Tomer commited on
Commit
898ada7
1 Parent(s): 285f378

first commit

Browse files
.gitattributes CHANGED
@@ -26,6 +26,7 @@
26
  saved_model/**/* filter=lfs diff=lfs merge=lfs -text
27
  *.tar.* filter=lfs diff=lfs merge=lfs -text
28
  *.tar filter=lfs diff=lfs merge=lfs -text
 
29
  *.tflite filter=lfs diff=lfs merge=lfs -text
30
  *.tgz filter=lfs diff=lfs merge=lfs -text
31
  *.wasm filter=lfs diff=lfs merge=lfs -text
 
26
  saved_model/**/* filter=lfs diff=lfs merge=lfs -text
27
  *.tar.* filter=lfs diff=lfs merge=lfs -text
28
  *.tar filter=lfs diff=lfs merge=lfs -text
29
+ *.csv filter=lfs diff=lfs merge=lfs -text
30
  *.tflite filter=lfs diff=lfs merge=lfs -text
31
  *.tgz filter=lfs diff=lfs merge=lfs -text
32
  *.wasm filter=lfs diff=lfs merge=lfs -text
config.json ADDED
@@ -0,0 +1,45 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "/mount/COnTRA/models/Finetuned_oSiS_150_ExR/best_model",
3
+ "activation_dropout": 0.0,
4
+ "activation_function": "gelu",
5
+ "architectures": [
6
+ "BartForConditionalGeneration"
7
+ ],
8
+ "attention_dropout": 0.0,
9
+ "bos_token_id": 0,
10
+ "classifier_dropout": 0.0,
11
+ "d_model": 256,
12
+ "decoder_attention_heads": 8,
13
+ "decoder_ffn_dim": 256,
14
+ "decoder_layerdrop": 0.0,
15
+ "decoder_layers": 6,
16
+ "decoder_start_token_id": 2,
17
+ "dropout": 0.1,
18
+ "encoder_attention_heads": 8,
19
+ "encoder_ffn_dim": 256,
20
+ "encoder_layerdrop": 0.0,
21
+ "encoder_layers": 6,
22
+ "eos_token_id": 2,
23
+ "forced_eos_token_id": 2,
24
+ "id2label": {
25
+ "0": "LABEL_0",
26
+ "1": "LABEL_1",
27
+ "2": "LABEL_2"
28
+ },
29
+ "init_std": 0.02,
30
+ "is_encoder_decoder": true,
31
+ "label2id": {
32
+ "LABEL_0": 0,
33
+ "LABEL_1": 1,
34
+ "LABEL_2": 2
35
+ },
36
+ "max_position_embeddings": 512,
37
+ "model_type": "bart",
38
+ "num_hidden_layers": 6,
39
+ "pad_token_id": 1,
40
+ "scale_embedding": false,
41
+ "torch_dtype": "float32",
42
+ "transformers_version": "4.18.0",
43
+ "use_cache": true,
44
+ "vocab_size": 108
45
+ }
configs.log ADDED
@@ -0,0 +1 @@
 
 
1
+ {"Description": "This is a finetuning configuration for codon optimization model based on BART language model.", "project_name": "SCPECBS3_changedmask", "dataset_path": "/mount/COnTRA/data/datasets/processed_data_SCPECBS3/homologs/SCPECBS3_SCPECBS3_ExprRefined150", "model_outpath": "/mount/COnTRA/models", "model_name": "Finetuned2Steps_homologs_150_ExR", "tokenizer_path": "/mount/COnTRA/tokenizers/contra_tokenizer_gen_exprrefined", "cai_refference_path": "/mount/COnTRA/data/datasets/processed_data_SCPECBS3/S_cerevisiae/S_cerevisiae.0.nt.fasta", "checkpoint_flag": true, "checkpoint_path": "/mount/COnTRA/models/Finetuned_oSiS_150_ExR/best_model", "finetune_flag": true, "dataset_single_species_flag": false, "special_token_th": 42, "sw_aa_size": 150, "model_config": {"vocab_size": 108, "max_position_embeddings": 512, "d_model": 256, "attention_layers": 6, "attention_heads": 8, "ffn_dim": 256}, "train_config": {"starting_lr": 0.0001, "mask_perc": 1.0, "batch_size": 32, "warmup_steps": 15000, "steps": 50000, "label_smoothing_factor": 0.15, "logging_steps": 500, "decoder_noise": 0.15, "seed": 1698782189}}
eval_config.json ADDED
@@ -0,0 +1,44 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "Description": "This is a pre-training configuration for codon optimization model based on BART language model.",
3
+ "project_name": "COnTRA_SCPECBS3_wintest",
4
+ "dataset_path": "/mount/COnTRA/data/datasets/processed_data_SCPECBS3/homologs/SCPECBS3_SCPECBS3_ExprRefined150",
5
