Update utils/vulnerability_classifier.py
Browse files
utils/vulnerability_classifier.py
CHANGED
|
@@ -8,6 +8,7 @@ from utils.preprocessing import processingpipeline
|
|
| 8 |
import streamlit as st
|
| 9 |
from transformers import pipeline
|
| 10 |
from setfit import SetFitModel
|
|
|
|
| 11 |
|
| 12 |
|
| 13 |
# labels dictionary
|
|
@@ -76,25 +77,14 @@ def load_vulnerabilityClassifier(config_file:str = None, classifier_name:str = N
|
|
| 76 |
classifier_name = config.get('vulnerability','MODEL')
|
| 77 |
|
| 78 |
logging.info("Loading vulnerability classifier")
|
| 79 |
-
|
| 80 |
-
# we are using the pipeline as the model is multilabel and DocumentClassifier
|
| 81 |
-
# from Haystack doesnt support multilabel
|
| 82 |
-
# in pipeline we use 'sigmoid' to explicitly tell pipeline to make it multilabel
|
| 83 |
-
# if not then it will automatically use softmax, which is not a desired thing.
|
| 84 |
-
# doc_classifier = TransformersDocumentClassifier(
|
| 85 |
-
# model_name_or_path=classifier_name,
|
| 86 |
-
# task="text-classification",
|
| 87 |
-
# top_k = None)
|
| 88 |
|
| 89 |
-
|
| 90 |
-
|
|
|
|
| 91 |
|
|
|
|
|
|
|
| 92 |
|
| 93 |
-
# doc_classifier = pipeline("text-classification",
|
| 94 |
-
# model=classifier_name,
|
| 95 |
-
# return_all_scores=True,
|
| 96 |
-
# function_to_apply= "sigmoid")
|
| 97 |
-
|
| 98 |
return doc_classifier
|
| 99 |
|
| 100 |
|
|
|
|
| 8 |
import streamlit as st
|
| 9 |
from transformers import pipeline
|
| 10 |
from setfit import SetFitModel
|
| 11 |
+
import os
|
| 12 |
|
| 13 |
|
| 14 |
# labels dictionary
|
|
|
|
| 77 |
classifier_name = config.get('vulnerability','MODEL')
|
| 78 |
|
| 79 |
logging.info("Loading vulnerability classifier")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 80 |
|
| 81 |
+
|
| 82 |
+
# Get token from environment variable
|
| 83 |
+
hf_token = os.environ["model_token"]
|
| 84 |
|
| 85 |
+
# Download model from HF Hub
|
| 86 |
+
doc_classifier = SetFitModel.from_pretrained(classifier_name, model_token=hf_token)
|
| 87 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 88 |
return doc_classifier
|
| 89 |
|
| 90 |
|