Fine tuning LayoutLMv2 base for VN InfoVQA
Browse files- .gitignore +1 -0
- config.json +132 -0
- preprocessor_config.json +8 -0
- pytorch_model.bin +3 -0
- runs/Dec27_14-06-06_68e1a976d12c/1640614008.5931008/events.out.tfevents.1640614008.68e1a976d12c.90.1 +3 -0
- runs/Dec27_14-06-06_68e1a976d12c/events.out.tfevents.1640614008.68e1a976d12c.90.0 +3 -0
- special_tokens_map.json +1 -0
- tokenizer.json +0 -0
- tokenizer_config.json +1 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
.gitignore
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
checkpoint-*/
|
config.json
ADDED
@@ -0,0 +1,132 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "microsoft/layoutlmv2-base-uncased",
|
3 |
+
"architectures": [
|
4 |
+
"LayoutLMv2ForQuestionAnswering"
|
5 |
+
],
|
6 |
+
"attention_probs_dropout_prob": 0.1,
|
7 |
+
"convert_sync_batchnorm": true,
|
8 |
+
"coordinate_size": 128,
|
9 |
+
"detectron2_config_args": {
|
10 |
+
"MODEL.ANCHOR_GENERATOR.SIZES": [
|
11 |
+
[
|
12 |
+
32
|
13 |
+
],
|
14 |
+
[
|
15 |
+
64
|
16 |
+
],
|
17 |
+
[
|
18 |
+
128
|
19 |
+
],
|
20 |
+
[
|
21 |
+
256
|
22 |
+
],
|
23 |
+
[
|
24 |
+
512
|
25 |
+
]
|
26 |
+
],
|
27 |
+
"MODEL.BACKBONE.NAME": "build_resnet_fpn_backbone",
|
28 |
+
"MODEL.FPN.IN_FEATURES": [
|
29 |
+
"res2",
|
30 |
+
"res3",
|
31 |
+
"res4",
|
32 |
+
"res5"
|
33 |
+
],
|
34 |
+
"MODEL.MASK_ON": true,
|
35 |
+
"MODEL.PIXEL_STD": [
|
36 |
+
57.375,
|
37 |
+
57.12,
|
38 |
+
58.395
|
39 |
+
],
|
40 |
+
"MODEL.POST_NMS_TOPK_TEST": 1000,
|
41 |
+
"MODEL.RESNETS.ASPECT_RATIOS": [
|
42 |
+
[
|
43 |
+
0.5,
|
44 |
+
1.0,
|
45 |
+
2.0
|
46 |
+
]
|
47 |
+
],
|
48 |
+
"MODEL.RESNETS.DEPTH": 101,
|
49 |
+
"MODEL.RESNETS.NUM_GROUPS": 32,
|
50 |
+
"MODEL.RESNETS.OUT_FEATURES": [
|
51 |
+
"res2",
|
52 |
+
"res3",
|
53 |
+
"res4",
|
54 |
+
"res5"
|
55 |
+
],
|
56 |
+
"MODEL.RESNETS.SIZES": [
|
57 |
+
[
|
58 |
+
32
|
59 |
+
],
|
60 |
+
[
|
61 |
+
64
|
62 |
+
],
|
63 |
+
[
|
64 |
+
128
|
65 |
+
],
|
66 |
+
[
|
67 |
+
256
|
68 |
+
],
|
69 |
+
[
|
70 |
+
512
|
71 |
+
]
|
72 |
+
],
|
73 |
+
"MODEL.RESNETS.STRIDE_IN_1X1": false,
|
74 |
+
"MODEL.RESNETS.WIDTH_PER_GROUP": 8,
|
75 |
+
"MODEL.ROI_BOX_HEAD.NAME": "FastRCNNConvFCHead",
|
76 |
+
"MODEL.ROI_BOX_HEAD.NUM_FC": 2,
|
77 |
+
"MODEL.ROI_BOX_HEAD.POOLER_RESOLUTION": 14,
|
78 |
+
"MODEL.ROI_HEADS.IN_FEATURES": [
|
79 |
+
"p2",
|
80 |
+
"p3",
|
81 |
+
"p4",
|
82 |
+
"p5"
|
83 |
+
],
|
84 |
+
"MODEL.ROI_HEADS.NAME": "StandardROIHeads",
|
85 |
+
"MODEL.ROI_HEADS.NUM_CLASSES": 5,
|
86 |
+
"MODEL.ROI_MASK_HEAD.NAME": "MaskRCNNConvUpsampleHead",
|
87 |
+
"MODEL.ROI_MASK_HEAD.NUM_CONV": 4,
|
88 |
+
"MODEL.ROI_MASK_HEAD.POOLER_RESOLUTION": 7,
|
89 |
+
"MODEL.RPN.IN_FEATURES": [
|
90 |
+
"p2",
|
91 |
+
"p3",
|
92 |
+
"p4",
|
93 |
+
"p5",
|
94 |
+
"p6"
|
95 |
+
],
|
96 |
+
"MODEL.RPN.POST_NMS_TOPK_TRAIN": 1000,
|
97 |
+
"MODEL.RPN.PRE_NMS_TOPK_TEST": 1000,
|
98 |
+
"MODEL.RPN.PRE_NMS_TOPK_TRAIN": 2000
