CSR-III Text

Item Name: CSR-III Text
Author(s): David Graff, Roni Rosenfeld, Doug Paul
LDC Catalog No.: LDC95T6
ISBN: 1-58563-046-2
ISLRN: 080-723-177-118-4
DOI: https://doi.org/10.35111/r8va-wb28
Member Year(s): 1995, 1998
DCMI Type(s): Software, Text
Data Source(s): newswire
Project(s): DARPA-CSR
Application(s): speech recognition
Language(s): English
Language ID(s): eng
License(s): CSR III Agreement
Online Documentation: LDC95T6 Documents
Licensing Instructions: Subscription & Standard Members, and Non-Members
Citation: Graff, David, Roni Rosenfeld, and Doug Paul. CSR-III Text LDC95T6. Web Download. Philadelphia: Linguistic Data Consortium, 1995.
Related Works: View

Introduction

CSR-III Text, the third ARPA Continuous Speech Recognition (CSR) language model training data, was developed by the Linguistic Data Consortium, BBN and Carnegie Mellon University and contains English newswire text, a language model and software to generate the language model. This corpus is a companion to CSR-III Speech, the 1994 CSR benchmark speech data collection (LDC95S23). The CSR program focused on the development of speaker-independent, large-vocabulary speech recognition systems. 

Data

This release includes source newswire text and derived statistical tables of unigram, bigram and trigram word frequencies.Also included are sofrware and related scripts, supportng data files used by the text processing programs and the CMU Standard Language Model Toolkit which was used to create the langauge model. 

Updates

No updates at this time. 

Additonal Licensing Instructions

This members-only corpus is available to current members. Contact ldc@ldc.upenn.edu for information about becoming a member. 

Available Media

View Fees





Login for the applicable fee