feat: remove v1.1 and fill missing values
Browse files
README.md
CHANGED
|
@@ -11,18 +11,18 @@ pinned: false
|
|
| 11 |
|
| 12 |
Pretrained pipelines reach state-of-the-art performance on most academic benchmarks and are used [in production by dozens of companies](https://herve.niderb.fr/consulting.html).
|
| 13 |
|
| 14 |
-
| Benchmark
|
| 15 |
-
| ---------------------- |
|
| 16 |
-
| AISHELL-4 |
|
| 17 |
-
| AliMeeting (channel 1) |
|
| 18 |
-
| AMI (IHM) |
|
| 19 |
-
| AMI (SDM) |
|
| 20 |
-
| AVA-AVD |
|
| 21 |
-
| DIHARD 3 (full) |
|
| 22 |
-
| Ego4D (dev.) |
|
| 23 |
-
| MSDWild |
|
| 24 |
-
| REPERE (phase2) |
|
| 25 |
-
| VoxConverse (v0.3) |
|
| 26 |
[Diarization error rate](http://pyannote.github.io/pyannote-metrics/reference.html#diarization) (in %)
|
| 27 |
|
| 28 |
Using one Nvidia Tesla V100 SXM2 GPU and one Intel Cascade Lake 6248 CPU,
|
|
|
|
| 11 |
|
| 12 |
Pretrained pipelines reach state-of-the-art performance on most academic benchmarks and are used [in production by dozens of companies](https://herve.niderb.fr/consulting.html).
|
| 13 |
|
| 14 |
+
| Benchmark | [v2.1](https://hf.co/pyannote/speaker-diarization-2.1) | [v3.1](https://hf.co/pyannote/speaker-diarization-3.1) | [Premium](https://forms.gle/eKhn7H2zTa68sMMx8) |
|
| 15 |
+
| ---------------------- | ------ | ------ | --------- |
|
| 16 |
+
| AISHELL-4 | 14.1 | 12.2 | 11.9 |
|
| 17 |
+
| AliMeeting (channel 1) | 27.4 | 24.4 | 22.5 |
|
| 18 |
+
| AMI (IHM) | 18.9 | 18.8 | 16.6 |
|
| 19 |
+
| AMI (SDM) | 27.1 | 22.4 | 20.9 |
|
| 20 |
+
| AVA-AVD | 66.3 | 50.0 | 39.8 |
|
| 21 |
+
| DIHARD 3 (full) | 26.9 | 21.7 | 17.2 |
|
| 22 |
+
| Ego4D (dev.) | 61.5 | 51.2 | 43.8 |
|
| 23 |
+
| MSDWild | 32.8 | 25.3 | 19.8 |
|
| 24 |
+
| REPERE (phase2) | 8.2 | 7.8 | 7.6 |
|
| 25 |
+
| VoxConverse (v0.3) | 11.2 | 11.3 | 9.4 |
|
| 26 |
[Diarization error rate](http://pyannote.github.io/pyannote-metrics/reference.html#diarization) (in %)
|
| 27 |
|
| 28 |
Using one Nvidia Tesla V100 SXM2 GPU and one Intel Cascade Lake 6248 CPU,
|