Commit
•
3491e56
1
Parent(s):
dfa054d
update model card README.md
Browse files
README.md
CHANGED
|
@@ -1,9 +1,6 @@
|
|
| 1 |
---
|
| 2 |
-
license: apache-2.0
|
| 3 |
tags:
|
| 4 |
- generated_from_trainer
|
| 5 |
-
metrics:
|
| 6 |
-
- wer
|
| 7 |
model-index:
|
| 8 |
- name: wav2vec2-large-asr-th
|
| 9 |
results: []
|
|
@@ -14,11 +11,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 14 |
|
| 15 |
# wav2vec2-large-asr-th
|
| 16 |
|
| 17 |
-
This model
|
| 18 |
-
It achieves the following results on the evaluation set:
|
| 19 |
-
- Loss: 0.5920
|
| 20 |
-
- Wer: 0.5256
|
| 21 |
-
- Cer: 0.1778
|
| 22 |
|
| 23 |
## Model description
|
| 24 |
|
|
@@ -37,39 +30,21 @@ More information needed
|
|
| 37 |
### Training hyperparameters
|
| 38 |
|
| 39 |
The following hyperparameters were used during training:
|
| 40 |
-
- learning_rate: 0.
|
| 41 |
-
- train_batch_size:
|
| 42 |
-
- eval_batch_size:
|
| 43 |
- seed: 42
|
| 44 |
- gradient_accumulation_steps: 2
|
| 45 |
-
- total_train_batch_size:
|
| 46 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 47 |
- lr_scheduler_type: linear
|
| 48 |
-
- lr_scheduler_warmup_steps:
|
| 49 |
-
- training_steps:
|
| 50 |
- mixed_precision_training: Native AMP
|
| 51 |
|
| 52 |
-
### Training results
|
| 53 |
-
|
| 54 |
-
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|
| 55 |
-
|:-------------:|:-----:|:----:|:---------------:|:------:|:------:|
|
| 56 |
-
| 3.6977 | 0.57 | 500 | 3.6087 | 1.0 | 0.9999 |
|
| 57 |
-
| 3.4771 | 1.14 | 1000 | 3.4975 | 1.0 | 0.9999 |
|
| 58 |
-
| 2.6942 | 1.71 | 1500 | 2.2851 | 1.0067 | 0.6597 |
|
| 59 |
-
| 1.698 | 2.28 | 2000 | 1.0650 | 0.7875 | 0.3045 |
|
| 60 |
-
| 1.5008 | 2.85 | 2500 | 0.8698 | 0.6838 | 0.2508 |
|
| 61 |
-
| 1.1706 | 3.42 | 3000 | 0.7382 | 0.6132 | 0.2140 |
|
| 62 |
-
| 1.1872 | 4.0 | 3500 | 0.6924 | 0.5840 | 0.2029 |
|
| 63 |
-
| 1.1422 | 4.57 | 4000 | 0.6531 | 0.5690 | 0.1959 |
|
| 64 |
-
| 0.9556 | 5.14 | 4500 | 0.6246 | 0.5432 | 0.1850 |
|
| 65 |
-
| 1.0091 | 5.71 | 5000 | 0.6052 | 0.5360 | 0.1822 |
|
| 66 |
-
| 1.0523 | 6.28 | 5500 | 0.5995 | 0.5293 | 0.1802 |
|
| 67 |
-
| 1.0205 | 6.85 | 6000 | 0.5920 | 0.5256 | 0.1778 |
|
| 68 |
-
|
| 69 |
-
|
| 70 |
### Framework versions
|
| 71 |
|
| 72 |
- Transformers 4.26.1
|
| 73 |
- Pytorch 1.13.1+cu116
|
| 74 |
-
- Datasets 2.
|
| 75 |
- Tokenizers 0.13.2
|
|
|
|
| 1 |
---
|
|
|
|
| 2 |
tags:
|
| 3 |
- generated_from_trainer
|
|
|
|
|
|
|
| 4 |
model-index:
|
| 5 |
- name: wav2vec2-large-asr-th
|
| 6 |
results: []
|
|
|
|
| 11 |
|
| 12 |
# wav2vec2-large-asr-th
|
| 13 |
|
| 14 |
+
This model was trained from scratch on the None dataset.
|
|
|
|
|
|
|
|
|
|
|
|
|
| 15 |
|
| 16 |
## Model description
|
| 17 |
|
|
|
|
| 30 |
### Training hyperparameters
|
| 31 |
|
| 32 |
The following hyperparameters were used during training:
|
| 33 |
+
- learning_rate: 0.0002
|
| 34 |
+
- train_batch_size: 24
|
| 35 |
+
- eval_batch_size: 16
|
| 36 |
- seed: 42
|
| 37 |
- gradient_accumulation_steps: 2
|
| 38 |
+
- total_train_batch_size: 48
|
| 39 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 40 |
- lr_scheduler_type: linear
|
| 41 |
+
- lr_scheduler_warmup_steps: 300
|
| 42 |
+
- training_steps: 3000
|
| 43 |
- mixed_precision_training: Native AMP
|
| 44 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 45 |
### Framework versions
|
| 46 |
|
| 47 |
- Transformers 4.26.1
|
| 48 |
- Pytorch 1.13.1+cu116
|
| 49 |
+
- Datasets 2.10.1
|
| 50 |
- Tokenizers 0.13.2
|