+ "tokenizer_path": "/mount/COnTRA/tokenizers/contra_tokenizer_gen_exprrefined",
6
+ "cai_refference_path": "/mount/COnTRA/data/datasets/processed_data_SCPECBS3/S_cerevisiae/S_cerevisiae.0.nt.fasta",
7
+ "checkpoint_flag": true,
8
+ "checkpoint_path": "/mount/COnTRA/models/Finetuned2Steps_homologs_150_ExR/best_model/",
9
+ "special_token_th": 42,
10
+ "mask_all": false,
11
+ "sw_aa_size": 150,
12
+ "eval_type": "model",
13
+ "model_type": "COBaBExRi",
14
+ "cai_query_species": "S_cerevisiae",
15
+ "outdir": "/mount/COnTRA/models/Finetuned2Steps_homologs_150_ExR/best_model/",
16
+ "outfile": "model_eval_mimic.csv",
17
+ "out_dict": [
18
+ "prot_len",
19
+ "num_of_correct_predicted_codons",
20
+ "accuracy",
21
+ "cross_entropy_loss",
22
+ "entropy",
23
+ "query_codons",
24
+ "subject_codons",
25
+ "pred_codons"
26
+ ],
27
+ "orig_dict": [
28
+ "qseqid",
29
+ "sseqid",
30
+ "query_species",
31
+ "subject_species",
32
+ "pident",
33
+ "length",
34
+ "mismatch",
35
+ "gapopen",
36
+ "qstart",
37
+ "qend",
38
+ "sstart",
39
+ "send",
40
+ "evalue",
41
+ "bitscore"
42
+ ],
43
+ "debug": false
44
+ }
mask/eval_config.json ADDED
@@ -0,0 +1,44 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "Description": "This is a pre-training configuration for codon optimization model based on BART language model.",
3
+ "project_name": "COnTRA_SCPECBS3_wintest",
4
+ "dataset_path": "/mount/COnTRA/data/datasets/processed_data_SCPECBS3/homologs/SCPECBS3_SCPECBS3_ExprRefined150",
5
+ "tokenizer_path": "/mount/COnTRA/tokenizers/contra_tokenizer_gen_exprrefined",
6
+ "cai_refference_path": "/mount/COnTRA/data/datasets/processed_data_SCPECBS3/S_cerevisiae/S_cerevisiae.0.nt.fasta",
7
+ "checkpoint_flag": true,
8
+ "checkpoint_path": "/mount/COnTRA/models/Finetuned2Steps_homologs_150_ExR/best_model/",
9
+ "special_token_th": 42,
10
+ "mask_all": true,
11
+ "sw_aa_size": 150,
12
+ "eval_type": "model",
13
+ "model_type": "COBaBExRi",
14
+ "cai_query_species": "S_cerevisiae",
15
+ "outdir": "/mount/COnTRA/models/Finetuned2Steps_homologs_150_ExR/best_model/mask/",
16
+ "outfile": "mask_model_eval.csv",
17
+ "out_dict": [
18
+ "prot_len",
19
+ "num_of_correct_predicted_codons",
20
+ "accuracy",
21
+ "cross_entropy_loss",
22
+ "entropy",
23
+ "query_codons",
24
+ "subject_codons",
25
+ "pred_codons"
26
+ ],
27
+ "orig_dict": [
28
+ "qseqid",
29
+ "sseqid",
30
+ "query_species",
31
+ "subject_species",
32
+ "pident",
33
+ "length",
34
+ "mismatch",
35
+ "gapopen",
36
+ "qstart",
37
+ "qend",
38
+ "sstart",
39
+ "send",
40
+ "evalue",
41
+ "bitscore"
42
+ ],
43
+ "debug": false
44
+ }
mask/mask_model_eval.csv ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aafb9fbb7065fe41cbe140ca00765beb26d62ee4edbe2a670adee13b4128d3bb
3
+ size 99270826
model_eval_mimic.csv ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7106fa168085d166531cba91c7d23b594007eed39a6d2767f173ce50ad6acc3c
3
+ size 79597253
pytorch_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aa03881611caa9243398982c68c05b6c25e1ade86aead717a2eb1120c32e52ef
3
+ size 26586259
special_tokens_map.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "sep_token": "<sep>", "pad_token": "<pad>", "cls_token": "<cls>", "mask_token": "<msk>"}
tokenizer.json ADDED
@@ -0,0 +1,1053 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "version": "1.0",
3
+ "truncation": null,
4
+ "padding": {
5
+ "strategy": "BatchLongest",
6
+ "direction": "Right",
7
+ "pad_to_multiple_of": null,
8
+ "pad_id": 0,
9
+ "pad_type_id": 0,
10
+ "pad_token": "<pad>"
11
+ },
12
+ "added_tokens": [
13
+ {
14
