Commit
•
b67dd98
1
Parent(s):
f110a79
update model card README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,84 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- ba
|
| 4 |
+
license: apache-2.0
|
| 5 |
+
tags:
|
| 6 |
+
- automatic-speech-recognition
|
| 7 |
+
- mozilla-foundation/common_voice_7_0
|
| 8 |
+
- generated_from_trainer
|
| 9 |
+
datasets:
|
| 10 |
+
- common_voice
|
| 11 |
+
model-index:
|
| 12 |
+
- name: wav2vec2-large-xls-r-300m-bashkir
|
| 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-large-xls-r-300m-bashkir
|
| 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_7_0 - BA dataset.
|
| 22 |
+
It achieves the following results on the evaluation set:
|
| 23 |
+
- Loss: 0.1892
|
| 24 |
+
- Wer: 0.2421
|
| 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: 0.0003
|
| 44 |
+
- train_batch_size: 32
|
| 45 |
+
- eval_batch_size: 32
|
| 46 |
+
- seed: 42
|
| 47 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 48 |
+
- lr_scheduler_type: linear
|
| 49 |
+
- lr_scheduler_warmup_steps: 2000
|
| 50 |
+
- num_epochs: 10.0
|
| 51 |
+
- mixed_precision_training: Native AMP
|
| 52 |
+
|
| 53 |
+
### Training results
|
| 54 |
+
|
| 55 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 56 |
+
|:-------------:|:-----:|:-----:|:---------------:|:------:|
|
| 57 |
+
| 1.4792 | 0.5 | 2000 | 0.4598 | 0.5404 |
|
| 58 |
+
| 1.449 | 1.0 | 4000 | 0.4650 | 0.5610 |
|
| 59 |
+
| 1.3742 | 1.49 | 6000 | 0.4001 | 0.4977 |
|
| 60 |
+
| 1.3375 | 1.99 | 8000 | 0.3916 | 0.4894 |
|
| 61 |
+
| 1.2961 | 2.49 | 10000 | 0.3641 | 0.4569 |
|
| 62 |
+
| 1.2714 | 2.99 | 12000 | 0.3491 | 0.4488 |
|
| 63 |
+
| 1.2399 | 3.48 | 14000 | 0.3151 | 0.3986 |
|
| 64 |
+
| 1.2067 | 3.98 | 16000 | 0.3081 | 0.3923 |
|
| 65 |
+
| 1.1842 | 4.48 | 18000 | 0.2875 | 0.3703 |
|
| 66 |
+
| 1.1644 | 4.98 | 20000 | 0.2840 | 0.3670 |
|
| 67 |
+
| 1.161 | 5.48 | 22000 | 0.2790 | 0.3597 |
|
| 68 |
+
| 1.1303 | 5.97 | 24000 | 0.2552 | 0.3272 |
|
| 69 |
+
| 1.0874 | 6.47 | 26000 | 0.2405 | 0.3142 |
|
| 70 |
+
| 1.0613 | 6.97 | 28000 | 0.2352 | 0.3055 |
|
| 71 |
+
| 1.0498 | 7.47 | 30000 | 0.2249 | 0.2910 |
|
| 72 |
+
| 1.021 | 7.96 | 32000 | 0.2118 | 0.2752 |
|
| 73 |
+
| 1.0002 | 8.46 | 34000 | 0.2046 | 0.2662 |
|
| 74 |
+
| 0.9762 | 8.96 | 36000 | 0.1969 | 0.2530 |
|
| 75 |
+
| 0.9568 | 9.46 | 38000 | 0.1917 | 0.2449 |
|
| 76 |
+
| 0.953 | 9.96 | 40000 | 0.1893 | 0.2425 |
|
| 77 |
+
|
| 78 |
+
|
| 79 |
+
### Framework versions
|
| 80 |
+
|
| 81 |
+
- Transformers 4.16.0.dev0
|
| 82 |
+
- Pytorch 1.10.1+cu102
|
| 83 |
+
- Datasets 1.17.1.dev0
|
| 84 |
+
- Tokenizers 0.11.0
|