Commit
•
78d1fc9
1
Parent(s):
2a2d52d
Training in progress, epoch 1
Browse files- README.md +77 -0
- config.json +44 -0
- preprocessor_config.json +14 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +7 -0
- tokenizer.json +0 -0
- tokenizer_config.json +38 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
README.md
CHANGED
@@ -1,3 +1,80 @@
|
|
1 |
---
|
2 |
license: mit
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
license: mit
|
3 |
+
base_model: microsoft/layoutlm-base-uncased
|
4 |
+
tags:
|
5 |
+
- generated_from_trainer
|
6 |
+
datasets:
|
7 |
+
- funsd
|
8 |
+
model-index:
|
9 |
+
- name: layoutlm-funsd
|
10 |
+
results: []
|
11 |
---
|
12 |
+
|
13 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
14 |
+
should probably proofread and complete it, then remove this comment. -->
|
15 |
+
|
16 |
+
# layoutlm-funsd
|
17 |
+
|
18 |
+
This model is a fine-tuned version of [microsoft/layoutlm-base-uncased](https://huggingface.co/microsoft/layoutlm-base-uncased) on the funsd dataset.
|
19 |
+
It achieves the following results on the evaluation set:
|
20 |
+
- Loss: 1.0844
|
21 |
+
- Answer: {'precision': 0.3143100511073254, 'recall': 0.4561186650185414, 'f1': 0.3721633888048412, 'number': 809}
|
22 |
+
- Header: {'precision': 0.275, 'recall': 0.18487394957983194, 'f1': 0.22110552763819097, 'number': 119}
|
23 |
+
- Question: {'precision': 0.44804716285924834, 'recall': 0.5708920187793427, 'f1': 0.5020644095788603, 'number': 1065}
|
24 |
+
- Overall Precision: 0.3826
|
25 |
+
- Overall Recall: 0.5013
|
26 |
+
- Overall F1: 0.4340
|
27 |
+
- Overall Accuracy: 0.5793
|
28 |
+
|
29 |
+
## Model description
|
30 |
+
|
31 |
+
More information needed
|
32 |
+
|
33 |
+
## Intended uses & limitations
|
34 |
+
|
35 |
+
More information needed
|
36 |
+
|
37 |
+
## Training and evaluation data
|
38 |
+
|
39 |
+
More information needed
|
40 |
+
|
41 |
+
## Training procedure
|
42 |
+
|
43 |
+
### Training hyperparameters
|
44 |
+
|
45 |
+
The following hyperparameters were used during training:
|
46 |
+
- learning_rate: 3e-05
|
47 |
+
- train_batch_size: 32
|
48 |
+
- eval_batch_size: 16
|
49 |
+
- seed: 42
|
50 |
+
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
|
51 |
+
- lr_scheduler_type: linear
|
52 |
+
- num_epochs: 15
|
53 |
+
|
54 |
+
### Training results
|
55 |
+
|
56 |
+
| Training Loss | Epoch | Step | Validation Loss | Answer | Header | Question | Overall Precision | Overall Recall | Overall F1 | Overall Accuracy |
|
57 |
+
|:-------------:|:-----:|:----:|:---------------:|:--------------------------------------------------------------------------------------------------------------:|:-----------------------------------------------------------------------------------------------------------:|:------------------------------------------------------------------------------------------------------------:|:-----------------:|:--------------:|:----------:|:----------------:|
|
58 |
+
