Update README.md
Browse files
README.md
CHANGED
|
@@ -15,24 +15,9 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 15 |
|
| 16 |
# Fine-tuning whisper
|
| 17 |
|
| 18 |
-
This model is a fine-tuned version of [openai/whisper-base](https://huggingface.co/openai/whisper-base) on the
|
| 19 |
-
|
| 20 |
-
- Loss: 0.0042
|
| 21 |
-
- Cer: 0.0
|
| 22 |
|
| 23 |
-
## Model description
|
| 24 |
-
|
| 25 |
-
More information needed
|
| 26 |
-
|
| 27 |
-
## Intended uses & limitations
|
| 28 |
-
|
| 29 |
-
More information needed
|
| 30 |
-
|
| 31 |
-
## Training and evaluation data
|
| 32 |
-
|
| 33 |
-
More information needed
|
| 34 |
-
|
| 35 |
-
## Training procedure
|
| 36 |
|
| 37 |
### Training hyperparameters
|
| 38 |
|
|
@@ -47,31 +32,6 @@ The following hyperparameters were used during training:
|
|
| 47 |
- training_steps: 100
|
| 48 |
- mixed_precision_training: Native AMP
|
| 49 |
|
| 50 |
-
### Training results
|
| 51 |
-
|
| 52 |
-
| Training Loss | Epoch | Step | Validation Loss | Cer |
|
| 53 |
-
|:-------------:|:-----:|:----:|:---------------:|:--------:|
|
| 54 |
-
| 4.2772 | 0.28 | 5 | 4.1382 | 173.1771 |
|
| 55 |
-
| 4.1353 | 0.56 | 10 | 3.2786 | 104.9479 |
|
| 56 |
-
| 2.5873 | 0.83 | 15 | 1.6016 | 47.3958 |
|
| 57 |
-
| 1.0092 | 1.11 | 20 | 0.6032 | 22.9167 |
|
| 58 |
-
| 0.3746 | 1.39 | 25 | 0.3756 | 16.1458 |
|
| 59 |
-
| 0.0944 | 1.67 | 30 | 0.1149 | 7.8125 |
|
| 60 |
-
| 0.0215 | 1.94 | 35 | 0.0449 | 2.6042 |
|
| 61 |
-
| 0.0038 | 2.22 | 40 | 0.0180 | 1.8229 |
|
| 62 |
-
| 0.0013 | 2.5 | 45 | 0.0118 | 0.7812 |
|
| 63 |
-
| 0.0011 | 2.78 | 50 | 0.0101 | 0.7812 |
|
| 64 |
-
| 0.0004 | 3.06 | 55 | 0.0081 | 0.2604 |
|
| 65 |
-
| 0.0003 | 3.33 | 60 | 0.0068 | 0.2604 |
|
| 66 |
-
| 0.0003 | 3.61 | 65 | 0.0058 | 0.2604 |
|
| 67 |
-
| 0.0002 | 3.89 | 70 | 0.0052 | 0.0 |
|
| 68 |
-
| 0.0002 | 4.17 | 75 | 0.0048 | 0.0 |
|
| 69 |
-
| 0.0002 | 4.44 | 80 | 0.0045 | 0.0 |
|
| 70 |
-
| 0.0002 | 4.72 | 85 | 0.0044 | 0.0 |
|
| 71 |
-
| 0.0002 | 5.0 | 90 | 0.0043 | 0.0 |
|
| 72 |
-
| 0.0002 | 5.28 | 95 | 0.0043 | 0.0 |
|
| 73 |
-
| 0.0002 | 5.56 | 100 | 0.0042 | 0.0 |
|
| 74 |
-
|
| 75 |
|
| 76 |
### Framework versions
|
| 77 |
|
|
|
|
| 15 |
|
| 16 |
# Fine-tuning whisper
|
| 17 |
|
| 18 |
+
This model is a fine-tuned version of [openai/whisper-base](https://huggingface.co/openai/whisper-base) on the TTS dataset
|
| 19 |
+
for Capstone Design II project in Sogang University.
|
|
|
|
|
|
|
| 20 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
|
| 22 |
### Training hyperparameters
|
| 23 |
|
|
|
|
| 32 |
- training_steps: 100
|
| 33 |
- mixed_precision_training: Native AMP
|
| 34 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 35 |
|
| 36 |
### Framework versions
|
| 37 |
|