abletobetable
commited on
Commit
•
0308fdc
1
Parent(s):
aba9679
Training in progress, step 400
Browse files- .gitignore +1 -0
- config.json +42 -0
- pytorch_model.bin +3 -0
- special_tokens_map.json +3 -0
- tokenizer.json +1050 -0
- tokenizer_config.json +5 -0
- training_args.bin +3 -0
.gitignore
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
checkpoint-*/
|
config.json
ADDED
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "distilbert-base-uncased",
|
3 |
+
"activation": "gelu",
|
4 |
+
"architectures": [
|
5 |
+
"DistilBertForSequenceClassification"
|
6 |
+
],
|
7 |
+
"attention_dropout": 0.6,
|
8 |
+
"dim": 256,
|
9 |
+
"dropout": 0.6,
|
10 |
+
"from_pretrained": false,
|
11 |
+
"hidden_dim": 1024,
|
12 |
+
"id2label": {
|
13 |
+
"0": "LABEL_0",
|
14 |
+
"1": "LABEL_1",
|
15 |
+
"2": "LABEL_2",
|
16 |
+
"3": "LABEL_3",
|
17 |
+
"4": "LABEL_4",
|
18 |
+
"5": "LABEL_5"
|
19 |
+
},
|
20 |
+
"initializer_range": 0.02,
|
21 |
+
"label2id": {
|
22 |
+
"LABEL_0": 0,
|
23 |
+
"LABEL_1": 1,
|
24 |
+
"LABEL_2": 2,
|
25 |
+
"LABEL_3": 3,
|
26 |
+
"LABEL_4": 4,
|
27 |
+
"LABEL_5": 5
|
28 |
+
},
|
29 |
+
"max_position_embeddings": 24,
|
30 |
+
"model_type": "distilbert",
|
31 |
+
"n_heads": 4,
|
32 |
+
"n_layers": 8,
|
33 |
+
"pad_token_id": 0,
|
34 |
+
"problem_type": "single_label_classification",
|
35 |
+
"qa_dropout": 0.6,
|
36 |
+
"seq_classif_dropout": 0.6,
|
37 |
+
"sinusoidal_pos_embds": false,
|
38 |
+
"tie_weights_": true,
|
39 |
+
"torch_dtype": "float32",
|
40 |
+
"transformers_version": "4.28.1",
|
41 |
+
"vocab_size": 256
|
42 |
+
}
|
pytorch_model.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:6ec9318347d9ad9b94769f181cefc615625abce3eb7ecc41cd181044e26a24be
|
3 |
+
size 25877037
|
special_tokens_map.json
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"pad_token": "[PAD]"
|
3 |
+
}
|
tokenizer.json
ADDED
@@ -0,0 +1,1050 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"version": "1.0",
|
3 |
+
"truncation": null,
|
4 |
+
"padding": null,
|
5 |
+
"added_tokens": [
|
6 |
+
{
|
7 |
+
"id": 0,
|
8 |
+
"content": "[PAD]",
|
9 |
+
"single_word": false,
|
10 |
+
"lstrip": false,
|
11 |
+
"rstrip": false,
|
12 |
+
"normalized": false,
|
13 |
+
"special": true
|
14 |
+
}
|
15 |
+
],
|
16 |
+
"normalizer": null,
|
17 |
+
"pre_tokenizer": null,
|
18 |
+
"post_processor": null,
|
19 |
+
"decoder": null,
|
20 |
+
"model": {
|
21 |
+
"type": "Unigram",
|
22 |
+
"unk_id": null,
|
23 |
+
"vocab": [
|
24 |
+
[
|
25 |
+
"[PAD]",
|
26 |
+
0.0
|
27 |
+
],
|
28 |
+
[
|
29 |
+
"е",
|
30 |
+
-3.1636796363706203
|
31 |
+
],
|
32 |
+
[
|
33 |
+
"с",
|
34 |
+