| 1.7974 | 1.0 | 5 | 1.6082 | {'precision': 0.015957446808510637, 'recall': 0.003708281829419036, 'f1': 0.006018054162487462, 'number': 809} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} | {'precision': 0.13218390804597702, 'recall': 0.0215962441314554, 'f1': 0.037126715092816794, 'number': 1065} | 0.0718 | 0.0130 | 0.0221 | 0.2950 |
|
59 |
+
| 1.6031 | 2.0 | 10 | 1.4809 | {'precision': 0.09702549575070822, 'recall': 0.16934487021013597, 'f1': 0.12336785231877535, 'number': 809} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} | {'precision': 0.2448499117127722, 'recall': 0.39061032863849765, 'f1': 0.301013024602026, 'number': 1065} | 0.1778 | 0.2775 | 0.2167 | 0.3926 |
|
60 |
+
| 1.4415 | 3.0 | 15 | 1.3965 | {'precision': 0.15503875968992248, 'recall': 0.32138442521631644, 'f1': 0.20917135961383748, 'number': 809} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} | {'precision': 0.25329341317365267, 'recall': 0.3971830985915493, 'f1': 0.3093235831809872, 'number': 1065} | 0.2041 | 0.3427 | 0.2558 | 0.4162 |
|
61 |
+
| 1.3417 | 4.0 | 20 | 1.2882 | {'precision': 0.1925233644859813, 'recall': 0.3819530284301607, 'f1': 0.25600662800331403, 'number': 809} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} | {'precision': 0.2921832884097035, 'recall': 0.5089201877934272, 'f1': 0.3712328767123288, 'number': 1065} | 0.2457 | 0.4270 | 0.3120 | 0.4305 |
|
62 |
+
| 1.2673 | 5.0 | 25 | 1.2461 | {'precision': 0.2402555910543131, 'recall': 0.4647713226205192, 'f1': 0.3167649536647009, 'number': 809} | {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119} | {'precision': 0.3362183754993342, 'recall': 0.47417840375586856, 'f1': 0.39345539540319435, 'number': 1065} | 0.2828 | 0.4420 | 0.3449 | 0.4621 |
|
63 |
+
| 1.1953 | 6.0 | 30 | 1.1667 | {'precision': 0.2396469789545146, 'recall': 0.4363411619283066, 'f1': 0.30937773882559155, 'number': 809} | {'precision': 0.1038961038961039, 'recall': 0.06722689075630252, 'f1': 0.08163265306122448, 'number': 119} | {'precision': 0.34711246200607904, 'recall': 0.536150234741784, 'f1': 0.42140221402214023, 'number': 1065} | 0.2917 | 0.4676 | 0.3593 | 0.5048 |
|
64 |
+
| 1.1257 | 7.0 | 35 | 1.1238 | {'precision': 0.271211022480058, 'recall': 0.4622991347342398, 'f1': 0.34186471663619744, 'number': 809} | {'precision': 0.17708333333333334, 'recall': 0.14285714285714285, 'f1': 0.15813953488372096, 'number': 119} | {'precision': 0.38812154696132595, 'recall': 0.5276995305164319, 'f1': 0.4472741742936729, 'number': 1065} | 0.3260 | 0.4782 | 0.3877 | 0.5539 |
|
65 |
+
| 1.0703 | 8.0 | 40 | 1.0882 | {'precision': 0.2758340113913751, 'recall': 0.41903584672435107, 'f1': 0.33267909715407257, 'number': 809} | {'precision': 0.1919191919191919, 'recall': 0.15966386554621848, 'f1': 0.17431192660550457, 'number': 119} | {'precision': 0.4045307443365696, 'recall': 0.5868544600938967, 'f1': 0.47892720306513414, 'number': 1065} | 0.3422 | 0.4932 | 0.4040 | 0.5809 |
|
66 |
+
| 1.0172 | 9.0 | 45 | 1.0768 | {'precision': 0.277602523659306, 'recall': 0.43510506798516685, 'f1': 0.3389504092441021, 'number': 809} | {'precision': 0.24096385542168675, 'recall': 0.16806722689075632, 'f1': 0.19801980198019803, 'number': 119} | {'precision': 0.40967092008059103, 'recall': 0.5727699530516432, 'f1': 0.4776820673453407, 'number': 1065} | 0.3458 | 0.4927 | 0.4064 | 0.5803 |
