https://github.com/namansnghl/sentence-generation-with-berp
Generate sentences with Restaurant Call Transcripts
https://github.com/namansnghl/sentence-generation-with-berp
n-gram-language-models natural-language-processing
Last synced: 4 months ago
JSON representation
Generate sentences with Restaurant Call Transcripts
- Host: GitHub
- URL: https://github.com/namansnghl/sentence-generation-with-berp
- Owner: namansnghl
- License: mit
- Created: 2024-02-26T14:03:01.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-03-01T06:11:25.000Z (over 1 year ago)
- Last Synced: 2025-01-08T08:47:51.133Z (5 months ago)
- Topics: n-gram-language-models, natural-language-processing
- Language: Python
- Homepage: https://www1.icsi.berkeley.edu/Speech/berp.html
- Size: 4.29 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Sentence-Generation-with-BeRP
This repository goes into exploring the transcripts of the Berkeley Restaurant Project (BeRP) for generating random but meaningful sentences.
## Technical Details:
- **BeRP Corpus**: The [BeRP dataset](https://github.com/wooters/berp-trans) comprises transcripts of conversations related to restaurant services in Berkeley, CA. This project utilizes the BeRP transcripts as the primary data source for building the n-gram model and generating sentences. These transcripts provide a rich and diverse collection of restaurant-related dialogue.
- **n-Gram Language Modeling**: The project implements an n-gram model modeling techniques to analyze the sequential word patterns within the transcripts, which predicts the next word in a sequence based on the preceding n-1 words. Experimenting with different n-gram sizes can offer varying levels of context dependence in the generated sentences.
## Key Files:
- `lm_model.py`: This file houses the core functionality of the trained model, which utilizes the n-gram language modeling approach and BeRP transcripts.
- `testing_ngram_lm.ipynb`: This Jupyter Notebook contains comprehensive unit tests to ensure the model's correctness across various scenarios, including edge cases.
- `test_minitrainingprovided.py`: This Python file encompasses all the unit tests used to evaluate the model's performance.