Commit
·
4c9bb37
1
Parent(s):
be3f1d1
Upload folder using huggingface_hub
Browse files- config.json +170 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- preprocessor_config.json +28 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +91 -0
- training_args.bin +3 -0
config.json
ADDED
@@ -0,0 +1,170 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "microsoft/dit-large-finetuned-rvlcdip",
|
3 |
+
"architectures": [
|
4 |
+
"BylawBeitForImageClassification"
|
5 |
+
],
|
6 |
+
"attention_probs_dropout_prob": 0.0,
|
7 |
+
"auxiliary_channels": 256,
|
8 |
+
"auxiliary_concat_input": false,
|
9 |
+
"auxiliary_loss_weight": 0.4,
|
10 |
+
"auxiliary_num_convs": 1,
|
11 |
+
"drop_path_rate": 0.1,
|
12 |
+
"hidden_act": "gelu",
|
13 |
+
"hidden_dropout_prob": 0.0,
|
14 |
+
"hidden_size": 768,
|
15 |
+
"id2label": {
|
16 |
+
"0": "template0",
|
17 |
+
"1": "template1",
|
18 |
+
"2": "template2",
|
19 |
+
"3": "template3",
|
20 |
+
"4": "template4",
|
21 |
+
"5": "template5",
|
22 |
+
"6": "template6",
|
23 |
+
"7": "template7",
|
24 |
+
"8": "template8",
|
25 |
+
"9": "template9",
|
26 |
+
"10": "template10",
|
27 |
+
"11": "template11",
|
28 |
+
"12": "template12",
|
29 |
+
"13": "template13",
|
30 |
+
"14": "template14",
|
31 |
+
"15": "template15",
|
32 |
+
"16": "template16",
|
33 |
+
"17": "template17",
|
34 |
+
"18": "template18",
|
35 |
+
"19": "template19",
|
36 |
+
"20": "template20",
|
37 |
+
"21": "template21",
|
38 |
+
"22": "template22",
|
39 |
+
"23": "template23",
|
40 |
+
"24": "template24",
|
41 |
+
"25": "template25",
|
42 |
+
"26": "template26",
|
43 |
+
"27": "template27",
|
44 |
+
"28": "template28",
|
45 |
+
"29": "template29",
|
46 |
+
"30": "template30",
|
47 |
+
"31": "template31",
|
48 |
+
"32": "template32",
|
49 |
+
"33": "template33",
|
50 |
+
"34": "template34",
|
51 |
+
"35": "template35",
|
52 |
+
"36": "template36",
|
53 |
+
"37": "template37",
|
54 |
+
"38": "template38",
|
55 |
+
"39": "template39",
|
56 |
+
"40": "template40",
|
57 |
+
"41": "template41",
|
58 |
+
"42": "template42",
|
59 |
+
"43": "template43",
|
60 |
+
"44": "template44",
|
61 |
+
"45": "template45",
|
62 |
+
"46": "template46",
|
63 |
+
"47": "template47",
|
64 |
+
"48": "template48",
|
65 |
+
"49": "template49",
|
66 |
+
"50": "template50",
|
67 |
+
"51": "template51",
|
68 |
+
"52": "template52",
|
69 |
+
"53": "template53",
|
70 |
+
"54": "template54",
|
71 |
+
"55": "template55",
|
72 |
+
"56": "template56",
|
73 |
+
"57": "template57",
|
74 |
+
"58": "template-1"
|
75 |
+
},
|
76 |
+
"image_size": 224,
|
77 |
+
"initializer_range": 0.02,
|
78 |
+
"intermediate_size": 3072,
|
79 |
+
"label2id": {
|
80 |
+
"template-1": 58,
|
81 |
+
"template0": 0,
|
82 |
+
"template1": 1,
|
83 |
+
"template10": 10,
|
84 |
+
"template11": 11,
|
85 |
+
"template12": 12,
|
86 |
+
"template13": 13,
|
87 |
+
"template14": 14,
|
88 |
+
"template15": 15,
|
89 |
+
"template16": 16,
