<!-- 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. -->
wav2vec2-base-stac-local
This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.9746
- Wer: 0.7828
- Cer: 0.3202
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: 1
- eval_batch_size: 1
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 2
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 30
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|---|---|---|---|---|---|
| 2.0603 | 1.0 | 2369 | 2.1282 | 0.9517 | 0.5485 |
| 1.6155 | 2.0 | 4738 | 1.6196 | 0.9060 | 0.4565 |
| 1.3462 | 3.0 | 7107 | 1.4331 | 0.8379 | 0.3983 |
| 1.1819 | 4.0 | 9476 | 1.3872 | 0.8233 | 0.3717 |
| 1.0189 | 5.0 | 11845 | 1.4066 | 0.8328 | 0.3660 |
| 0.9026 | 6.0 | 14214 | 1.3502 | 0.8198 | 0.3508 |
| 0.777 | 7.0 | 16583 | 1.3016 | 0.7922 | 0.3433 |
| 0.7109 | 8.0 | 18952 | 1.2662 | 0.8302 | 0.3510 |
| 0.6766 | 9.0 | 21321 | 1.4321 | 0.8103 | 0.3368 |
| 0.6078 | 10.0 | 23690 | 1.3592 | 0.7871 | 0.3360 |
| 0.5958 | 11.0 | 26059 | 1.4389 | 0.7819 | 0.3397 |
| 0.5094 | 12.0 | 28428 | 1.3391 | 0.8017 | 0.3239 |
| 0.4567 | 13.0 | 30797 | 1.4718 | 0.8026 | 0.3347 |
| 0.4448 | 14.0 | 33166 | 1.7450 | 0.8043 | 0.3424 |
| 0.3976 | 15.0 | 35535 | 1.4581 | 0.7888 | 0.3283 |
| 0.3449 | 16.0 | 37904 | 1.5688 | 0.8078 | 0.3397 |
| 0.3046 | 17.0 | 40273 | 1.8630 | 0.8060 | 0.3448 |
| 0.2983 | 18.0 | 42642 | 1.8400 | 0.8190 | 0.3425 |
| 0.2728 | 19.0 | 45011 | 1.6726 | 0.8034 | 0.3280 |
| 0.2579 | 20.0 | 47380 | 1.6661 | 0.8138 | 0.3249 |
| 0.2169 | 21.0 | 49749 | 1.7389 | 0.8138 | 0.3277 |
| 0.2498 | 22.0 | 52118 | 1.7205 | 0.7948 | 0.3207 |
| 0.1831 | 23.0 | 54487 | 1.8641 | 0.8103 | 0.3229 |
| 0.1927 | 24.0 | 56856 | 1.8724 | 0.7784 | 0.3251 |
| 0.1649 | 25.0 | 59225 | 1.9187 | 0.7974 | 0.3277 |
| 0.1594 | 26.0 | 61594 | 1.9022 | 0.7828 | 0.3220 |
| 0.1338 | 27.0 | 63963 | 1.9303 | 0.7862 | 0.3212 |
| 0.1441 | 28.0 | 66332 | 1.9528 | 0.7845 | 0.3207 |
| 0.129 | 29.0 | 68701 | 1.9676 | 0.7819 | 0.3212 |
| 0.1169 | 30.0 | 71070 | 1.9746 | 0.7828 | 0.3202 |
Framework versions
- Transformers 4.17.0
- Pytorch 1.8.1+cu102
- Datasets 1.18.3
- Tokenizers 0.12.1
