Commit
•
6e7a45b
1
Parent(s):
464889d
update model card README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,80 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- ka
|
| 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-georgian
|
| 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-georgian
|
| 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 - KA dataset.
|
| 22 |
+
It achieves the following results on the evaluation set:
|
| 23 |
+
- Loss: 0.3666
|
| 24 |
+
- Wer: 0.4211
|
| 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: 500
|
| 50 |
+
- num_epochs: 100.0
|
| 51 |
+
- mixed_precision_training: Native AMP
|
| 52 |
+
|
| 53 |
+
### Training results
|
| 54 |
+
|
| 55 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 56 |
+
|:-------------:|:-----:|:----:|:---------------:|:------:|
|
| 57 |
+
| 1.8805 | 5.95 | 500 | 0.7547 | 0.8438 |
|
| 58 |
+
| 1.2123 | 11.9 | 1000 | 0.4732 | 0.6542 |
|
| 59 |
+
| 1.0822 | 17.86 | 1500 | 0.4027 | 0.5778 |
|
| 60 |
+
| 0.9938 | 23.81 | 2000 | 0.3847 | 0.5524 |
|
| 61 |
+
| 0.9383 | 29.76 | 2500 | 0.3845 | 0.5204 |
|
| 62 |
+
| 0.8932 | 35.71 | 3000 | 0.3833 | 0.5297 |
|
| 63 |
+
| 0.8495 | 41.67 | 3500 | 0.3759 | 0.5036 |
|
| 64 |
+
| 0.8201 | 47.62 | 4000 | 0.3616 | 0.4859 |
|
| 65 |
+
| 0.7794 | 53.57 | 4500 | 0.3874 | 0.4938 |
|
| 66 |
+
| 0.735 | 59.52 | 5000 | 0.3748 | 0.4782 |
|
| 67 |
+
| 0.7082 | 65.48 | 5500 | 0.3615 | 0.4675 |
|
| 68 |
+
| 0.669 | 71.43 | 6000 | 0.3797 | 0.4601 |
|
| 69 |
+
| 0.6457 | 77.38 | 6500 | 0.3812 | 0.4515 |
|
| 70 |
+
| 0.6098 | 83.33 | 7000 | 0.3660 | 0.4343 |
|
| 71 |
+
| 0.5874 | 89.29 | 7500 | 0.3640 | 0.4257 |
|
| 72 |
+
| 0.5627 | 95.24 | 8000 | 0.3661 | 0.4239 |
|
| 73 |
+
|
| 74 |
+
|
| 75 |
+
### Framework versions
|
| 76 |
+
|
| 77 |
+
- Transformers 4.16.0.dev0
|
| 78 |
+
- Pytorch 1.10.1+cu102
|
| 79 |
+
- Datasets 1.17.1.dev0
|
| 80 |
+
- Tokenizers 0.11.0
|