System Detail
wespeaker/res293-voxceleb
wespeaker_resnet293
Submitted by: Public checkpoint
- Global position
- #13 / 25
- Scenario-macro EER
- 11.3130
- Scenario-macro minDCF
- 0.439607
- Trial-set coverage
- 26 / 26
- Scenario top-three
- 0
Publicly ranked
This result is included in the reviewed public ranking.
Model Provenance
Source and architecture
- Displayed name:
wespeaker/res293-voxceleb - Submitted by: Public checkpoint
- Metadata: OpenSVBench model record
- Source: WeSpeaker
- Architecture: ResNet
- Model family: ResNet
- Submitted alias:
wespeaker-voxceleb-resnet293-LM - Created at:
2026-05-29T14:55:53.658417+00:00
Training And Links
Training data and references
- Training data: VoxCeleb2 dev, 5,994 speakers
- Training setup: ResNet293-TSTP-emb256 r-vector, large-margin fine-tuned; official card reports 28.62M parameters and 28.10G FLOPs.
- Score processing: None
- Submission mode:
full-core - Model source: Open link
- Paper: Deep Residual Learning for Image Recognition
Higher-Ranked Scenarios
Scenario ranks above the global position
Show 3 supporting trial sets
- VoxKnesset: #10 on its trial-set ranking.
- VoxCeleb1-O: #11 on its trial-set ranking.
- TidyVoiceX2-ASV: #11 on its trial-set ranking.
Lower-Ranked Scenarios
Scenario ranks below the global position
Show 3 supporting trial sets
- CN-Celeb: #18 on its trial-set ranking.
- CN-Celeb Genre: #17 on its trial-set ranking.
- CN-Celeb Short: #17 on its trial-set ranking.
Scenario Rankings
Full ranking across scenarios
This table shows where the model sits on each scenario, using source-balanced scenario scores and the linked trial sets as evidence.
| Scenario | Rank | Lens | Evidence | Score |
|---|---|---|---|---|
|
Aging Robustness
How stable identity representations remain across age-derived and longitudinal recording gaps.
|
#10/25
Competitive
|
Speaker aging and time gaps
0.8655 EER from leader
|
2.6572
0.106953 minDCF
|
|
|
Cross-Lingual Robustness
How well speaker identity survives enrollment-test language mismatch.
|
#11/25
Competitive
|
Language mismatch
3.0906 EER from leader
|
4.7483
0.298819 minDCF
|
|
|
Speaking-Style Robustness
Whether the model can preserve identity across emotion-driven change, whispered speech, and noise-induced Lombard speaking style.
|
#12/25
Competitive
|
Speaking style shift
2.4089 EER from leader
|
4.3256
0.263660 minDCF
|
|
|
Overlap Robustness
How well speaker identity survives light, mid, and heavy overlap in both meeting and domestic recordings.
|
#12/25
Competitive
|
Overlapping speakers
4.6378 EER from leader
|
12.3294
0.554516 minDCF
|
|
|
Noise / Reverb Robustness
How reliably the model preserves identity when room acoustics, distractor noise, and microphone placement deviate from an easier in-corpus reference condition.
|
#13/25
Competitive
|
Noise and reverberation
3.1000 EER from leader
|
5.3600
0.148780 minDCF
|
|
|
Channel / Device Robustness
How resilient the model is to device and channel mismatch.
|
#13/25
Competitive
|
Device and channel variation
10.0021 EER from leader
|
13.9682
0.677811 minDCF
|
|
|
Accent / Dialect Robustness
How reliably the model tracks identity across accent and dialect mismatch.
|
#13/25
Competitive
|
Accent and dialect variation
5.9504 EER from leader
|
21.7119
0.663725 minDCF
|
|
|
Distance Robustness
How much performance changes across meeting and domestic distance mismatch.
|
#14/25
Needs work
|
Distance mismatch
7.0758 EER from leader
|
11.7700
0.508920 minDCF
|
|
|
Short-Duration Robustness
How much performance holds up when speech evidence is limited by duration.
|
#14/25
Needs work
|
Short-duration speech
5.4860 EER from leader
|
14.0805
0.568931 minDCF
|
|
|
In-The-Wild Robustness
How strong the model is on unconstrained celebrity and media speech across official CN-Celeb and VoxCeleb protocols.
|
#15/25
Needs work
|
Open-domain media speech
4.2484 EER from leader
|
6.6344
0.231795 minDCF
|
|
|
Genre-Shift Robustness
Whether performance holds when CN-Celeb enrollment and test speech come from different source genres.
|
#17/25
Needs work
|
Source-genre variation
13.6175 EER from leader
|
26.8575
0.811767 minDCF
|
Variant Compare
Compare ResNet variants
Compare the family variants on the overall board, a Scenario, or a Trial Set.
