keldrenloy
commited on
Commit
·
f3b7368
1
Parent(s):
c2da883
New Fine-Tuned Model
Browse files- config.json +166 -0
- merges.txt +0 -0
- optimizer.pt +3 -0
- pytorch_model.bin +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +51 -0
- tokenizer_config.json +86 -0
- trainer_state.json +148 -0
- training_args.bin +3 -0
- vocab.json +0 -0
config.json
ADDED
@@ -0,0 +1,166 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "C:/Users/keldr/Desktop/Workspace/Repo/12Sep2022/model/layoutlmv3/checkpoint-1500",
|
3 |
+
"architectures": [
|
4 |
+
"LayoutLMv3ForTokenClassification"
|
5 |
+
],
|
6 |
+
"attention_probs_dropout_prob": 0.1,
|
7 |
+
"bos_token_id": 0,
|
8 |
+
"classifier_dropout": null,
|
9 |
+
"coordinate_size": 128,
|
10 |
+
"eos_token_id": 2,
|
11 |
+
"has_relative_attention_bias": true,
|
12 |
+
"has_spatial_attention_bias": true,
|
13 |
+
"hidden_act": "gelu",
|
14 |
+
"hidden_dropout_prob": 0.1,
|
15 |
+
"hidden_size": 768,
|
16 |
+
"id2label": {
|
17 |
+
"0": "O",
|
18 |
+
"1": "B-MENU.NM",
|
19 |
+
"2": "B-MENU.NUM",
|
20 |
+
"3": "B-MENU.UNITPRICE",
|
21 |
+
"4": "B-MENU.CNT",
|
22 |
+
"5": "B-MENU.DISCOUNTPRICE",
|
23 |
+
"6": "B-MENU.PRICE",
|
24 |
+
"7": "B-MENU.ITEMSUBTOTAL",
|
25 |
+
"8": "B-MENU.VATYN",
|
26 |
+
"9": "B-MENU.ETC",
|
27 |
+
"10": "B-MENU.SUB_NM",
|
28 |
+
"11": "B-MENU.SUB_UNITPRICE",
|
29 |
+
"12": "B-MENU.SUB_CNT",
|
30 |
+
"13": "B-MENU.SUB_PRICE",
|
31 |
+
"14": "B-MENU.SUB_ETC",
|
32 |
+
"15": "B-VOID_MENU.NM",
|
33 |
+
"16": "B-VOID_MENU.PRICE",
|
34 |
+
"17": "B-SUB_TOTAL.SUBTOTAL_PRICE",
|
35 |
+
"18": "B-SUB_TOTAL.DISCOUNT_PRICE",
|
36 |
+
"19": "B-SUB_TOTAL.SERVICE_PRICE",
|
37 |
+
"20": "B-SUB_TOTAL.OTHERSVC_PRICE",
|
38 |
+
"21": "B-SUB_TOTAL.TAX_PRICE",
|
39 |
+
"22": "B-SUB_TOTAL.ETC",
|
40 |
+
"23": "B-TOTAL.TOTAL_PRICE",
|
41 |
+
"24": "B-TOTAL.TOTAL_ETC",
|
42 |
+
"25": "B-TOTAL.CASHPRICE",
|
43 |
+
"26": "B-TOTAL.CHANGEPRICE",
|
44 |
+
"27": "B-TOTAL.CREDITCARDPRICE",
|
45 |
+
"28": "B-TOTAL.EMONEYPRICE",
|
46 |
+
"29": "B-TOTAL.MENUTYPE_CNT",
|
47 |
+
"30": "B-TOTAL.MENUQTY_CNT",
|
48 |
+
"31": "I-MENU.NM",
|
49 |
+
"32": "I-MENU.NUM",
|
50 |
+
"33": "I-MENU.UNITPRICE",
|
51 |
+
"34": "I-MENU.CNT",
|
52 |
+
"35": "I-MENU.DISCOUNTPRICE",
|
53 |
+
"36": "I-MENU.PRICE",
|
54 |
+
"37": "I-MENU.ITEMSUBTOTAL",
|
55 |
+
"38": "I-MENU.VATYN",
|
56 |
+
"39": "I-MENU.ETC",
|
57 |
+
"40": "I-MENU.SUB_NM",
|
58 |
+
"41": "I-MENU.SUB_UNITPRICE",
|
59 |
+
"42": "I-MENU.SUB_CNT",
|
60 |
+
"43": "I-MENU.SUB_PRICE",
|
61 |
+
"44": "I-MENU.SUB_ETC",
|
62 |
+
"45": "I-VOID_MENU.NM",
|
63 |
+
"46": "I-VOID_MENU.PRICE",
