update model card README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,83 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
tags:
|
| 4 |
+
- generated_from_trainer
|
| 5 |
+
model-index:
|
| 6 |
+
- name: wav2vec-base-Millad_TIMIT
|
| 7 |
+
results: []
|
| 8 |
+
---
|
| 9 |
+
|
| 10 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 11 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 12 |
+
|
| 13 |
+
# wav2vec-base-Millad_TIMIT
|
| 14 |
+
|
| 15 |
+
This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the None dataset.
|
| 16 |
+
It achieves the following results on the evaluation set:
|
| 17 |
+
- Loss: 1.1736
|
| 18 |
+
- Wer: 0.6360
|
| 19 |
+
|
| 20 |
+
## Model description
|
| 21 |
+
|
| 22 |
+
More information needed
|
| 23 |
+
|
| 24 |
+
## Intended uses & limitations
|
| 25 |
+
|
| 26 |
+
More information needed
|
| 27 |
+
|
| 28 |
+
## Training and evaluation data
|
| 29 |
+
|
| 30 |
+
More information needed
|
| 31 |
+
|
| 32 |
+
## Training procedure
|
| 33 |
+
|
| 34 |
+
### Training hyperparameters
|
| 35 |
+
|
| 36 |
+
The following hyperparameters were used during training:
|
| 37 |
+
- learning_rate: 0.0001
|
| 38 |
+
- train_batch_size: 8
|
| 39 |
+
- eval_batch_size: 8
|
| 40 |
+
- seed: 42
|
| 41 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 42 |
+
- lr_scheduler_type: linear
|
| 43 |
+
- lr_scheduler_warmup_steps: 5000
|
| 44 |
+
- num_epochs: 60
|
| 45 |
+
- mixed_precision_training: Native AMP
|
| 46 |
+
|
| 47 |
+
### Training results
|
| 48 |
+
|
| 49 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 50 |
+
|:-------------:|:-----:|:-----:|:---------------:|:------:|
|
| 51 |
+
| 3.0219 | 2.36 | 2000 | 2.4563 | 1.0141 |
|
| 52 |
+
| 0.7358 | 4.73 | 4000 | 2.1121 | 0.8905 |
|
| 53 |
+
| 0.5107 | 7.09 | 6000 | 2.0443 | 0.8587 |
|
| 54 |
+
| 0.4013 | 9.46 | 8000 | 1.9790 | 0.8799 |
|
| 55 |
+
| 0.3409 | 11.82 | 10000 | 1.9401 | 0.8216 |
|
| 56 |
+
| 0.2997 | 14.18 | 12000 | 1.7686 | 0.8375 |
|
| 57 |
+
| 0.2596 | 16.55 | 14000 | 1.6557 | 0.8604 |
|
| 58 |
+
| 0.2352 | 18.91 | 16000 | 1.5478 | 0.7562 |
|
| 59 |
+
| 0.2132 | 21.28 | 18000 | 1.6100 | 0.7385 |
|
| 60 |
+
| 0.1967 | 23.64 | 20000 | 1.3982 | 0.7650 |
|
| 61 |
+
| 0.1854 | 26.0 | 22000 | 1.3530 | 0.6837 |
|
| 62 |
+
| 0.1675 | 28.37 | 24000 | 1.3607 | 0.7120 |
|
| 63 |
+
| 0.1544 | 30.73 | 26000 | 1.1866 | 0.6979 |
|
| 64 |
+
| 0.1495 | 33.1 | 28000 | 1.2665 | 0.6943 |
|
| 65 |
+
| 0.1323 | 35.46 | 30000 | 1.4958 | 0.6820 |
|
| 66 |
+
| 0.1247 | 37.83 | 32000 | 1.2287 | 0.6431 |
|
| 67 |
+
| 0.1156 | 40.19 | 34000 | 1.3678 | 0.6749 |
|
| 68 |
+
| 0.1069 | 42.55 | 36000 | 1.2598 | 0.6396 |
|
| 69 |
+
| 0.1026 | 44.92 | 38000 | 1.1801 | 0.6979 |
|
| 70 |
+
| 0.0928 | 47.28 | 40000 | 1.2173 | 0.6378 |
|
| 71 |
+
| 0.0877 | 49.65 | 42000 | 1.2012 | 0.6572 |
|
| 72 |
+
| 0.0812 | 52.01 | 44000 | 1.0857 | 0.6237 |
|
| 73 |
+
| 0.0752 | 54.37 | 46000 | 1.2142 | 0.6254 |
|
| 74 |
+
| 0.0695 | 56.74 | 48000 | 1.1816 | 0.6343 |
|
| 75 |
+
| 0.064 | 59.1 | 50000 | 1.1736 | 0.6360 |
|
| 76 |
+
|
| 77 |
+
|
| 78 |
+
### Framework versions
|
| 79 |
+
|
| 80 |
+
- Transformers 4.17.0
|
| 81 |
+
- Pytorch 1.12.0+cu113
|
| 82 |
+
- Datasets 1.18.3
|
| 83 |
+
- Tokenizers 0.12.1
|