Commit ·
d80d762
1
Parent(s): 2290244
update model card README.md
Browse files
README.md
CHANGED
|
@@ -12,6 +12,10 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 12 |
# nick_asr_v2
|
| 13 |
|
| 14 |
This model is a fine-tuned version of [ntoldalagi/nick_asr_v2](https://huggingface.co/ntoldalagi/nick_asr_v2) on an unknown dataset.
|
|
|
|
|
|
|
|
|
|
|
|
|
| 15 |
|
| 16 |
## Model description
|
| 17 |
|
|
@@ -31,11 +35,11 @@ More information needed
|
|
| 31 |
|
| 32 |
The following hyperparameters were used during training:
|
| 33 |
- learning_rate: 5e-05
|
| 34 |
-
- train_batch_size:
|
| 35 |
-
- eval_batch_size:
|
| 36 |
- seed: 42
|
| 37 |
-
- gradient_accumulation_steps:
|
| 38 |
-
- total_train_batch_size:
|
| 39 |
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
| 40 |
- lr_scheduler_type: linear
|
| 41 |
- num_epochs: 20
|
|
@@ -43,6 +47,25 @@ The following hyperparameters were used during training:
|
|
| 43 |
|
| 44 |
### Training results
|
| 45 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 46 |
|
| 47 |
|
| 48 |
### Framework versions
|
|
@@ -50,4 +73,4 @@ The following hyperparameters were used during training:
|
|
| 50 |
- Transformers 4.18.0
|
| 51 |
- Pytorch 1.10.0+cu111
|
| 52 |
- Datasets 2.0.0
|
| 53 |
-
- Tokenizers 0.
|
|
|
|
| 12 |
# nick_asr_v2
|
| 13 |
|
| 14 |
This model is a fine-tuned version of [ntoldalagi/nick_asr_v2](https://huggingface.co/ntoldalagi/nick_asr_v2) on an unknown dataset.
|
| 15 |
+
It achieves the following results on the evaluation set:
|
| 16 |
+
- Loss: 1.4562
|
| 17 |
+
- Wer: 0.6422
|
| 18 |
+
- Cer: 0.2409
|
| 19 |
|
| 20 |
## Model description
|
| 21 |
|
|
|
|
| 35 |
|
| 36 |
The following hyperparameters were used during training:
|
| 37 |
- learning_rate: 5e-05
|
| 38 |
+
- train_batch_size: 4
|
| 39 |
+
- eval_batch_size: 4
|
| 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 |
- num_epochs: 20
|
|
|
|
| 47 |
|
| 48 |
### Training results
|
| 49 |
|
| 50 |
+
| Training Loss | Epoch | Step | Cer | Validation Loss | Wer |
|
| 51 |
+
|:-------------:|:-----:|:-----:|:------:|:---------------:|:------:|
|
| 52 |
+
| 0.2616 | 0.44 | 300 | 0.2905 | 1.2200 | 0.7496 |
|
| 53 |
+
| 0.441 | 0.87 | 600 | 0.2866 | 1.1936 | 0.7385 |
|
| 54 |
+
| 0.4366 | 1.31 | 900 | 0.2795 | 1.1584 | 0.7274 |
|
| 55 |
+
| 0.3982 | 1.75 | 1200 | 0.2808 | 1.2033 | 0.7274 |
|
| 56 |
+
| 0.3891 | 2.18 | 1500 | 0.2753 | 1.2044 | 0.7166 |
|
| 57 |
+
| 0.3508 | 2.91 | 2000 | 1.2382 | 0.7220 | 0.2743 |
|
| 58 |
+
| 0.2783 | 4.37 | 3000 | 1.3327 | 0.7177 | 0.2705 |
|
| 59 |
+
| 0.2495 | 5.82 | 4000 | 1.2286 | 0.6749 | 0.2638 |
|
| 60 |
+
| 0.1982 | 7.28 | 5000 | 1.3073 | 0.6721 | 0.2585 |
|
| 61 |
+
| 0.1717 | 8.73 | 6000 | 1.2941 | 0.6627 | 0.2500 |
|
| 62 |
+
| 0.1508 | 10.19 | 7000 | 1.3625 | 0.6584 | 0.2490 |
|
| 63 |
+
| 0.1329 | 11.64 | 8000 | 1.3863 | 0.6584 | 0.2474 |
|
| 64 |
+
| 0.1303 | 13.1 | 9000 | 1.3714 | 0.6534 | 0.2449 |
|
| 65 |
+
| 0.1159 | 14.56 | 10000 | 1.4043 | 0.6473 | 0.2442 |
|
| 66 |
+
| 0.1015 | 16.01 | 11000 | 1.4245 | 0.6498 | 0.2419 |
|
| 67 |
+
| 0.098 | 17.47 | 12000 | 1.4410 | 0.6440 | 0.2425 |
|
| 68 |
+
| 0.0869 | 18.92 | 13000 | 1.4562 | 0.6422 | 0.2409 |
|
| 69 |
|
| 70 |
|
| 71 |
### Framework versions
|
|
|
|
| 73 |
- Transformers 4.18.0
|
| 74 |
- Pytorch 1.10.0+cu111
|
| 75 |
- Datasets 2.0.0
|
| 76 |
+
- Tokenizers 0.12.1
|