hungphongtrn/vietmed_asr_v3
收藏Hugging Face2024-05-31 更新2024-07-06 收录
下载链接：
https://hf-mirror.com/datasets/hungphongtrn/vietmed_asr_v3
下载链接
链接失效反馈官方服务：
资源简介：
---
dataset_info:
features:
- name: hyp
sequence: string
- name: id
dtype: string
- name: gt
sequence: string
- name: onehot_gt
sequence: string
splits:
- name: w2v2Viet_Paramshare_LossRem_WERtest_29_0
num_bytes: 1962748
num_examples: 3435
- name: XLSR53Viet_Paramshare_LossRem_WERtest28_8
num_bytes: 1965552
num_examples: 3435
- name: XLSR53Viet
num_bytes: 1963894
num_examples: 3435
download_size: 991089
dataset_size: 5892194
configs:
- config_name: default
data_files:
- split: w2v2Viet_Paramshare_LossRem_WERtest_29_0
path: data/w2v2Viet_Paramshare_LossRem_WERtest_29_0-*
- split: XLSR53Viet_Paramshare_LossRem_WERtest28_8
path: data/XLSR53Viet_Paramshare_LossRem_WERtest28_8-*
- split: XLSR53Viet
path: data/XLSR53Viet-*
---
The dataset is primarily used for speech recognition tasks, featuring four main characteristics: hyp (hypothetical string sequence), id (identifier string), gt (ground truth string sequence), and onehot_gt (one-hot encoded ground truth string sequence). The dataset is divided into three parts: w2v2Viet_Paramshare_LossRem_WERtest_29_0, XLSR53Viet_Paramshare_LossRem_WERtest28_8, and XLSR53Viet, each with specific byte counts and example quantities. The total download size of the dataset is 991089 bytes, and the total dataset size is 5892194 bytes. The configuration file defines the default configuration and specifies the path for each data file.
提供机构：
hungphongtrn
