Model save
Browse files
README.md
CHANGED
|
@@ -4,6 +4,8 @@ license: apache-2.0
|
|
| 4 |
base_model: openai/whisper-large-v3
|
| 5 |
tags:
|
| 6 |
- generated_from_trainer
|
|
|
|
|
|
|
| 7 |
model-index:
|
| 8 |
- name: whisper-large-v3-lora-bn-en-banking
|
| 9 |
results: []
|
|
@@ -15,6 +17,10 @@ should probably proofread and complete it, then remove this comment. -->
|
|
| 15 |
# whisper-large-v3-lora-bn-en-banking
|
| 16 |
|
| 17 |
This model is a fine-tuned version of [openai/whisper-large-v3](https://huggingface.co/openai/whisper-large-v3) on an unknown dataset.
|
|
|
|
|
|
|
|
|
|
|
|
|
| 18 |
|
| 19 |
## Model description
|
| 20 |
|
|
@@ -33,20 +39,34 @@ More information needed
|
|
| 33 |
### Training hyperparameters
|
| 34 |
|
| 35 |
The following hyperparameters were used during training:
|
| 36 |
-
- learning_rate:
|
| 37 |
-
- train_batch_size:
|
| 38 |
-
- eval_batch_size:
|
| 39 |
- seed: 42
|
| 40 |
- gradient_accumulation_steps: 8
|
| 41 |
-
- total_train_batch_size:
|
| 42 |
-
- optimizer: Use OptimizerNames.
|
|
|
|
| 43 |
- lr_scheduler_type: linear
|
| 44 |
- lr_scheduler_warmup_ratio: 0.1
|
| 45 |
-
- num_epochs:
|
| 46 |
- mixed_precision_training: Native AMP
|
| 47 |
|
| 48 |
### Training results
|
| 49 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 50 |
|
| 51 |
|
| 52 |
### Framework versions
|
|
|
|
| 4 |
base_model: openai/whisper-large-v3
|
| 5 |
tags:
|
| 6 |
- generated_from_trainer
|
| 7 |
+
metrics:
|
| 8 |
+
- wer
|
| 9 |
model-index:
|
| 10 |
- name: whisper-large-v3-lora-bn-en-banking
|
| 11 |
results: []
|
|
|
|
| 17 |
# whisper-large-v3-lora-bn-en-banking
|
| 18 |
|
| 19 |
This model is a fine-tuned version of [openai/whisper-large-v3](https://huggingface.co/openai/whisper-large-v3) on an unknown dataset.
|
| 20 |
+
It achieves the following results on the evaluation set:
|
| 21 |
+
- Loss: 4.6914
|
| 22 |
+
- Wer: 98.5075
|
| 23 |
+
- Cer: 94.7368
|
| 24 |
|
| 25 |
## Model description
|
| 26 |
|
|
|
|
| 39 |
### Training hyperparameters
|
| 40 |
|
| 41 |
The following hyperparameters were used during training:
|
| 42 |
+
- learning_rate: 1e-05
|
| 43 |
+
- train_batch_size: 1
|
| 44 |
+
- eval_batch_size: 8
|
| 45 |
- seed: 42
|
| 46 |
- gradient_accumulation_steps: 8
|
| 47 |
+
- total_train_batch_size: 8
|
| 48 |
+
- optimizer: Use OptimizerNames.ADAFACTOR and the args are:
|
| 49 |
+
No additional optimizer arguments
|
| 50 |
- lr_scheduler_type: linear
|
| 51 |
- lr_scheduler_warmup_ratio: 0.1
|
| 52 |
+
- num_epochs: 100
|
| 53 |
- mixed_precision_training: Native AMP
|
| 54 |
|
| 55 |
### Training results
|
| 56 |
|
| 57 |
+
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer |
|
| 58 |
+
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|
|
| 59 |
+
| 2.3917 | 1.0 | 1 | 4.8753 | 98.5075 | 94.7368 |
|
| 60 |
+
| 2.3917 | 2.0 | 2 | 4.8739 | 98.5075 | 94.7368 |
|
| 61 |
+
| 2.3911 | 3.0 | 3 | 4.8704 | 98.5075 | 94.7368 |
|
| 62 |
+
| 2.3893 | 4.0 | 4 | 4.8605 | 98.5075 | 94.7368 |
|
| 63 |
+
| 2.3856 | 5.0 | 5 | 4.8488 | 98.5075 | 94.7368 |
|
| 64 |
+
| 2.3801 | 6.0 | 6 | 4.8313 | 98.5075 | 94.7368 |
|
| 65 |
+
| 2.3728 | 7.0 | 7 | 4.8101 | 98.5075 | 94.7368 |
|
| 66 |
+
| 2.3642 | 8.0 | 8 | 4.7857 | 98.5075 | 94.7368 |
|
| 67 |
+
| 2.354 | 9.0 | 9 | 4.7571 | 98.5075 | 94.7368 |
|
| 68 |
+
| 2.3421 | 10.0 | 10 | 4.7262 | 98.5075 | 94.7368 |
|
| 69 |
+
| 2.3302 | 11.0 | 11 | 4.6914 | 98.5075 | 94.7368 |
|
| 70 |
|
| 71 |
|
| 72 |
### Framework versions
|