Add Res2TCNGuard submission (EER 56.10%, scoring-reproduced)

submissions/res2tcnguard.yaml

@@ -0,0 +1,61 @@
+schema_version: 4
+
+system:
+  name: "Res2TCNGuard"
+  slug: "res2tcnguard"
+  description: >
+    TCN-based anti-spoofing countermeasure: sinc-convolution front-end, Res2Net
+    encoder, and dual temporal convolutional networks. ASVspoof2019 LA
+    pretrained, FP32, deterministic first-64600-sample window (no random crop).
+  code: "https://github.com/lab260ru/Res2TCNGuard"
+  checkpoint: "https://huggingface.co/SpeechAntiSpoofingBenchmarks/Res2TCNGuard/blob/523ef915d798df3deee2be4fd06af7e1965ede14/best_1.495.pth"
+  params_millions: 0.172102
+  paper:
+    arxiv_id: "10.48084/etasr.8906"
+    url: "https://etasr.com/index.php/ETASR/article/view/8906"
+    bibtex: |
+      @article{Borodin_Kudryavtsev_Mkrtchian_Gorodnichev_2024,
+        place={Greece},
+        title={Capsule-based and TCN-based Approaches for Spoofing Detection in Voice Biometry},
+        volume={14},
+        number={6},
+        url={https://etasr.com/index.php/ETASR/article/view/8906},
+        DOI={10.48084/etasr.8906},
+        journal={Engineering, Technology & Applied Science Research},
+        author={Borodin, Kirill and Kudryavtsev, Vasiliy and Mkrtchian, Grach and Gorodnichev, Mikhail},
+        year={2024},
+        month={Dec.},
+        pages={18409--18414}
+      }
+
+dataset:
+  id: SpeechAntiSpoofingBenchmarks/CD-ADD
+  revision: c2de87d49b268b624e6af7440dc2890703098965
+  split: test
+
+scores:
+  eer_percent: 56.0992700729927
+  n_trials: 20786
+  n_skipped: 0
+
+artifact:
+  scores_url: https://huggingface.co/SpeechAntiSpoofingBenchmarks/Res2TCNGuard/resolve/44f5a8347036114f1a4b1b7808d925e5cd0b37ce/.eval_results/SpeechAntiSpoofingBenchmarks/CD-ADD/scores.txt
+  scores_sha256: d3d6558e5a5e8809fcfcc6cae2fb174ceb98385cb17729d4b2f95c59bc0d7c2e
+  bench_version: speech-spoof-bench==0.3.0
+
+reproduction:
+  reproduced_by: SpeechAntiSpoofingBenchmarks
+  reproduced_at: 2026-05-31
+  reproduced_bench_version: speech-spoof-bench==0.3.0
+  match: scoring
+
+submitter:
+  hf_username: korallll
+  contact: k.n.borodin@mtuci.ru
+
+submitted_at: 2026-05-31
+notes: >
+  Deterministic first-64600-sample window (no random crop). Cross-dataset
+  generalization: model trained on ASVspoof2019 LA, evaluated on CD-ADD
+  (modern neural-TTS deepfakes). EER > 50% indicates the model does not
+  generalize to this out-of-domain set.