File size: 2,136 Bytes
d8485dc |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 |
---
license: mit
base_model: pyannote/segmentation-3.0
tags:
- speaker-diarization
- speaker-segmentation
- generated_from_trainer
datasets:
- diarizers-community/callhome
model-index:
- name: speaker-segmentation-fine-tuned-callhome-jpn
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# speaker-segmentation-fine-tuned-callhome-jpn
This model is a fine-tuned version of [pyannote/segmentation-3.0](https://huggingface.co/pyannote/segmentation-3.0) on the diarizers-community/callhome jpn dataset.
It achieves the following results on the evaluation set:
- Loss: 0.7512
- Der: 0.2263
- False Alarm: 0.0467
- Missed Detection: 0.1359
- Confusion: 0.0437
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.001
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- num_epochs: 5.0
### Training results
| Training Loss | Epoch | Step | Validation Loss | Der | False Alarm | Missed Detection | Confusion |
|:-------------:|:-----:|:----:|:---------------:|:------:|:-----------:|:----------------:|:---------:|
| 0.5796 | 1.0 | 328 | 0.7613 | 0.2340 | 0.0514 | 0.1341 | 0.0484 |
| 0.5518 | 2.0 | 656 | 0.7646 | 0.2328 | 0.0480 | 0.1391 | 0.0458 |
| 0.5304 | 3.0 | 984 | 0.7723 | 0.2321 | 0.0426 | 0.1421 | 0.0474 |
| 0.5043 | 4.0 | 1312 | 0.7504 | 0.2272 | 0.0490 | 0.1336 | 0.0446 |
| 0.5086 | 5.0 | 1640 | 0.7512 | 0.2263 | 0.0467 | 0.1359 | 0.0437 |
### Framework versions
- Transformers 4.40.1
- Pytorch 2.2.2+cu118
- Datasets 2.18.0
- Tokenizers 0.19.1
|