+ "id": 0,
15
+ "content": "<pad>",
16
+ "single_word": false,
17
+ "lstrip": false,
18
+ "rstrip": false,
19
+ "normalized": false,
20
+ "special": true
21
+ },
22
+ {
23
+ "id": 1,
24
+ "content": "<gap>",
25
+ "single_word": false,
26
+ "lstrip": false,
27
+ "rstrip": false,
28
+ "normalized": false,
29
+ "special": true
30
+ },
31
+ {
32
+ "id": 2,
33
+ "content": "<s>",
34
+ "single_word": false,
35
+ "lstrip": false,
36
+ "rstrip": false,
37
+ "normalized": false,
38
+ "special": true
39
+ },
40
+ {
41
+ "id": 3,
42
+ "content": "</s>",
43
+ "single_word": false,
44
+ "lstrip": false,
45
+ "rstrip": false,
46
+ "normalized": false,
47
+ "special": true
48
+ },
49
+ {
50
+ "id": 4,
51
+ "content": "<unk>",
52
+ "single_word": false,
53
+ "lstrip": false,
54
+ "rstrip": false,
55
+ "normalized": false,
56
+ "special": true
57
+ },
58
+ {
59
+ "id": 5,
60
+ "content": "<cls>",
61
+ "single_word": false,
62
+ "lstrip": false,
63
+ "rstrip": false,
64
+ "normalized": false,
65
+ "special": true
66
+ },
67
+ {
68
+ "id": 6,
69
+ "content": "<sep>",
70
+ "single_word": false,
71
+ "lstrip": false,
72
+ "rstrip": false,
73
+ "normalized": false,
74
+ "special": true
75
+ },
76
+ {
77
+ "id": 7,
78
+ "content": "<expr_top10>",
79
+ "single_word": false,
80
+ "lstrip": false,
81
+ "rstrip": false,
82
+ "normalized": false,
83
+ "special": true
84
+ },
85
+ {
86
+ "id": 8,
87
+ "content": "<expr_pre75_90>",
88
+ "single_word": false,
89
+ "lstrip": false,
90
+ "rstrip": false,
91
+ "normalized": false,
92
+ "special": true
93
+ },
94
+ {
95
+ "id": 9,
96
+ "content": "<expr_pre50_75>",
97
+ "single_word": false,
98
+ "lstrip": false,
99
+ "rstrip": false,
100
+ "normalized": false,
101
+ "special": true
102
+ },
103
+ {
104
+ "id": 10,
105
+ "content": "<expr_pre25_50>",
106
+ "single_word": false,
107
+ "lstrip": false,
108
+ "rstrip": false,
109
+ "normalized": false,
110
+ "special": true
111
+ },
112
+ {
113
+ "id": 11,
114
+ "content": "<expr_low25>",
115
+ "single_word": false,
116
+ "lstrip": false,
117
+ "rstrip": false,
118
+ "normalized": false,
119
+ "special": true
120
+ },
121
+ {
122
+ "id": 12,
123
+ "content": "<expr_unk>",
124
+ "single_word": false,
125
+ "lstrip": false,
126
+ "rstrip": false,
127
+ "normalized": false,
128
+ "special": true
129
+ },
130
+ {
131
+ "id": 13,
132
+ "content": "<seqpos_begin>",
133
+ "single_word": false,
134
+ "lstrip": false,
135
+ "rstrip": false,
136
+ "normalized": false,
137
+ "special": true
138
+ },
139
+ {
140
+ "id": 14,
141
+ "content": "<seqpos_end>",
142
+ "single_word": false,
143
+ "lstrip": false,
144
+ "rstrip": false,
145
+ "normalized": false,
146
+ "special": true
147
+ },
148
+ {
149
+ "id": 15,
150
+ "content": "<seqpos_mid>",
151
+ "single_word": false,
152
+ "lstrip": false,
153
+ "rstrip": false,
154
+ "normalized": false,
155
+ "special": true
156
+ },
157
+ {
158
+ "id": 16,
159
+ "content": "<sim_high>",
160
+ "single_word": false,
161
+ "lstrip": false,
162
+ "rstrip": false,
163
+ "normalized": false,
164
+ "special": true
165
+ },
166
+ {
167
+ "id": 17,
168
+ "content": "<sim_low>",
169
+ "single_word": false,
170
+ "lstrip": false,
171
+ "rstrip": false,
172
+ "normalized": false,
173
+ "special": true
174
+ },
175
+ {
176
+ "id": 18,
177
+ "content": "<S_cerevisiae>",
178
+ "single_word": false,
179
+ "lstrip": false,
180
+ "rstrip": false,
181
+ "normalized": false,
182
+ "special": true
183
+ },
184
+ {
185
+ "id": 19,
186
+ "content": "<S_pombe>",
187
+ "single_word": false,
188
+ "lstrip": false,
189
