<!-- 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. -->
mascir_fr_hubert_version1000
This model is a fine-tuned version of facebook/hubert-large-ls960-ft on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.8026
- Wer: 0.5
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: 0.0001
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1000
- num_epochs: 50
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 5.9386 | 2.0 | 500 | 2.9031 | 0.9856 |
| 2.0513 | 4.0 | 1000 | 1.0727 | 0.9144 |
| 1.0528 | 6.0 | 1500 | 0.7645 | 0.7567 |
| 0.7915 | 8.0 | 2000 | 0.6926 | 0.6744 |
| 0.6418 | 10.0 | 2500 | 0.6881 | 0.6633 |
| 0.5558 | 12.0 | 3000 | 0.6724 | 0.5978 |
| 0.4792 | 14.0 | 3500 | 0.6674 | 0.6011 |
| 0.4236 | 16.0 | 4000 | 0.6907 | 0.5778 |
| 0.3808 | 18.0 | 4500 | 0.7231 | 0.5444 |
| 0.3364 | 20.0 | 5000 | 0.7069 | 0.5456 |
| 0.3193 | 22.0 | 5500 | 0.7189 | 0.5456 |
| 0.2827 | 24.0 | 6000 | 0.7432 | 0.5322 |
| 0.2769 | 26.0 | 6500 | 0.7838 | 0.5656 |
| 0.2543 | 28.0 | 7000 | 0.8012 | 0.5333 |
| 0.2365 | 30.0 | 7500 | 0.8180 | 0.5178 |
| 0.2274 | 32.0 | 8000 | 0.7943 | 0.5233 |
| 0.2095 | 34.0 | 8500 | 0.7664 | 0.5222 |
| 0.2055 | 36.0 | 9000 | 0.7621 | 0.5122 |
| 0.2044 | 38.0 | 9500 | 0.7712 | 0.5056 |
| 0.1946 | 40.0 | 10000 | 0.7987 | 0.4989 |
| 0.1891 | 42.0 | 10500 | 0.7978 | 0.5044 |
| 0.1878 | 44.0 | 11000 | 0.7894 | 0.4967 |
| 0.1742 | 46.0 | 11500 | 0.7964 | 0.4944 |
| 0.1701 | 48.0 | 12000 | 0.7990 | 0.4956 |
| 0.163 | 50.0 | 12500 | 0.8026 | 0.5 |
Framework versions
- Transformers 4.31.0
- Pytorch 2.0.1+cu118
- Datasets 2.14.3
- Tokenizers 0.13.3