|
64 |
+
"47": "I-SUB_TOTAL.SUBTOTAL_PRICE",
|
65 |
+
"48": "I-SUB_TOTAL.DISCOUNT_PRICE",
|
66 |
+
"49": "I-SUB_TOTAL.SERVICE_PRICE",
|
67 |
+
"50": "I-SUB_TOTAL.OTHERSVC_PRICE",
|
68 |
+
"51": "I-SUB_TOTAL.TAX_PRICE",
|
69 |
+
"52": "I-SUB_TOTAL.ETC",
|
70 |
+
"53": "I-TOTAL.TOTAL_PRICE",
|
71 |
+
"54": "I-TOTAL.TOTAL_ETC",
|
72 |
+
"55": "I-TOTAL.CASHPRICE",
|
73 |
+
"56": "I-TOTAL.CHANGEPRICE",
|
74 |
+
"57": "I-TOTAL.CREDITCARDPRICE",
|
75 |
+
"58": "I-TOTAL.EMONEYPRICE",
|
76 |
+
"59": "I-TOTAL.MENUTYPE_CNT",
|
77 |
+
"60": "I-TOTAL.MENUQTY_CNT"
|
78 |
+
},
|
79 |
+
"initializer_range": 0.02,
|
80 |
+
"input_size": 224,
|
81 |
+
"intermediate_size": 3072,
|
82 |
+
"label2id": {
|
83 |
+
"B-MENU.CNT": 4,
|
84 |
+
"B-MENU.DISCOUNTPRICE": 5,
|
85 |
+
"B-MENU.ETC": 9,
|
86 |
+
"B-MENU.ITEMSUBTOTAL": 7,
|
87 |
+
"B-MENU.NM": 1,
|
88 |
+
"B-MENU.NUM": 2,
|
89 |
+
"B-MENU.PRICE": 6,
|
90 |
+
"B-MENU.SUB_CNT": 12,
|
91 |
+
"B-MENU.SUB_ETC": 14,
|
92 |
+
"B-MENU.SUB_NM": 10,
|
93 |
+
"B-MENU.SUB_PRICE": 13,
|
94 |
+
"B-MENU.SUB_UNITPRICE": 11,
|
95 |
+
"B-MENU.UNITPRICE": 3,
|
96 |
+
"B-MENU.VATYN": 8,
|
97 |
+
"B-SUB_TOTAL.DISCOUNT_PRICE": 18,
|
98 |
+
"B-SUB_TOTAL.ETC": 22,
|
99 |
+
"B-SUB_TOTAL.OTHERSVC_PRICE": 20,
|
100 |
+
"B-SUB_TOTAL.SERVICE_PRICE": 19,
|
101 |
+
"B-SUB_TOTAL.SUBTOTAL_PRICE": 17,
|
102 |
+
"B-SUB_TOTAL.TAX_PRICE": 21,
|
103 |
+
"B-TOTAL.CASHPRICE": 25,
|
104 |
+
"B-TOTAL.CHANGEPRICE": 26,
|
105 |
+
"B-TOTAL.CREDITCARDPRICE": 27,
|
106 |
+
"B-TOTAL.EMONEYPRICE": 28,
|
107 |
+
"B-TOTAL.MENUQTY_CNT": 30,
|
108 |
+
"B-TOTAL.MENUTYPE_CNT": 29,
|
109 |
+
"B-TOTAL.TOTAL_ETC": 24,
|
110 |
+
"B-TOTAL.TOTAL_PRICE": 23,
|
111 |
+
"B-VOID_MENU.NM": 15,
|
112 |
+
"B-VOID_MENU.PRICE": 16,
|
113 |
+
"I-MENU.CNT": 34,
|
114 |
+
"I-MENU.DISCOUNTPRICE": 35,
|
115 |
+
"I-MENU.ETC": 39,
|
116 |
+
"I-MENU.ITEMSUBTOTAL": 37,
|
117 |
+
"I-MENU.NM": 31,
|
118 |
+
"I-MENU.NUM": 32,
|
119 |
+
"I-MENU.PRICE": 36,
|
120 |
+
"I-MENU.SUB_CNT": 42,
|
121 |
+
"I-MENU.SUB_ETC": 44,
|
122 |
+
"I-MENU.SUB_NM": 40,
|
123 |
+
"I-MENU.SUB_PRICE": 43,
|
124 |
+
"I-MENU.SUB_UNITPRICE": 41,
|
125 |
+
"I-MENU.UNITPRICE": 33,
|
126 |
+
"I-MENU.VATYN": 38,
|
127 |
+
"I-SUB_TOTAL.DISCOUNT_PRICE": 48,
|
128 |
+
"I-SUB_TOTAL.ETC": 52,
|
129 |
+
"I-SUB_TOTAL.OTHERSVC_PRICE": 50,
|
130 |
+
"I-SUB_TOTAL.SERVICE_PRICE": 49,
|
131 |
+
"I-SUB_TOTAL.SUBTOTAL_PRICE": 47,
|
132 |
+
"I-SUB_TOTAL.TAX_PRICE": 51,
|
133 |
+
"I-TOTAL.CASHPRICE": 55,
|
134 |
+
"I-TOTAL.CHANGEPRICE": 56,
|
135 |
+
"I-TOTAL.CREDITCARDPRICE": 57,
|
136 |
+
"I-TOTAL.EMONEYPRICE": 58,
|
137 |
+