Show
Variant Compare
Compare ResNet variants
Compare the family variants on the overall board, a Scenario, or a Trial Set.
| Variant | Submitted by | Source | Training Data | Training Setup | Rank | EER | minDCF | Open |
|---|---|---|---|---|---|---|---|---|
|
Agora Voiceprint (ResNet293)
ResNet293 + TSTP (256-d embedding)
|
agora | custom_local_torch | VoxCeleb2 dev + VoxBlink2 + 3D-Speaker train split | AAM-Softmax / ArcMargin Speed perturbation MUSAN noise + RIR reverberation augmentation Non-VAD diff-chunk Knowledge distillation | #5 | 7.8931 | 0.297508 | Open |
|
wespeaker/res152-voxceleb
ResNet
|
Public checkpoint | WeSpeaker | VoxCeleb2 dev, 5,994 speakers | ResNet152-TSTP-emb256 r-vector, ArcMargin, 150-epoch VoxCeleb2 recipe with speed perturbation and MUSAN/RIRS augmentation. | #12 | 11.2275 | 0.442748 | Open |
|
wespeaker/res293-voxceleb
ResNet
Current
|
Public checkpoint | WeSpeaker | VoxCeleb2 dev, 5,994 speakers | ResNet293-TSTP-emb256 r-vector, large-margin fine-tuned; official card reports 28.62M parameters and 28.10G FLOPs. | #13 | 11.3130 | 0.439607 | Open |
|
wespeaker/res34-voxceleb
ResNet
|
Public checkpoint | WeSpeaker | VoxCeleb2 dev, 5,994 speakers | ResNet34-TSTP-emb256 r-vector, large-margin fine-tuned; official card reports 6.63M parameters and 4.55G FLOPs. | #15 | 12.0516 | 0.477054 | Open |
|
wespeaker/res34-cnceleb
ResNet
|
Public checkpoint | WeSpeaker | CN-Celeb train | ResNet34 r-vector with TSTP pooling and large-margin fine-tuning on the CN-Celeb WeSpeaker recipe. | #22 | 14.4910 | 0.607964 | Open |
Trial-Set Results
Raw ranking by trial set
Expand this section to inspect the detailed trial-set rankings.
Show
Raw ranking by trial set
Expand this section to inspect the detailed trial-set rankings.
| Trial Set | Rank | Scenarios | Trials | Score |
|---|---|---|---|---|
|
TidyVoiceX2-ASV
TidyVoiceX2-ASV
|
#11/25
|
Cross-lingual
|
200000 |
4.7483
0.298819 minDCF
|
|
HI-MIA
HI-MIA
|
#14/25
|
Short-duration
|
660000 |
6.3817
0.301015 minDCF
|
|
GSC Short
Google Speech Commands
|
#15/25
|
Short-duration
|
220000 |
11.4500
0.608915 minDCF
|
|
GLOBE
GLOBE
|
#17/25
|
Accent/dialect
|
56848 |
25.4837
0.531618 minDCF
|
|
CN-Celeb
CN-Celeb
|
#18/25
|
In-the-wild
|
3484292 |
12.4334
0.412299 minDCF
|
|
CN-Celeb Genre
CN-Celeb
|
#17/25
|
Genre shift
|
440000 |
26.8575
0.811767 minDCF
|
|
CN-Celeb Short
CN-Celeb
|
#17/25
|
Short-duration
|
546964 |
26.0196
0.911865 minDCF
|
|
VoxCeleb1-O
VoxCeleb1
|
#11/25
|
In-the-wild
|
37611 |
0.5159
0.037014 minDCF
|
|
VoxCeleb1-E
VoxCeleb1
|
#12/25
|
In-the-wild
|
579818 |
0.7016
0.041837 minDCF
|
|
VoxCeleb1-H
VoxCeleb1
|
#13/25
|
In-the-wild
|
550894 |
1.2890
0.075021 minDCF
|
|
VoxCeleb Short
VoxCeleb1
|
#12/25
|
Short-duration
|
394724 |
12.4707
0.453929 minDCF
|
|
3D-Speaker Device
3D-Speaker
|
#14/25
|
Channel/device
|
180000 |
18.9420
0.831067 minDCF
|
|
3D-Speaker Distance
3D-Speaker
|
#15/25
|
Distance
|
175163 |
18.0185
0.789752 minDCF
|
|
3D-Speaker Dialect
3D-Speaker
|
#12/25
|
Accent/dialect
|
180000 |
17.9400
0.795833 minDCF
|
|
FFSVC 2022 Cross-Channel
FFSVC 2022
|
#13/25
|
Channel/device
|
72000 |
8.9944
0.524556 minDCF
|
|
FFSVC 2022 Cross-Domain
FFSVC 2022
|
#13/25
|
Distance
|
66546 |
9.1379
0.532346 minDCF
|
|
Whisper40 Whisper
Whisper40
|
#12/25
|
Speaking style
|
17600 |
9.5562
0.533062 minDCF
|
|
Lombard Grid Lombard
Lombard Grid
|
#14/25
|
Speaking style
|
29524 |
0.1490
0.014754 minDCF
|
|
VOiCES Noise/Reverb
VOiCES
|
#13/25
|
Noise/reverb
|
55000 |
5.3600
0.148780 minDCF
|
|
CHiME-6 Domestic Far-Field
CHiME-6
|
#12/25
|
Distance
|
36487 |
11.1336
0.435876 minDCF
|
|
CHiME-6 Overlap
CHiME-6
|
#12/25
|
Overlap
|
39600 |
15.1389
0.626972 minDCF
|
|
ESD
ESD
|
#13/25
|
Speaking style
|
437408 |
3.2714
0.243165 minDCF
|
|
AliMeeting Near/Far
AliMeeting
|
#14/25
|
Distance
|
220000 |
8.7900
0.277705 minDCF
|
|
AliMeeting Overlap
AliMeeting
|
#17/25
|
Overlap
|
165000 |
9.5200
0.482060 minDCF
|
|
VoxKnesset
VoxKnesset
|
#10/25
|
Aging
|
158312 |
3.5733
0.143699 minDCF
|
|
VoxPopuli Aging
VoxPopuli
|
#12/25
|
Aging
|
146575 |
1.7411
0.070206 minDCF
|