-3.359173210921522
|
35 |
+
],
|
36 |
+
[
|
37 |
+
"м",
|
38 |
+
-3.4767311793904128
|
39 |
+
],
|
40 |
+
[
|
41 |
+
"в",
|
42 |
+
-3.6053862526055447
|
43 |
+
],
|
44 |
+
[
|
45 |
+
"и",
|
46 |
+
-3.6510660137914854
|
47 |
+
],
|
48 |
+
[
|
49 |
+
"о",
|
50 |
+
-3.7152455221804175
|
51 |
+
],
|
52 |
+
[
|
53 |
+
"л",
|
54 |
+
-3.8174792636171873
|
55 |
+
],
|
56 |
+
[
|
57 |
+
"а",
|
58 |
+
-3.9030196208874592
|
59 |
+
],
|
60 |
+
[
|
61 |
+
"т",
|
62 |
+
-3.919992837958791
|
63 |
+
],
|
64 |
+
[
|
65 |
+
"у",
|
66 |
+
-3.9304056679642887
|
67 |
+
],
|
68 |
+
[
|
69 |
+
"й",
|
70 |
+
-4.05418950185703
|
71 |
+
],
|
72 |
+
[
|
73 |
+
"к",
|
74 |
+
-4.102994397537685
|
75 |
+
],
|
76 |
+
[
|
77 |
+
"х",
|
78 |
+
-4.103691610663569
|
79 |
+
],
|
80 |
+
[
|
81 |
+
"р",
|
82 |
+
-4.163663784247136
|
83 |
+
],
|
84 |
+
[
|
85 |
+
"я",
|
86 |
+
-4.18517039517814
|
87 |
+
],
|
88 |
+
[
|
89 |
+
"д",
|
90 |
+
-4.293637726858796
|
91 |
+
],
|
92 |
+
[
|
93 |
+
"по",
|
94 |
+
-4.318874404725777
|
95 |
+
],
|
96 |
+
[
|
97 |
+
"н",
|
98 |
+
-4.319659895767661
|
99 |
+
],
|
100 |
+
[
|
101 |
+
"ра",
|
102 |
+
-4.326548476322547
|
103 |
+
],
|
104 |
+
[
|
105 |
+
"ли",
|
106 |
+
-4.463338075253289
|
107 |
+
],
|
108 |
+
[
|
109 |
+
"на",
|
110 |
+
-4.548396485176257
|
111 |
+
],
|
112 |
+
[
|
113 |
+
"ка",
|
114 |
+
-4.557609822166179
|
115 |
+
],
|
116 |
+
[
|
117 |
+
"г",
|
118 |
+
-4.578202766924267
|
119 |
+
],
|
120 |
+
[
|
121 |
+
"п",
|
122 |
+
-4.5948832974278995
|
123 |
+
],
|
124 |
+
[
|
125 |
+
"ла",
|
126 |
+
-4.622429724624767
|
127 |
+
],
|
128 |
+
[
|
129 |
+
"з",
|
130 |
+
-4.668711901804957
|
131 |
+
],
|
132 |
+
[
|
133 |
+
"за",
|
134 |
+
-4.679470001112353
|
135 |
+
],
|
136 |
+
[
|
137 |
+
"ж",
|
138 |
+
-4.692888448272974
|
139 |
+
],
|
140 |
+
[
|
141 |
+
"ны",
|
142 |
+
-4.717019931310672
|
143 |
+
],
|
144 |
+
[
|
145 |
+
"ни",
|
146 |
+
-4.717178138341595
|
147 |
+
],
|
148 |
+
[
|
149 |
+
"б",
|
150 |
+
-4.724329749814947
|
151 |
+
],
|
152 |
+
[
|
153 |
+
"ре",
|
154 |
+
-4.817940624969843
|
155 |
+
],
|
156 |
+
[
|
157 |
+
"ся",
|
158 |
+
-4.837228337071683
|
159 |
+
],
|
160 |
+
[
|
161 |
+
"не",
|
162 |
+
-4.847884203151359
|
163 |
+
],
|
164 |
+
[
|
165 |
+
"ч",
|
166 |
+
-4.8875128612907455
|
167 |
+
],
|
168 |
+
[
|
169 |
+
"ро",
|
170 |
+
-4.926997101838515
|
171 |
+
],
|
172 |
+
[
|
173 |
+
"ло",
|
174 |
+
-4.974133703658512
|
175 |
+
],
|
176 |
+
[
|
177 |
+
"та",
|
178 |
+
-5.037108135230574
|
179 |
+
],
|
180 |
+
[
|
181 |
+
"ми",
|
182 |
+
-5.050561901133188
|
183 |
+
],
|
184 |
+
[
|
185 |
+
"вы",
|
186 |
+
-5.066939597672465
|
187 |
+
],
|
