bmuscato commited on
Commit
5b9f8b2
·
verified ·
1 Parent(s): 5d36935

Upload folder using huggingface_hub

Browse files
checkpoint-1648/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-1648/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b3479ea0d9ff366b09f80feb55ad22c35a2c05212c6090d84559215850da6c6a
3
+ size 498616156
checkpoint-1648/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d37038fdc0fae7b16bbb2b9904902c2a94f20af2f7ac5349b268532a090d1e67
3
+ size 113433150
checkpoint-1648/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bd1e45f76387ed9c0e2abe9ba2abc20ec46c157b1d7885ad885413f78bc1562f
3
+ size 14244
checkpoint-1648/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:01585781256a71b3e15c278036c76b5f22548b21cba6dd13efe9cd989e1bc677
3
+ size 1064
checkpoint-1648/special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "mask_token": {
6
+ "content": "<mask>",
7
+ "lstrip": true,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "pad_token": "<pad>",
13
+ "sep_token": "</s>",
14
+ "unk_token": "<unk>"
15
+ }
checkpoint-1648/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
checkpoint-1648/tokenizer_config.json ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "added_tokens_decoder": {
4
+ "0": {
5
+ "content": "<s>",
6
+ "lstrip": false,
7
+ "normalized": true,
8
+ "rstrip": false,
9
+ "single_word": false,
10
+ "special": true
11
+ },
12
+ "1": {
13
+ "content": "<pad>",
14
+ "lstrip": false,
15
+ "normalized": true,
16
+ "rstrip": false,
17
+ "single_word": false,
18
+ "special": true
19
+ },
20
+ "2": {
21
+ "content": "</s>",
22
+ "lstrip": false,
23
+ "normalized": true,
24
+ "rstrip": false,
25
+ "single_word": false,
26
+ "special": true
27
+ },
28
+ "3": {
29
+ "content": "<unk>",
30
+ "lstrip": false,
31
+ "normalized": true,
32
+ "rstrip": false,
33
+ "single_word": false,
34
+ "special": true
35
+ },
36
+ "50264": {
37
+ "content": "<mask>",
38
+ "lstrip": true,
39
+ "normalized": false,
40
+ "rstrip": false,
41
+ "single_word": false,
42
+ "special": true
43
+ }
44
+ },
45
+ "bos_token": "<s>",
46
+ "clean_up_tokenization_spaces": false,
47
+ "cls_token": "<s>",
48
+ "eos_token": "</s>",
49
+ "errors": "replace",
50
+ "extra_special_tokens": {},
51
+ "mask_token": "<mask>",
52
+ "model_max_length": 512,
53
+ "pad_token": "<pad>",
54
+ "sep_token": "</s>",
55
+ "tokenizer_class": "RobertaTokenizer",
56
+ "trim_offsets": true,
57
+ "unk_token": "<unk>"
58
+ }
checkpoint-1648/trainer_state.json ADDED
@@ -0,0 +1,235 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": 1648,
3
+ "best_metric": 0.5107126832008362,
4
+ "best_model_checkpoint": "./multitask_model2/checkpoint-1648",
5
+ "epoch": 8.0,
6
+ "eval_steps": 500,
7
+ "global_step": 1648,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 1.0,
14
+ "grad_norm": 1.1497278213500977,
15
+ "learning_rate": 1.900485436893204e-05,
16
+ "loss": 0.5967,
17
+ "step": 206
18
+ },
19
+ {
20
+ "epoch": 1.0,
21