"I-TOTAL.MENUQTY_CNT": 60,
|
138 |
+
"I-TOTAL.MENUTYPE_CNT": 59,
|
139 |
+
"I-TOTAL.TOTAL_ETC": 54,
|
140 |
+
"I-TOTAL.TOTAL_PRICE": 53,
|
141 |
+
"I-VOID_MENU.NM": 45,
|
142 |
+
"I-VOID_MENU.PRICE": 46,
|
143 |
+
"O": 0
|
144 |
+
},
|
145 |
+
"layer_norm_eps": 1e-05,
|
146 |
+
"max_2d_position_embeddings": 1024,
|
147 |
+
"max_position_embeddings": 514,
|
148 |
+
"max_rel_2d_pos": 256,
|
149 |
+
"max_rel_pos": 128,
|
150 |
+
"model_type": "layoutlmv3",
|
151 |
+
"num_attention_heads": 12,
|
152 |
+
"num_channels": 3,
|
153 |
+
"num_hidden_layers": 12,
|
154 |
+
"pad_token_id": 1,
|
155 |
+
"patch_size": 16,
|
156 |
+
"rel_2d_pos_bins": 64,
|
157 |
+
"rel_pos_bins": 32,
|
158 |
+
"second_input_size": 112,
|
159 |
+
"shape_size": 128,
|
160 |
+
"text_embed": true,
|
161 |
+
"torch_dtype": "float32",
|
162 |
+
"transformers_version": "4.21.3",
|
163 |
+
"type_vocab_size": 1,
|
164 |
+
"visual_embed": true,
|
165 |
+
"vocab_size": 50265
|
166 |
+
}
|
merges.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|
optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cb397d2fd4fc9fd20fb3481f804497d11012a396d8a3eed3ebf0377d4d5129e3
|
3 |
+
size 1007841057
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:b80ac1d866bc5529d1bd6d76d63c8bbe757dcc75b70b996b85775378119abc2a
|
3 |
+
size 503934991
|
rng_state.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5e1995b85ba43838fb22c00186b84295b5b5154a22388f95fa5950ce217416fe
|
3 |
+
size 14503
|
scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:60ba5b1e7321bddf1c4a55ff6a28dee5eee0bf2ab6739b8d32643b3646d24439
|
3 |
+
size 623
|
special_tokens_map.json
ADDED
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"bos_token": {
|
3 |
+
"content": "<s>",
|
4 |
+
"lstrip": false,
|
5 |
+
"normalized": true,
|
6 |
+
"rstrip": false,
|
7 |
+
"single_word": false
|
8 |
+
},
|
9 |
+
"cls_token": {
|
10 |
+
"content": "<s>",
|
11 |
+
"lstrip": false,
|
12 |
+
"normalized": true,
|
13 |
+
"rstrip": false,
|
14 |
+
"single_word": false
|
15 |
+
},
|
16 |
+
"eos_token": {
|
17 |
+
"content": "</s>",
|
18 |
+
"lstrip": false,
|
19 |
+
"normalized": true,
|
20 |
+
"rstrip": false,
|
21 |
+
"single_word": false
|
22 |
+
},
|
23 |
+
"mask_token": {
|
24 |
+
"content": "<mask>",
|
25 |
+
"lstrip": true,
|
26 |
+
"normalized": true,
|
27 |
+
"rstrip": false,
|
28 |
+
"single_word": false
|
29 |
+
},
|
30 |
+
"pad_token": {
|
31 |
+
"content": "<pad>",
|
32 |
+
"lstrip": false,
|
33 |
+
"normalized": true,
|
34 |
+
"rstrip": false,
|
35 |
+
"single_word": false
|
36 |
+
},
|
37 |
+
"sep_token": {
|
38 |
+
"content": "</s>",