188 |
+
[
|
189 |
+
"ю",
|
190 |
+
-5.0733887719846615
|
191 |
+
],
|
192 |
+
[
|
193 |
+
"ва",
|
194 |
+
-5.113926274278541
|
195 |
+
],
|
196 |
+
[
|
197 |
+
"ш",
|
198 |
+
-5.117244871533596
|
199 |
+
],
|
200 |
+
[
|
201 |
+
"ки",
|
202 |
+
-5.119883285875997
|
203 |
+
],
|
204 |
+
[
|
205 |
+
"но",
|
206 |
+
-5.122834314800923
|
207 |
+
],
|
208 |
+
[
|
209 |
+
"ль",
|
210 |
+
-5.140078050336989
|
211 |
+
],
|
212 |
+
[
|
213 |
+
"во",
|
214 |
+
-5.172027764098102
|
215 |
+
],
|
216 |
+
[
|
217 |
+
"ет",
|
218 |
+
-5.177442189854449
|
219 |
+
],
|
220 |
+
[
|
221 |
+
"про",
|
222 |
+
-5.180757178756543
|
223 |
+
],
|
224 |
+
[
|
225 |
+
"до",
|
226 |
+
-5.195432996961147
|
227 |
+
],
|
228 |
+
[
|
229 |
+
"го",
|
230 |
+
-5.202898545471496
|
231 |
+
],
|
232 |
+
[
|
233 |
+
"ы",
|
234 |
+
-5.220288580282363
|
235 |
+
],
|
236 |
+
[
|
237 |
+
"те",
|
238 |
+
-5.223318808225296
|
239 |
+
],
|
240 |
+
[
|
241 |
+
"да",
|
242 |
+
-5.258319969867449
|
243 |
+
],
|
244 |
+
[
|
245 |
+
"сь",
|
246 |
+
-5.296768127931534
|
247 |
+
],
|
248 |
+
[
|
249 |
+
"при",
|
250 |
+
-5.300719847408468
|
251 |
+
],
|
252 |
+
[
|
253 |
+
"ст",
|
254 |
+
-5.312071600207284
|
255 |
+
],
|
256 |
+
[
|
257 |
+
"об",
|
258 |
+
-5.321440340809567
|
259 |
+
],
|
260 |
+
[
|
261 |
+
"ри",
|
262 |
+
-5.351944024735441
|
263 |
+
],
|
264 |
+
[
|
265 |
+
"ру",
|
266 |
+
-5.382717583024741
|
267 |
+
],
|
268 |
+
[
|
269 |
+
"де",
|
270 |
+
-5.42335709376504
|
271 |
+
],
|
272 |
+
[
|
273 |
+
"ти",
|
274 |
+
-5.430152754589945
|
275 |
+
],
|
276 |
+
[
|
277 |
+
"от",
|
278 |
+
-5.452013842602508
|
279 |
+
],
|
280 |
+
[
|
281 |
+
"па",
|
282 |
+
-5.461349448256417
|
283 |
+
],
|
284 |
+
[
|
285 |
+
"ди",
|
286 |
+
-5.462112441879387
|
287 |
+
],
|
288 |
+
[
|
289 |
+
"со",
|
290 |
+
-5.47073375693526
|
291 |
+
],
|
292 |
+
[
|
293 |
+
"ку",
|
294 |
+
-5.4745889291199274
|
295 |
+
],
|
296 |
+
[
|
297 |
+
"ко",
|
298 |
+
-5.504473412608788
|
299 |
+
],
|
300 |
+
[
|
301 |
+
"то",
|
302 |
+
-5.504860293985376
|
303 |
+
],
|
304 |
+
[
|
305 |
+
"га",
|
306 |
+
-5.682461175015035
|
307 |
+
],
|
308 |
+
[
|
309 |
+
"бо",
|
310 |
+
-5.702286932153875
|
311 |
+
],
|
312 |
+
[
|
313 |
+
"ля",
|
314 |
+
-5.707186880493859
|
315 |
+
],
|
316 |
+
[
|
317 |
+
"чи",
|
318 |
+
-5.726149608807444
|
319 |
+
],
|
320 |
+
[
|
321 |
+
"ры",
|
322 |
+
-5.7581398104776715
|
323 |
+
],
|
324 |
+
[
|
325 |
+
"ь",
|
326 |
+
-5.769722615320581
|
327 |
+
],
|
328 |
+
[
|
329 |
+
"раз",
|
330 |
+
-5.777657793853896
|
331 |
+
],
|
332 |
+
[
|
333 |
+
"ин",
|
334 |
+
-5.783559540075619
|
335 |
+
],
|
336 |
+
[
|
337 |
+
"ной",