+ "eval_dis_accuracy": 0.6557971014492754,
22
+ "eval_dis_f1": 0.6988906497622821,
23
+ "eval_dis_precision": 0.7078651685393258,
24
+ "eval_dis_recall": 0.6901408450704225,
25
+ "eval_loss": 0.5416484475135803,
26
+ "eval_runtime": 2.1702,
27
+ "eval_samples_per_second": 508.718,
28
+ "eval_steps_per_second": 16.128,
29
+ "eval_target_accuracy": 0.7644927536231884,
30
+ "eval_target_f1": 0.7200199762781696,
31
+ "eval_target_precision": 0.7536231884057971,
32
+ "eval_target_recall": 0.7086333006968841,
33
+ "eval_target_soft_ce": 0.543969452381134,
34
+ "step": 206
35
+ },
36
+ {
37
+ "epoch": 2.0,
38
+ "grad_norm": 1.0708316564559937,
39
+ "learning_rate": 1.800485436893204e-05,
40
+ "loss": 0.5368,
41
+ "step": 412
42
+ },
43
+ {
44
+ "epoch": 2.0,
45
+ "eval_dis_accuracy": 0.6875,
46
+ "eval_dis_f1": 0.7537473233404711,
47
+ "eval_dis_precision": 0.6929133858267716,
48
+ "eval_dis_recall": 0.8262910798122066,
49
+ "eval_loss": 0.5247225761413574,
50
+ "eval_runtime": 1.5657,
51
+ "eval_samples_per_second": 705.133,
52
+ "eval_steps_per_second": 22.355,
53
+ "eval_target_accuracy": 0.792572463768116,
54
+ "eval_target_f1": 0.7548633944423059,
55
+ "eval_target_precision": 0.7887541264423719,
56
+ "eval_target_recall": 0.7414977250475148,
57
+ "eval_target_soft_ce": 0.5287814140319824,
58
+ "step": 412
59
+ },
60
+ {
61
+ "epoch": 3.0,
62
+ "grad_norm": 2.7234156131744385,
63
+ "learning_rate": 1.700485436893204e-05,
64
+ "loss": 0.5226,
65
+ "step": 618
66
+ },
67
+ {
68
+ "epoch": 3.0,
69
+ "eval_dis_accuracy": 0.6856884057971014,
70
+ "eval_dis_f1": 0.7457875457875458,
71
+ "eval_dis_precision": 0.7011019283746557,
72
+ "eval_dis_recall": 0.7965571205007824,
73
+ "eval_loss": 0.5280918478965759,
74
+ "eval_runtime": 1.5817,
75
+ "eval_samples_per_second": 697.973,
76
+ "eval_steps_per_second": 22.128,
77
+ "eval_target_accuracy": 0.7889492753623188,
78
+ "eval_target_f1": 0.7471213067050722,
79
+ "eval_target_precision": 0.7893963191172325,
80
+ "eval_target_recall": 0.7328010712434487,
81
+ "eval_target_soft_ce": 0.5373103022575378,
82
+ "step": 618
83
+ },
84
+ {
85
+ "epoch": 4.0,
86
+ "grad_norm": 1.301528811454773,
87
+ "learning_rate": 1.600485436893204e-05,
88
+ "loss": 0.5116,
89
+ "step": 824
90
+ },
91
+ {
92
+ "epoch": 4.0,
93
+ "eval_dis_accuracy": 0.697463768115942,
94
+ "eval_dis_f1": 0.7569141193595342,
95
+ "eval_dis_precision": 0.7074829931972789,
96
+ "eval_dis_recall": 0.8137715179968701,
97
+ "eval_loss": 0.5131542086601257,
98
+ "eval_runtime": 2.1729,
99
+ "eval_samples_per_second": 508.077,
100
+ "eval_steps_per_second": 16.108,
101
+ "eval_target_accuracy": 0.7789855072463768,
102
+ "eval_target_f1": 0.7520288681051623,
103
+ "eval_target_precision": 0.7591876208897486,
104
+ "eval_target_recall": 0.746961930541957,
105
+ "eval_target_soft_ce": 0.5194346308708191,
106
+ "step": 824
107
+ },
108
+ {
109
+ "epoch": 5.0,
110
+ "grad_norm": 1.33175790309906,
111
+ "learning_rate": 1.5004854368932041e-05,
112