+ "rstrip": false,
190
+ "normalized": false,
191
+ "special": true
192
+ },
193
+ {
194
+ "id": 20,
195
+ "content": "<E_coli>",
196
+ "single_word": false,
197
+ "lstrip": false,
198
+ "rstrip": false,
199
+ "normalized": false,
200
+ "special": true
201
+ },
202
+ {
203
+ "id": 21,
204
+ "content": "<B_subtilis>",
205
+ "single_word": false,
206
+ "lstrip": false,
207
+ "rstrip": false,
208
+ "normalized": false,
209
+ "special": true
210
+ },
211
+ {
212
+ "id": 22,
213
+ "content": "<mask_*>",
214
+ "single_word": false,
215
+ "lstrip": false,
216
+ "rstrip": false,
217
+ "normalized": false,
218
+ "special": true
219
+ },
220
+ {
221
+ "id": 23,
222
+ "content": "<mask_S>",
223
+ "single_word": false,
224
+ "lstrip": false,
225
+ "rstrip": false,
226
+ "normalized": false,
227
+ "special": true
228
+ },
229
+ {
230
+ "id": 24,
231
+ "content": "<mask_W>",
232
+ "single_word": false,
233
+ "lstrip": false,
234
+ "rstrip": false,
235
+ "normalized": false,
236
+ "special": true
237
+ },
238
+ {
239
+ "id": 25,
240
+ "content": "<mask_D>",
241
+ "single_word": false,
242
+ "lstrip": false,
243
+ "rstrip": false,
244
+ "normalized": false,
245
+ "special": true
246
+ },
247
+ {
248
+ "id": 26,
249
+ "content": "<mask_N>",
250
+ "single_word": false,
251
+ "lstrip": false,
252
+ "rstrip": false,
253
+ "normalized": false,
254
+ "special": true
255
+ },
256
+ {
257
+ "id": 27,
258
+ "content": "<mask_V>",
259
+ "single_word": false,
260
+ "lstrip": false,
261
+ "rstrip": false,
262
+ "normalized": false,
263
+ "special": true
264
+ },
265
+ {
266
+ "id": 28,
267
+ "content": "<mask_H>",
268
+ "single_word": false,
269
+ "lstrip": false,
270
+ "rstrip": false,
271
+ "normalized": false,
272
+ "special": true
273
+ },
274
+ {
275
+ "id": 29,
276
+ "content": "<mask_A>",
277
+ "single_word": false,
278
+ "lstrip": false,
279
+ "rstrip": false,
280
+ "normalized": false,
281
+ "special": true
282
+ },
283
+ {
284
+ "id": 30,
285
+ "content": "<mask_K>",
286
+ "single_word": false,
287
+ "lstrip": false,
288
+ "rstrip": false,
289
+ "normalized": false,
290
+ "special": true
291
+ },
292
+ {
293
+ "id": 31,
294
+ "content": "<mask_G>",
295
+ "single_word": false,
296
+ "lstrip": false,
297
+ "rstrip": false,
298
+ "normalized": false,
299
+ "special": true
300
+ },
301
+ {
302
+ "id": 32,
303
+ "content": "<mask_Q>",
304
+ "single_word": false,
305
+ "lstrip": false,
306
+ "rstrip": false,
307
+ "normalized": false,
308
+ "special": true
309
+ },
310
+ {
311
+ "id": 33,
312
+ "content": "<mask_F>",
313
+ "single_word": false,
314
+ "lstrip": false,
315
+ "rstrip": false,
316
+ "normalized": false,
317
+ "special": true
318
+ },
319
+ {
320
+ "id": 34,
321
+ "content": "<mask_I>",
322
+ "single_word": false,
323
+ "lstrip": false,
324
+ "rstrip": false,
325
+ "normalized": false,
326
+ "special": true
327
+ },
328
+ {
329
+ "id": 35,
330
+ "content": "<mask_T>",
331
+ "single_word": false,
332
+ "lstrip": false,
333
+ "rstrip": false,
334
+ "normalized": false,
335
+ "special": true
336
+ },
337
+ {
338
+ "id": 36,
339
+ "content": "<mask_C>",
340
+ "single_word": false,
341
+ "lstrip": false,
342
+ "rstrip": false,
343
+ "normalized": false,
344
+ "special": true
345
+ },
346
+ {
347
+ "id": 37,
348
+ "content": "<mask_L>",
349
+ "single_word": false,
350
+ "lstrip": false,
351
+ "rstrip": false,
352
+ "normalized": false,
353
+ "special": true
354
+ },
355
+ {
356
+ "id": 38,
357
+ "content": "<mask_P>",
358
+ "single_word": false,
359
+ "lstrip": false,
360
+ "rstrip": false,
361
+ "normalized": false,