|
338 |
+
-5.7929381147642065
|
339 |
+
],
|
340 |
+
[
|
341 |
+
"ен",
|
342 |
+
-5.810652857528928
|
343 |
+
],
|
344 |
+
[
|
345 |
+
"под",
|
346 |
+
-5.843061465852916
|
347 |
+
],
|
348 |
+
[
|
349 |
+
"ан",
|
350 |
+
-5.865100339367238
|
351 |
+
],
|
352 |
+
[
|
353 |
+
"пи",
|
354 |
+
-5.879046574282177
|
355 |
+
],
|
356 |
+
[
|
357 |
+
"ар",
|
358 |
+
-5.893862057404114
|
359 |
+
],
|
360 |
+
[
|
361 |
+
"ть",
|
362 |
+
-5.897306442912067
|
363 |
+
],
|
364 |
+
[
|
365 |
+
"ф",
|
366 |
+
-5.902229127271682
|
367 |
+
],
|
368 |
+
[
|
369 |
+
"жи",
|
370 |
+
-5.925370546484607
|
371 |
+
],
|
372 |
+
[
|
373 |
+
"ци",
|
374 |
+
-5.928100752513754
|
375 |
+
],
|
376 |
+
[
|
377 |
+
"че",
|
378 |
+
-5.945348977901575
|
379 |
+
],
|
380 |
+
[
|
381 |
+
"ты",
|
382 |
+
-5.972166760117742
|
383 |
+
],
|
384 |
+
[
|
385 |
+
"ба",
|
386 |
+
-5.977038105260874
|
387 |
+
],
|
388 |
+
[
|
389 |
+
"вер",
|
390 |
+
-5.983428566220837
|
391 |
+
],
|
392 |
+
[
|
393 |
+
"ная",
|
394 |
+
-5.990849208890614
|
395 |
+
],
|
396 |
+
[
|
397 |
+
"ча",
|
398 |
+
-6.013202406107329
|
399 |
+
],
|
400 |
+
[
|
401 |
+
"пе",
|
402 |
+
-6.032341480893762
|
403 |
+
],
|
404 |
+
[
|
405 |
+
"са",
|
406 |
+
-6.034450309096376
|
407 |
+
],
|
408 |
+
[
|
409 |
+
"ем",
|
410 |
+
-6.037807542535035
|
411 |
+
],
|
412 |
+
[
|
413 |
+
"ще",
|
414 |
+
-6.038370941834772
|
415 |
+
],
|
416 |
+
[
|
417 |
+
"ше",
|
418 |
+
-6.062354644467771
|
419 |
+
],
|
420 |
+
[
|
421 |
+
"из",
|
422 |
+
-6.0662889525004005
|
423 |
+
],
|
424 |
+
[
|
425 |
+
"ств",
|
426 |
+
-6.069384268201753
|
427 |
+
],
|
428 |
+
[
|
429 |
+
"ного",
|
430 |
+
-6.082294432517265
|
431 |
+
],
|
432 |
+
[
|
433 |
+
"ско",
|
434 |
+
-6.082319149708081
|
435 |
+
],
|
436 |
+
[
|
437 |
+
"э",
|
438 |
+
-6.100856930219239
|
439 |
+
],
|
440 |
+
[
|
441 |
+
"ком",
|
442 |
+
-6.103344547667658
|
443 |
+
],
|
444 |
+
[
|
445 |
+
"ски",
|
446 |
+
-6.1068794661162125
|
447 |
+
],
|
448 |
+
[
|
449 |
+
"ё",
|
450 |
+
-6.119189253137699
|
451 |
+
],
|
452 |
+
[
|
453 |
+
"ов",
|
454 |
+
-6.137817559051781
|
455 |
+
],
|
456 |
+
[
|
457 |
+
"би",
|
458 |
+
-6.178944074692679
|
459 |
+
],
|
460 |
+
[
|
461 |
+
"ста",
|
462 |
+
-6.185855399748316
|
463 |
+
],
|
464 |
+
[
|
465 |
+
"пу",
|
466 |
+
-6.204423533569658
|
467 |
+
],
|
468 |
+
[
|
469 |
+
"ков",
|
470 |
+
-6.213633369881911
|
471 |
+
],
|
472 |
+
[
|
473 |
+
"рас",
|
474 |
+
-6.215736329120726
|
475 |
+
],
|
476 |
+
[
|
477 |
+
"пере",
|
478 |
+
-6.217680605082461
|
479 |
+
],
|
480 |
+
[
|
481 |
+
"ши",
|
482 |
+
-6.237591353842036
|
483 |
+
],
|
484 |
+
[
|
485 |
+
"ить",
|
486 |
