Instructions to use subhamoybasu/laya-evidence-v2-seed14 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Laya
How to use subhamoybasu/laya-evidence-v2-seed14 with Laya:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
Laya Evidence v2, seed 14
This is the released recipe trained with seed 14: one of the three seed runs whose mean the release reports (seeds 13, 14 and 15). The released model is seed 13, subhamoybasu/laya-evidence-v2. This run is published so that its row of the results table can be re-evaluated on the test split, not only recomputed from the saved scores. It is not a different model; it differs from the released one only in its random seed, its threshold and the exact weights that came out.
| Companion to | subhamoybasu/laya-evidence-v2, the released model; its card has the task, the usage, the limitations and the full attribution |
| Lock entry | ours_s14 in configs/final_v2.lock.json (https://github.com/subhamoy-basu/laya-evidence) |
| Training run | v2_joint_C_cuad_aug_fever_s14, seed 14, code at commit edd840a8b9db02e84736c053a686c71375fb1644 |
| Weights | evidence.safetensors, sha256 28e293542c888ce25efd7bbb7030fdc346a5878105144cad04d9cb396386ed65; LoRA adapter lora_adapter/adapter_model.safetensors, sha256 c4182e736467476d8d0022a28536611cea54478980d54acead51f3c596986a3c |
| Default threshold ฯ | 0.85 (presets: "abstract" 0.62, "contract" 0.89) |
| Claims per encoder pass | up to 4 |
| License | CC BY-SA 4.0 (see LICENSE and NOTICE): trained on FEVER (CC BY-SA 3.0); the code is Apache-2.0 |
Test results
From the committed evaluation reports of this run (https://github.com/subhamoy-basu/laya-evidence/blob/main/results/v2/final/summary.md has every number and the other runs). Every run of the release was scored on the test split after the lock, and every access is logged.
| Test set | ContractNLI mAP (macro) | ContractNLI evidence F1 | ContractNLI false alarms | SciFact evidence F1 |
|---|---|---|---|---|
| Official split | 0.860 | 0.753 | 0.124 | 0.662 |
| Clean subset | 0.848 | 0.741 | 0.125 | 0.533 |
How to use
Requires Python 3.11 or later.
pip install "laya-evidence[lora] @ git+https://github.com/subhamoy-basu/laya-evidence"
from laya_evidence import EvidenceExtractor
ex = EvidenceExtractor.from_pretrained("subhamoybasu/laya-evidence-v2-seed14")
doc = ("3. The Receiving Party may disclose Confidential Information to its employees and advisers who "
"need to know it for the Purpose. 4. The Receiving Party shall not reverse engineer, decompile or "
"disassemble any samples or software provided under this Agreement.")
claims = ['Receiving Party may share some Confidential Information with some of its employees.', 'Receiving Party may independently develop information similar to Confidential Information.']
for r in ex.find(doc, claims):
print(r.claim_id, r.status, [(s.label, s.text) for s in r.spans])
The API is the released model's; see its card. It also ships calibration.json, fitted on this run's own dev scores, so each span reports evidence_probability as the released model does.
Attribution
Training data attribution.
ContractNLI: "ContractNLI: A Dataset for Document-level Natural Language Inference for Contracts", Yuta Koreeda and Christopher D. Manning; licensor Hitachi America, Ltd.; https://stanfordnlp.github.io/contract-nli/; licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/) and provided without warranties. Modified: document-level labels and evidence-span indices were converted to one label per span (supports, contradicts or none).
SciFact: "Fact or Fiction: Verifying Scientific Claims", David Wadden, Shanchuan Lin, Kyle Lo, Lucy Lu Wang, Madeleine van Zuylen, Arman Cohan and Hannaneh Hajishirzi; https://github.com/allenai/scifact; claims and evidence annotations licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/) and provided without warranties. Modified: rationale annotations were converted to one label per abstract sentence. Contains information from the SciFact corpus (https://github.com/allenai/scifact), whose abstracts are drawn from S2ORC (https://github.com/allenai/s2orc), which is made available under the ODC Attribution License (https://opendatacommons.org/licenses/by/1-0/).