362
+ "special": true
363
+ },
364
+ {
365
+ "id": 39,
366
+ "content": "<mask_E>",
367
+ "single_word": false,
368
+ "lstrip": false,
369
+ "rstrip": false,
370
+ "normalized": false,
371
+ "special": true
372
+ },
373
+ {
374
+ "id": 40,
375
+ "content": "<mask_M>",
376
+ "single_word": false,
377
+ "lstrip": false,
378
+ "rstrip": false,
379
+ "normalized": false,
380
+ "special": true
381
+ },
382
+ {
383
+ "id": 41,
384
+ "content": "<mask_Y>",
385
+ "single_word": false,
386
+ "lstrip": false,
387
+ "rstrip": false,
388
+ "normalized": false,
389
+ "special": true
390
+ },
391
+ {
392
+ "id": 42,
393
+ "content": "<mask_R>",
394
+ "single_word": false,
395
+ "lstrip": false,
396
+ "rstrip": false,
397
+ "normalized": false,
398
+ "special": true
399
+ },
400
+ {
401
+ "id": 43,
402
+ "content": "TTT",
403
+ "single_word": false,
404
+ "lstrip": false,
405
+ "rstrip": false,
406
+ "normalized": true,
407
+ "special": false
408
+ },
409
+ {
410
+ "id": 44,
411
+ "content": "TTC",
412
+ "single_word": false,
413
+ "lstrip": false,
414
+ "rstrip": false,
415
+ "normalized": true,
416
+ "special": false
417
+ },
418
+ {
419
+ "id": 45,
420
+ "content": "TTA",
421
+ "single_word": false,
422
+ "lstrip": false,
423
+ "rstrip": false,
424
+ "normalized": true,
425
+ "special": false
426
+ },
427
+ {
428
+ "id": 46,
429
+ "content": "TTG",
430
+ "single_word": false,
431
+ "lstrip": false,
432
+ "rstrip": false,
433
+ "normalized": true,
434
+ "special": false
435
+ },
436
+ {
437
+ "id": 47,
438
+ "content": "TCT",
439
+ "single_word": false,
440
+ "lstrip": false,
441
+ "rstrip": false,
442
+ "normalized": true,
443
+ "special": false
444
+ },
445
+ {
446
+ "id": 48,
447
+ "content": "TCC",
448
+ "single_word": false,
449
+ "lstrip": false,
450
+ "rstrip": false,
451
+ "normalized": true,
452
+ "special": false
453
+ },
454
+ {
455
+ "id": 49,
456
+ "content": "TCA",
457
+ "single_word": false,
458
+ "lstrip": false,
459
+ "rstrip": false,
460
+ "normalized": true,
461
+ "special": false
462
+ },
463
+ {
464
+ "id": 50,
465
+ "content": "TCG",
466
+ "single_word": false,
467
+ "lstrip": false,
468
+ "rstrip": false,
469
+ "normalized": true,
470
+ "special": false
471
+ },
472
+ {
473
+ "id": 51,
474
+ "content": "TAT",
475
+ "single_word": false,
476
+ "lstrip": false,
477
+ "rstrip": false,
478
+ "normalized": true,
479
+ "special": false
480
+ },
481
+ {
482
+ "id": 52,
483
+ "content": "TAC",
484
+ "single_word": false,
485
+ "lstrip": false,
486
+ "rstrip": false,
487
+ "normalized": true,
488
+ "special": false
489
+ },
490
+ {
491
+ "id": 53,
492
+ "content": "TAA",
493
+ "single_word": false,
494
+ "lstrip": false,
495
+ "rstrip": false,
496
+ "normalized": true,
497
+ "special": false
498
+ },
499
+ {
500
+ "id": 54,
501
+ "content": "TAG",
502
+ "single_word": false,
503
+ "lstrip": false,
504
+ "rstrip": false,
505
+ "normalized": true,
506
+ "special": false
507
+ },
508
+ {
509
+ "id": 55,
510
+ "content": "TGT",
511
+ "single_word": false,
512
+ "lstrip": false,
513
+ "rstrip": false,
514
+ "normalized": true,
515
+ "special": false
516
+ },
517
+ {
518
+ "id": 56,
519
+ "content": "TGC",
520
+ "single_word": false,
521
+ "lstrip": false,
522
+ "rstrip": false,
523
+ "normalized": true,
524
+ "special": false
525
+ },
526
+ {
527
+ "id": 57,
528
+ "content": "TGA",
529
+ "single_word": false,
530
+ "lstrip": false,
531
+ "rstrip": false,
532
+ "normalized": true,
533
+ "special": false
534
+ },
535
+ {
536
+ "id": 58,
537
+ "content": "TGG",
538
+ "single_word": false,
