|
|
--- |
|
|
dataset_info: |
|
|
features: |
|
|
- name: sample_id |
|
|
dtype: int64 |
|
|
- name: noise_type |
|
|
dtype: string |
|
|
- name: clean_transcription |
|
|
dtype: string |
|
|
- name: baseline_transcription |
|
|
dtype: string |
|
|
- name: grpo_transcription |
|
|
dtype: string |
|
|
- name: wer_baseline |
|
|
dtype: float32 |
|
|
- name: wer_grpo |
|
|
dtype: float32 |
|
|
- name: wer_diff |
|
|
dtype: float32 |
|
|
- name: clean_audio |
|
|
dtype: |
|
|
audio: |
|
|
sampling_rate: 16000 |
|
|
- name: baseline_predicted_audio |
|
|
dtype: |
|
|
audio: |
|
|
sampling_rate: 16000 |
|
|
- name: grpo_predicted_audio |
|
|
dtype: |
|
|
audio: |
|
|
sampling_rate: 16000 |
|
|
- name: dirty_audio |
|
|
dtype: |
|
|
audio: |
|
|
sampling_rate: 16000 |
|
|
splits: |
|
|
- name: train |
|
|
num_bytes: 11891496.0 |
|
|
num_examples: 10 |
|
|
download_size: 11152088 |
|
|
dataset_size: 11891496.0 |
|
|
configs: |
|
|
- config_name: default |
|
|
data_files: |
|
|
- split: train |
|
|
path: data/train-* |
|
|
--- |
|
|
|
