Commit
96e2104
·
verified ·
1 Parent(s): 07e5c8f

Add cambridgeltl-SapBERT-from-PubMedBERT-fulltext-E5-Pairwise

Browse files

Upload fine-tuned rerankers for BioASQ 14B

Co-authored-by: André Ribeiro <andrepedro2004@hotmail.com>
Co-authored-by: Rúben Garrido <rubengarrido@ua.pt>

cambridgeltl-SapBERT-from-PubMedBERT-fulltext-E5-Pairwise/config.json ADDED
@@ -0,0 +1,35 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_cross_attention": false,
3
+ "architectures": [
4
+ "BertForSequenceClassification"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.1,
7
+ "bos_token_id": null,
8
+ "classifier_dropout": null,
9
+ "dtype": "bfloat16",
10
+ "eos_token_id": null,
11
+ "gradient_checkpointing": false,
12
+ "hidden_act": "gelu",
13
+ "hidden_dropout_prob": 0.1,
14
+ "hidden_size": 768,
15
+ "id2label": {
16
+ "0": "LABEL_0"
17
+ },
18
+ "initializer_range": 0.02,
19
+ "intermediate_size": 3072,
20
+ "is_decoder": false,
21
+ "label2id": {
22
+ "LABEL_0": 0
23
+ },
24
+ "layer_norm_eps": 1e-12,
25
+ "max_position_embeddings": 512,
26
+ "model_type": "bert",
27
+ "num_attention_heads": 12,
28
+ "num_hidden_layers": 12,
29
+ "pad_token_id": 0,
30
+ "tie_word_embeddings": true,
31
+ "transformers_version": "5.2.0",
32
+ "type_vocab_size": 2,
33
+ "use_cache": false,
34
+ "vocab_size": 30522
35
+ }
cambridgeltl-SapBERT-from-PubMedBERT-fulltext-E5-Pairwise/model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8c7ada747cf88c76d74582b4c7e4d8b533935bd9e9dfcd6b40becc8f33a23397
3
+ size 218989610
cambridgeltl-SapBERT-from-PubMedBERT-fulltext-E5-Pairwise/ranx_results.json ADDED
@@ -0,0 +1,31 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model": "cambridgeltl-SapBERT-from-PubMedBERT-fulltext",
3
+ "val_files": [
4
+ "../../data/val_data/13B1_golden.json",
5
+ "../../data/val_data/13B2_golden.json"
6
+ ],
7
+ "total": {
8
+ "ndcg@5": 0.34799635548798247,
9
+ "mrr": 0.4906069571530938,
10
+ "recall@10": 0.2883412555378209,
11
+ "recall@100": 0.9155763792983985,
12
+ "map@10": 0.1675449376040265,
13
+ "map-bioasq@10": 0.25943456271397447
14
+ },
15
+ "13B1_golden.json": {
16
+ "ndcg@5": 0.24936963598469228,
17
+ "mrr": 0.408355155418017,
18
+ "recall@10": 0.21811569115908816,
19
+ "recall@100": 0.9182208355190018,
20
+ "map@10": 0.11149176957961644,
21
+ "map-bioasq@10": 0.1728135698723934
22
+ },
23
+ "13B2_golden.json": {
24
+ "ndcg@5": 0.4466230749912727,
25
+ "mrr": 0.5728587588881707,
26
+ "recall@10": 0.35856681991655365,
27
+ "recall@100": 0.9129319230777954,
28
+ "map@10": 0.2235981056284365,
29
+ "map-bioasq@10": 0.34605555555555556
30
+ }
31
+ }
cambridgeltl-SapBERT-from-PubMedBERT-fulltext-E5-Pairwise/tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
cambridgeltl-SapBERT-from-PubMedBERT-fulltext-E5-Pairwise/tokenizer_config.json ADDED
@@ -0,0 +1,15 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "backend": "tokenizers",
3
+ "cls_token": "[CLS]",
4
+ "do_lower_case": false,
5
+ "full_tokenizer_file": null,
6
+ "is_local": false,
7
+ "mask_token": "[MASK]",
