Trial Set
HI-MIA
No upstream SV trial keys are used; trials are generated from the test split.
Notice: This is not the original official SV trial list.
- Source dataset
- HI-MIA
- Trials
- 660000
- Protocol source
- Self-constructed protocol
- Source split
- Public test split
Source Dataset
About HI-MIA
HI-MIA was collected for far-field, text-dependent speaker verification in home-like rooms. Its 340 participants were recorded with close-talk microphones and microphone arrays at different positions and distances. The source task uses wake-up phrases; OpenSVBench uses the public test split to construct a separate short-duration, text-independent trial set.
- Language
- Mandarin Chinese
- Speakers
- 340 participants
- Recording
- Close-talk microphones and far-field microphone arrays
- Original task
- Text-dependent speaker verification / wake-up word
Direct Ranking
Systems on this trial set
25 reviewed systems currently have valid results.
| Rank | System | Submitted by | Trial-Set EER ↓ | Trial-Set minDCF ↓ |
|---|---|---|---|---|
| #1 |
MECT-b2-czur-finetuned
MECT-b2-czur-finetuned
|
czur | 2.3383 | 0.120418 |
| #2 |
Agora Voiceprint (ResNet293)
agora_vp
|
agora | 2.4227 | 0.143702 |
| #3 |
iic/eres2net-large-3dspeaker
eres2net_large_3dspeaker
|
Public checkpoint | 2.8963 | 0.178460 |
| #4 |
W2V-BERT 2.0 + Adapter-MFA (cross-device consistency FT + weight soup)
ActionPower VoicePrint
|
ActionPower | 3.4767 | 0.210408 |
| #5 |
W2V-BERT 2.0 + MHFA (VoxCeleb2 + VoxBlink2)
Wav2Vec2-BERT
|
pengjy | 3.6067 | 0.210263 |
| #6 |
w2v-bert2.0 + cc + ecapa-tdnn + astp + sf2
w2v-bert2.0 + cc + ecapa-tdnn + astp + sf2
|
JinghanPeng | 3.6515 | 0.192412 |
| #7 |
iic/eres2netv2-zh
eresv2_zh
|
Public checkpoint | 3.7090 | 0.208377 |
| #8 |
AntSpeaker/MECT-B2
antspeaker-mect-b2-vb2-full-core
|
Public checkpoint | 3.7550 | 0.188360 |
| #9 |
palabraai/redimnet2-b6
redimnet2
|
Public checkpoint | 5.7113 | 0.276288 |
| #10 |
wespeaker/res152-voxceleb
wespeaker_resnet152
|
Public checkpoint | 5.8453 | 0.284905 |
| #11 |
palabraai/redimnet2-b6-vb2+vox2
redimnet2-b6-vb2vox2-lm-full-core
|
Public checkpoint | 5.8667 | 0.240098 |
| #12 |
iic/campplus
campplus
|
Public checkpoint | 5.8722 | 0.341062 |
| #13 |
wespeaker/samresnet100
vb2-sam100
|
Public checkpoint | 5.9033 | 0.273133 |
| #14 |
wespeaker/res293-voxceleb
wespeaker_resnet293
|
Public checkpoint | 6.3817 | 0.301015 |
| #15 |
wespeaker/w2vbert2
vb2-w2vbert
|
Public checkpoint | 7.0400 | 0.349597 |
| #16 |
wespeaker/res34-voxceleb
wespeaker_resnet34_voxceleb
|
Public checkpoint | 7.0523 | 0.323703 |
| #17 |
iic/eres2net-en
eres_en
|
Public checkpoint | 7.0967 | 0.323343 |
| #18 |
speechbrain/ecapa
ecapa
|
Public checkpoint | 7.6850 | 0.369167 |
| #19 |
wespeaker/res34-cnceleb
wespeaker_r34
|
Public checkpoint | 7.7542 | 0.426640 |
| #20 |
wespeaker/campplus-voxceleb
campplus_wespeaker
|
Public checkpoint | 7.9797 | 0.395343 |
| #21 |
wespeaker/ecapa512-voxceleb
wespeaker_ecapa512
|
Public checkpoint | 8.5340 | 0.404212 |
| #22 |
wespeaker/ecapa1024-voxceleb
wespeaker_ecapa1024
|
Public checkpoint | 8.6250 | 0.381742 |
| #23 |
Dual-LoRA W2V-BERT2-Adapter-MFA (TidyVoice)
duallora-w2vbert2-tidyvoice-full-core
|
sgqt | 10.5467 | 0.426292 |
| #24 |
microsoft/unispeech-sat-base-plus-sv
unispeech_sat_base_plus
|
Public checkpoint | 23.7600 | 0.944420 |
| #25 |
microsoft/wavlm-base-plus-sv
wavlm_base
|
Public checkpoint | 25.0650 | 0.910383 |