+ "loss": 0.5048,
113
+ "step": 1030
114
+ },
115
+ {
116
+ "epoch": 5.0,
117
+ "eval_dis_accuracy": 0.7128623188405797,
118
+ "eval_dis_f1": 0.7759717314487633,
119
+ "eval_dis_precision": 0.7074742268041238,
120
+ "eval_dis_recall": 0.8591549295774648,
121
+ "eval_loss": 0.5118098855018616,
122
+ "eval_runtime": 2.184,
123
+ "eval_samples_per_second": 505.485,
124
+ "eval_steps_per_second": 16.025,
125
+ "eval_target_accuracy": 0.7898550724637681,
126
+ "eval_target_f1": 0.761273268884473,
127
+ "eval_target_precision": 0.7739093242087254,
128
+ "eval_target_recall": 0.7535708114957093,
129
+ "eval_target_soft_ce": 0.5185291171073914,
130
+ "step": 1030
131
+ },
132
+ {
133
+ "epoch": 6.0,
134
+ "grad_norm": 1.5620982646942139,
135
+ "learning_rate": 1.4004854368932039e-05,
136
+ "loss": 0.4983,
137
+ "step": 1236
138
+ },
139
+ {
140
+ "epoch": 6.0,
141
+ "eval_dis_accuracy": 0.7001811594202898,
142
+ "eval_dis_f1": 0.7578639356254572,
143
+ "eval_dis_precision": 0.7115384615384616,
144
+ "eval_dis_recall": 0.810641627543036,
145
+ "eval_loss": 0.5142112374305725,
146
+ "eval_runtime": 2.1937,
147
+ "eval_samples_per_second": 503.249,
148
+ "eval_steps_per_second": 15.954,
149
+ "eval_target_accuracy": 0.7961956521739131,
150
+ "eval_target_f1": 0.755370956702432,
151
+ "eval_target_precision": 0.8000787100546474,
152
+ "eval_target_recall": 0.7401586707366239,
153
+ "eval_target_soft_ce": 0.5227047801017761,
154
+ "step": 1236
155
+ },
156
+ {
157
+ "epoch": 7.0,
158
+ "grad_norm": 1.2635632753372192,
159
+ "learning_rate": 1.300485436893204e-05,
160
+ "loss": 0.4893,
161
+ "step": 1442
162
+ },
163
+ {
164
+ "epoch": 7.0,
165
+ "eval_dis_accuracy": 0.6947463768115942,
166
+ "eval_dis_f1": 0.7437262357414449,
167
+ "eval_dis_precision": 0.7233727810650887,
168
+ "eval_dis_recall": 0.7652582159624414,
169
+ "eval_loss": 0.5122794508934021,
170
+ "eval_runtime": 2.4324,
171
+ "eval_samples_per_second": 453.876,
172
+ "eval_steps_per_second": 14.389,
173
+ "eval_target_accuracy": 0.7980072463768116,
174
+ "eval_target_f1": 0.7658736989258561,
175
+ "eval_target_precision": 0.7890072298807826,
176
+ "eval_target_recall": 0.754542705753614,
177
+ "eval_target_soft_ce": 0.5191038846969604,
178
+ "step": 1442
179
+ },
180
+ {
181
+ "epoch": 8.0,
182
+ "grad_norm": 1.2441052198410034,
183
+ "learning_rate": 1.200485436893204e-05,
184
+ "loss": 0.4884,
185
+ "step": 1648
186
+ },
187
+ {
188
+ "epoch": 8.0,
189
+ "eval_dis_accuracy": 0.7146739130434783,
190
+ "eval_dis_f1": 0.7664936990363233,
191
+ "eval_dis_precision": 0.7281690140845071,
192
+ "eval_dis_recall": 0.809076682316119,
193
+ "eval_loss": 0.5107126832008362,
194
+ "eval_runtime": 2.1956,
195
+ "eval_samples_per_second": 502.813,
196
+ "eval_steps_per_second": 15.941,
197
+ "eval_target_accuracy": 0.7898550724637681,
198
+ "eval_target_f1": 0.7573626373626374,
199
+ "eval_target_precision": 0.7777777777777778,
200
+ "eval_target_recall": 0.7470771180095606,
201