+
-6.252525660416677
|
487 |
+
],
|
488 |
+
[
|
489 |
+
"ну",
|
490 |
+
-6.263488690915469
|
491 |
+
],
|
492 |
+
[
|
493 |
+
"ду",
|
494 |
+
-6.272614811909408
|
495 |
+
],
|
496 |
+
[
|
497 |
+
"це",
|
498 |
+
-6.290245699117212
|
499 |
+
],
|
500 |
+
[
|
501 |
+
"сто",
|
502 |
+
-6.298801808923643
|
503 |
+
],
|
504 |
+
[
|
505 |
+
"енны",
|
506 |
+
-6.302090293473124
|
507 |
+
],
|
508 |
+
[
|
509 |
+
"ц",
|
510 |
+
-6.308440789473085
|
511 |
+
],
|
512 |
+
[
|
513 |
+
"сти",
|
514 |
+
-6.334320371974208
|
515 |
+
],
|
516 |
+
[
|
517 |
+
"же",
|
518 |
+
-6.335635473316215
|
519 |
+
],
|
520 |
+
[
|
521 |
+
"ал",
|
522 |
+
-6.358520860862264
|
523 |
+
],
|
524 |
+
[
|
525 |
+
"му",
|
526 |
+
-6.364234522588495
|
527 |
+
],
|
528 |
+
[
|
529 |
+
"ров",
|
530 |
+
-6.383232073647166
|
531 |
+
],
|
532 |
+
[
|
533 |
+
"кон",
|
534 |
+
-6.390409708582491
|
535 |
+
],
|
536 |
+
[
|
537 |
+
"нов",
|
538 |
+
-6.403321122273851
|
539 |
+
],
|
540 |
+
[
|
541 |
+
"ном",
|
542 |
+
-6.404496872419566
|
543 |
+
],
|
544 |
+
[
|
545 |
+
"мен",
|
546 |
+
-6.4064413418034265
|
547 |
+
],
|
548 |
+
[
|
549 |
+
"су",
|
550 |
+
-6.409942213032797
|
551 |
+
],
|
552 |
+
[
|
553 |
+
"ги",
|
554 |
+
-6.410630152721253
|
555 |
+
],
|
556 |
+
[
|
557 |
+
"ное",
|
558 |
+
-6.420773351327056
|
559 |
+
],
|
560 |
+
[
|
561 |
+
"ск",
|
562 |
+
-6.445759163556911
|
563 |
+
],
|
564 |
+
[
|
565 |
+
"тор",
|
566 |
+
-6.446487040563173
|
567 |
+
],
|
568 |
+
[
|
569 |
+
"щи",
|
570 |
+
-6.44829973786927
|
571 |
+
],
|
572 |
+
[
|
573 |
+
"тер",
|
574 |
+
-6.46119939486005
|
575 |
+
],
|
576 |
+
[
|
577 |
+
"ке",
|
578 |
+
-6.476807169306808
|
579 |
+
],
|
580 |
+
[
|
581 |
+
"пред",
|
582 |
+
-6.480919764600885
|
583 |
+
],
|
584 |
+
[
|
585 |
+
"ту",
|
586 |
+
-6.51078072140959
|
587 |
+
],
|
588 |
+
[
|
589 |
+
"ша",
|
590 |
+
-6.516865779803864
|
591 |
+
],
|
592 |
+
[
|
593 |
+
"бу",
|
594 |
+
-6.527673142443851
|
595 |
+
],
|
596 |
+
[
|
597 |
+
"ения",
|
598 |
+
-6.538926614200793
|
599 |
+
],
|
600 |
+
[
|
601 |
+
"ную",
|
602 |
+
-6.552064033974947
|
603 |
+
],
|
604 |
+
[
|
605 |
+
"пре",
|
606 |
+
-6.553167434011977
|
607 |
+
],
|
608 |
+
[
|
609 |
+
"вал",
|
610 |
+
-6.557256564987665
|
611 |
+
],
|
612 |
+
[
|
613 |
+
"ом",
|
614 |
+
-6.568516369824362
|
615 |
+
],
|
616 |
+
[
|
617 |
+
"ать",
|
618 |
+
-6.576879449964455
|
619 |
+
],
|
620 |
+
[
|
621 |
+
"кой",
|
622 |
+
-6.586751821830786
|
623 |
+
],
|
624 |
+
[
|
625 |
+
"ным",
|
626 |
+
-6.59444236346413
|
627 |
+
],
|
628 |
+
[
|
629 |
+
"ей",
|
630 |
+
-6.599767946641588
|
631 |
+
],
|
632 |
+
[
|
633 |
+
"ил",
|
634 |
+
