meteor

README.md

@@ -4,7 +4,7 @@ emoji: 🤗
 colorFrom: blue
 colorTo: red
 sdk: gradio
-sdk_version: 3.19.1
+sdk_version: 3.0.2
 app_file: app.py
 pinned: false
 tags:
@@ -116,9 +116,6 @@ While the correlation between METEOR and human judgments was measured for Chines
 
 Furthermore, while the alignment and matching done in METEOR is based on unigrams, using multiple word entities (e.g. bigrams) could contribute to improving its accuracy -- this has been proposed in [more recent publications](https://www.cs.cmu.edu/~alavie/METEOR/pdf/meteor-naacl-2010.pdf) on the subject.
 
-Scores differ by up to **±10 points** across v1.0↔v1.5 and flag combinations (`-l`, `-norm`, `-vOut`). 
-Pin the Java package and document your flags. This uses the NLTK implementation (METEOR v1.0).
-[Lübbers, 2024](https://github.com/cluebbers/Reproducibility-METEOR-NLP)
 
 ## Citation
 

meteor.py

@@ -15,18 +15,12 @@
 
 import datasets
 import numpy as np
+from datasets.config import importlib_metadata, version
 from nltk.translate import meteor_score
-from packaging import version
 
 import evaluate
 
 
-if evaluate.config.PY_VERSION < version.parse("3.8"):
-    import importlib_metadata
-else:
-    import importlib.metadata as importlib_metadata
-
-
 NLTK_VERSION = version.parse(importlib_metadata.version("nltk"))
 if NLTK_VERSION >= version.Version("3.6.4"):
     from nltk import word_tokenize
@@ -120,9 +114,7 @@ class Meteor(evaluate.Metric):
         import nltk
 
         nltk.download("wordnet")
-        if NLTK_VERSION >= version.Version("3.9.0"):
-            nltk.download("punkt_tab")
-        elif NLTK_VERSION >= version.Version("3.6.5"):
+        if NLTK_VERSION >= version.Version("3.6.5"):
             nltk.download("punkt")
         if NLTK_VERSION >= version.Version("3.6.6"):
             nltk.download("omw-1.4")

requirements.txt

@@ -1,2 +1,2 @@
-git+https://github.com/huggingface/evaluate@7c4656a407213b71cb7e6f6634b7935c18f5140d
+git+https://github.com/huggingface/evaluate@940d6dee3b4a23eabb0c81e4117c9533cd7c458a
 nltk