jaeseungyang commited on
Commit
e74a555
·
verified ·
1 Parent(s): 11eb2f0

Upload 3 files

Browse files
docling-layout-old/config.json ADDED
@@ -0,0 +1,155 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_dropout": 0.0,
3
+ "activation_function": "silu",
4
+ "anchor_image_size": null,
5
+ "architectures": [
6
+ "RTDetrForObjectDetection"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "auxiliary_loss": true,
10
+ "backbone": null,
11
+ "backbone_config": {
12
+ "depths": [
13
+ 3,
14
+ 4,
15
+ 6,
16
+ 3
17
+ ],
18
+ "downsample_in_bottleneck": false,
19
+ "downsample_in_first_stage": false,
20
+ "embedding_size": 64,
21
+ "hidden_act": "relu",
22
+ "hidden_sizes": [
23
+ 256,
24
+ 512,
25
+ 1024,
26
+ 2048
27
+ ],
28
+ "layer_type": "bottleneck",
29
+ "model_type": "rt_detr_resnet",
30
+ "num_channels": 3,
31
+ "out_features": [
32
+ "stage2",
33
+ "stage3",
34
+ "stage4"
35
+ ],
36
+ "out_indices": [
37
+ 2,
38
+ 3,
39
+ 4
40
+ ],
41
+ "stage_names": [
42
+ "stem",
43
+ "stage1",
44
+ "stage2",
45
+ "stage3",
46
+ "stage4"
47
+ ]
48
+ },
49
+ "backbone_kwargs": null,
50
+ "batch_norm_eps": 1e-05,
51
+ "box_noise_scale": 1.0,
52
+ "d_model": 256,
53
+ "decoder_activation_function": "relu",
54
+ "decoder_attention_heads": 8,
55
+ "decoder_ffn_dim": 1024,
56
+ "decoder_in_channels": [
57
+ 256,
58
+ 256,
59
+ 256
60
+ ],
61
+ "decoder_layers": 6,
62
+ "decoder_n_points": 4,
63
+ "disable_custom_kernels": true,
64
+ "dropout": 0.0,
65
+ "encode_proj_layers": [
66
+ 2
67
+ ],
68
+ "encoder_activation_function": "gelu",
69
+ "encoder_attention_heads": 8,
70
+ "encoder_ffn_dim": 1024,
71
+ "encoder_hidden_dim": 256,
72
+ "encoder_in_channels": [
73
+ 512,
74
+ 1024,
75
+ 2048
76
+ ],
77
+ "encoder_layers": 1,
78
+ "eos_coefficient": 0.0001,
79
+ "eval_size": null,
80
+ "feat_strides": [
81
+ 8,
82
+ 16,
83
+ 32
84
+ ],
85
+ "focal_loss_alpha": 0.75,
86
+ "focal_loss_gamma": 2.0,
87
+ "freeze_backbone_batch_norms": true,
88
+ "hidden_expansion": 1.0,
89
+ "id2label": {
90
+ "0": "background",
91
+ "1": "Caption",
92
+ "2": "Footnote",
93
+ "3": "Formula",
94
+ "4": "List-item",
95
+ "5": "Page-footer",
96
+ "6": "Page-header",
97
+ "7": "Picture",
98
+ "8": "Section-header",
99
+ "9": "Table",
100
+ "10": "Text",
101
+ "11": "Title",
102
+ "12": "Document Index",
103
+ "13": "Code",
104
+ "14": "Checkbox-Selected",
105
+ "15": "Checkbox-Unselected",
106
+ "16": "Form",
107
+ "17": "Key-Value Region"
108
+ },
109
+ "initializer_bias_prior_prob": null,
110
+ "initializer_range": 0.01,
111
+ "is_encoder_decoder": true,
112
+ "label2id": {
113
+ "Caption": "1",
114
+ "Checkbox-Selected": "14",
115
+ "Checkbox-Unselected": "15",
116
+ "Code": "13",
117
+ "Document Index": "12",
118
+ "Footnote": "2",
119
+ "Form": "16",
120
+ "Formula": "3",
121
+ "Key-Value Region": "17",
122
+ "List-item": "4",
123
+ "Page-footer": "5",
124
+ "Page-header": "6",
125
+ "Picture": "7",
126
+ "Section-header": "8",
127
+ "Table": "9",
128
+ "Text": "10",
129
+ "Title": "11",
130
+ "background": "0"
131
+ },
132
+ "label_noise_ratio": 0.5,
133
+ "layer_norm_eps": 1e-05,
134
+ "learn_initial_query": false,
135
+ "matcher_alpha": 0.25,
136
+ "matcher_bbox_cost": 5.0,
137
+ "matcher_class_cost": 2.0,
138
+ "matcher_gamma": 2.0,
139
+ "matcher_giou_cost": 2.0,
140
+ "model_type": "rt_detr",
141
+ "normalize_before": false,
142
+ "num_denoising": 100,
143
+ "num_feature_levels": 3,
144
+ "num_queries": 300,
145
+ "positional_encoding_temperature": 10000,
146
+ "torch_dtype": "float32",
147
+ "transformers_version": "4.51.3",
148
+ "use_focal_loss": true,
149
+ "use_pretrained_backbone": false,
150
+ "use_timm_backbone": false,
151
+ "weight_loss_bbox": 5.0,
152
+ "weight_loss_giou": 2.0,
153
+ "weight_loss_vfl": 1.0,
154
+ "with_box_refine": true
155
+ }
docling-layout-old/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f234e230f8d332f880a81e5e652d3bae326d2e52802a07439e678b16587f3c75
3
+ size 171666216
docling-layout-old/preprocessor_config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "do_convert_annotations": true,
3
+ "do_normalize": false,
4
+ "do_pad": false,
5
+ "do_rescale": true,
6
+ "do_resize": true,
7
+ "format": "coco_detection",
8
+ "image_mean": [
9
+ 0.485,
10
+ 0.456,
11
+ 0.406
12
+ ],
13
+ "image_processor_type": "RTDetrImageProcessor",
14
+ "image_std": [
15
+ 0.229,
16
+ 0.224,
17
+ 0.225
18
+ ],
19
+ "pad_size": null,
20
+ "resample": 2,
21
+ "rescale_factor": 0.00392156862745098,
22
+ "size": {
23
+ "height": 640,
24
+ "width": 640
25
+ }
26
+ }