+ "eval_target_soft_ce": 0.5190978050231934,
202
+ "step": 1648
203
+ }
204
+ ],
205
+ "logging_steps": 500,
206
+ "max_steps": 4120,
207
+ "num_input_tokens_seen": 0,
208
+ "num_train_epochs": 20,
209
+ "save_steps": 500,
210
+ "stateful_callbacks": {
211
+ "EarlyStoppingCallback": {
212
+ "args": {
213
+ "early_stopping_patience": 5,
214
+ "early_stopping_threshold": 0.0
215
+ },
216
+ "attributes": {
217
+ "early_stopping_patience_counter": 0
218
+ }
219
+ },
220
+ "TrainerControl": {
221
+ "args": {
222
+ "should_epoch_stop": false,
223
+ "should_evaluate": false,
224
+ "should_log": false,
225
+ "should_save": true,
226
+ "should_training_stop": false
227
+ },
228
+ "attributes": {}
229
+ }
230
+ },
231
+ "total_flos": 0.0,
232
+ "train_batch_size": 32,
233
+ "trial_name": null,
234
+ "trial_params": null
235
+ }
checkpoint-1648/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f32b0cf3c27c444da01e7312faa99aadf87ec2a1877a2c08fc94b51f49f45092
3
+ size 5368
checkpoint-1648/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
disagreement_head.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:163a1c0d42a61bbbf4e1fbec5712f2e1e550726cfe8b855f52712644e88211f7
3
+ size 4756
emissions.csv ADDED
@@ -0,0 +1,21 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ timestamp,project_name,run_id,experiment_id,duration,emissions,emissions_rate,cpu_power,gpu_power,ram_power,cpu_energy,gpu_energy,ram_energy,energy_consumed,water_consumed,country_name,country_iso_code,region,cloud_provider,cloud_region,os,python_version,codecarbon_version,cpu_count,cpu_model,gpu_count,gpu_model,longitude,latitude,ram_total_size,tracking_mode,on_cloud,pue,wue
2
+ 2025-12-27T13:52:17,codecarbon,44b99aef-804f-4d4b-860e-78d6254daae5,5b0fa12a-3dd7-45bb-9766-cc326314d9f1,83.89951640740037,0.0034680981025371,4.133633006532143e-05,46.006028622,336.9653671208123,70.0,0.0010280316728148,0.0078945779823271,0.001563962295998,0.0104865719511401,0.0,Italy,ITA,tuscany,,,Linux-5.4.0-216-generic-x86_64-with-glibc2.31,3.10.18,3.0.8,96,Intel(R) Xeon(R) Gold 6342 CPU @ 2.80GHz,2,2 x NVIDIA RTX A6000,10.4326,43.7667,1007.494571685791,machine,N,1.0,0.0
3
+ 2025-12-27T13:59:53,codecarbon,99188916-f18f-4b73-9fcc-0b32a5ce8651,5b0fa12a-3dd7-45bb-9766-cc326314d9f1,83.63794799894094,0.0034706976701121,4.149668605160009e-05,46.006432318,335.68992000960066,70.0,0.0010284683618572,0.0079014546544868,0.0015645093100248,0.010494432326369,0.0,Italy,ITA,tuscany,,,Linux-5.4.0-216-generic-x86_64-with-glibc2.31,3.10.18,3.0.8,96,Intel(R) Xeon(R) Gold 6342 CPU @ 2.80GHz,2,2 x NVIDIA RTX A6000,10.4326,43.7667,1007.494571685791,machine,N,1.0,0.0
4
+ 2025-12-27T14:02:07,codecarbon,ab41bf6e-ca05-4720-afc7-80381bcbde52,5b0fa12a-3dd7-45bb-9766-cc326314d9f1,91.69486468099058,0.0049698273067074,5.419962528978762e-05,46.045785088,455.0391749686571,70.0,0.0013202338105998,0.0117002618602128,0.0020068911261183,0.0150273867969311,0.0,Italy,ITA,tuscany,,,Linux-5.4.0-216-generic-x86_64-with-glibc2.31,3.10.18,3.0.8,96,Intel(R) Xeon(R) Gold 6342 CPU @ 2.80GHz,2,2 x NVIDIA RTX A6000,10.4326,43.7667,1007.494571685791,machine,N,1.0,0.0
