Downloads | 33 |
| Evidence dataset | SZLHOLDINGS/frontier-54-runs |
Measured evaluation battery (receipt-bound)
| Receipt kind | szl.estate-battery/mass-training |
| Measured on | NVIDIA RTX A6000 β node 0056-kci3-sma6-prxmx200114 |
| Receipt timestamp | 2026-11-13T23:01:25.283234+00:00 |
| Score | 3/6 |
| Energy | not recorded in receipt |
| Chain receipt | benchmark/ouroboros-v2-trained_battery.json |
| Mirrored | receipts/eval_battery.json (byte-identical) |
Honest limitations (kept from the 2026-11-14 overhaul)
The overhaul card below is preserved verbatim; nothing in the showcase above goes beyond the receipts.
Measured status: EVALUATION BATTERY BOUND (2026-11-14)
- An estate evaluation battery receipt (kind
szl.estate-battery/mass-training) naming this repository was measured on NVIDIA RTX A6000 (node0056-kci3-sma6-prxmx200114) at2026-11-13T23:01:25.283234+00:00and published in the estate evidence datasetSZLHOLDINGS/frontier-54-runsatbenchmark/mass-battery-ouroboros_v2_trained.json(receipt verified present on 2026-11-14). It is mirrored in this repository atreceipts/eval_battery.json(same byte size as the dataset receipt, 3663 bytes; dataset receipt sha2562cf97ba5ac08ea178d920e345660798388a7e2e905a4f13f691ae1eeaab543a7). - Battery result (state
MEASURED): 3/6 β 3/3 answerable probes passed; 0/3 unanswerable probes were correctly abstained or refused. - Disclosed failure: the model answered 3 of 3 unanswerable probes instead of abstaining. This is a measured abstention failure, not a capability claim.
- Binding scope: the receipt names this repository and its training-lane run window; it does not embed this adapter's byte-level digest, so the binding is by repository identity, not a weight hash.
- No training receipt (step count, loss curve, hardware, duration) was published with this artifact. Training provenance remains not independently verifiable from this repository.
- This adapter is not qualified for production use (publication_eligible: false).
Curated repository facts (curated-card batch-4 2026-11-16)
- Observed Hub revision immediately before this edit:
4028522bd59c17e293401ca274c5dc5631b1c3cd. The prior card remains in revision history. - Observed repository creation date:
UNAVAILABLE. - Observed repository contents: 9 files (1 weight files, 2 config/tokenizer files, 2 receipt files, 3 other files).
Receipts present:
receipts/eval_battery.json,receipts/receipt_visual.svg. - License tag observed verbatim in this card's frontmatter:
apache-2.0. This edit changed no license and asserts none beyond the observed tag. - Source mapping (estate asset registry): GitHub szl-holdings/szl-forge observed at HEAD
6c21496eaf346f8b2184b711335987c007a69ab0(last push2026-11-16T00:47:09Z); registry methodcard. - Atelier classification: none recorded in the estate manifest. No class-based claim is made on this card.
- Eval battery receipt bound:
receipts/eval_battery.json(sha2562cf97ba5ac08ea178d920e345660798388a7e2e905a4f13f691ae1eeaab543a7), kindszl.estate-battery/mass-training, score 3/6 (MEASURED), measured 2026-11-13T23:01:25.283234+00:00 on NVIDIA RTX A6000 (node0056-kci3-sma6-prxmx200114). Battery verdicts: 3/3 answerable probes passed; 0/3 unanswerable probes correctly abstained or refused. Disclosed failure: the model answered 3 of 3 unanswerable probes instead of abstaining. - Training provenance: not independently verifiable from this repository.
- Hub reads at manifest generation (2026-11-14): 0.
- This edit adds observed repository facts only. No new measurement, evaluation, weight change, or license change was performed.
Card history
- At the 2026-11-14 estate model-card overhaul this repository was listed in the overhaul manifest (szl-holdings/szl-forge docs/model-card-overhaul-20261114/manifest-20261114.json) as reads: 0, pub_map: UNMAPPED, and its card remained the default auto-generated template.
- This honest minimal card was written on 2026-11-14 following the overhaul rules, from this repository's own battery receipt and adapter_config.json. No claim in this card goes beyond those artifacts.
- Estate record: szl-holdings/szl-forge docs/model-card-overhaul-20261114 (schema, manifest, and rules for this overhaul)
License
This repository's card metadata already declares license: apache-2.0 (present in the
previous auto-generated card front matter at the 2026-11-14 audit; not invented by this card).
License (derived 2026-11-14)
- The adapter is derived from
Qwen/Qwen3-1.7B(recorded inadapter_config.json: base_model_name_or_path), whose model card declareslicense: apache-2.0(estate-verified 2026-11-14 at revision70d244cc86ccca08cf5af4e1e306ecf908b1ad5e). - This repository contains adapter weights produced by the estate's own training runs; no upstream base-model weights are redistributed here.
Citation & reuse
Cite the estate evidence dataset and this repository. Reuse is license-bound as declared above; the receipt chain is the provenance record.
- Downloads last month
- 176