File size: 4,679 Bytes
bf05986 |
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 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 |
---
license: apache-2.0
base_model: facebook/wav2vec2-large-xlsr-53
tags:
- automatic-speech-recognition
- ./sample_speech.py
- generated_from_trainer
metrics:
- wer
model-index:
- name: en-xlsr
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. -->
# en-xlsr
This model is a fine-tuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on the ./SAMPLE_SPEECH.PY - NA dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4835
- Cer: 0.1119
- Wer: 0.2446
## 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: 0.0003
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 2
- total_train_batch_size: 32
- total_eval_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 10
- num_epochs: 10
### Training results
| Training Loss | Epoch | Step | Validation Loss | Cer | Wer |
|:-------------:|:-----:|:----:|:---------------:|:------:|:------:|
| 2.9534 | 0.22 | 100 | 2.9533 | 1.0 | 1.0 |
| 2.933 | 0.44 | 200 | 2.9231 | 1.0 | 1.0 |
| 2.904 | 0.65 | 300 | 2.8851 | 1.0 | 1.0 |
| 2.3607 | 0.87 | 400 | 2.1546 | 0.6799 | 0.9976 |
| 1.1725 | 1.09 | 500 | 0.9899 | 0.2665 | 0.6191 |
| 0.9865 | 1.31 | 600 | 0.8060 | 0.2126 | 0.5064 |
| 0.8959 | 1.53 | 700 | 0.7131 | 0.1980 | 0.4607 |
| 0.7743 | 1.74 | 800 | 0.6663 | 0.1799 | 0.4370 |
| 0.7805 | 1.96 | 900 | 0.6159 | 0.1683 | 0.3997 |
| 0.6562 | 2.18 | 1000 | 0.6186 | 0.1537 | 0.3705 |
| 0.6223 | 2.4 | 1100 | 0.5698 | 0.1496 | 0.3552 |
| 0.5627 | 2.62 | 1200 | 0.5555 | 0.1446 | 0.3372 |
| 0.5476 | 2.84 | 1300 | 0.5435 | 0.1416 | 0.3307 |
| 0.5002 | 3.05 | 1400 | 0.5304 | 0.1436 | 0.3393 |
| 0.5174 | 3.27 | 1500 | 0.5377 | 0.1485 | 0.3357 |
| 0.4745 | 3.49 | 1600 | 0.5289 | 0.1340 | 0.3132 |
| 0.5239 | 3.71 | 1700 | 0.5112 | 0.1395 | 0.3239 |
| 0.5115 | 3.93 | 1800 | 0.5079 | 0.1342 | 0.3094 |
| 0.4471 | 4.14 | 1900 | 0.5131 | 0.1301 | 0.2965 |
| 0.4455 | 4.36 | 2000 | 0.5015 | 0.1278 | 0.2931 |
| 0.4199 | 4.58 | 2100 | 0.4954 | 0.1299 | 0.2962 |
| 0.4699 | 4.8 | 2200 | 0.4827 | 0.1268 | 0.2890 |
| 0.3521 | 5.02 | 2300 | 0.4857 | 0.1217 | 0.2782 |
| 0.3976 | 5.23 | 2400 | 0.4936 | 0.1231 | 0.2802 |
| 0.365 | 5.45 | 2500 | 0.4906 | 0.1221 | 0.2774 |
| 0.3857 | 5.67 | 2600 | 0.4843 | 0.1202 | 0.2757 |
| 0.3578 | 5.89 | 2700 | 0.4857 | 0.1196 | 0.2708 |
| 0.3298 | 6.11 | 2800 | 0.4867 | 0.1197 | 0.2689 |
| 0.3099 | 6.32 | 2900 | 0.4924 | 0.1237 | 0.2770 |
| 0.3606 | 6.54 | 3000 | 0.4851 | 0.1189 | 0.2684 |
| 0.3807 | 6.76 | 3100 | 0.4700 | 0.1196 | 0.2656 |
| 0.3286 | 6.98 | 3200 | 0.4770 | 0.1205 | 0.2730 |
| 0.3318 | 7.2 | 3300 | 0.4845 | 0.1166 | 0.2579 |
| 0.2936 | 7.42 | 3400 | 0.4909 | 0.1159 | 0.2570 |
| 0.3119 | 7.63 | 3500 | 0.4899 | 0.1150 | 0.2539 |
| 0.3142 | 7.85 | 3600 | 0.4782 | 0.1143 | 0.2550 |
| 0.2935 | 8.07 | 3700 | 0.4885 | 0.1153 | 0.2527 |
| 0.2805 | 8.29 | 3800 | 0.4906 | 0.1143 | 0.2529 |
| 0.254 | 8.51 | 3900 | 0.4822 | 0.1144 | 0.2538 |
| 0.2855 | 8.72 | 4000 | 0.4852 | 0.1123 | 0.2476 |
| 0.2661 | 8.94 | 4100 | 0.4847 | 0.1132 | 0.2496 |
| 0.2524 | 9.16 | 4200 | 0.4900 | 0.1116 | 0.2442 |
| 0.253 | 9.38 | 4300 | 0.4888 | 0.1120 | 0.2458 |
| 0.2591 | 9.6 | 4400 | 0.4813 | 0.1125 | 0.2458 |
| 0.2583 | 9.81 | 4500 | 0.4844 | 0.1114 | 0.2435 |
### Framework versions
- Transformers 4.34.0
- Pytorch 2.1.0+cu121
- Datasets 2.14.5
- Tokenizers 0.14.1
|