5
+ 2025-12-27T14:05:08,codecarbon,c89e01a7-0716-4607-8702-30453df3175e,5b0fa12a-3dd7-45bb-9766-cc326314d9f1,91.75349273905158,0.0048127107100098,5.245261587694847e-05,46.024751818,437.5443765632813,70.0,0.0011263668058828,0.0117137229820798,0.0017122197628021,0.0145523095507648,0.0,Italy,ITA,tuscany,,,Linux-5.4.0-216-generic-x86_64-with-glibc2.31,3.10.18,3.0.8,96,Intel(R) Xeon(R) Gold 6342 CPU @ 2.80GHz,2,2 x NVIDIA RTX A6000,10.4326,43.7667,1007.494571685791,machine,N,1.0,0.0
6
+ 2025-12-27T14:09:50,codecarbon,45c5843e-c3d2-4aa3-9c24-d39b91544917,5b0fa12a-3dd7-45bb-9766-cc326314d9f1,89.67684659175575,0.0047066118667172,5.248413660377287e-05,46.03500112046154,465.7865820406681,70.0,0.0011067691513161,0.0114421072092341,0.0016826195565673,0.0142314959171175,0.0,Italy,ITA,tuscany,,,Linux-5.4.0-216-generic-x86_64-with-glibc2.31,3.10.18,3.0.8,96,Intel(R) Xeon(R) Gold 6342 CPU @ 2.80GHz,2,2 x NVIDIA RTX A6000,10.4326,43.7667,1007.494571685791,machine,N,1.0,0.0
7
+ 2025-12-27T14:13:48,codecarbon,18c5560a-c017-4119-b2be-c21f49fdbd1e,5b0fa12a-3dd7-45bb-9766-cc326314d9f1,83.28071516565979,0.0035179738233526,4.224235846623975e-05,46.008629467750005,334.57483409322487,70.0,0.0010241245246591,0.0080552161663689,0.0015580416811216,0.0106373823721497,0.0,Italy,ITA,tuscany,,,Linux-5.4.0-216-generic-x86_64-with-glibc2.31,3.10.18,3.0.8,96,Intel(R) Xeon(R) Gold 6342 CPU @ 2.80GHz,2,2 x NVIDIA RTX A6000,10.4326,43.7667,1007.494571685791,machine,N,1.0,0.0
8
+ 2025-12-27T14:20:33,codecarbon,6511ba02-6693-4ea6-8824-1f42a1294606,5b0fa12a-3dd7-45bb-9766-cc326314d9f1,83.45597690157592,0.0034181521389267,4.095754750984404e-05,46.01146614400001,325.506046808152,70.0,0.0010258111518348,0.0077492464771609,0.0015604915063,0.0103355491352957,0.0,Italy,ITA,tuscany,,,Linux-5.4.0-216-generic-x86_64-with-glibc2.31,3.10.18,3.0.8,96,Intel(R) Xeon(R) Gold 6342 CPU @ 2.80GHz,2,2 x NVIDIA RTX A6000,10.4326,43.7667,1007.494571685791,machine,N,1.0,0.0
9
+ 2025-12-27T14:26:28,codecarbon,ee459da1-3a5a-43ae-a72c-e62c37f08cfb,5b0fa12a-3dd7-45bb-9766-cc326314d9f1,117.67954310774805,0.004692493413635,3.9875183823058846e-05,46.00980869500001,338.93103355875405,70.0,0.0014520560536897,0.0105277406444201,0.002209008911609,0.0141888056097189,0.0,Italy,ITA,tuscany,,,Linux-5.4.0-216-generic-x86_64-with-glibc2.31,3.10.18,3.0.8,96,Intel(R) Xeon(R) Gold 6342 CPU @ 2.80GHz,2,2 x NVIDIA RTX A6000,10.4326,43.7667,1007.494571685791,machine,N,1.0,0.0
10
+ 2025-12-27T14:30:38,codecarbon,4a08e70b-f576-4133-8950-146172b9f5ed,5b0fa12a-3dd7-45bb-9766-cc326314d9f1,102.52209948003292,0.0040673024848744,3.967244628721806e-05,46.00773841272729,334.9938548419866,70.0,0.0012611206037528,0.0091187856283454,0.0019184944744687,0.012298400706567,0.0,Italy,ITA,tuscany,,,Linux-5.4.0-216-generic-x86_64-with-glibc2.31,3.10.18,3.0.8,96,Intel(R) Xeon(R) Gold 6342 CPU @ 2.80GHz,2,2 x NVIDIA RTX A6000,10.4326,43.7667,1007.494571685791,machine,N,1.0,0.0