-6.62139500267716
|
635 |
+
],
|
636 |
+
[
|
637 |
+
"лю",
|
638 |
+
-6.623376951115419
|
639 |
+
],
|
640 |
+
[
|
641 |
+
"ня",
|
642 |
+
-6.6306206476008995
|
643 |
+
],
|
644 |
+
[
|
645 |
+
"фи",
|
646 |
+
-6.631450294454992
|
647 |
+
],
|
648 |
+
[
|
649 |
+
"щ",
|
650 |
+
-6.652009873573957
|
651 |
+
],
|
652 |
+
[
|
653 |
+
"лен",
|
654 |
+
-6.671904516820077
|
655 |
+
],
|
656 |
+
[
|
657 |
+
"ой",
|
658 |
+
-6.698990150040521
|
659 |
+
],
|
660 |
+
[
|
661 |
+
"ение",
|
662 |
+
-6.701814987349791
|
663 |
+
],
|
664 |
+
[
|
665 |
+
"тра",
|
666 |
+
-6.73501512479902
|
667 |
+
],
|
668 |
+
[
|
669 |
+
"ник",
|
670 |
+
-6.7420250834690965
|
671 |
+
],
|
672 |
+
[
|
673 |
+
"зна",
|
674 |
+
-6.762833698209765
|
675 |
+
],
|
676 |
+
[
|
677 |
+
"нны",
|
678 |
+
-6.763688359226229
|
679 |
+
],
|
680 |
+
[
|
681 |
+
"став",
|
682 |
+
-6.777063140328193
|
683 |
+
],
|
684 |
+
[
|
685 |
+
"ают",
|
686 |
+
-6.781120088269675
|
687 |
+
],
|
688 |
+
[
|
689 |
+
"ща",
|
690 |
+
-6.801003632576518
|
691 |
+
],
|
692 |
+
[
|
693 |
+
"нул",
|
694 |
+
-6.833802047509305
|
695 |
+
],
|
696 |
+
[
|
697 |
+
"ят",
|
698 |
+
-6.835997636374822
|
699 |
+
],
|
700 |
+
[
|
701 |
+
"ца",
|
702 |
+
-6.841362171180944
|
703 |
+
],
|
704 |
+
[
|
705 |
+
"вать",
|
706 |
+
-6.845472275365411
|
707 |
+
],
|
708 |
+
[
|
709 |
+
"гу",
|
710 |
+
-6.855305485842229
|
711 |
+
],
|
712 |
+
[
|
713 |
+
"тель",
|
714 |
+
-6.865088978982657
|
715 |
+
],
|
716 |
+
[
|
717 |
+
"дер",
|
718 |
+
-6.869778946510176
|
719 |
+
],
|
720 |
+
[
|
721 |
+
"ую",
|
722 |
+
-6.8832869290382055
|
723 |
+
],
|
724 |
+
[
|
725 |
+
"зы",
|
726 |
+
-6.887226954371981
|
727 |
+
],
|
728 |
+
[
|
729 |
+
"ют",
|
730 |
+
-6.901154517295787
|
731 |
+
],
|
732 |
+
[
|
733 |
+
"ого",
|
734 |
+
-6.909640881743588
|
735 |
+
],
|
736 |
+
[
|
737 |
+
"ды",
|
738 |
+
-6.9173006816290155
|
739 |
+
],
|
740 |
+
[
|
741 |
+
"вали",
|
742 |
+
-6.9235986241477425
|
743 |
+
],
|
744 |
+
[
|
745 |
+
"воз",
|
746 |
+
-6.924974892692157
|
747 |
+
],
|
748 |
+
[
|
749 |
+
"сту",
|
750 |
+
-6.938330038589573
|
751 |
+
],
|
752 |
+
[
|
753 |
+
"вши",
|
754 |
+
-6.944084484532091
|
755 |
+
],
|
756 |
+
[
|
757 |
+
"мер",
|
758 |
+
-6.949151102970281
|
759 |
+
],
|
760 |
+
[
|
761 |
+
"стр",
|
762 |
+
-6.96429947532023
|
763 |
+
],
|
764 |
+
[
|
765 |
+
"вед",
|
766 |
+
-6.973302237822137
|
767 |
+
],
|
768 |
+
[
|
769 |
+
"вод",
|
770 |
+
-7.0079649322914825
|
771 |
+
],
|
772 |
+
[
|
773 |
+
"ская",
|
774 |
+
-7.012072077773152
|
775 |
+
],
|
776 |
+
[
|
777 |
+
"лов",
|
778 |
+
-7.017291086492113
|
779 |
+
],
|
780 |
+
[
|
781 |
+
"зи",
|
782 |
