System Detail

palabraai/redimnet2-b6-vb2+vox2

redimnet2-b6-vb2vox2-lm-full-core

Submitted by: Public checkpoint

ranked review: approved PalabraAI ReDimNet2-B6 ReDimNet2

Global position: #1 / 19
Scenario-macro EER: 9.1442
Scenario-macro minDCF: 0.324160
Trial-set coverage: 26 / 26
Scenario top-three: 9

Publicly ranked This result is included in the reviewed public ranking.

Global leaderboard Compare variants Submit another run

Model Provenance

Source and architecture

Displayed name: palabraai/redimnet2-b6-vb2+vox2
Submitted by: Public checkpoint
Metadata: OpenSVBench model record
Source: PalabraAI
Architecture: ReDimNet2-B6
Model family: ReDimNet2
Submitted alias: b6-vb2+vox2_v0-lm.pt
Created at: 2026-10-16T09:04:30.545942+00:00

Training And Links

Training data and references

Training data: VoxBlink2 + VoxCeleb2
Training setup: ReDimNet2-B6 large-margin checkpoint trained on VoxBlink2 and VoxCeleb2; the official release reports 12.3M parameters and 13.05 GMACs.
Submission mode: full-core
Model / checkpoint: Open link
Paper: ReDimNet2: Scaling Speaker Verification via Time-Pooled Dimension Reshaping

Open Model / Checkpoint Open Paper

Higher-Ranked Scenarios

Scenario ranks above the global position

Aging #1 Noise/reverb #1 Speaking style #1

Show 3 supporting trial sets

Lombard Grid Lombard: #1 on its trial-set ranking.
VoxCeleb1-H: #1 on its trial-set ranking.
VOiCES Noise/Reverb: #1 on its trial-set ranking.

Lower-Ranked Scenarios

Scenario ranks below the global position

Genre shift #4 Accent/dialect #4 Channel/device #2

Show 3 supporting trial sets

GLOBE: #7 on its trial-set ranking.
CN-Celeb Short: #6 on its trial-set ranking.
HI-MIA: #5 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.	#1/19 Leader	Speaker aging and time gaps 0.0000 EER from leader	#1 2.2933 #2 1.5159	1.9046 0.052327 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.	#1/19 Leader	Noise and reverberation 0.0000 EER from leader	#1 2.2600	2.2600 0.073780 minDCF
Speaking-Style Robustness Whether the model can preserve identity across emotion-driven change, whispered speech, and noise-induced Lombard speaking style.	#1/19 Leader	Speaking style shift 0.0000 EER from leader	#1 2.3127 #1 4.7875 #1 0.0373	2.3792 0.145230 minDCF
Cross-Lingual Robustness How well speaker identity survives enrollment-test language mismatch.	#1/19 Leader	Language mismatch 0.0000 EER from leader	#1 3.4191	3.4191 0.259714 minDCF
In-The-Wild Robustness How strong the model is on unconstrained celebrity and media speech across official CN-Celeb and VoxCeleb protocols.	#1/19 Leader	Open-domain media speech 0.0000 EER from leader	#4 7.8288 #2 0.1701 #2 0.3387 #1 0.6788	4.1123 0.141370 minDCF
Overlap Robustness How well speaker identity survives light, mid, and heavy overlap in both meeting and domestic recordings.	#1/19 Leader	Overlapping speakers 0.0000 EER from leader	#2 6.1000 #1 11.8889	8.9944 0.440323 minDCF
Short-Duration Robustness How much performance holds up when speech evidence is limited by duration.	#1/19 Leader	Short-duration speech 0.0000 EER from leader	#5 5.8667 #4 8.1985 #6 25.1046 #1 9.5363	12.1765 0.431643 minDCF
Distance Robustness How much performance changes across meeting and domestic distance mismatch.	#2/19 Top 3	Distance mismatch 0.2720 EER from leader	#3 15.4780 #4 8.3550 #2 7.3741 #3 6.9360	9.5358 0.334705 minDCF
Channel / Device Robustness How resilient the model is to device and channel mismatch.	#2/19 Top 3	Device and channel variation 4.7906 EER from leader	#3 15.3933 #3 6.8083	11.1008 0.464227 minDCF
Accent / Dialect Robustness How reliably the model tracks identity across accent and dialect mismatch.	#4/19 Competitive	Accent and dialect variation 0.7439 EER from leader	#7 25.2515 #3 17.1067	21.1791 0.581639 minDCF
Genre-Shift Robustness Whether performance holds when CN-Celeb enrollment and test speech come from different source genres.	#4/19 Competitive	Source-genre variation 9.2520 EER from leader	#4 23.5245	23.5245 0.640800 minDCF

Variant Compare

Compare ReDimNet2-B6 variants

Expand this section to compare checkpoints in the same architecture group by source, training data, training setup, and leaderboard result.

