Hugging Face's logo

hf-tiny-v2
/
tiny-random-DFineModel

Safetensors
d_fine
Model card Files
xet
 Community
hf-transformers-bot commited on
Commit
0879ec1
·
verified ·
1 Parent(s): da8cd56

Update tiny models for DFineModel

Browse files
Files changed (3) hide show
  1. config.json +224 -0
  2. model.safetensors +3 -0
  3. preprocessor_config.json +29 -0
config.json ADDED
@@ -0,0 +1,224 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "activation_dropout": 0.0,
3
+ "activation_function": "silu",
4
+ "anchor_image_size": null,
5
+ "architectures": [
6
+ "DFineModel"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "auxiliary_loss": true,
10
+ "backbone_config": {
11
+ "depths": [
12
+ 1,
13
+ 1,
14
+ 2,
15
+ 1
16
+ ],
17
+ "embedding_size": 64,
18
+ "embeddings_size": 10,
19
+ "hidden_act": "relu",
20
+ "hidden_sizes": [
21
+ 64,
22
+ 128,
23
+ 256,
24
+ 512
25
+ ],
26
+ "initializer_range": 0.02,
27
+ "model_type": "hgnet_v2",
28
+ "num_channels": 3,
29
+ "out_features": [
30
+ "stage2",
31
+ "stage3",
32
+ "stage4"
33
+ ],
34
+ "out_indices": [
35
+ 2,
36
+ 3,
37
+ 4
38
+ ],
39
+ "stage_downsample": [
40
+ false,
41
+ true,
42
+ true,
43
+ true
44
+ ],
45
+ "stage_downsample_strides": [
46
+ 2,
47
+ 2,
48
+ 2,
49
+ 2
50
+ ],
51
+ "stage_in_channels": [
52
+ 16,
53
+ 64,
54
+ 128,
55
+ 256
56
+ ],
57
+ "stage_kernel_size": [
58
+ 3,
59
+ 3,
60
+ 5,
61
+ 5
62
+ ],
63
+ "stage_light_block": [
64
+ false,
65
+ false,
66
+ true,
67
+ true
68
+ ],
69
+ "stage_mid_channels": [
70
+ 16,
71
+ 32,
72
+ 64,
73
+ 128
74
+ ],
75
+ "stage_names": [
76
+ "stem",
77
+ "stage1",
78
+ "stage2",
79
+ "stage3",
80
+ "stage4"
81
+ ],
82
+ "stage_num_blocks": [
83
+ 1,
84
+ 1,
85
+ 2,
86
+ 1
87
+ ],
88
+ "stage_numb_of_layers": [
89
+ 3,
90
+ 3,
91
+ 3,
92
+ 3
93
+ ],
94
+ "stage_out_channels": [
95
+ 64,
96
+ 128,
97
+ 256,
98
+ 512
99
+ ],
100
+ "stem_channels": [
101
+ 3,
102
+ 16,
103
+ 16
104
+ ],
105
+ "stem_strides": [
106
+ 2,
107
+ 1,
108
+ 1,
109
+ 2,
110
+ 1
111
+ ],
112
+ "use_lab": true,
113
+ "use_learnable_affine_block": false
114
+ },
115
+ "batch_norm_eps": 1e-05,
116
+ "box_noise_scale": 1.0,
117
+ "d_model": 32,
118
+ "decoder_activation_function": "relu",
119
+ "decoder_attention_heads": 2,
120
+ "decoder_ffn_dim": 64,
121
+ "decoder_in_channels": [
122
+ 32,
123
+ 32,
124
+ 32
125
+ ],
126
+ "decoder_layers": 2,
127
+ "decoder_method": "default",
128
+ "decoder_n_levels": 3,
129
+ "decoder_n_points": [
130
+ 3,
131
+ 6,
132
+ 3
133
+ ],
134
+ "decoder_offset_scale": 0.5,
135
+ "depth_mult": 0.34,
136
+ "disable_custom_kernels": true,
137
+ "dropout": 0.0,
138
+ "dtype": "float32",
139
+ "encode_proj_layers": [
140
+ 2
141
+ ],
142
+ "encoder_activation_function": "gelu",
143
+ "encoder_attention_heads": 2,