539
+ "lstrip": false,
540
+ "rstrip": false,
541
+ "normalized": true,
542
+ "special": false
543
+ },
544
+ {
545
+ "id": 59,
546
+ "content": "CTT",
547
+ "single_word": false,
548
+ "lstrip": false,
549
+ "rstrip": false,
550
+ "normalized": true,
551
+ "special": false
552
+ },
553
+ {
554
+ "id": 60,
555
+ "content": "CTC",
556
+ "single_word": false,
557
+ "lstrip": false,
558
+ "rstrip": false,
559
+ "normalized": true,
560
+ "special": false
561
+ },
562
+ {
563
+ "id": 61,
564
+ "content": "CTA",
565
+ "single_word": false,
566
+ "lstrip": false,
567
+ "rstrip": false,
568
+ "normalized": true,
569
+ "special": false
570
+ },
571
+ {
572
+ "id": 62,
573
+ "content": "CTG",
574
+ "single_word": false,
575
+ "lstrip": false,
576
+ "rstrip": false,
577
+ "normalized": true,
578
+ "special": false
579
+ },
580
+ {
581
+ "id": 63,
582
+ "content": "CCT",
583
+ "single_word": false,
584
+ "lstrip": false,
585
+ "rstrip": false,
586
+ "normalized": true,
587
+ "special": false
588
+ },
589
+ {
590
+ "id": 64,
591
+ "content": "CCC",
592
+ "single_word": false,
593
+ "lstrip": false,
594
+ "rstrip": false,
595
+ "normalized": true,
596
+ "special": false
597
+ },
598
+ {
599
+ "id": 65,
600
+ "content": "CCA",
601
+ "single_word": false,
602
+ "lstrip": false,
603
+ "rstrip": false,
604
+ "normalized": true,
605
+ "special": false
606
+ },
607
+ {
608
+ "id": 66,
609
+ "content": "CCG",
610
+ "single_word": false,
611
+ "lstrip": false,
612
+ "rstrip": false,
613
+ "normalized": true,
614
+ "special": false
615
+ },
616
+ {
617
+ "id": 67,
618
+ "content": "CAT",
619
+ "single_word": false,
620
+ "lstrip": false,
621
+ "rstrip": false,
622
+ "normalized": true,
623
+ "special": false
624
+ },
625
+ {
626
+ "id": 68,
627
+ "content": "CAC",
628
+ "single_word": false,
629
+ "lstrip": false,
630
+ "rstrip": false,
631
+ "normalized": true,
632
+ "special": false
633
+ },
634
+ {
635
+ "id": 69,
636
+ "content": "CAA",
637
+ "single_word": false,
638
+ "lstrip": false,
639
+ "rstrip": false,
640
+ "normalized": true,
641
+ "special": false
642
+ },
643
+ {
644
+ "id": 70,
645
+ "content": "CAG",
646
+ "single_word": false,
647
+ "lstrip": false,
648
+ "rstrip": false,
649
+ "normalized": true,
650
+ "special": false
651
+ },
652
+ {
653
+ "id": 71,
654
+ "content": "CGT",
655
+ "single_word": false,
656
+ "lstrip": false,
657
+ "rstrip": false,
658
+ "normalized": true,
659
+ "special": false
660
+ },
661
+ {
662
+ "id": 72,
663
+ "content": "CGC",
664
+ "single_word": false,
665
+ "lstrip": false,
666
+ "rstrip": false,
667
+ "normalized": true,
668
+ "special": false
669
+ },
670
+ {
671
+ "id": 73,
672
+ "content": "CGA",
673
+ "single_word": false,
674
+ "lstrip": false,
675
+ "rstrip": false,
676
+ "normalized": true,
677
+ "special": false
678
+ },
679
+ {
680
+ "id": 74,
681
+ "content": "CGG",
682
+ "single_word": false,
683
+ "lstrip": false,
684
+ "rstrip": false,
685
+ "normalized": true,
686
+ "special": false
687
+ },
688
+ {
689
+ "id": 75,
690
+ "content": "ATT",
691
+ "single_word": false,
692
+ "lstrip": false,
693
+ "rstrip": false,
694
+ "normalized": true,
695
+ "special": false
696
+ },
697
+ {
698
+ "id": 76,
699
+ "content": "ATC",
700
+ "single_word": false,
701
+ "lstrip": false,
702
+ "rstrip": false,
703
+ "normalized": true,
704
+ "special": false
705
+ },
706
+ {
707
+ "id": 77,
708
+ "content": "ATA",
709
+ "single_word": false,
710
+ "lstrip": false,
711
+ "rstrip": false,
712
+ "normalized": true,
713
+ "special": false
714
+ },
715
+ {
716
+ "id": 78,
