|
|
--- |
|
|
library_name: peft |
|
|
language: |
|
|
- it |
|
|
license: apache-2.0 |
|
|
base_model: openai/whisper-medium |
|
|
tags: |
|
|
- generated_from_trainer |
|
|
datasets: |
|
|
- b-brave-clean |
|
|
metrics: |
|
|
- wer |
|
|
model-index: |
|
|
- name: Whisper Medium |
|
|
results: |
|
|
- task: |
|
|
type: automatic-speech-recognition |
|
|
name: Automatic Speech Recognition |
|
|
dataset: |
|
|
name: b-brave-clean |
|
|
type: b-brave-clean |
|
|
config: default |
|
|
split: test |
|
|
args: default |
|
|
metrics: |
|
|
- type: wer |
|
|
value: 57.44985673352435 |
|
|
name: Wer |
|
|
--- |
|
|
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
|
|
# Whisper Medium |
|
|
|
|
|
This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the b-brave-clean dataset. |
|
|
It achieves the following results on the evaluation set: |
|
|
- Loss: 0.6278 |
|
|
- Wer: 57.4499 |
|
|
- Cer: 38.7444 |
|
|
- Lr: 0.0000 |
|
|
|
|
|
## Model description |
|
|
|
|
|
More information needed |
|
|
|
|
|
## Intended uses & limitations |
|
|
|
|
|
More information needed |
|
|
|
|
|
## Training and evaluation data |
|
|
|
|
|
More information needed |
|
|
|
|
|
## Training procedure |
|
|
|
|
|
### Training hyperparameters |
|
|
|
|
|
The following hyperparameters were used during training: |
|
|
- learning_rate: 3e-05 |
|
|
- train_batch_size: 8 |
|
|
- eval_batch_size: 8 |
|
|
- seed: 42 |
|
|
- gradient_accumulation_steps: 2 |
|
|
- total_train_batch_size: 16 |
|
|
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments |
|
|
- lr_scheduler_type: linear |
|
|
- lr_scheduler_warmup_ratio: 0.5 |
|
|
- num_epochs: 16 |
|
|
- mixed_precision_training: Native AMP |
|
|
|
|
|
### Training results |
|
|
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Wer | Cer | Lr | |
|
|
|:-------------:|:-----:|:----:|:---------------:|:--------:|:-------:|:------:| |
|
|
| 4.3003 | 1.0 | 251 | 1.8420 | 77.7937 | 47.7279 | 0.0000 | |
|
|
| 2.9936 | 2.0 | 502 | 1.2434 | 77.3639 | 47.9117 | 0.0000 | |
|
|
| 2.8135 | 3.0 | 753 | 1.1431 | 74.9284 | 47.7279 | 0.0000 | |
|
|
| 2.4032 | 4.0 | 1004 | 1.0400 | 74.6418 | 47.1237 | 0.0000 | |
|
|
| 1.9765 | 5.0 | 1255 | 0.9656 | 74.2120 | 47.2551 | 0.0000 | |
|
|
| 1.8144 | 6.0 | 1506 | 0.8685 | 107.3066 | 74.4944 | 0.0000 | |
|
|
| 1.4479 | 7.0 | 1757 | 0.8158 | 63.8968 | 42.3431 | 0.0000 | |
|
|
| 1.2718 | 8.0 | 2008 | 0.7787 | 67.1920 | 45.3901 | 0.0000 | |
|
|
| 1.135 | 9.0 | 2259 | 0.7262 | 111.1748 | 98.4502 | 0.0000 | |
|
|
| 0.833 | 10.0 | 2510 | 0.7015 | 94.4126 | 61.3869 | 0.0000 | |
|
|
| 0.7509 | 11.0 | 2761 | 0.6963 | 62.1777 | 41.7914 | 0.0000 | |
|
|
| 0.7171 | 12.0 | 3012 | 0.6650 | 61.8911 | 42.9997 | 0.0000 | |
|
|
| 0.4706 | 13.0 | 3263 | 0.6258 | 77.6504 | 57.8671 | 0.0000 | |
|
|
| 0.4988 | 14.0 | 3514 | 0.6249 | 59.5989 | 40.5306 | 0.0000 | |
|
|
| 0.4503 | 15.0 | 3765 | 0.6201 | 59.1691 | 40.0841 | 0.0000 | |
|
|
| 0.3741 | 16.0 | 4016 | 0.6278 | 57.4499 | 38.7444 | 0.0000 | |
|
|
|
|
|
|
|
|
### Framework versions |
|
|
|
|
|
- PEFT 0.14.0 |
|
|
- Transformers 4.48.3 |
|
|
- Pytorch 2.2.0 |
|
|
- Datasets 3.2.0 |
|
|
- Tokenizers 0.21.0 |