11
+ 2025-12-27T14:36:24,codecarbon,be8ee28c-e9a0-463f-bb42-367b8ef9c090,5b0fa12a-3dd7-45bb-9766-cc326314d9f1,98.85874194465578,0.0044613508702703,4.512854182150056e-05,46.00674199000001,373.11043835124434,70.0,0.0012180895478665,0.0104188105572688,0.0018529942165232,0.0134898943216587,0.0,Italy,ITA,tuscany,,,Linux-5.4.0-216-generic-x86_64-with-glibc2.31,3.10.18,3.0.8,96,Intel(R) Xeon(R) Gold 6342 CPU @ 2.80GHz,2,2 x NVIDIA RTX A6000,10.4326,43.7667,1007.494571685791,machine,N,1.0,0.0
12
+ 2025-12-27T14:42:21,codecarbon,46bac4f2-5244-4b6b-aa7b-33dbdecc3f55,5b0fa12a-3dd7-45bb-9766-cc326314d9f1,132.19837917760017,0.0060829987807942,4.6014170662577335e-05,46.007512444000014,393.3383935597069,70.0,0.0016313887048931,0.0142800644796068,0.0024818570701405,0.0183933102546405,0.0,Italy,ITA,tuscany,,,Linux-5.4.0-216-generic-x86_64-with-glibc2.31,3.10.18,3.0.8,96,Intel(R) Xeon(R) Gold 6342 CPU @ 2.80GHz,2,2 x NVIDIA RTX A6000,10.4326,43.7667,1007.494571685791,machine,N,1.0,0.0
13
+ 2025-12-27T14:46:16,codecarbon,fe23013c-99ef-4688-b087-7a1a19fb8f1b,5b0fa12a-3dd7-45bb-9766-cc326314d9f1,118.31770533509552,0.0055101232043169,4.657057190816318e-05,46.01693938323078,402.6200207827816,70.0,0.0014602054385564,0.0129794953835755,0.0022213917229272,0.0166610925450593,0.0,Italy,ITA,tuscany,,,Linux-5.4.0-216-generic-x86_64-with-glibc2.31,3.10.18,3.0.8,96,Intel(R) Xeon(R) Gold 6342 CPU @ 2.80GHz,2,2 x NVIDIA RTX A6000,10.4326,43.7667,1007.494571685791,machine,N,1.0,0.0
14
+ 2025-12-27T14:50:13,codecarbon,23fc06d5-3a60-42b3-a957-e16d1219f2e1,5b0fa12a-3dd7-45bb-9766-cc326314d9f1,118.57160128466784,0.0055450454157226,4.67653751458582e-05,46.01624443646155,397.10806524950726,70.0,0.0014624585131554,0.0130795162969405,0.0022247128685988,0.0167666876786948,0.0,Italy,ITA,tuscany,,,Linux-5.4.0-216-generic-x86_64-with-glibc2.31,3.10.18,3.0.8,96,Intel(R) Xeon(R) Gold 6342 CPU @ 2.80GHz,2,2 x NVIDIA RTX A6000,10.4326,43.7667,1007.494571685791,machine,N,1.0,0.0
15
+ 2025-12-27T14:56:48,codecarbon,c78442fe-0c9e-4836-a80a-e749281c61b9,5b0fa12a-3dd7-45bb-9766-cc326314d9f1,122.42325865477324,0.0060867232832244,4.971868377061151e-05,46.002658915,270.1728538000382,70.0,0.0015044357518107,0.0146114353002531,0.0022887010704829,0.0184045721225468,0.0,Italy,ITA,tuscany,,,Linux-5.4.0-216-generic-x86_64-with-glibc2.31,3.10.18,3.0.8,96,Intel(R) Xeon(R) Gold 6342 CPU @ 2.80GHz,2,2 x NVIDIA RTX A6000,10.4326,43.7667,1007.494571685791,machine,N,1.0,0.0
16
+ 2025-12-27T15:02:45,codecarbon,a4be2972-d51e-4a68-892b-811763dc7934,5b0fa12a-3dd7-45bb-9766-cc326314d9f1,106.3050300180912,0.0052708492423759,4.958231272291553e-05,46.00569627775002,430.4303165231557,70.0,0.0013132348580643,0.0126263676010438,0.0019979916312524,0.0159375940903607,0.0,Italy,ITA,tuscany,,,Linux-5.4.0-216-generic-x86_64-with-glibc2.31,3.10.18,3.0.8,96,Intel(R) Xeon(R) Gold 6342 CPU @ 2.80GHz,2,2 x NVIDIA RTX A6000,10.4326,43.7667,1007.494571685791,machine,N,1.0,0.0