717
+ "content": "ATG",
718
+ "single_word": false,
719
+ "lstrip": false,
720
+ "rstrip": false,
721
+ "normalized": true,
722
+ "special": false
723
+ },
724
+ {
725
+ "id": 79,
726
+ "content": "ACT",
727
+ "single_word": false,
728
+ "lstrip": false,
729
+ "rstrip": false,
730
+ "normalized": true,
731
+ "special": false
732
+ },
733
+ {
734
+ "id": 80,
735
+ "content": "ACC",
736
+ "single_word": false,
737
+ "lstrip": false,
738
+ "rstrip": false,
739
+ "normalized": true,
740
+ "special": false
741
+ },
742
+ {
743
+ "id": 81,
744
+ "content": "ACA",
745
+ "single_word": false,
746
+ "lstrip": false,
747
+ "rstrip": false,
748
+ "normalized": true,
749
+ "special": false
750
+ },
751
+ {
752
+ "id": 82,
753
+ "content": "ACG",
754
+ "single_word": false,
755
+ "lstrip": false,
756
+ "rstrip": false,
757
+ "normalized": true,
758
+ "special": false
759
+ },
760
+ {
761
+ "id": 83,
762
+ "content": "AAT",
763
+ "single_word": false,
764
+ "lstrip": false,
765
+ "rstrip": false,
766
+ "normalized": true,
767
+ "special": false
768
+ },
769
+ {
770
+ "id": 84,
771
+ "content": "AAC",
772
+ "single_word": false,
773
+ "lstrip": false,
774
+ "rstrip": false,
775
+ "normalized": true,
776
+ "special": false
777
+ },
778
+ {
779
+ "id": 85,
780
+ "content": "AAA",
781
+ "single_word": false,
782
+ "lstrip": false,
783
+ "rstrip": false,
784
+ "normalized": true,
785
+ "special": false
786
+ },
787
+ {
788
+ "id": 86,
789
+ "content": "AAG",
790
+ "single_word": false,
791
+ "lstrip": false,
792
+ "rstrip": false,
793
+ "normalized": true,
794
+ "special": false
795
+ },
796
+ {
797
+ "id": 87,
798
+ "content": "AGT",
799
+ "single_word": false,
800
+ "lstrip": false,
801
+ "rstrip": false,
802
+ "normalized": true,
803
+ "special": false
804
+ },
805
+ {
806
+ "id": 88,
807
+ "content": "AGC",
808
+ "single_word": false,
809
+ "lstrip": false,
810
+ "rstrip": false,
811
+ "normalized": true,
812
+ "special": false
813
+ },
814
+ {
815
+ "id": 89,
816
+ "content": "AGA",
817
+ "single_word": false,
818
+ "lstrip": false,
819
+ "rstrip": false,
820
+ "normalized": true,
821
+ "special": false
822
+ },
823
+ {
824
+ "id": 90,
825
+ "content": "AGG",
826
+ "single_word": false,
827
+ "lstrip": false,
828
+ "rstrip": false,
829
+ "normalized": true,
830
+ "special": false
831
+ },
832
+ {
833
+ "id": 91,
834
+ "content": "GTT",
835
+ "single_word": false,
836
+ "lstrip": false,
837
+ "rstrip": false,
838
+ "normalized": true,
839
+ "special": false
840
+ },
841
+ {
842
+ "id": 92,
843
+ "content": "GTC",
844
+ "single_word": false,
845
+ "lstrip": false,
846
+ "rstrip": false,
847
+ "normalized": true,
848
+ "special": false
849
+ },
850
+ {
851
+ "id": 93,
852
+ "content": "GTA",
853
+ "single_word": false,
854
+ "lstrip": false,
855
+ "rstrip": false,
856
+ "normalized": true,
857
+ "special": false
858
+ },
859
+ {
860
+ "id": 94,
861
+ "content": "GTG",
862
+ "single_word": false,
863
+ "lstrip": false,
864
+ "rstrip": false,
865
+ "normalized": true,
866
+ "special": false
867
+ },
868
+ {
869
+ "id": 95,
870
+ "content": "GCT",
871
+ "single_word": false,
872
+ "lstrip": false,
873
+ "rstrip": false,
874
+ "normalized": true,
875
+ "special": false
876
+ },
877
+ {
878
+ "id": 96,
879
+ "content": "GCC",
880
+ "single_word": false,
881
+ "lstrip": false,
882
+ "rstrip": false,
883
+ "normalized": true,
884
+ "special": false
885
+ },
886
+ {
887
+ "id": 97,
888
+ "content": "GCA",
889
+ "single_word": false,
890
+ "lstrip": false,
891
+ "rstrip": false,
892
+ "normalized": true,
893
