End of training
Browse files- preprocessor_config.json +1 -1
- tokenizer_config.json +7 -0
preprocessor_config.json
CHANGED
@@ -1,5 +1,5 @@
|
|
1 |
{
|
2 |
-
"apply_ocr":
|
3 |
"do_resize": true,
|
4 |
"feature_extractor_type": "LayoutLMv2FeatureExtractor",
|
5 |
"image_processor_type": "LayoutLMv2ImageProcessor",
|
|
|
1 |
{
|
2 |
+
"apply_ocr": false,
|
3 |
"do_resize": true,
|
4 |
"feature_extractor_type": "LayoutLMv2FeatureExtractor",
|
5 |
"image_processor_type": "LayoutLMv2ImageProcessor",
|
tokenizer_config.json
CHANGED
@@ -54,9 +54,11 @@
|
|
54 |
"do_basic_tokenize": true,
|
55 |
"do_lower_case": true,
|
56 |
"mask_token": "[MASK]",
|
|
|
57 |
"model_max_length": 512,
|
58 |
"never_split": null,
|
59 |
"only_label_first_subword": true,
|
|
|
60 |
"pad_token": "[PAD]",
|
61 |
"pad_token_box": [
|
62 |
0,
|
@@ -65,6 +67,8 @@
|
|
65 |
0
|
66 |
],
|
67 |
"pad_token_label": -100,
|
|
|
|
|
68 |
"processor_class": "LayoutLMv2Processor",
|
69 |
"sep_token": "[SEP]",
|
70 |
"sep_token_box": [
|
@@ -73,8 +77,11 @@
|
|
73 |
1000,
|
74 |
1000
|
75 |
],
|
|
|
76 |
"strip_accents": null,
|
77 |
"tokenize_chinese_chars": true,
|
78 |
"tokenizer_class": "LayoutLMv2Tokenizer",
|
|
|
|
|
79 |
"unk_token": "[UNK]"
|
80 |
}
|
|
|
54 |
"do_basic_tokenize": true,
|
55 |
"do_lower_case": true,
|
56 |
"mask_token": "[MASK]",
|
57 |
+
"max_length": 512,
|
58 |
"model_max_length": 512,
|
59 |
"never_split": null,
|
60 |
"only_label_first_subword": true,
|
61 |
+
"pad_to_multiple_of": null,
|
62 |
"pad_token": "[PAD]",
|
63 |
"pad_token_box": [
|
64 |
0,
|
|
|
67 |
0
|
68 |
],
|
69 |
"pad_token_label": -100,
|
70 |
+
"pad_token_type_id": 0,
|
71 |
+
"padding_side": "right",
|
72 |
"processor_class": "LayoutLMv2Processor",
|
73 |
"sep_token": "[SEP]",
|
74 |
"sep_token_box": [
|
|
|
77 |
1000,
|
78 |
1000
|
79 |
],
|
80 |
+
"stride": 0,
|
81 |
"strip_accents": null,
|
82 |
"tokenize_chinese_chars": true,
|
83 |
"tokenizer_class": "LayoutLMv2Tokenizer",
|
84 |
+
"truncation_side": "right",
|
85 |
+
"truncation_strategy": "longest_first",
|
86 |
"unk_token": "[UNK]"
|
87 |
}
|