tedlium

Dataset Card for tedlium

Dataset Summary

The TED-LIUM corpus is English-language TED talks, with transcriptions, sampled at 16kHz. The three releases of the corpus range from 118 to 452 hours of transcribed speech data.

Example

from datasets import load_dataset

tedlium = load_dataset("LIUM/tedlium", "release1") # for Release 1

# see structure
print(tedlium)

# load audio sample on the fly
audio_input = tedlium["train"][0]["audio"] # first decoded… See the full description on the dataset page: https://huggingface.co/datasets/LIUM/tedlium.

TED-LIUM语料库为英语语言的TED演讲集，包含转录文本，采样频率为16kHz。该语料库共分为三个版本，转录语音数据时长介于118至452小时之间。

示例

从数据集中加载数据集：

python
from datasets import load_dataset

# 加载Release 1版本的TED-LIUM语料库
tedlium = load_dataset("LIUM/tedlium", "release1")

# 查看数据集结构
print(tedlium)

# 动态加载音频样本
audio_input = tedlium["train"][0]["audio"]
# 查看数据集的完整描述，请访问：https://huggingface.co/datasets/LIUM/tedlium.