17
+ 2025-12-27T15:09:54,codecarbon,96194419-9194-4a03-a092-8692ed770e2f,5b0fa12a-3dd7-45bb-9766-cc326314d9f1,213.8194051720202,0.0109155758725974,5.105044541591432e-05,46.283023441000005,335.09607115710946,70.0,0.0065053860413555,0.0224906413258452,0.0040096628963999,0.0330056902636006,0.0,Italy,ITA,tuscany,,,Linux-5.4.0-216-generic-x86_64-with-glibc2.31,3.10.18,3.0.8,96,Intel(R) Xeon(R) Gold 6342 CPU @ 2.80GHz,2,2 x NVIDIA RTX A6000,10.4326,43.7667,1007.494571685791,machine,N,1.0,0.0
18
+ 2025-12-27T15:24:25,codecarbon,0fa7a59d-823e-44fc-a239-7ccb9af65c7b,5b0fa12a-3dd7-45bb-9766-cc326314d9f1,209.9902435876429,0.0095149210260789,4.531125286355391e-05,46.010721026800006,383.3990292654128,70.0,0.0025934692527804,0.0222321161189853,0.0039449110211701,0.0287704963929358,0.0,Italy,ITA,tuscany,,,Linux-5.4.0-216-generic-x86_64-with-glibc2.31,3.10.18,3.0.8,96,Intel(R) Xeon(R) Gold 6342 CPU @ 2.80GHz,2,2 x NVIDIA RTX A6000,10.4326,43.7667,1007.494571685791,machine,N,1.0,0.0
19
+ 2025-12-27T15:33:18,codecarbon,d4708ef7-03d0-4be8-a7ad-6d0701126a9d,5b0fa12a-3dd7-45bb-9766-cc326314d9f1,315.73992678523064,0.0145499318302652,4.608201432871753e-05,46.011900119500005,386.87574628907896,70.0,0.0039872089575019,0.0340754142047501,0.005932356942445,0.0439949801046971,0.0,Italy,ITA,tuscany,,,Linux-5.4.0-216-generic-x86_64-with-glibc2.31,3.10.18,3.0.8,96,Intel(R) Xeon(R) Gold 6342 CPU @ 2.80GHz,2,2 x NVIDIA RTX A6000,10.4326,43.7667,1007.494571685791,machine,N,1.0,0.0
20
+ 2025-12-27T15:39:49,codecarbon,c196d60d-c19a-4982-a0f7-127dfb13a65a,5b0fa12a-3dd7-45bb-9766-cc326314d9f1,218.5689760390669,0.009840141514632,4.502076046178325e-05,46.007411382000015,361.9919306181327,70.0,0.0026962526667695,0.0229559236424847,0.0041016971256118,0.0297538734348661,0.0,Italy,ITA,tuscany,,,Linux-5.4.0-216-generic-x86_64-with-glibc2.31,3.10.18,3.0.8,96,Intel(R) Xeon(R) Gold 6342 CPU @ 2.80GHz,2,2 x NVIDIA RTX A6000,10.4326,43.7667,1007.494571685791,machine,N,1.0,0.0
21
+ 2025-12-27T15:47:54,codecarbon,8459530a-74c3-4609-bc8d-b0dc078ce64e,5b0fa12a-3dd7-45bb-9766-cc326314d9f1,302.51271716691554,0.013778324491281824,4.5546265361397834e-05,46.00380238300001,267.05554895924075,70.0,0.0038275114928972417,0.03215940017196317,0.005674940702660632,0.04166185236752104,0.0,Italy,ITA,tuscany,,,Linux-5.4.0-216-generic-x86_64-with-glibc2.31,3.10.18,3.0.8,96,Intel(R) Xeon(R) Gold 6342 CPU @ 2.80GHz,2,2 x NVIDIA RTX A6000,10.4326,43.7667,1007.494571685791,machine,N,1.0,0.0
encoder/config.json ADDED
@@ -0,0 +1,26 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "RobertaModel"
4
+ ],
5
+ "attention_probs_dropout_prob": 0.1,
6
+ "bos_token_id": 0,
7
+ "classifier_dropout": null,
8
+ "dtype": "float32",
9
+ "eos_token_id": 2,
10
+ "hidden_act": "gelu",
11
+ "hidden_dropout_prob": 0.1,
12
+ "hidden_size": 768,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 3072,
15
+ "layer_norm_eps": 1e-05,
16
+ "max_position_embeddings": 514,
17