|
39 |
+
"lstrip": false,
|
40 |
+
"normalized": true,
|
41 |
+
"rstrip": false,
|
42 |
+
"single_word": false
|
43 |
+
},
|
44 |
+
"unk_token": {
|
45 |
+
"content": "<unk>",
|
46 |
+
"lstrip": false,
|
47 |
+
"normalized": true,
|
48 |
+
"rstrip": false,
|
49 |
+
"single_word": false
|
50 |
+
}
|
51 |
+
}
|
tokenizer_config.json
ADDED
@@ -0,0 +1,86 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"add_prefix_space": true,
|
3 |
+
"apply_ocr": false,
|
4 |
+
"bos_token": {
|
5 |
+
"__type": "AddedToken",
|
6 |
+
"content": "<s>",
|
7 |
+
"lstrip": false,
|
8 |
+
"normalized": true,
|
9 |
+
"rstrip": false,
|
10 |
+
"single_word": false
|
11 |
+
},
|
12 |
+
"cls_token": {
|
13 |
+
"__type": "AddedToken",
|
14 |
+
"content": "<s>",
|
15 |
+
"lstrip": false,
|
16 |
+
"normalized": true,
|
17 |
+
"rstrip": false,
|
18 |
+
"single_word": false
|
19 |
+
},
|
20 |
+
"cls_token_box": [
|
21 |
+
0,
|
22 |
+
0,
|
23 |
+
0,
|
24 |
+
0
|
25 |
+
],
|
26 |
+
"eos_token": {
|
27 |
+
"__type": "AddedToken",
|
28 |
+
"content": "</s>",
|
29 |
+
"lstrip": false,
|
30 |
+
"normalized": true,
|
31 |
+
"rstrip": false,
|
32 |
+
"single_word": false
|
33 |
+
},
|
34 |
+
"errors": "replace",
|
35 |
+
"mask_token": {
|
36 |
+
"__type": "AddedToken",
|
37 |
+
"content": "<mask>",
|
38 |
+
"lstrip": true,
|
39 |
+
"normalized": true,
|
40 |
+
"rstrip": false,
|
41 |
+
"single_word": false
|
42 |
+
},
|
43 |
+
"model_max_length": 512,
|
44 |
+
"name_or_path": "microsoft/layoutlmv3-base",
|
45 |
+
"only_label_first_subword": true,
|
46 |
+
"pad_token": {
|
47 |
+
"__type": "AddedToken",
|
48 |
+
"content": "<pad>",
|
49 |
+
"lstrip": false,
|
50 |
+
"normalized": true,
|
51 |
+
"rstrip": false,
|
52 |
+
"single_word": false
|
53 |
+
},
|
54 |
+
"pad_token_box": [
|
55 |
+
0,
|
56 |
+
0,
|
57 |
+
0,
|
58 |
+
0
|
59 |
+
],
|
60 |
+
"pad_token_label": -100,
|
61 |
+
"sep_token": {
|
62 |
+
"__type": "AddedToken",
|
63 |
+
"content": "</s>",
|
64 |
+
"lstrip": false,
|
65 |
+
"normalized": true,
|
66 |
+
"rstrip": false,
|
67 |
+
"single_word": false
|
68 |
+
},
|
69 |
+
"sep_token_box": [
|
70 |
+
0,
|
71 |
+
0,
|
72 |
+
0,
|
73 |
+
0
|
74 |
+
],
|
75 |
+
"special_tokens_map_file": null,
|
76 |
+
"tokenizer_class": "LayoutLMv3Tokenizer",
|
77 |
+
"tokenizer_file": null,
|
78 |
+
"unk_token": {
|
79 |
+
"__type": "AddedToken",
|
80 |
+
"content": "<unk>",
|
81 |
+
"lstrip": false,
|
82 |
+
"normalized": true,
|
83 |
+
"rstrip": false,
|
84 |
+
"single_word": false
|
85 |
+
}
|
86 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,148 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 0.8999713384924047,