|
67 |
+
| 0.9713 | 10.0 | 50 | 1.0884 | {'precision': 0.3041700735895339, 'recall': 0.45982694684796044, 'f1': 0.3661417322834645, 'number': 809} | {'precision': 0.2631578947368421, 'recall': 0.16806722689075632, 'f1': 0.20512820512820512, 'number': 119} | {'precision': 0.4506024096385542, 'recall': 0.5267605633802817, 'f1': 0.4857142857142857, 'number': 1065} | 0.3746 | 0.4782 | 0.4201 | 0.5781 |
|
68 |
+
| 0.9434 | 11.0 | 55 | 1.1220 | {'precision': 0.29082426127527217, 'recall': 0.4622991347342398, 'f1': 0.35704057279236273, 'number': 809} | {'precision': 0.2727272727272727, 'recall': 0.17647058823529413, 'f1': 0.21428571428571427, 'number': 119} | {'precision': 0.4404934687953556, 'recall': 0.5699530516431925, 'f1': 0.4969300040933278, 'number': 1065} | 0.3656 | 0.5028 | 0.4233 | 0.5669 |
|
69 |
+
| 0.9288 | 12.0 | 60 | 1.0876 | {'precision': 0.298372513562387, 'recall': 0.4079110012360939, 'f1': 0.34464751958224543, 'number': 809} | {'precision': 0.23958333333333334, 'recall': 0.19327731092436976, 'f1': 0.21395348837209302, 'number': 119} | {'precision': 0.4299933642999336, 'recall': 0.6084507042253521, 'f1': 0.5038880248833593, 'number': 1065} | 0.3695 | 0.5023 | 0.4258 | 0.5784 |
|
70 |
+
| 0.9043 | 13.0 | 65 | 1.1185 | {'precision': 0.31703204047217537, 'recall': 0.4647713226205192, 'f1': 0.3769423558897243, 'number': 809} | {'precision': 0.2894736842105263, 'recall': 0.18487394957983194, 'f1': 0.22564102564102564, 'number': 119} | {'precision': 0.4605263157894737, 'recall': 0.5258215962441315, 'f1': 0.49101271372205174, 'number': 1065} | 0.3866 | 0.4807 | 0.4285 | 0.5679 |
|
71 |
+
| 0.8884 | 14.0 | 70 | 1.1097 | {'precision': 0.31260364842454397, 'recall': 0.46600741656365885, 'f1': 0.37419354838709684, 'number': 809} | {'precision': 0.29333333333333333, 'recall': 0.18487394957983194, 'f1': 0.2268041237113402, 'number': 119} | {'precision': 0.4597791798107255, 'recall': 0.5474178403755868, 'f1': 0.4997856836690956, 'number': 1065} | 0.3852 | 0.4927 | 0.4324 | 0.5710 |
|
72 |
+
| 0.8759 | 15.0 | 75 | 1.0844 | {'precision': 0.3143100511073254, 'recall': 0.4561186650185414, 'f1': 0.3721633888048412, 'number': 809} | {'precision': 0.275, 'recall': 0.18487394957983194, 'f1': 0.22110552763819097, 'number': 119} | {'precision': 0.44804716285924834, 'recall': 0.5708920187793427, 'f1': 0.5020644095788603, 'number': 1065} | 0.3826 | 0.5013 | 0.4340 | 0.5793 |
|
73 |
+
|
74 |
+
|
75 |
+
### Framework versions
|
76 |
+
|
77 |
+
- Transformers 4.33.0
|
78 |
+
- Pytorch 2.0.1+cu118
|
79 |
+
- Datasets 2.14.5
|
80 |
+
- Tokenizers 0.13.3
|
config.json
ADDED
@@ -0,0 +1,44 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "microsoft/layoutlm-base-uncased",
|
3 |
+
"architectures": [
|
4 |
+
"LayoutLMForTokenClassification"
|
5 |
+
],
|
6 |
+
"attention_probs_dropout_prob": 0.1,
|
7 |
+
"hidden_act": "gelu",
|
8 |
+
"hidden_dropout_prob": 0.1,
|
9 |
+
"hidden_size": 768,
|
