|
--- |
|
base_model: facebook/w2v-bert-2.0 |
|
language: eve |
|
tags: |
|
- generated_from_trainer |
|
datasets: |
|
- audiofolder |
|
metrics: |
|
- wer |
|
- cer |
|
model-index: |
|
- name: wav2vec-bert-2.0-even-pakendorf |
|
results: |
|
- task: |
|
name: Automatic Speech Recognition |
|
type: automatic-speech-recognition |
|
dataset: |
|
name: audiofolder |
|
type: audiofolder |
|
config: default |
|
split: train |
|
args: default |
|
metrics: |
|
- name: Wer |
|
type: wer |
|
value: 0.5968606805108706 |
|
--- |
|
|
|
<!-- 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. --> |
|
|
|
# wav2vec-bert-2.0-even-pakendorf-0406-1347 |
|
|
|
This model is a fine-tuned version of [facebook/w2v-bert-2.0](https://huggingface.co/facebook/w2v-bert-2.0) on the audiofolder dataset. |
|
It achieves the following results on the evaluation set: |
|
- Cer: 0.2128 |
|
- Loss: inf |
|
- Wer: 0.5969 |
|
|
|
## Model description |
|
|
|
``` |
|
from transformers import AutoModelForCTC, Wav2Vec2BertProcessor |
|
|
|
model = AutoModelForCTC.from_pretrained("tbkazakova/wav2vec-bert-2.0-even-pakendorf") |
|
processor = Wav2Vec2BertProcessor.from_pretrained("tbkazakova/wav2vec-bert-2.0-even-pakendorf") |
|
|
|
data, sampling_rate = librosa.load('audio.wav') |
|
librosa.resample(data, orig_sr=sampling_rate, target_sr=16000) |
|
logits = model(torch.tensor(processor(data, |
|
sampling_rate=16000).input_features[0]).unsqueeze(0)).logits |
|
|
|
pred_ids = torch.argmax(logits, dim=-1)[0] |
|
print(processor.decode(pred_ids)) |
|
``` |
|
|
|
## 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: 5e-05 |
|
- train_batch_size: 8 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- gradient_accumulation_steps: 2 |
|
- total_train_batch_size: 16 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_steps: 500 |
|
- num_epochs: 10 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Cer | Validation Loss | Wer | |
|
|:-------------:|:------:|:----:|:------:|:---------------:|:------:| |
|
| 4.5767 | 0.5051 | 200 | 0.4932 | inf | 0.9973 | |
|
| 1.8775 | 1.0101 | 400 | 0.3211 | inf | 0.8494 | |
|
| 1.6006 | 1.5152 | 600 | 0.3017 | inf | 0.8040 | |
|
| 1.4476 | 2.0202 | 800 | 0.2896 | inf | 0.7534 | |
|
| 1.2213 | 2.5253 | 1000 | 0.2610 | inf | 0.7080 | |
|
| 1.1485 | 3.0303 | 1200 | 0.2684 | inf | 0.6800 | |
|
| 0.9554 | 3.5354 | 1400 | 0.2459 | inf | 0.6732 | |
|
| 0.9379 | 4.0404 | 1600 | 0.2275 | inf | 0.6251 | |
|
| 0.7644 | 4.5455 | 1800 | 0.2235 | inf | 0.6224 | |
|
| 0.7891 | 5.0505 | 2000 | 0.2180 | inf | 0.6053 | |
|
| 0.633 | 5.5556 | 2200 | 0.2130 | inf | 0.5996 | |
|
| 0.6197 | 6.0606 | 2400 | 0.2126 | inf | 0.6032 | |
|
| 0.5212 | 6.5657 | 2600 | 0.2196 | inf | 0.6019 | |
|
| 0.4881 | 7.0707 | 2800 | 0.2125 | inf | 0.5894 | |
|
| 0.4 | 7.5758 | 3000 | 0.2066 | inf | 0.5852 | |
|
| 0.4008 | 8.0808 | 3200 | 0.2076 | inf | 0.5790 | |
|
| 0.3304 | 8.5859 | 3400 | 0.2096 | inf | 0.5884 | |
|
| 0.3446 | 9.0909 | 3600 | 0.2124 | inf | 0.5983 | |
|
| 0.3237 | 9.5960 | 3800 | 0.2128 | inf | 0.5969 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.41.2 |
|
- Pytorch 2.3.0+cu121 |
|
- Datasets 2.19.2 |
|
- Tokenizers 0.19.1 |