Show

Variant	Submitted by	Source	Training Data	Training Setup	Global Rank	Macro EER	Macro minDCF	Open
palabraai/redimnet2-b6-vb2+vox2 Current	Public checkpoint	PalabraAI	VoxBlink2 + VoxCeleb2	ReDimNet2-B6 large-margin checkpoint trained on VoxBlink2 and VoxCeleb2; the official release reports 12.3M parameters and 13.05 GMACs.	#1 ranked	9.1442	0.324160	Open
palabraai/redimnet2-b6	Public checkpoint	PalabraAI	VoxCeleb2 dev	ReDimNet2-B6 large-margin checkpoint; local model card reports 12.3M parameters and 13.05 GMACs.	#3 ranked	10.5277	0.419432	Open

Trial-Set Results

Raw ranking by trial set

Expand this section to inspect the detailed trial-set rankings.

Show

Trial Set	Rank	Scenarios	Trials	Score
TidyVoiceX2-ASV TidyVoiceX2-ASV	#1/19	Cross-lingual	200000	3.4191 0.259714 minDCF
HI-MIA HI-MIA	#5/19	Short-duration	660000	5.8667 0.240098 minDCF
GSC Short Google Speech Commands	#4/19	Short-duration	220000	8.1985 0.487250 minDCF
GLOBE GLOBE	#7/19	Accent/dialect	56848	25.2515 0.507624 minDCF
CN-Celeb CN-Celeb	#4/19	In-the-wild	3484292	7.8288 0.259349 minDCF
CN-Celeb Genre CN-Celeb	#4/19	Genre shift	440000	23.5245 0.640800 minDCF
CN-Celeb Short CN-Celeb	#6/19	Short-duration	546964	25.1046 0.644055 minDCF
VoxCeleb1-O VoxCeleb1	#2/19	In-the-wild	37611	0.1701 0.011700 minDCF
VoxCeleb1-E VoxCeleb1	#2/19	In-the-wild	579818	0.3387 0.019785 minDCF
VoxCeleb1-H VoxCeleb1	#1/19	In-the-wild	550894	0.6788 0.038685 minDCF
VoxCeleb Short VoxCeleb1	#1/19	Short-duration	394724	9.5363 0.355169 minDCF
3D-Speaker Device 3D-Speaker	#3/19	Channel/device	180000	15.3933 0.565787 minDCF
3D-Speaker Distance 3D-Speaker	#3/19	Distance	175163	15.4780 0.513483 minDCF
3D-Speaker Dialect 3D-Speaker	#3/19	Accent/dialect	180000	17.1067 0.655653 minDCF
FFSVC 2022 Cross-Channel FFSVC 2022	#3/19	Channel/device	72000	6.8083 0.362667 minDCF
FFSVC 2022 Cross-Domain FFSVC 2022	#3/19	Distance	66546	6.9360 0.373640 minDCF
Whisper40 Whisper Whisper40	#1/19	Speaking style	17600	4.7875 0.269813 minDCF
Lombard Grid Lombard Lombard Grid	#1/19	Speaking style	29524	0.0373 0.001453 minDCF
VOiCES Noise/Reverb VOiCES	#1/19	Noise/reverb	55000	2.2600 0.073780 minDCF
CHiME-6 Domestic Far-Field CHiME-6	#2/19	Distance	36487	7.3741 0.266385 minDCF
CHiME-6 Overlap CHiME-6	#1/19	Overlap	39600	11.8889 0.584639 minDCF
ESD ESD	#1/19	Speaking style	437408	2.3127 0.164424 minDCF
AliMeeting Near/Far AliMeeting	#4/19	Distance	220000	8.3550 0.185310 minDCF
AliMeeting Overlap AliMeeting	#2/19	Overlap	165000	6.1000 0.296007 minDCF
VoxKnesset VoxKnesset	#1/19	Aging	158312	2.2933 0.078447 minDCF
VoxPopuli Aging VoxPopuli	#2/19	Aging	146575	1.5159 0.026206 minDCF