|
--- |
|
title: iBLEU |
|
emoji: 📊 |
|
colorFrom: red |
|
colorTo: indigo |
|
sdk: gradio |
|
sdk_version: 3.16.1 |
|
app_file: app.py |
|
pinned: false |
|
--- |
|
|
|
iBLEU measures the adequacy and dissimilarity of generated paraphrases. |
|
|
|
Example: |
|
``` |
|
>>> inputs = ["greetings general kenobi", "foo foo bar bar"] |
|
>>> predictions = ["hello there general kenobi", "foo bar foobar"] |
|
>>> references = [["hello there general kenobi", "hello there !"], ["foo bar foobar", "foo bar foobar"]] |
|
>>> ibleu = evaluate.load("rahular/ibleu") |
|
>>> results = ibleu.compute(inputs=inputs, predictions=predictions, references=references) |
|
>>> print(results) |
|
{'score': 60.41585343630594} |
|
``` |