|
90 |
+
"template17": 17,
|
91 |
+
"template18": 18,
|
92 |
+
"template19": 19,
|
93 |
+
"template2": 2,
|
94 |
+
"template20": 20,
|
95 |
+
"template21": 21,
|
96 |
+
"template22": 22,
|
97 |
+
"template23": 23,
|
98 |
+
"template24": 24,
|
99 |
+
"template25": 25,
|
100 |
+
"template26": 26,
|
101 |
+
"template27": 27,
|
102 |
+
"template28": 28,
|
103 |
+
"template29": 29,
|
104 |
+
"template3": 3,
|
105 |
+
"template30": 30,
|
106 |
+
"template31": 31,
|
107 |
+
"template32": 32,
|
108 |
+
"template33": 33,
|
109 |
+
"template34": 34,
|
110 |
+
"template35": 35,
|
111 |
+
"template36": 36,
|
112 |
+
"template37": 37,
|
113 |
+
"template38": 38,
|
114 |
+
"template39": 39,
|
115 |
+
"template4": 4,
|
116 |
+
"template40": 40,
|
117 |
+
"template41": 41,
|
118 |
+
"template42": 42,
|
119 |
+
"template43": 43,
|
120 |
+
"template44": 44,
|
121 |
+
"template45": 45,
|
122 |
+
"template46": 46,
|
123 |
+
"template47": 47,
|
124 |
+
"template48": 48,
|
125 |
+
"template49": 49,
|
126 |
+
"template5": 5,
|
127 |
+
"template50": 50,
|
128 |
+
"template51": 51,
|
129 |
+
"template52": 52,
|
130 |
+
"template53": 53,
|
131 |
+
"template54": 54,
|
132 |
+
"template55": 55,
|
133 |
+
"template56": 56,
|
134 |
+
"template57": 57,
|
135 |
+
"template6": 6,
|
136 |
+
"template7": 7,
|
137 |
+
"template8": 8,
|
138 |
+
"template9": 9
|
139 |
+
},
|
140 |
+
"layer_norm_eps": 1e-12,
|
141 |
+
"layer_scale_init_value": 0.1,
|
142 |
+
"model_type": "beit",
|
143 |
+
"num_attention_heads": 12,
|
144 |
+
"num_channels": 3,
|
145 |
+
"num_hidden_layers": 12,
|
146 |
+
"out_indices": [
|
147 |
+
3,
|
148 |
+
5,
|
149 |
+
7,
|
150 |
+
11
|
151 |
+
],
|
152 |
+
"patch_size": 16,
|
153 |
+
"pool_scales": [
|
154 |
+
1,
|
155 |
+
2,
|
156 |
+
3,
|
157 |
+
6
|
158 |
+
],
|
159 |
+
"problem_type": "single_label_classification",
|
160 |
+
"semantic_loss_ignore_index": 255,
|
161 |
+
"torch_dtype": "float32",
|
162 |
+
"transformers_version": "4.35.2",
|
163 |
+
"use_absolute_position_embeddings": false,
|
164 |
+
"use_auxiliary_head": true,
|
165 |
+
"use_mask_token": false,
|
166 |
+
"use_mean_pooling": true,
|
167 |
+
"use_relative_position_bias": false,
|
168 |
+
"use_shared_relative_position_bias": false,
|
169 |
+
"vocab_size": 8192
|
170 |
+
}
|
model.safetensors
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5e32039855078db466c1a4800dae1d616eb6f97a3c43df68c1c81bd8f958f3f4
|
3 |
+
size 342832060
|
optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:48940705d7d85a05669abffdd78add551f9682e8a663c3530e7d3f030379ee99
|
3 |
+
size 685791930
|
preprocessor_config.json
ADDED
@@ -0,0 +1,28 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"crop_size": {
|
3 |
+
"height": 224,
|
4 |
+
"width": 224
|
5 |
+
},
|
6 |
+
"do_center_crop": false,
|
7 |
+
"do_normalize": true,
|
8 |
+
"do_reduce_labels": false,
|
9 |
+
"do_rescale": true,
|
10 |
+
"do_resize": true,
|
11 |
+
