Spaces:
Sleeping
Sleeping
Add option to choose device for the pipeline
Browse files
commafixer/src/baseline.py
CHANGED
@@ -2,8 +2,8 @@ from transformers import AutoTokenizer, AutoModelForTokenClassification, pipelin
|
|
2 |
|
3 |
|
4 |
class BaselineCommaFixer:
|
5 |
-
def __init__(self):
|
6 |
-
self._ner = _create_baseline_pipeline()
|
7 |
|
8 |
def fix_commas(self, s: str) -> str:
|
9 |
return _fix_commas_based_on_pipeline_output(
|
@@ -12,10 +12,10 @@ class BaselineCommaFixer:
|
|
12 |
)
|
13 |
|
14 |
|
15 |
-
def _create_baseline_pipeline(model_name="oliverguhr/fullstop-punctuation-multilang-large") -> NerPipeline:
|
16 |
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
17 |
model = AutoModelForTokenClassification.from_pretrained(model_name)
|
18 |
-
return pipeline('ner', model=model, tokenizer=tokenizer)
|
19 |
|
20 |
|
21 |
def _remove_punctuation(s: str) -> str:
|
|
|
2 |
|
3 |
|
4 |
class BaselineCommaFixer:
|
5 |
+
def __init__(self, device=-1):
|
6 |
+
self._ner = _create_baseline_pipeline(device=device)
|
7 |
|
8 |
def fix_commas(self, s: str) -> str:
|
9 |
return _fix_commas_based_on_pipeline_output(
|
|
|
12 |
)
|
13 |
|
14 |
|
15 |
+
def _create_baseline_pipeline(model_name="oliverguhr/fullstop-punctuation-multilang-large", device=-1) -> NerPipeline:
|
16 |
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
17 |
model = AutoModelForTokenClassification.from_pretrained(model_name)
|
18 |
+
return pipeline('ner', model=model, tokenizer=tokenizer, device=device)
|
19 |
|
20 |
|
21 |
def _remove_punctuation(s: str) -> str:
|