|
99 |
+
},
|
100 |
+
"fast_qkv": true,
|
101 |
+
"gradient_checkpointing": false,
|
102 |
+
"has_relative_attention_bias": true,
|
103 |
+
"has_spatial_attention_bias": true,
|
104 |
+
"has_visual_segment_embedding": true,
|
105 |
+
"hidden_act": "gelu",
|
106 |
+
"hidden_dropout_prob": 0.1,
|
107 |
+
"hidden_size": 768,
|
108 |
+
"image_feature_pool_shape": [
|
109 |
+
7,
|
110 |
+
7,
|
111 |
+
256
|
112 |
+
],
|
113 |
+
"initializer_range": 0.02,
|
114 |
+
"intermediate_size": 3072,
|
115 |
+
"layer_norm_eps": 1e-12,
|
116 |
+
"max_2d_position_embeddings": 1024,
|
117 |
+
"max_position_embeddings": 512,
|
118 |
+
"max_rel_2d_pos": 256,
|
119 |
+
"max_rel_pos": 128,
|
120 |
+
"model_type": "layoutlmv2",
|
121 |
+
"num_attention_heads": 12,
|
122 |
+
"num_hidden_layers": 12,
|
123 |
+
"output_past": true,
|
124 |
+
"pad_token_id": 0,
|
125 |
+
"rel_2d_pos_bins": 64,
|
126 |
+
"rel_pos_bins": 32,
|
127 |
+
"shape_size": 128,
|
128 |
+
"torch_dtype": "float32",
|
129 |
+
"transformers_version": "4.15.0",
|
130 |
+
"type_vocab_size": 2,
|
131 |
+
"vocab_size": 30522
|
132 |
+
}
|
preprocessor_config.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"apply_ocr": false,
|
3 |
+
"do_resize": true,
|
4 |
+
"feature_extractor_type": "LayoutLMv2FeatureExtractor",
|
5 |
+
"ocr_lang": null,
|
6 |
+
"resample": 2,
|
7 |
+
"size": 224
|
8 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:25a6f19166f4a859bd33e28956ed7330861147581965ca8fee1a7a920a0af460
|
3 |
+
size 802263195
|
runs/Dec27_14-06-06_68e1a976d12c/1640614008.5931008/events.out.tfevents.1640614008.68e1a976d12c.90.1
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:929aaffcb9f285fd44923150c62e702b3a9914f9e9cf62ed08c23f9d1c325e01
|
3 |
+
size 4812
|
runs/Dec27_14-06-06_68e1a976d12c/events.out.tfevents.1640614008.68e1a976d12c.90.0
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:3a8f58f1434b1242e2be3ca4d802288f4929ed146d28bad6232c8120db40fa2c
|
3 |
+
size 7515
|
special_tokens_map.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}
|
tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|
tokenizer_config.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"do_lower_case": true, "unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "cls_token_box": [0, 0, 0, 0], "sep_token_box": [1000, 1000, 1000, 1000], "pad_token_box": [0, 0, 0, 0], "pad_token_label": -100, "only_label_first_subword": true, "tokenize_chinese_chars": true, "strip_accents": null, "model_max_length": 512, "special_tokens_map_file": null, "name_or_path": "microsoft/layoutlmv2-base-uncased", "do_basic_tokenize": true, "never_split": null, "additional_special_tokens": null, "tokenizer_class": "LayoutLMv2Tokenizer"}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:770ec9ec9199e97fab3a73bb9a9f7d2a18c77c66f9c2d90fc85bf77e8e86c27f
|
3 |
+
size 2991
|
vocab.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|