|
3 |
+
"best_model_checkpoint": "C:\\Users\\keldr\\Desktop\\Workspace\\Repo\\12Sep2022\\model\\layoutlmv3\\sg_receipt\\checkpoint-1000",
|
4 |
+
"epoch": 1.1350737797956867,
|
5 |
+
"global_step": 1000,
|
6 |
+
"is_hyper_param_search": false,
|
7 |
+
"is_local_process_zero": true,
|
8 |
+
"is_world_process_zero": true,
|
9 |
+
"log_history": [
|
10 |
+
{
|
11 |
+
"epoch": 0.11,
|
12 |
+
"eval_accuracy": 0.8920392584514721,
|
13 |
+
"eval_f1": 0.8600230414746544,
|
14 |
+
"eval_loss": 0.5883092880249023,
|
15 |
+
"eval_precision": 0.8575531303848363,
|
16 |
+
"eval_recall": 0.8625072212593876,
|
17 |
+
"eval_runtime": 12.796,
|
18 |
+
"eval_samples_per_second": 9.456,
|
19 |
+
"eval_steps_per_second": 9.456,
|
20 |
+
"step": 100
|
21 |
+
},
|
22 |
+
{
|
23 |
+
"epoch": 0.23,
|
24 |
+
"eval_accuracy": 0.8964013086150491,
|
25 |
+
"eval_f1": 0.8633629332569465,
|
26 |
+
"eval_loss": 0.5547744631767273,
|
27 |
+
"eval_precision": 0.85625,
|
28 |
+
"eval_recall": 0.8705950317735413,
|
29 |
+
"eval_runtime": 12.4455,
|
30 |
+
"eval_samples_per_second": 9.722,
|
31 |
+
"eval_steps_per_second": 9.722,
|
32 |
+
"step": 200
|
33 |
+
},
|
34 |
+
{
|
35 |
+
"epoch": 0.34,
|
36 |
+
"eval_accuracy": 0.8927662668120684,
|
37 |
+
"eval_f1": 0.871264367816092,
|
38 |
+
"eval_loss": 0.5231363773345947,
|
39 |
+
"eval_precision": 0.866781017724414,
|
40 |
+
"eval_recall": 0.8757943385326401,
|
41 |
+
"eval_runtime": 12.511,
|
42 |
+
"eval_samples_per_second": 9.672,
|
43 |
+
"eval_steps_per_second": 9.672,
|
44 |
+
"step": 300
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.45,
|
48 |
+
"eval_accuracy": 0.8993093420574336,
|
49 |
+
"eval_f1": 0.878761822871883,
|
50 |
+
"eval_loss": 0.502784788608551,
|
51 |
+
"eval_precision": 0.8720136518771331,
|
52 |
+
"eval_recall": 0.8856152512998267,
|
53 |
+
"eval_runtime": 12.5449,
|
54 |
+
"eval_samples_per_second": 9.645,
|
55 |
+
"eval_steps_per_second": 9.645,
|
56 |
+
"step": 400
|
57 |
+
},
|
58 |
+
{
|
59 |
+
"epoch": 0.57,
|
60 |
+
"learning_rate": 9.432463110102157e-06,
|
61 |
+
"loss": 0.4376,
|
62 |
+
"step": 500
|
63 |
+
},
|
64 |
+
{
|
65 |
+
"epoch": 0.57,
|
66 |
+
"eval_accuracy": 0.9043984005816067,
|
67 |
+
"eval_f1": 0.8830795748348176,
|
68 |
+
"eval_loss": 0.4736110270023346,
|
69 |
+
"eval_precision": 0.8782857142857143,
|
70 |
+
"eval_recall": 0.8879260543038706,
|
71 |
+
"eval_runtime": 12.6247,
|
72 |
+
"eval_samples_per_second": 9.584,
|
73 |
+
"eval_steps_per_second": 9.584,
|
74 |
+
"step": 500
|
75 |
+
},
|
76 |
+
{
|
77 |
+
"epoch": 0.68,
