Datasets:
submissions: add res2tcnguard (#11)
Browse files- submissions: add res2tcnguard (4a4f83f3f1f1e05e3f6f942ad62153ca0b27c22b)
- fill reproduction (match: scoring) for res2tcnguard (d3ffea67a6654b7dcbb53166cbe2c7c9f27673ba)
submissions/res2tcnguard.yaml
ADDED
|
@@ -0,0 +1,44 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
schema_version: 4
|
| 2 |
+
system:
|
| 3 |
+
name: Res2TCNGuard
|
| 4 |
+
slug: res2tcnguard
|
| 5 |
+
description: 'TCN-based anti-spoofing countermeasure: sinc-convolution front-end,
|
| 6 |
+
Res2Net encoder, and dual temporal convolutional networks. ASVspoof2019 LA pretrained,
|
| 7 |
+
FP32, deterministic first-64600-sample window (no random crop).
|
| 8 |
+
|
| 9 |
+
'
|
| 10 |
+
code: https://github.com/lab260ru/Res2TCNGuard
|
| 11 |
+
checkpoint: https://huggingface.co/SpeechAntiSpoofingBenchmarks/Res2TCNGuard/blob/523ef915d798df3deee2be4fd06af7e1965ede14/best_1.495.pth
|
| 12 |
+
paper:
|
| 13 |
+
arxiv_id: 10.48084/etasr.8906
|
| 14 |
+
url: https://etasr.com/index.php/ETASR/article/view/8906
|
| 15 |
+
bibtex: "@article{Borodin_Kudryavtsev_Mkrtchian_Gorodnichev_2024,\n place={Greece},\n\
|
| 16 |
+
\ title={Capsule-based and TCN-based Approaches for Spoofing Detection in Voice\
|
| 17 |
+
\ Biometry},\n volume={14},\n number={6},\n url={https://etasr.com/index.php/ETASR/article/view/8906},\n\
|
| 18 |
+
\ DOI={10.48084/etasr.8906},\n journal={Engineering, Technology & Applied\
|
| 19 |
+
\ Science Research},\n author={Borodin, Kirill and Kudryavtsev, Vasiliy and\
|
| 20 |
+
\ Mkrtchian, Grach and Gorodnichev, Mikhail},\n year={2024},\n month={Dec.},\n\
|
| 21 |
+
\ pages={18409--18414}\n}\n"
|
| 22 |
+
params_millions: 0.172102
|
| 23 |
+
dataset:
|
| 24 |
+
id: SpeechAntiSpoofingBenchmarks/CFAD
|
| 25 |
+
revision: 9b7a2999069867d6f0329e1b03abf324ee0d5d30
|
| 26 |
+
split: test
|
| 27 |
+
scores:
|
| 28 |
+
eer_percent: 38.99233296823658
|
| 29 |
+
n_trials: 62999
|
| 30 |
+
n_skipped: 0
|
| 31 |
+
artifact:
|
| 32 |
+
scores_url: https://huggingface.co/SpeechAntiSpoofingBenchmarks/Res2TCNGuard/resolve/e307477042796796a45e234a9803dd6965f39f9c/.eval_results/SpeechAntiSpoofingBenchmarks/CFAD/scores.txt
|
| 33 |
+
scores_sha256: ca559ad1f423fd61b1b31c4041add16826371597c29a30155a5f2b337d809a4c
|
| 34 |
+
bench_version: speech-spoof-bench==0.3.4
|
| 35 |
+
reproduction:
|
| 36 |
+
reproduced_by: SpeechAntiSpoofingBenchmarks
|
| 37 |
+
reproduced_at: '2026-06-08'
|
| 38 |
+
reproduced_bench_version: speech-spoof-bench==0.3.4
|
| 39 |
+
match: scoring
|
| 40 |
+
submitter:
|
| 41 |
+
hf_username: korallll
|
| 42 |
+
contact: k.n.borodin@mtuci.ru
|
| 43 |
+
submitted_at: '2026-06-08'
|
| 44 |
+
notes: Deterministic first-64600-sample window (no random crop).
|