+
-7.018282598482377
|
783 |
+
],
|
784 |
+
[
|
785 |
+
"мат",
|
786 |
+
-7.032293879761621
|
787 |
+
],
|
788 |
+
[
|
789 |
+
"сы",
|
790 |
+
-7.045030730480992
|
791 |
+
],
|
792 |
+
[
|
793 |
+
"вает",
|
794 |
+
-7.061487717557234
|
795 |
+
],
|
796 |
+
[
|
797 |
+
"фе",
|
798 |
+
-7.061960024820606
|
799 |
+
],
|
800 |
+
[
|
801 |
+
"ность",
|
802 |
+
-7.062069414670495
|
803 |
+
],
|
804 |
+
[
|
805 |
+
"альны",
|
806 |
+
-7.0660079813306345
|
807 |
+
],
|
808 |
+
[
|
809 |
+
"тельны",
|
810 |
+
-7.070899992884047
|
811 |
+
],
|
812 |
+
[
|
813 |
+
"ически",
|
814 |
+
-7.074910446308693
|
815 |
+
],
|
816 |
+
[
|
817 |
+
"кор",
|
818 |
+
-7.077010664573005
|
819 |
+
],
|
820 |
+
[
|
821 |
+
"стро",
|
822 |
+
-7.082808344718582
|
823 |
+
],
|
824 |
+
[
|
825 |
+
"бы",
|
826 |
+
-7.089034041901167
|
827 |
+
],
|
828 |
+
[
|
829 |
+
"енно",
|
830 |
+
-7.092846340929076
|
831 |
+
],
|
832 |
+
[
|
833 |
+
"ально",
|
834 |
+
-7.10679840132555
|
835 |
+
],
|
836 |
+
[
|
837 |
+
"жен",
|
838 |
+
-7.115828365681051
|
839 |
+
],
|
840 |
+
[
|
841 |
+
"пер",
|
842 |
+
-7.14065227425605
|
843 |
+
],
|
844 |
+
[
|
845 |
+
"нь",
|
846 |
+
-7.146275463065723
|
847 |
+
],
|
848 |
+
[
|
849 |
+
"ство",
|
850 |
+
-7.191344050847166
|
851 |
+
],
|
852 |
+
[
|
853 |
+
"лени",
|
854 |
+
-7.192798006712747
|
855 |
+
],
|
856 |
+
[
|
857 |
+
"ками",
|
858 |
+
-7.223665578070644
|
859 |
+
],
|
860 |
+
[
|
861 |
+
"сер",
|
862 |
+
-7.273427212354391
|
863 |
+
],
|
864 |
+
[
|
865 |
+
"ция",
|
866 |
+
-7.309302043202202
|
867 |
+
],
|
868 |
+
[
|
869 |
+
"пы",
|
870 |
+
-7.311239771943473
|
871 |
+
],
|
872 |
+
[
|
873 |
+
"ическ",
|
874 |
+
-7.362626913781078
|
875 |
+
],
|
876 |
+
[
|
877 |
+
"ман",
|
878 |
+
-7.372506880655845
|
879 |
+
],
|
880 |
+
[
|
881 |
+
"цы",
|
882 |
+
-7.380877909236978
|
883 |
+
],
|
884 |
+
[
|
885 |
+
"лось",
|
886 |
+
-7.388421354263928
|
887 |
+
],
|
888 |
+
[
|
889 |
+
"фа",
|
890 |
+
-7.396928557939997
|
891 |
+
],
|
892 |
+
[
|
893 |
+
"полн",
|
894 |
+
-7.404319309649214
|
895 |
+
],
|
896 |
+
[
|
897 |
+
"бес",
|
898 |
+
-7.4114860814725665
|
899 |
+
],
|
900 |
+
[
|
901 |
+
"ания",
|
902 |
+
-7.424202557681505
|
903 |
+
],
|
904 |
+
[
|
905 |
+
"бер",
|
906 |
+
-7.440982773293651
|
907 |
+
],
|
908 |
+
[
|
909 |
+
"итель",
|
910 |
+
-7.461792961820363
|
911 |
+
],
|
912 |
+
[
|
913 |
+
"вес",
|
914 |
+
-7.467919044018323
|
915 |
+
],
|
916 |
+
[
|
917 |
+
"мол",
|
918 |
+
-7.473308124014182
|
919 |
+
],
|
920 |
+
[
|
921 |
+
"ант",
|
922 |
+
-7.478987381298589
|
923 |
+
],
|
924 |
+
[
|
925 |
+
"фор",
|
926 |
+
-7.484167266702338
|
927 |
+
],
|
928 |