CUAD: "CUAD: An Expert-Annotated NLP Dataset for Legal Contract Review", Dan Hendrycks, Collin Burns, Anya Chen and Spencer Ball; licensor The Atticus Project; https://huggingface.co/datasets/theatticusproject/cuad; licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/) and provided without warranties. Modified: 31 of its 41 clause categories were turned into claims, written by the generator model Qwen3-32B-AWQ from the category questions and reviewed by the coding assistant (two rewritten), and each segment overlapping an annotated clause of a category was labelled as supporting its claim.
FEVER: "FEVER: a Large-scale Dataset for Fact Extraction and VERification", James Thorne, Andreas Vlachos, Christos Christodoulopoulos and Arpit Mittal; https://fever.ai/; the annotations incorporate material from Wikipedia and are made available under the license terms of the Wikipedia articles or, where those are unavailable, CC BY-SA 3.0 (https://creativecommons.org/licenses/by-sa/3.0/), and are provided without warranties; taken from the MultiVerS pretraining data (https://github.com/dwadden/multivers). Modified: 75,000 claims with their Wikipedia introductions were sampled (25,000 per label), each sentence became one segment, and evidence sentences were labelled supports or contradicts. These weights were trained on FEVER and are therefore released under CC BY-SA 4.0 (https://creativecommons.org/licenses/by-sa/4.0/).
Base model: Laya by Convai Innovations (https://huggingface.co/convaiinnovations/laya), Apache-2.0, loaded at revision 55cf4c4ebb4ebe31b2550e8bdf3bd21b99753851. Laya's weights are not included in this repository. This model adds trained layers on top of Laya's frozen encoder; Laya itself is not modified. Laya Evidence is an independent project, not affiliated with or endorsed by Convai Innovations. LoRA adapters on Laya's encoder (lora_adapter/) were also trained; they modify the encoder only on the evidence path.
Citation
Please cite the released model's repository (the BibTeX of subhamoybasu/laya-evidence-v2):
@software{basu-2026-laya-evidence,
title = "{Laya Evidence}",
author = "Basu, Subhamoy",
year = "2026",
version = "2.1.1",
url = "https://github.com/subhamoy-basu/laya-evidence",
note = "Model weights: \url{https://huggingface.co/subhamoybasu/laya-evidence-v2}. Apache License 2.0; model weights: CC BY-SA 4.0"
}
@inproceedings{hendrycks-etal-2021-cuad,
title = "{CUAD}: An Expert-Annotated {NLP} Dataset for Legal Contract Review",
author = "Hendrycks, Dan and
Burns, Collin and
Chen, Anya and
Ball, Spencer",
editor = "Vanschoren, Joaquin and
Yeung, Serena",
booktitle = "Proceedings of the Neural Information Processing Systems Track on Datasets and Benchmarks",
volume = "1",
year = "2021",
eprint = "2103.06268",
archivePrefix = "arXiv",
primaryClass = "cs.CL",
url = "https://datasets-benchmarks-proceedings.neurips.cc/paper/2021/hash/6ea9ab1baa0efb9e19094440c317e21b-Abstract-round1.html"
}
@inproceedings{thorne-etal-2018-fever,
title = "{FEVER}: a Large-scale Dataset for Fact Extraction and {VER}ification",
author = "Thorne, James and
Vlachos, Andreas and
Christodoulopoulos, Christos and
Mittal, Arpit",
editor = "Walker, Marilyn and
Ji, Heng and
Stent, Amanda",
booktitle = "Proceedings of the 2018 Conference of the North {A}merican Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers)",
month = jun,
year = "2018",
address = "New Orleans, Louisiana",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/N18-1074/",
doi = "10.18653/v1/N18-1074",
pages = "809--819"
}
Model tree for subhamoybasu/laya-evidence-v2-seed14
Base model
convaiinnovations/laya