|
78 |
+
"eval_accuracy": 0.9091239549254816,
|
79 |
+
"eval_f1": 0.8862722573233774,
|
80 |
+
"eval_loss": 0.46922487020492554,
|
81 |
+
"eval_precision": 0.8812107367218732,
|
82 |
+
"eval_recall": 0.8913922588099364,
|
83 |
+
"eval_runtime": 14.0441,
|
84 |
+
"eval_samples_per_second": 8.616,
|
85 |
+
"eval_steps_per_second": 8.616,
|
86 |
+
"step": 600
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 0.79,
|
90 |
+
"eval_accuracy": 0.9156670301708469,
|
91 |
+
"eval_f1": 0.9011494252873562,
|
92 |
+
"eval_loss": 0.4721052646636963,
|
93 |
+
"eval_precision": 0.8965122927387078,
|
94 |
+
"eval_recall": 0.9058347775852109,
|
95 |
+
"eval_runtime": 13.5885,
|
96 |
+
"eval_samples_per_second": 8.905,
|
97 |
+
"eval_steps_per_second": 8.905,
|
98 |
+
"step": 700
|
99 |
+
},
|
100 |
+
{
|
101 |
+
"epoch": 0.91,
|
102 |
+
"eval_accuracy": 0.9094874591057798,
|
103 |
+
"eval_f1": 0.8888888888888888,
|
104 |
+
"eval_loss": 0.4439501464366913,
|
105 |
+
"eval_precision": 0.8858290304073436,
|
106 |
+
"eval_recall": 0.8919699595609474,
|
107 |
+
"eval_runtime": 12.9799,
|
108 |
+
"eval_samples_per_second": 9.322,
|
109 |
+
"eval_steps_per_second": 9.322,
|
110 |
+
"step": 800
|
111 |
+
},
|
112 |
+
{
|
113 |
+
"epoch": 1.02,
|
114 |
+
"eval_accuracy": 0.905852417302799,
|
115 |
+
"eval_f1": 0.8834812482107071,
|
116 |
+
"eval_loss": 0.4436013102531433,
|
117 |
+
"eval_precision": 0.8757094211123723,
|
118 |
+
"eval_recall": 0.8913922588099364,
|
119 |
+
"eval_runtime": 12.664,
|
120 |
+
"eval_samples_per_second": 9.555,
|
121 |
+
"eval_steps_per_second": 9.555,
|
122 |
+
"step": 900
|
123 |
+
},
|
124 |
+
{
|
125 |
+
"epoch": 1.14,
|
126 |
+
"learning_rate": 8.864926220204314e-06,
|
127 |
+
"loss": 0.2859,
|
128 |
+
"step": 1000
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"epoch": 1.14,
|
132 |
+
"eval_accuracy": 0.9240276263177026,
|
133 |
+
"eval_f1": 0.8999713384924047,
|
134 |
+
"eval_loss": 0.4385237395763397,
|
135 |
+
"eval_precision": 0.8930602957906713,
|
136 |
+
"eval_recall": 0.9069901790872328,
|
137 |
+
"eval_runtime": 12.782,
|
138 |
+
"eval_samples_per_second": 9.466,
|
139 |
+
"eval_steps_per_second": 9.466,
|
140 |
+
"step": 1000
|
141 |
+
}
|
142 |
+
],
|
143 |
+
"max_steps": 8810,
|
144 |
+
"num_train_epochs": 10,
|
145 |
+
"total_flos": 265547209728000.0,
|
146 |
+
"trial_name": null,
|
147 |
+
"trial_params": null
|
148 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a688d53928461862e70d05d28157eda3c5436d6cbbfd7d76e319a14254943534
|
3 |
+
size 3439
|
vocab.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|