update model card README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,85 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
tags:
|
| 4 |
+
- generated_from_trainer
|
| 5 |
+
datasets:
|
| 6 |
+
- common_voice
|
| 7 |
+
model-index:
|
| 8 |
+
- name: wav2vec2-xls-r-300m-ab-CV8
|
| 9 |
+
results: []
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 13 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 14 |
+
|
| 15 |
+
# wav2vec2-xls-r-300m-ab-CV8
|
| 16 |
+
|
| 17 |
+
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 dataset.
|
| 18 |
+
It achieves the following results on the evaluation set:
|
| 19 |
+
- Loss: 0.2105
|
| 20 |
+
- Wer: 0.5474
|
| 21 |
+
|
| 22 |
+
## Model description
|
| 23 |
+
|
| 24 |
+
More information needed
|
| 25 |
+
|
| 26 |
+
## Intended uses & limitations
|
| 27 |
+
|
| 28 |
+
More information needed
|
| 29 |
+
|
| 30 |
+
## Training and evaluation data
|
| 31 |
+
|
| 32 |
+
More information needed
|
| 33 |
+
|
| 34 |
+
## Training procedure
|
| 35 |
+
|
| 36 |
+
### Training hyperparameters
|
| 37 |
+
|
| 38 |
+
The following hyperparameters were used during training:
|
| 39 |
+
- learning_rate: 0.0001
|
| 40 |
+
- train_batch_size: 16
|
| 41 |
+
- eval_batch_size: 8
|
| 42 |
+
- seed: 42
|
| 43 |
+
- gradient_accumulation_steps: 2
|
| 44 |
+
- total_train_batch_size: 32
|
| 45 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 46 |
+
- lr_scheduler_type: linear
|
| 47 |
+
- lr_scheduler_warmup_steps: 300
|
| 48 |
+
- num_epochs: 15
|
| 49 |
+
- mixed_precision_training: Native AMP
|
| 50 |
+
|
| 51 |
+
### Training results
|
| 52 |
+
|
| 53 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 54 |
+
|:-------------:|:-----:|:-----:|:---------------:|:------:|
|
| 55 |
+
| 4.7729 | 0.63 | 500 | 3.0624 | 1.0021 |
|
| 56 |
+
| 2.7348 | 1.26 | 1000 | 1.0460 | 0.9815 |
|
| 57 |
+
| 1.2756 | 1.9 | 1500 | 0.4618 | 0.8309 |
|
| 58 |
+
| 1.0419 | 2.53 | 2000 | 0.3725 | 0.7449 |
|
| 59 |
+
| 0.9491 | 3.16 | 2500 | 0.3368 | 0.7345 |
|
| 60 |
+
| 0.9006 | 3.79 | 3000 | 0.3014 | 0.6936 |
|
| 61 |
+
| 0.8519 | 4.42 | 3500 | 0.2852 | 0.6767 |
|
| 62 |
+
| 0.8243 | 5.06 | 4000 | 0.2701 | 0.6504 |
|
| 63 |
+
| 0.7902 | 5.69 | 4500 | 0.2641 | 0.6221 |
|
| 64 |
+
| 0.7767 | 6.32 | 5000 | 0.2549 | 0.6192 |
|
| 65 |
+
| 0.7516 | 6.95 | 5500 | 0.2515 | 0.6179 |
|
| 66 |
+
| 0.737 | 7.59 | 6000 | 0.2408 | 0.5963 |
|
| 67 |
+
| 0.7217 | 8.22 | 6500 | 0.2429 | 0.6261 |
|
| 68 |
+
| 0.7101 | 8.85 | 7000 | 0.2366 | 0.5687 |
|
| 69 |
+
| 0.6922 | 9.48 | 7500 | 0.2277 | 0.5680 |
|
| 70 |
+
| 0.6866 | 10.11 | 8000 | 0.2242 | 0.5847 |
|
| 71 |
+
| 0.6703 | 10.75 | 8500 | 0.2222 | 0.5803 |
|
| 72 |
+
| 0.6649 | 11.38 | 9000 | 0.2247 | 0.5765 |
|
| 73 |
+
| 0.6513 | 12.01 | 9500 | 0.2182 | 0.5644 |
|
| 74 |
+
| 0.6369 | 12.64 | 10000 | 0.2128 | 0.5508 |
|
| 75 |
+
| 0.6425 | 13.27 | 10500 | 0.2132 | 0.5514 |
|
| 76 |
+
| 0.6399 | 13.91 | 11000 | 0.2116 | 0.5495 |
|
| 77 |
+
| 0.6208 | 14.54 | 11500 | 0.2105 | 0.5474 |
|
| 78 |
+
|
| 79 |
+
|
| 80 |
+
### Framework versions
|
| 81 |
+
|
| 82 |
+
- Transformers 4.11.3
|
| 83 |
+
- Pytorch 1.10.0+cu111
|
| 84 |
+
- Datasets 1.18.1
|
| 85 |
+
- Tokenizers 0.10.3
|