pierreguillou
commited on
Commit
•
ed9f2e9
1
Parent(s):
ea1aefa
Update app.py
Browse files
app.py
CHANGED
@@ -80,6 +80,15 @@ feature_extractor = LayoutLMv2FeatureExtractor(apply_ocr=False)
|
|
80 |
from transformers import AutoTokenizer
|
81 |
tokenizer_layoutxlm = AutoTokenizer.from_pretrained(tokenizer_id_layoutxlm)
|
82 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
83 |
# APP outputs by model
|
84 |
def app_outputs_by_model(uploaded_pdf, model_id, model, tokenizer, max_length, id2label, cls_box, sep_box):
|
85 |
filename, msg, images = pdf_to_images(uploaded_pdf)
|
|
|
80 |
from transformers import AutoTokenizer
|
81 |
tokenizer_layoutxlm = AutoTokenizer.from_pretrained(tokenizer_id_layoutxlm)
|
82 |
|
83 |
+
# get labels
|
84 |
+
id2label_lilt = model_lilt.config.id2label
|
85 |
+
label2id_lilt = model_lilt.config.label2id
|
86 |
+
num_labels_lilt = len(id2label_lilt)
|
87 |
+
|
88 |
+
id2label_layoutxlm = model_layoutxlm.config.id2label
|
89 |
+
label2id_layoutxlm = model_layoutxlm.config.label2id
|
90 |
+
num_labels_layoutxlm = len(id2label_layoutxlm)
|
91 |
+
|
92 |
# APP outputs by model
|
93 |
def app_outputs_by_model(uploaded_pdf, model_id, model, tokenizer, max_length, id2label, cls_box, sep_box):
|
94 |
filename, msg, images = pdf_to_images(uploaded_pdf)
|