144
+ "encoder_ffn_dim": 64,
145
+ "encoder_hidden_dim": 32,
146
+ "encoder_in_channels": [
147
+ 128,
148
+ 256,
149
+ 512
150
+ ],
151
+ "encoder_layers": 1,
152
+ "eos_coefficient": 0.0001,
153
+ "eval_idx": -1,
154
+ "eval_size": null,
155
+ "feat_strides": [
156
+ 8,
157
+ 16,
158
+ 32
159
+ ],
160
+ "focal_loss_alpha": 0.75,
161
+ "focal_loss_gamma": 2.0,
162
+ "freeze_backbone_batch_norms": true,
163
+ "head_dim": 16,
164
+ "hidden_expansion": 1.0,
165
+ "id2label": {
166
+ "0": "LABEL_0",
167
+ "1": "LABEL_1",
168
+ "2": "LABEL_2",
169
+ "3": "LABEL_3",
170
+ "4": "LABEL_4",
171
+ "5": "LABEL_5",
172
+ "6": "LABEL_6",
173
+ "7": "LABEL_7",
174
+ "8": "LABEL_8",
175
+ "9": "LABEL_9"
176
+ },
177
+ "image_size": 64,
178
+ "initializer_bias_prior_prob": null,
179
+ "initializer_range": 0.01,
180
+ "is_encoder_decoder": true,
181
+ "label2id": {
182
+ "LABEL_0": 0,
183
+ "LABEL_1": 1,
184
+ "LABEL_2": 2,
185
+ "LABEL_3": 3,
186
+ "LABEL_4": 4,
187
+ "LABEL_5": 5,
188
+ "LABEL_6": 6,
189
+ "LABEL_7": 7,
190
+ "LABEL_8": 8,
191
+ "LABEL_9": 9
192
+ },
193
+ "label_noise_ratio": 0.5,
194
+ "layer_norm_eps": 1e-05,
195
+ "layer_scale": 1,
196
+ "learn_initial_query": false,
197
+ "lqe_hidden_dim": 64,
198
+ "lqe_layers": 2,
199
+ "matcher_alpha": 0.25,
200
+ "matcher_bbox_cost": 5.0,
201
+ "matcher_class_cost": 2.0,
202
+ "matcher_gamma": 2.0,
203
+ "matcher_giou_cost": 2.0,
204
+ "max_num_bins": 32,
205
+ "model_type": "d_fine",
206
+ "normalize_before": false,
207
+ "num_denoising": 0,
208
+ "num_feature_levels": 3,
209
+ "num_queries": 30,
210
+ "positional_encoding_temperature": 10000,
211
+ "reg_max": 32,
212
+ "reg_scale": 4.0,
213
+ "tie_word_embeddings": true,
214
+ "top_prob_values": 4,
215
+ "transformers_version": "5.16.0.dev0",
216
+ "up": 0.5,
217
+ "use_focal_loss": true,
218
+ "weight_loss_bbox": 5.0,
219
+ "weight_loss_ddf": 1.5,
220
+ "weight_loss_fgl": 0.15,
221
+ "weight_loss_giou": 2.0,
222
+ "weight_loss_vfl": 1.0,
223
+ "with_box_refine": true
224
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:dd022cdd94d92779d9a7c999da9508f091395706a835bd21536c07b13de5510f
3
+ size 3436952
preprocessor_config.json ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "crop_size": {
3
+ "height": 64,
4
+ "width": 64
5
+ },
6
+ "do_convert_annotations": true,
7
+ "do_normalize": false,
8
+ "do_pad": false,
9
+ "do_rescale": true,
10
+ "do_resize": true,
11
+ "format": "coco_detection",
12
+ "image_mean": [
13
+ 0.485,
14
+ 0.456,
15
+ 0.406
16
+ ],
17
+ "image_processor_type": "RTDetrImageProcessor",
18
+ "image_std": [
19
+ 0.229,
20
+ 0.224,
21
+ 0.225
22
+ ],
23
+ "resample": 2,
24
+ "rescale_factor": 0.00392156862745098,
25
+ "size": {
26
+ "height": 64,
27
+ "width": 64
28
+ }
29
+ }