Spaces:
Runtime error
Runtime error
davebulaval
commited on
Commit
Β·
7217d6a
1
Parent(s):
1982c24
uniformization of interface and add .to for tokenizer output
Browse files- meaningbert.py +14 -16
meaningbert.py
CHANGED
@@ -64,8 +64,8 @@ _KWARGS_DESCRIPTION = """
|
|
64 |
MeaningBERT metric for assessing meaning preservation between sentences.
|
65 |
|
66 |
Args:
|
67 |
-
|
68 |
-
|
69 |
device (str): Device to use for model inference. By default, set to "cuda".
|
70 |
|
71 |
Returns:
|
@@ -75,10 +75,10 @@ Returns:
|
|
75 |
|
76 |
Examples:
|
77 |
|
78 |
-
>>>
|
79 |
-
>>>
|
80 |
>>> meaning_bert = evaluate.load("davebulaval/meaningbert", device="cuda:0")
|
81 |
-
>>> results = meaning_bert.compute(
|
82 |
"""
|
83 |
|
84 |
_HASH = "21845c0cc85a2e8e16c89bb0053f489095cf64c5b19e9c3865d3e10047aba51b"
|
@@ -110,19 +110,17 @@ class MeaningBERT(evaluate.Metric):
|
|
110 |
|
111 |
def _compute(
|
112 |
self,
|
113 |
-
|
114 |
-
|
115 |
device: str = "cuda",
|
116 |
) -> Dict:
|
117 |
-
assert len(
|
118 |
-
|
119 |
-
), "The number of
|
120 |
hashcode = _HASH
|
121 |
|
122 |
# Index of sentence with perfect match between two sentences
|
123 |
-
matching_index = [
|
124 |
-
i for i, item in enumerate(documents) if item in simplifications
|
125 |
-
]
|
126 |
|
127 |
# We load the MeaningBERT pretrained model
|
128 |
scorer = AutoModelForSequenceClassification.from_pretrained(
|
@@ -135,12 +133,12 @@ class MeaningBERT(evaluate.Metric):
|
|
135 |
|
136 |
# We tokenize the text as a pair and return Pytorch Tensors
|
137 |
tokenize_text = tokenizer(
|
138 |
-
|
139 |
-
|
140 |
truncation=True,
|
141 |
padding=True,
|
142 |
return_tensors="pt",
|
143 |
-
)
|
144 |
|
145 |
with filter_logging_context():
|
146 |
# We process the text
|
|
|
64 |
MeaningBERT metric for assessing meaning preservation between sentences.
|
65 |
|
66 |
Args:
|
67 |
+
references (list of str): References sentences.
|
68 |
+
predictions (list of str): Predictions sentences (same number of element as documents).
|
69 |
device (str): Device to use for model inference. By default, set to "cuda".
|
70 |
|
71 |
Returns:
|
|
|
75 |
|
76 |
Examples:
|
77 |
|
78 |
+
>>> references = ["hello there", "general kenobi"]
|
79 |
+
>>> predictions = ["hello there", "general kenobi"]
|
80 |
>>> meaning_bert = evaluate.load("davebulaval/meaningbert", device="cuda:0")
|
81 |
+
>>> results = meaning_bert.compute(references=references, predictions=predictions)
|
82 |
"""
|
83 |
|
84 |
_HASH = "21845c0cc85a2e8e16c89bb0053f489095cf64c5b19e9c3865d3e10047aba51b"
|
|
|
110 |
|
111 |
def _compute(
|
112 |
self,
|
113 |
+
references: List,
|
114 |
+
predictions: List,
|
115 |
device: str = "cuda",
|
116 |
) -> Dict:
|
117 |
+
assert len(references) == len(
|
118 |
+
predictions
|
119 |
+
), "The number of references is different of the number of predictions."
|
120 |
hashcode = _HASH
|
121 |
|
122 |
# Index of sentence with perfect match between two sentences
|
123 |
+
matching_index = [i for i, item in enumerate(references) if item in predictions]
|
|
|
|
|
124 |
|
125 |
# We load the MeaningBERT pretrained model
|
126 |
scorer = AutoModelForSequenceClassification.from_pretrained(
|
|
|
133 |
|
134 |
# We tokenize the text as a pair and return Pytorch Tensors
|
135 |
tokenize_text = tokenizer(
|
136 |
+
references,
|
137 |
+
predictions,
|
138 |
truncation=True,
|
139 |
padding=True,
|
140 |
return_tensors="pt",
|
141 |
+
).to(device)
|
142 |
|
143 |
with filter_logging_context():
|
144 |
# We process the text
|