Commit
•
a6051e6
1
Parent(s):
2dd76c6
update model card README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,88 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- ca
|
| 4 |
+
license: apache-2.0
|
| 5 |
+
tags:
|
| 6 |
+
- automatic-speech-recognition
|
| 7 |
+
- mozilla-foundation/common_voice_8_0
|
| 8 |
+
- collectivat/tv3_parla
|
| 9 |
+
- projecte-aina/parlament_parla
|
| 10 |
+
- generated_from_trainer
|
| 11 |
+
model-index:
|
| 12 |
+
- name: wav2vec2-xls-r-300m-ca
|
| 13 |
+
results: []
|
| 14 |
+
---
|
| 15 |
+
|
| 16 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 17 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 18 |
+
|
| 19 |
+
# wav2vec2-xls-r-300m-ca
|
| 20 |
+
|
| 21 |
+
This model is a fine-tuned version of [facebook/wav2vec2-xls-r-300m](https://huggingface.co/facebook/wav2vec2-xls-r-300m) on the MOZILLA-FOUNDATION/COMMON_VOICE_8_0 - CA dataset.
|
| 22 |
+
It achieves the following results on the evaluation set:
|
| 23 |
+
- Loss: 0.2758
|
| 24 |
+
- Wer: 0.1792
|
| 25 |
+
|
| 26 |
+
## Model description
|
| 27 |
+
|
| 28 |
+
More information needed
|
| 29 |
+
|
| 30 |
+
## Intended uses & limitations
|
| 31 |
+
|
| 32 |
+
More information needed
|
| 33 |
+
|
| 34 |
+
## Training and evaluation data
|
| 35 |
+
|
| 36 |
+
More information needed
|
| 37 |
+
|
| 38 |
+
## Training procedure
|
| 39 |
+
|
| 40 |
+
### Training hyperparameters
|
| 41 |
+
|
| 42 |
+
The following hyperparameters were used during training:
|
| 43 |
+
- learning_rate: 7.5e-05
|
| 44 |
+
- train_batch_size: 32
|
| 45 |
+
- eval_batch_size: 32
|
| 46 |
+
- seed: 42
|
| 47 |
+
- gradient_accumulation_steps: 4
|
| 48 |
+
- total_train_batch_size: 128
|
| 49 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 50 |
+
- lr_scheduler_type: linear
|
| 51 |
+
- lr_scheduler_warmup_steps: 2000
|
| 52 |
+
- num_epochs: 6.0
|
| 53 |
+
- mixed_precision_training: Native AMP
|
| 54 |
+
|
| 55 |
+
### Training results
|
| 56 |
+
|
| 57 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 58 |
+
|:-------------:|:-----:|:-----:|:---------------:|:------:|
|
| 59 |
+
| 6.2099 | 0.09 | 500 | 3.4125 | 1.0 |
|
| 60 |
+
| 2.9961 | 0.18 | 1000 | 2.9224 | 1.0 |
|
| 61 |
+
| 2.2147 | 0.26 | 1500 | 0.6521 | 0.5568 |
|
| 62 |
+
| 1.3017 | 0.35 | 2000 | 0.3153 | 0.2761 |
|
| 63 |
+
| 1.1196 | 0.44 | 2500 | 0.2444 | 0.2367 |
|
| 64 |
+
| 1.0712 | 0.53 | 3000 | 0.2324 | 0.2132 |
|
| 65 |
+
| 1.052 | 0.62 | 3500 | 0.2173 | 0.2032 |
|
| 66 |
+
| 1.2813 | 2.13 | 4000 | 0.3326 | 0.2099 |
|
| 67 |
+
| 1.2365 | 2.4 | 4500 | 0.3224 | 0.2003 |
|
| 68 |
+
| 1.2193 | 2.66 | 5000 | 0.3198 | 0.1957 |
|
| 69 |
+
| 1.2072 | 2.93 | 5500 | 0.3063 | 0.1933 |
|
| 70 |
+
| 1.213 | 3.2 | 6000 | 0.3051 | 0.1980 |
|
| 71 |
+
| 1.2074 | 3.46 | 6500 | 0.3012 | 0.1879 |
|
| 72 |
+
| 1.1918 | 3.73 | 7000 | 0.2947 | 0.1829 |
|
| 73 |
+
| 1.1893 | 4.0 | 7500 | 0.2895 | 0.1807 |
|
| 74 |
+
| 1.1751 | 4.26 | 8000 | 0.2878 | 0.1776 |
|
| 75 |
+
| 1.1628 | 4.53 | 8500 | 0.2835 | 0.1731 |
|
| 76 |
+
| 1.1577 | 4.79 | 9000 | 0.2816 | 0.1761 |
|
| 77 |
+
| 1.1448 | 5.06 | 9500 | 0.2757 | 0.1740 |
|
| 78 |
+
| 1.1407 | 5.33 | 10000 | 0.2768 | 0.1798 |
|
| 79 |
+
| 1.1401 | 5.59 | 10500 | 0.2780 | 0.1816 |
|
| 80 |
+
| 1.1333 | 5.86 | 11000 | 0.2748 | 0.1750 |
|
| 81 |
+
|
| 82 |
+
|
| 83 |
+
### Framework versions
|
| 84 |
+
|
| 85 |
+
- Transformers 4.16.0.dev0
|
| 86 |
+
- Pytorch 1.10.1+cu102
|
| 87 |
+
- Datasets 1.18.1
|
| 88 |
+
- Tokenizers 0.11.0
|