File size: 4,762 Bytes
c53b4c9 0c6089d c53b4c9 0c6089d c53b4c9 0c6089d c53b4c9 0c6089d c53b4c9 0c6089d c53b4c9 0c6089d c53b4c9 0c6089d |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 |
---
language:
- cs
license: apache-2.0
tags:
- automatic-speech-recognition
- mozilla-foundation/common_voice_8_0
- generated_from_trainer
- robust-speech-event
- xlsr-fine-tuning-week
datasets:
- common_voice
model-index:
- name: Czech comodoro Wav2Vec2 XLSR 300M CV8
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: Common Voice 8
type: mozilla-foundation/common_voice_8_0
args: cs
metrics:
- name: Test WER
type: wer
value: 47.46
- name: Test CER
type: cer
value: 10.88
---
<!-- 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-xls-r-300m-cs-cv8
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the common_voice 8.0 dataset.
It achieves the following results on the evaluation set:
- WER: 0.47455377483706096
- CER: 0.10877155235645618
## Model description
Fine-tuned [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) on Czech using the [Common Voice](https://huggingface.co/datasets/common_voice) dataset.
When using this model, make sure that your speech input is sampled at 16kHz.
The model can be used directly (without a language model) as follows:
```python
import torch
import torchaudio
from datasets import load_dataset
from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
test_dataset = load_dataset("mozilla-foundation/common_voice_8_0", "cs", split="test[:2%]")
processor = Wav2Vec2Processor.from_pretrained("comodoro/wav2vec2-xls-r-300m-cs-cv8")
model = Wav2Vec2ForCTC.from_pretrained("comodoro/wav2vec2-xls-r-300m-cs-cv8")
resampler = torchaudio.transforms.Resample(48_000, 16_000)
# Preprocessing the datasets.
# We need to read the aduio files as arrays
def speech_file_to_array_fn(batch):
speech_array, sampling_rate = torchaudio.load(batch["path"])
batch["speech"] = resampler(speech_array).squeeze().numpy()
return batch
test_dataset = test_dataset.map(speech_file_to_array_fn)
inputs = processor(test_dataset[:2]["speech"], sampling_rate=16_000, return_tensors="pt", padding=True)
with torch.no_grad():
logits = model(inputs.input_values, attention_mask=inputs.attention_mask).logits
predicted_ids = torch.argmax(logits, dim=-1)
print("Prediction:", processor.batch_decode(predicted_ids))
print("Reference:", test_dataset[:2]["sentence"])
```
## Evaluation
The model can be evaluated using the attached `eval.py` script.
## Training and evaluation data
The Common Voice 8.0 `train` and `validation` datasets were used for training
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 7e-05
- train_batch_size: 32
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 20
- total_train_batch_size: 640
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- num_epochs: 150
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|:-------------:|:------:|:----:|:---------------:|:------:|:------:|
| 7.2926 | 8.06 | 250 | 3.8497 | 1.0 | 1.0 |
| 3.417 | 16.13 | 500 | 3.2852 | 1.0 | 0.9857 |
| 2.0264 | 24.19 | 750 | 0.7099 | 0.7342 | 0.1768 |
| 0.4018 | 32.25 | 1000 | 0.6188 | 0.6415 | 0.1551 |
| 0.2444 | 40.32 | 1250 | 0.6632 | 0.6362 | 0.1600 |
| 0.1882 | 48.38 | 1500 | 0.6070 | 0.5783 | 0.1388 |
| 0.153 | 56.44 | 1750 | 0.6425 | 0.5720 | 0.1377 |
| 0.1214 | 64.51 | 2000 | 0.6363 | 0.5546 | 0.1337 |
| 0.1011 | 72.57 | 2250 | 0.6310 | 0.5222 | 0.1224 |
| 0.0879 | 80.63 | 2500 | 0.6353 | 0.5258 | 0.1253 |
| 0.0782 | 88.7 | 2750 | 0.6078 | 0.4904 | 0.1127 |
| 0.0709 | 96.76 | 3000 | 0.6465 | 0.4960 | 0.1154 |
| 0.0661 | 104.82 | 3250 | 0.6622 | 0.4945 | 0.1166 |
| 0.0616 | 112.89 | 3500 | 0.6440 | 0.4786 | 0.1104 |
| 0.0579 | 120.95 | 3750 | 0.6815 | 0.4887 | 0.1144 |
| 0.0549 | 129.03 | 4000 | 0.6603 | 0.4780 | 0.1105 |
| 0.0527 | 137.09 | 4250 | 0.6652 | 0.4749 | 0.1090 |
| 0.0506 | 145.16 | 4500 | 0.6958 | 0.4846 | 0.1133 |
### Framework versions
- Transformers 4.16.0.dev0
- Pytorch 1.10.1+cu102
- Datasets 1.17.1.dev0
- Tokenizers 0.11.0
|