ARTICLEMetrics for text generation NLP models evaluation

Covers: implementation of NLP Model evaluation
Why this is worth your time
Overview and codes of common evaluation metrics being used in practical examples of text generation models.
How to use this item?

Read the entire article. Make use of the provided Python codes as well.

Author(s) / creator(s) / reference(s)
Divish Dayal
Shortlist
publicShare

NLP Model Evaluation

Luan NguyenTotal time needed: ~18 minutes
Learning Objectives
This list provides basic information about metrics relate to NLP model evaluation.
Potential Use Cases
Understand and build NLP models evaluation
Target Audience
BEGINNER
Go through the following annotated items in order:
BOOK_CHAPTER 1. Extrinsic vs. Intrinsic
It provides a good overview and comparison between the two types of model evaluation.
5 minutes
VIDEO 2. Perplexity
It explains the basic concept of perplexity which is an important metrics in evaluating NLP models.
9 minutes
BOOK_CHAPTER 3. Entropy
It provides a definition of entropy.
10 minutes
VIDEO 4. Cross entropy
It provides clear explaination of entropy and cross entropy and how this loss metrics is used in ML.
10 minutes
ARTICLE 5. Relationship between perplexity and entropy
It provides a good explanation of how perplexity and entorpy are related.
10 minutes
ARTICLE 6. Metrics for text generation NLP models evaluation
Overview and codes of common evaluation metrics being used in practical examples of text generation models.
10 minutes
ARTICLE 7. Metrics for text comparison NLP models evaluation
It provides an overview of common evaluation metrics to compare generated text and target text of NLP models.
4 minutes

Concepts Convered