+ "model_type": "roberta",
18
+ "num_attention_heads": 12,
19
+ "num_hidden_layers": 12,
20
+ "pad_token_id": 1,
21
+ "position_embedding_type": "absolute",
22
+ "transformers_version": "4.57.1",
23
+ "type_vocab_size": 1,
24
+ "use_cache": true,
25
+ "vocab_size": 50265
26
+ }
encoder/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d4fee71e440a1067ba46f522a2085bf3ef792f116a2edb3bd1b583c52296cac4
3
+ size 498604904
multitask_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_type": "multitask_roberta",
3
+ "base_model": "FacebookAI/roberta-base",
4
+ "task_labels_map": {
5
+ "disagreement": 1,
6
+ "target": 2
7
+ },
8
+ "hidden_size": 768,
9
+ "dropout_prob": 0.4
10
+ }
target_head.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:320e3d120ccc52b181034fd24996726028726451e05448d34e8744d7b0cfdd74
3
+ size 7728
tokenizer/merges.txt ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer/sentencepiece.bpe.model ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:cfc8146abe2a0488e9e2a0c56de7952f7c11ab059eca145a0a727afce0db2865
3
+ size 5069051
tokenizer/special_tokens_map.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token": "<s>",
3
+ "cls_token": "<s>",
4
+ "eos_token": "</s>",
5
+ "mask_token": {
6
+ "content": "<mask>",
7
+ "lstrip": true,
8
+ "normalized": false,
9
+ "rstrip": false,
10
+ "single_word": false
11
+ },
12
+ "pad_token": "<pad>",
13
+ "sep_token": "</s>",
14
+ "unk_token": "<unk>"
15
+ }
tokenizer/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer/tokenizer_config.json ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": false,
3
+ "added_tokens_decoder": {
4
+ "0": {
5
+ "content": "<s>",
6
+ "lstrip": false,
7
+ "normalized": true,
8
+ "rstrip": false,
9
+ "single_word": false,
10
+ "special": true
11
+ },
12
+ "1": {
13
+ "content": "<pad>",
14
+ "lstrip": false,
15
+ "normalized": true,
16
+ "rstrip": false,
17
+ "single_word": false,
18
+ "special": true
19
+ },
20
+ "2": {
21
+ "content": "</s>",
22
+ "lstrip": false,
23
+ "normalized": true,
24
+ "rstrip": false,
25
+ "single_word": false,
26
+ "special": true
27
+ },
28
+ "3": {
29
+ "content": "<unk>",
30
+ "lstrip": false,
31
+ "normalized": true,
32
+ "rstrip": false,
33
+ "single_word": false,
34
+ "special": true
35
+ },
36
+ "50264": {
37
+ "content": "<mask>",
38
+ "lstrip": true,
39
+ "normalized": false,
40
+ "rstrip": false,
41
+ "single_word": false,
42
+ "special": true
43
+ }
44
+ },
45
+ "bos_token": "<s>",
46
+ "clean_up_tokenization_spaces": false,
47
+ "cls_token": "<s>",
48
+ "eos_token": "</s>",
49
+ "errors": "replace",
50
+ "extra_special_tokens": {},
51
+ "mask_token": "<mask>",
52
+ "model_max_length": 512,
53
+ "pad_token": "<pad>",
54
+ "sep_token": "</s>",
55
+ "tokenizer_class": "RobertaTokenizer",
56
+ "trim_offsets": true,
57
+ "unk_token": "<unk>"
58
+ }
tokenizer/vocab.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer/vocab.txt ADDED
The diff for this file is too large to render. See raw diff