update model card README.md
Browse files
README.md
CHANGED
|
@@ -1,9 +1,5 @@
|
|
| 1 |
---
|
| 2 |
-
language:
|
| 3 |
-
- fa
|
| 4 |
tags:
|
| 5 |
-
- automatic-speech-recognition
|
| 6 |
-
- common_voice
|
| 7 |
- generated_from_trainer
|
| 8 |
datasets:
|
| 9 |
- common_voice
|
|
@@ -17,10 +13,10 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 17 |
|
| 18 |
# common6
|
| 19 |
|
| 20 |
-
This model is a fine-tuned version of [common6/checkpoint-3500](https://huggingface.co/common6/checkpoint-3500) on the
|
| 21 |
It achieves the following results on the evaluation set:
|
| 22 |
-
- Loss: 0.
|
| 23 |
-
- Wer: 0.
|
| 24 |
|
| 25 |
## Model description
|
| 26 |
|
|
@@ -48,23 +44,33 @@ The following hyperparameters were used during training:
|
|
| 48 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 49 |
- lr_scheduler_type: linear
|
| 50 |
- lr_scheduler_warmup_steps: 100
|
| 51 |
-
- num_epochs:
|
| 52 |
- mixed_precision_training: Native AMP
|
| 53 |
|
| 54 |
### Training results
|
| 55 |
|
| 56 |
-
| Training Loss | Epoch | Step
|
| 57 |
-
|
| 58 |
-
| 1.0344 | 10.0 | 500
|
| 59 |
-
| 0.9651 | 20.0 | 1000
|
| 60 |
-
| 0.9125 | 30.0 | 1500
|
| 61 |
-
| 0.8831 | 40.0 | 2000
|
| 62 |
-
| 0.8399 | 50.0 | 2500
|
| 63 |
-
| 0.819 | 60.0 | 3000
|
| 64 |
-
| 0.8029 | 70.0 | 3500
|
| 65 |
-
| 0.
|
| 66 |
-
| 0.
|
| 67 |
-
| 0.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 68 |
|
| 69 |
|
| 70 |
### Framework versions
|
|
|
|
| 1 |
---
|
|
|
|
|
|
|
| 2 |
tags:
|
|
|
|
|
|
|
| 3 |
- generated_from_trainer
|
| 4 |
datasets:
|
| 5 |
- common_voice
|
|
|
|
| 13 |
|
| 14 |
# common6
|
| 15 |
|
| 16 |
+
This model is a fine-tuned version of [common6/checkpoint-3500](https://huggingface.co/common6/checkpoint-3500) on the common_voice dataset.
|
| 17 |
It achieves the following results on the evaluation set:
|
| 18 |
+
- Loss: 0.3706
|
| 19 |
+
- Wer: 0.3421
|
| 20 |
|
| 21 |
## Model description
|
| 22 |
|
|
|
|
| 44 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 45 |
- lr_scheduler_type: linear
|
| 46 |
- lr_scheduler_warmup_steps: 100
|
| 47 |
+
- num_epochs: 200.0
|
| 48 |
- mixed_precision_training: Native AMP
|
| 49 |
|
| 50 |
### Training results
|
| 51 |
|
| 52 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer |
|
| 53 |
+
|:-------------:|:-----:|:-----:|:---------------:|:------:|
|
| 54 |
+
| 1.0344 | 10.0 | 500 | 0.4043 | 0.4511 |
|
| 55 |
+
| 0.9651 | 20.0 | 1000 | 0.3793 | 0.4159 |
|
| 56 |
+
| 0.9125 | 30.0 | 1500 | 0.3756 | 0.4046 |
|
| 57 |
+
| 0.8831 | 40.0 | 2000 | 0.3650 | 0.3876 |
|
| 58 |
+
| 0.8399 | 50.0 | 2500 | 0.3605 | 0.3772 |
|
| 59 |
+
| 0.819 | 60.0 | 3000 | 0.3622 | 0.3714 |
|
| 60 |
+
| 0.8029 | 70.0 | 3500 | 0.3561 | 0.3664 |
|
| 61 |
+
| 0.8104 | 80.0 | 4000 | 0.3595 | 0.3660 |
|
| 62 |
+
| 0.8118 | 90.0 | 4500 | 0.3460 | 0.3592 |
|
| 63 |
+
| 0.7831 | 100.0 | 5000 | 0.3566 | 0.3593 |
|
| 64 |
+
| 0.744 | 110.0 | 5500 | 0.3578 | 0.3535 |
|
| 65 |
+
| 0.7388 | 120.0 | 6000 | 0.3538 | 0.3520 |
|
| 66 |
+
| 0.714 | 130.0 | 6500 | 0.3682 | 0.3506 |
|
| 67 |
+
| 0.7291 | 140.0 | 7000 | 0.3625 | 0.3505 |
|
| 68 |
+
| 0.697 | 150.0 | 7500 | 0.3619 | 0.3479 |
|
| 69 |
+
| 0.6811 | 160.0 | 8000 | 0.3631 | 0.3440 |
|
| 70 |
+
| 0.6841 | 170.0 | 8500 | 0.3672 | 0.3460 |
|
| 71 |
+
| 0.6616 | 180.0 | 9000 | 0.3677 | 0.3410 |
|
| 72 |
+
| 0.6471 | 190.0 | 9500 | 0.3707 | 0.3420 |
|
| 73 |
+
| 0.6759 | 200.0 | 10000 | 0.3706 | 0.3421 |
|
| 74 |
|
| 75 |
|
| 76 |
### Framework versions
|