"image_mean": [
|
12 |
+
0.5,
|
13 |
+
0.5,
|
14 |
+
0.5
|
15 |
+
],
|
16 |
+
"image_processor_type": "BeitImageProcessor",
|
17 |
+
"image_std": [
|
18 |
+
0.5,
|
19 |
+
0.5,
|
20 |
+
0.5
|
21 |
+
],
|
22 |
+
"resample": 2,
|
23 |
+
"rescale_factor": 0.00392156862745098,
|
24 |
+
"size": {
|
25 |
+
"height": 224,
|
26 |
+
"width": 224
|
27 |
+
}
|
28 |
+
}
|
rng_state.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:01601d9ff8c9c4d2e271bdb8df55c2baf63fd82f3808f580f938cf0105dd4f43
|
3 |
+
size 14244
|
scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:007ead22b25c9cd78601b26c5ac19643a48ce5843bfc4d487a7585267e7f41b1
|
3 |
+
size 1064
|
trainer_state.json
ADDED
@@ -0,0 +1,91 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 6.0,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 324,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.46,
|
13 |
+
"learning_rate": 3.08641975308642e-06,
|
14 |
+
"loss": 4.0718,
|
15 |
+
"step": 25
|
16 |
+
},
|
17 |
+
{
|
18 |
+
"epoch": 0.93,
|
19 |
+
"learning_rate": 6.17283950617284e-06,
|
20 |
+
"loss": 3.9483,
|
21 |
+
"step": 50
|
22 |
+
},
|
23 |
+
{
|
24 |
+
"epoch": 1.39,
|
25 |
+
"learning_rate": 9.25925925925926e-06,
|
26 |
+
"loss": 3.8844,
|
27 |
+
"step": 75
|
28 |
+
},
|
29 |
+
{
|
30 |
+
"epoch": 1.85,
|
31 |
+
"learning_rate": 1.234567901234568e-05,
|
32 |
+
"loss": 3.8757,
|
33 |
+
"step": 100
|
34 |
+
},
|
35 |
+
{
|
36 |
+
"epoch": 2.31,
|
37 |
+
"learning_rate": 1.54320987654321e-05,
|
38 |
+
"loss": 3.8588,
|
39 |
+
"step": 125
|
40 |
+
},
|
41 |
+
{
|
42 |
+
"epoch": 2.78,
|
43 |
+
"learning_rate": 1.851851851851852e-05,
|
44 |
+
"loss": 3.8919,
|
45 |
+
"step": 150
|
46 |
+
},
|
47 |
+
{
|
48 |
+
"epoch": 3.24,
|
49 |
+
"learning_rate": 1.971677559912854e-05,
|
50 |
+
"loss": 3.8707,
|
51 |
+
"step": 175
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 3.7,
|
55 |
+
"learning_rate": 1.917211328976035e-05,
|
56 |
+
"loss": 3.8775,
|
57 |
+
"step": 200
|
58 |
+
},
|
59 |
+
{
|
60 |
+
"epoch": 4.17,
|
61 |
+
"learning_rate": 1.862745098039216e-05,
|
62 |
+
"loss": 3.8929,
|
63 |
+
"step": 225
|
64 |
+
},
|
65 |
+
{
|
66 |
+
"epoch": 4.63,
|
67 |
+
"learning_rate": 1.8082788671023966e-05,
|
68 |
+
"loss": 3.8813,
|
69 |
+
"step": 250
|
70 |
+
},
|
71 |
+
{
|
72 |
+
"epoch": 5.09,
|
73 |
+
"learning_rate": 1.7538126361655776e-05,
|
74 |
+
"loss": 3.8796,
|
75 |
+
"step": 275
|
76 |
+
},
|
77 |
+
{
|
78 |
+
"epoch": 5.56,
|
79 |
+
"learning_rate": 1.6993464052287582e-05,
|
80 |
+
"loss": 3.8681,
|
81 |
+
"step": 300
|
82 |
+
}
|
83 |
+
],
|
84 |
+
"logging_steps": 25,
|
85 |
+
"max_steps": 1080,
|
86 |
+
"num_train_epochs": 20,
|
87 |
+
"save_steps": 500,
|
88 |
+
"total_flos": 1.5808841019347927e+18,
|
89 |
+
"trial_name": null,
|
90 |
+
"trial_params": null
|
91 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:257df100e856919d305c2bec3e070a52103c6a22d4e0994e60ee1316e1958b8f
|
3 |
+
size 4728
|