10 |
+
"id2label": {
|
11 |
+
"0": "O",
|
12 |
+
"1": "B-HEADER",
|
13 |
+
"2": "I-HEADER",
|
14 |
+
"3": "B-QUESTION",
|
15 |
+
"4": "I-QUESTION",
|
16 |
+
"5": "B-ANSWER",
|
17 |
+
"6": "I-ANSWER"
|
18 |
+
},
|
19 |
+
"initializer_range": 0.02,
|
20 |
+
"intermediate_size": 3072,
|
21 |
+
"label2id": {
|
22 |
+
"B-ANSWER": 5,
|
23 |
+
"B-HEADER": 1,
|
24 |
+
"B-QUESTION": 3,
|
25 |
+
"I-ANSWER": 6,
|
26 |
+
"I-HEADER": 2,
|
27 |
+
"I-QUESTION": 4,
|
28 |
+
"O": 0
|
29 |
+
},
|
30 |
+
"layer_norm_eps": 1e-12,
|
31 |
+
"max_2d_position_embeddings": 1024,
|
32 |
+
"max_position_embeddings": 512,
|
33 |
+
"model_type": "layoutlm",
|
34 |
+
"num_attention_heads": 12,
|
35 |
+
"num_hidden_layers": 12,
|
36 |
+
"output_past": true,
|
37 |
+
"pad_token_id": 0,
|
38 |
+
"position_embedding_type": "absolute",
|
39 |
+
"torch_dtype": "float32",
|
40 |
+
"transformers_version": "4.33.0",
|
41 |
+
"type_vocab_size": 2,
|
42 |
+
"use_cache": true,
|
43 |
+
"vocab_size": 30522
|
44 |
+
}
|
preprocessor_config.json
ADDED
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"apply_ocr": true,
|
3 |
+
"do_resize": true,
|
4 |
+
"feature_extractor_type": "LayoutLMv2FeatureExtractor",
|
5 |
+
"image_processor_type": "LayoutLMv2ImageProcessor",
|
6 |
+
"ocr_lang": null,
|
7 |
+
"processor_class": "LayoutLMv2Processor",
|
8 |
+
"resample": 2,
|
9 |
+
"size": {
|
10 |
+
"height": 224,
|
11 |
+
"width": 224
|
12 |
+
},
|
13 |
+
"tesseract_config": ""
|
14 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c1e900a711707a090d4d8ed042a8c9002f840b888eb7aa3ee4b8e6ad2135ebe3
|
3 |
+
size 450603969
|
special_tokens_map.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"cls_token": "[CLS]",
|
3 |
+
"mask_token": "[MASK]",
|
4 |
+
"pad_token": "[PAD]",
|
5 |
+
"sep_token": "[SEP]",
|
6 |
+
"unk_token": "[UNK]"
|
7 |
+
}
|
tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer_config.json
ADDED
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"additional_special_tokens": null,
|
3 |
+
"apply_ocr": false,
|
4 |
+
"clean_up_tokenization_spaces": true,
|
5 |
+
"cls_token": "[CLS]",
|
6 |
+
"cls_token_box": [
|
7 |
+
0,
|
8 |
+
0,
|
9 |
+
0,
|
10 |
+
0
|
11 |
+
],
|
12 |
+
"do_basic_tokenize": true,
|
13 |
+
"do_lower_case": true,
|
14 |
+
"mask_token": "[MASK]",
|
15 |
+
"model_max_length": 512,
|
16 |
+
"never_split": null,
|
17 |
+
"only_label_first_subword": true,
|
18 |
+
"pad_token": "[PAD]",
|
19 |
+
"pad_token_box": [
|
20 |
+
0,
|
21 |
+
0,
|
22 |
+
0,
|
23 |
+
0
|
24 |
+
],
|
25 |
+
"pad_token_label": -100,
|
26 |
+
"processor_class": "LayoutLMv2Processor",
|
27 |
+
"sep_token": "[SEP]",
|
28 |
+
"sep_token_box": [
|
29 |
+
1000,
|
30 |
+
1000,
|
31 |
+
1000,
|
32 |
+
1000
|
33 |
+
],
|
34 |
+
"strip_accents": null,
|
35 |
+
"tokenize_chinese_chars": true,
|
36 |
+
"tokenizer_class": "LayoutLMv2Tokenizer",
|
37 |
+
"unk_token": "[UNK]"
|
38 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:c171eb48151a6f712517bf002dd7d43be136c8408d649c4ae7c8a23f1af5d84c
|
3 |
+
size 4091
|
vocab.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|