+
[
|
929 |
+
"ключ",
|
930 |
+
-7.498912525005472
|
931 |
+
],
|
932 |
+
[
|
933 |
+
"ности",
|
934 |
+
-7.505175854659576
|
935 |
+
],
|
936 |
+
[
|
937 |
+
"ешь",
|
938 |
+
-7.508768160872986
|
939 |
+
],
|
940 |
+
[
|
941 |
+
"авто",
|
942 |
+
-7.538910090995358
|
943 |
+
],
|
944 |
+
[
|
945 |
+
"теле",
|
946 |
+
-7.540809540515368
|
947 |
+
],
|
948 |
+
[
|
949 |
+
"тур",
|
950 |
+
-7.541974194216193
|
951 |
+
],
|
952 |
+
[
|
953 |
+
"без",
|
954 |
+
-7.575807709091023
|
955 |
+
],
|
956 |
+
[
|
957 |
+
"вид",
|
958 |
+
-7.58781483111255
|
959 |
+
],
|
960 |
+
[
|
961 |
+
"пле",
|
962 |
+
-7.603493590785373
|
963 |
+
],
|
964 |
+
[
|
965 |
+
"клад",
|
966 |
+
-7.615845269944964
|
967 |
+
],
|
968 |
+
[
|
969 |
+
"нуть",
|
970 |
+
-7.641275309552205
|
971 |
+
],
|
972 |
+
[
|
973 |
+
"ость",
|
974 |
+
-7.646320761168189
|
975 |
+
],
|
976 |
+
[
|
977 |
+
"ание",
|
978 |
+
-7.662126846510481
|
979 |
+
],
|
980 |
+
[
|
981 |
+
"очны",
|
982 |
+
-7.664907785029916
|
983 |
+
],
|
984 |
+
[
|
985 |
+
"луч",
|
986 |
+
-7.718489513949259
|
987 |
+
],
|
988 |
+
[
|
989 |
+
"олог",
|
990 |
+
-7.732846447649657
|
991 |
+
],
|
992 |
+
[
|
993 |
+
"ительно",
|
994 |
+
-7.745348765681442
|
995 |
+
],
|
996 |
+
[
|
997 |
+
"вел",
|
998 |
+
-7.818080089048806
|
999 |
+
],
|
1000 |
+
[
|
1001 |
+
"твор",
|
1002 |
+
-7.8208575242033165
|
1003 |
+
],
|
1004 |
+
[
|
1005 |
+
"ющих",
|
1006 |
+
-7.8747498410196
|
1007 |
+
],
|
1008 |
+
[
|
1009 |
+
"ичны",
|
1010 |
+
-7.924801361025127
|
1011 |
+
],
|
1012 |
+
[
|
1013 |
+
"плат",
|
1014 |
+
-7.960841164508131
|
1015 |
+
],
|
1016 |
+
[
|
1017 |
+
"ировать",
|
1018 |
+
-7.963776371734233
|
1019 |
+
],
|
1020 |
+
[
|
1021 |
+
"благо",
|
1022 |
+
-7.98839112202585
|
1023 |
+
],
|
1024 |
+
[
|
1025 |
+
"ющий",
|
1026 |
+
-8.02454317720849
|
1027 |
+
],
|
1028 |
+
[
|
1029 |
+
"сход",
|
1030 |
+
-8.135419072857303
|
1031 |
+
],
|
1032 |
+
[
|
1033 |
+
"смотр",
|
1034 |
+
-8.14697766082976
|
1035 |
+
],
|
1036 |
+
[
|
1037 |
+
"следова",
|
1038 |
+
-8.638979587708416
|
1039 |
+
],
|
1040 |
+
[
|
1041 |
+
"электро",
|
1042 |
+
-8.73535669065444
|
1043 |
+
],
|
1044 |
+
[
|
1045 |
+
"ъ",
|
1046 |
+
-9.09744307828603
|
1047 |
+
]
|
1048 |
+
]
|
1049 |
+
}
|
1050 |
+
}
|
tokenizer_config.json
ADDED
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"clean_up_tokenization_spaces": true,
|
3 |
+
"model_max_length": 1000000000000000019884624838656,
|
4 |
+
"tokenizer_class": "PreTrainedTokenizerFast"
|
5 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cb23c241811238ccf93a19cb434413d32b9f0f9e831a03f4ed2c9e3ef355d8da
|
3 |
+
size 3579
|