|
--- |
|
library_name: transformers |
|
language: |
|
- de |
|
license: apache-2.0 |
|
base_model: mkenfenheuer/whisper-small-de-v2 |
|
tags: |
|
- generated_from_trainer |
|
- automatic-speech-recognition |
|
- german |
|
datasets: |
|
- mkenfenheuer/whisper-small-de-v2 |
|
metrics: |
|
- wer |
|
model-index: |
|
- name: Whisper Small DE v3.0 - Maximilian Kenfenheuer |
|
results: |
|
- task: |
|
name: Automatic Speech Recognition |
|
type: automatic-speech-recognition |
|
dataset: |
|
name: Common Voice 17.0 |
|
type: mkenfenheuer/whisper-small-de-v2 |
|
config: de |
|
split: test |
|
args: 'config: de, split: validated' |
|
metrics: |
|
- name: Wer |
|
type: wer |
|
value: 2.451703737799017 |
|
--- |
|
|
|
This model is a converted version of [mkenfenheuer/whisper-small-de-v3](https://huggingface.co/mkenfenheuer/whisper-small-de-v3/) converted to ctranslate2 with int8 quantization. |
|
|
|
<!-- 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 Small DE v3.0 - Maximilian Kenfenheuer |
|
|
|
This model is a fine-tuned version of [mkenfenheuer/whisper-small-de-v2](https://huggingface.co/mkenfenheuer/whisper-small-de-v2) on the Common Voice 17.0 dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.0402 |
|
- Wer: 2.4517 |
|
|
|
## 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: 1e-05 |
|
- train_batch_size: 16 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- 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 |
|
- training_steps: 1001 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Wer | |
|
|:-------------:|:------:|:----:|:---------------:|:------:| |
|
| 0.0694 | 0.9990 | 1000 | 0.0402 | 2.4517 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.48.0 |
|
- Pytorch 2.5.1+cu121 |
|
- Datasets 3.2.0 |
|
- Tokenizers 0.21.0 |
|
|
