Commit
·
90fe690
1
Parent(s):
38cfbcc
update model card README.md
Browse files
README.md
CHANGED
|
@@ -1,8 +1,7 @@
|
|
| 1 |
---
|
|
|
|
| 2 |
tags:
|
| 3 |
- generated_from_trainer
|
| 4 |
-
datasets:
|
| 5 |
-
- common_voice
|
| 6 |
model-index:
|
| 7 |
- name: test_bug2
|
| 8 |
results: []
|
|
@@ -13,7 +12,10 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 13 |
|
| 14 |
# test_bug2
|
| 15 |
|
| 16 |
-
This model is a fine-tuned version of [
|
|
|
|
|
|
|
|
|
|
| 17 |
|
| 18 |
## Model description
|
| 19 |
|
|
@@ -33,24 +35,44 @@ More information needed
|
|
| 33 |
|
| 34 |
The following hyperparameters were used during training:
|
| 35 |
- learning_rate: 0.0003
|
| 36 |
-
- train_batch_size:
|
| 37 |
- eval_batch_size: 8
|
| 38 |
- seed: 42
|
| 39 |
-
- gradient_accumulation_steps:
|
| 40 |
-
- total_train_batch_size:
|
| 41 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 42 |
- lr_scheduler_type: linear
|
| 43 |
- lr_scheduler_warmup_steps: 500
|
| 44 |
-
- num_epochs:
|
| 45 |
- mixed_precision_training: Native AMP
|
| 46 |
|
| 47 |
### Training results
|
| 48 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 49 |
|
| 50 |
|
| 51 |
### Framework versions
|
| 52 |
|
| 53 |
-
- Transformers 4.
|
| 54 |
-
- Pytorch 1.
|
| 55 |
- Datasets 1.18.3
|
| 56 |
- Tokenizers 0.12.1
|
|
|
|
| 1 |
---
|
| 2 |
+
license: cc-by-nc-4.0
|
| 3 |
tags:
|
| 4 |
- generated_from_trainer
|
|
|
|
|
|
|
| 5 |
model-index:
|
| 6 |
- name: test_bug2
|
| 7 |
results: []
|
|
|
|
| 12 |
|
| 13 |
# test_bug2
|
| 14 |
|
| 15 |
+
This model is a fine-tuned version of [nguyenvulebinh/wav2vec2-base-vietnamese-250h](https://huggingface.co/nguyenvulebinh/wav2vec2-base-vietnamese-250h) on the None dataset.
|
| 16 |
+
It achieves the following results on the evaluation set:
|
| 17 |
+
- Loss: 0.3549
|
| 18 |
+
- Wer: 0.2334
|
| 19 |
|
| 20 |
## Model description
|
| 21 |
|
|
|
|
| 35 |
|
| 36 |
The following hyperparameters were used during training:
|
| 37 |
- learning_rate: 0.0003
|
| 38 |
+
- train_batch_size: 4
|
| 39 |
- eval_batch_size: 8
|
| 40 |
- seed: 42
|
| 41 |
+
- gradient_accumulation_steps: 4
|
| 42 |
+
- total_train_batch_size: 16
|
| 43 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 44 |
- lr_scheduler_type: linear
|
| 45 |
- lr_scheduler_warmup_steps: 500
|
| 46 |
+
- num_epochs: 5
|
| 47 |
- mixed_precision_training: Native AMP
|
| 48 |
|
| 49 |
### Training results
|
| 50 |
|
| 51 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 52 |
+
|:-------------:|:-----:|:----:|:---------------:|:------:|
|
| 53 |
+
| 0.6139 | 0.27 | 50 | 0.4498 | 0.2591 |
|
| 54 |
+
| 0.3695 | 0.53 | 100 | 0.3626 | 0.2381 |
|
| 55 |
+
| 0.3065 | 0.8 | 150 | 0.3484 | 0.2281 |
|
| 56 |
+
| 0.268 | 1.07 | 200 | 0.3606 | 0.2488 |
|
| 57 |
+
| 0.282 | 1.34 | 250 | 0.3440 | 0.2409 |
|
| 58 |
+
| 0.2688 | 1.6 | 300 | 0.3707 | 0.2459 |
|
| 59 |
+
| 0.2683 | 1.87 | 350 | 0.3736 | 0.2474 |
|
| 60 |
+
| 0.2599 | 2.14 | 400 | 0.4010 | 0.2664 |
|
| 61 |
+
| 0.2683 | 2.41 | 450 | 0.3890 | 0.2627 |
|
| 62 |
+
| 0.2623 | 2.67 | 500 | 0.4109 | 0.2790 |
|
| 63 |
+
| 0.2633 | 2.94 | 550 | 0.4251 | 0.2800 |
|
| 64 |
+
| 0.2431 | 3.21 | 600 | 0.4424 | 0.2941 |
|
| 65 |
+
| 0.2263 | 3.48 | 650 | 0.4179 | 0.2677 |
|
| 66 |
+
| 0.2268 | 3.74 | 700 | 0.4049 | 0.2715 |
|
| 67 |
+
| 0.1965 | 4.01 | 750 | 0.3953 | 0.2599 |
|
| 68 |
+
| 0.1851 | 4.28 | 800 | 0.3549 | 0.2467 |
|
| 69 |
+
| 0.1724 | 4.54 | 850 | 0.3586 | 0.2450 |
|
| 70 |
+
| 0.1587 | 4.81 | 900 | 0.3549 | 0.2334 |
|
| 71 |
|
| 72 |
|
| 73 |
### Framework versions
|
| 74 |
|
| 75 |
+
- Transformers 4.16.0
|
| 76 |
+
- Pytorch 1.13.1+cu116
|
| 77 |
- Datasets 1.18.3
|
| 78 |
- Tokenizers 0.12.1
|