+ "special": false
894
+ },
895
+ {
896
+ "id": 98,
897
+ "content": "GCG",
898
+ "single_word": false,
899
+ "lstrip": false,
900
+ "rstrip": false,
901
+ "normalized": true,
902
+ "special": false
903
+ },
904
+ {
905
+ "id": 99,
906
+ "content": "GAT",
907
+ "single_word": false,
908
+ "lstrip": false,
909
+ "rstrip": false,
910
+ "normalized": true,
911
+ "special": false
912
+ },
913
+ {
914
+ "id": 100,
915
+ "content": "GAC",
916
+ "single_word": false,
917
+ "lstrip": false,
918
+ "rstrip": false,
919
+ "normalized": true,
920
+ "special": false
921
+ },
922
+ {
923
+ "id": 101,
924
+ "content": "GAA",
925
+ "single_word": false,
926
+ "lstrip": false,
927
+ "rstrip": false,
928
+ "normalized": true,
929
+ "special": false
930
+ },
931
+ {
932
+ "id": 102,
933
+ "content": "GAG",
934
+ "single_word": false,
935
+ "lstrip": false,
936
+ "rstrip": false,
937
+ "normalized": true,
938
+ "special": false
939
+ },
940
+ {
941
+ "id": 103,
942
+ "content": "GGT",
943
+ "single_word": false,
944
+ "lstrip": false,
945
+ "rstrip": false,
946
+ "normalized": true,
947
+ "special": false
948
+ },
949
+ {
950
+ "id": 104,
951
+ "content": "GGC",
952
+ "single_word": false,
953
+ "lstrip": false,
954
+ "rstrip": false,
955
+ "normalized": true,
956
+ "special": false
957
+ },
958
+ {
959
+ "id": 105,
960
+ "content": "GGA",
961
+ "single_word": false,
962
+ "lstrip": false,
963
+ "rstrip": false,
964
+ "normalized": true,
965
+ "special": false
966
+ },
967
+ {
968
+ "id": 106,
969
+ "content": "GGG",
970
+ "single_word": false,
971
+ "lstrip": false,
972
+ "rstrip": false,
973
+ "normalized": true,
974
+ "special": false
975
+ },
976
+ {
977
+ "id": 107,
978
+ "content": "<msk>",
979
+ "single_word": false,
980
+ "lstrip": false,
981
+ "rstrip": false,
982
+ "normalized": false,
983
+ "special": true
984
+ }
985
+ ],
986
+ "normalizer": null,
987
+ "pre_tokenizer": {
988
+ "type": "Whitespace"
989
+ },
990
+ "post_processor": {
991
+ "type": "TemplateProcessing",
992
+ "single": [
993
+ {
994
+ "SpecialToken": {
995
+ "id": "<cls>",
996
+ "type_id": 0
997
+ }
998
+ },
999
+ {
1000
+ "Sequence": {
1001
+ "id": "A",
1002
+ "type_id": 0
1003
+ }
1004
+ },
1005
+ {
1006
+ "SpecialToken": {
1007
+ "id": "<sep>",
1008
+ "type_id": 0
1009
+ }
1010
+ }
1011
+ ],
1012
+ "pair": [
1013
+ {
1014
+ "Sequence": {
1015
+ "id": "A",
1016
+ "type_id": 0
1017
+ }
1018
+ },
1019
+ {
1020
+ "Sequence": {
1021
+ "id": "B",
1022
+ "type_id": 0
1023
+ }
1024
+ }
1025
+ ],
1026
+ "special_tokens": {
1027
+ "<cls>": {
1028
+ "id": "<cls>",
1029
+ "ids": [
1030
+ 5
1031
+ ],
1032
+ "tokens": [
1033
+ "<cls>"
1034
+ ]
1035
+ },
1036
+ "<sep>": {
1037
+ "id": "<sep>",
1038
+ "ids": [
1039
+ 6
1040
+ ],
1041
+ "tokens": [
1042
+ "<sep>"
1043
+ ]
1044
+ }
1045
+ }
1046
+ },
1047
+ "decoder": null,
1048
+ "model": {
1049
+ "type": "WordLevel",
1050
+ "vocab": {},
1051
+ "unk_token": "<unk>"
1052
+ }
1053
+ }
tokenizer_config.json ADDED
@@ -0,0 +1 @@
 
 
1
+ {"bos_token": "<s>", "cls_token": "<cls>", "eos_token": "</s>", "mask_token": "<msk>", "model_max_length": 512, "pad_token": "<pad>", "padding_side": "right", "sep_token": "<sep>", "unk_token": "<unk>", "special_tokens_map_file": "/mount/COnTRA/tokenizers/contra_tokenizer_gen_exprrefined/special_tokens_map.json", "name_or_path": "/mount/COnTRA/tokenizers/contra_tokenizer_gen_exprrefined", "tokenizer_class": "PreTrainedTokenizerFast"}
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:59309a23b7c1e53e3ea3a85bc3aef890382181df0a37f954f00a102ee47f91f7
3
+ size 3119