Skip to main content

Approximate the WER of an ASR transcript

Project description

# Word Error Rate for automatic speech recognition

This repository contains a simple python package to approximate the WER of a transcript. It computes the minimum-edit distance
between the ground-truth sentence and the hypothesis sentence of a speech-to-text API. The minimum-edit distance is calculated
using the
[Wagner-Fisher](https://en.wikipedia.org/wiki/Wagner%E2%80%93Fischer_algorithm)
algorithm. Because this algorithm computes the character-level minimum-edit distance, every word in a sentence is assigned a
unique integer, and the edit-distance is computed over a string of integers.

# Installation

You should be able to install this package using pip:

```bash
$ pip install jiwer
```

# Usage

The most simple use-case is computing the edit distance between two strings:

```python
from jiwer import wer

ground_truth = "hello world"
hypothesis = "hello duck"

error = wer(ground_truth, hypothesis)
```

You can also compute the WER over multiple sentences:

```python
from jiwer import wer

ground_truth = ["hello world", "i like monthy python"]
hypothesis = ["hello duck", "i like python"]

error = wer(ground_truth, hypothesis)
```

When the amount of ground-truth sentences and hypothesis sentences differ, a minimum alignment is done over the merged sentence:

```python
ground_truth = ["hello world", "i like monthy python", "what do you mean, african or european swallow?"]
hypothesis = ["hello", "i like", "python", "what you mean swallow"]

# is equivelent to

ground_truth = "hello world i like monhty python what do you mean african or european swallow"
hypothesis = "hello i like python what you mean swallow"
```

# Additional preprocessing

Some additional preprocessing can be done on the input. By default, whitespace is removed, everything is set to lower-case,
`.` and `,` are removed, everything between `[]` and `<>` (common for Kaldi models) is removed and each word is tokenized by
splitting by one or more spaces. Additionally, common abbreviations, such as `won't`, `let's`,`n't` will be expanded if
`standardize=True` is passed along the `wer` method.

```python
from jiwer import wer

ground_truth = "he's my neminis"
hypothesis = "he is my <unk> [laughter]"

wer(ground_truth, hypothesis, standardize=True)

# is equivelent to

ground_truth = "he is my neminis"
hypothesis = "he is my"

wer(ground_truth, hypothesis)
```

Also, there is an option give a list of words to remove from the
transcription, such as "yhe", or "so".

```python
from jiwer import wer

ground_truth = "yhe about that bug"
hypothesis = "yeah about that bug"

wer(ground_truth, hypothesis, words_to_filter=["yhe", "yeah"])

# is equivelent to

ground_truth = "about that bug"
hypothesis = "about that bug"

wer(ground_truth, hypothesis)

```





Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

jiwer-1.2.tar.gz (5.2 kB view hashes)

Uploaded Source

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page