8
+ "model_max_length": 1000000000000000019884624838656,
9
+ "pad_token": "[PAD]",
10
+ "sep_token": "[SEP]",
11
+ "strip_accents": null,
12
+ "tokenize_chinese_chars": true,
13
+ "tokenizer_class": "BertTokenizer",
14
+ "unk_token": "[UNK]"
15
+ }
cambridgeltl-SapBERT-from-PubMedBERT-fulltext-E5-Pairwise/trainer_state.json ADDED
@@ -0,0 +1,963 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 5.0,
6
+ "eval_steps": 500,
7
+ "global_step": 6375,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.0392156862745098,
14
+ "grad_norm": 5.03125,
15
+ "learning_rate": 1.9846274509803922e-05,
16
+ "loss": 0.9589825439453125,
17
+ "step": 50
18
+ },
19
+ {
20
+ "epoch": 0.0784313725490196,
21
+ "grad_norm": 5.59375,
22
+ "learning_rate": 1.9689411764705883e-05,
23
+ "loss": 0.9116700744628906,
24
+ "step": 100
25
+ },
26
+ {
27
+ "epoch": 0.11764705882352941,
28
+ "grad_norm": 5.28125,
29
+ "learning_rate": 1.9532549019607844e-05,
30
+ "loss": 0.8409913635253906,
31
+ "step": 150
32
+ },
33
+ {
34
+ "epoch": 0.1568627450980392,
35
+ "grad_norm": 5.4375,
36
+ "learning_rate": 1.9375686274509805e-05,
37
+ "loss": 0.7898799896240234,
38
+ "step": 200
39
+ },
40
+ {
41
+ "epoch": 0.19607843137254902,
42
+ "grad_norm": 6.34375,
43
+ "learning_rate": 1.9218823529411766e-05,
44
+ "loss": 0.7475868225097656,
45
+ "step": 250
46
+ },
47
+ {
48
+ "epoch": 0.23529411764705882,
49
+ "grad_norm": 7.28125,
50
+ "learning_rate": 1.9061960784313727e-05,
51
+ "loss": 0.753358154296875,
52
+ "step": 300
53
+ },
54
+ {
55
+ "epoch": 0.27450980392156865,
56
+ "grad_norm": 6.40625,
57
+ "learning_rate": 1.8905098039215688e-05,
58
+ "loss": 0.6765160369873047,
59
+ "step": 350
60
+ },
61
+ {
62
+ "epoch": 0.3137254901960784,
63
+ "grad_norm": 7.25,
64
+ "learning_rate": 1.874823529411765e-05,
65
+ "loss": 0.7208175659179688,
66
+ "step": 400
67
+ },
68
+ {
69
+ "epoch": 0.35294117647058826,
70
+ "grad_norm": 7.4375,
71
+ "learning_rate": 1.8591372549019607e-05,
72
+ "loss": 0.6581819152832031,
73
+ "step": 450
74
+ },
75
+ {
76
+ "epoch": 0.39215686274509803,
77
+ "grad_norm": 7.0625,
78
+ "learning_rate": 1.843450980392157e-05,
79
+ "loss": 0.6623909759521485,
80
+ "step": 500
81
+ },
82
+ {
83
+ "epoch": 0.43137254901960786,
84
+ "grad_norm": 9.625,
85
+ "learning_rate": 1.8277647058823532e-05,
86
+ "loss": 0.6893426513671875,
87
+ "step": 550
88
+ },
89
+ {
90
+ "epoch": 0.47058823529411764,
91
+ "grad_norm": 10.0625,
92
+ "learning_rate": 1.8120784313725493e-05,
93
+ "loss": 0.6717083740234375,
94
+ "step": 600
95
+ },
96
+ {
97
+ "epoch": 0.5098039215686274,
98
+ "grad_norm": 7.625,
99
+ "learning_rate": 1.796392156862745e-05,
100
+ "loss": 0.6486688232421876,
101
+ "step": 650
102
+ },
103
+ {
104
+ "epoch": 0.5490196078431373,
105
+ "grad_norm": 8.375,
106
+ "learning_rate": 1.7807058823529412e-05,
107
+ "loss": 0.6297515106201171,
108
+ "step": 700
109
+ },
110
+ {
111
+ "epoch": 0.5882352941176471,
112
+ "grad_norm": 7.625,
113
+ "learning_rate": 1.7650196078431373e-05,
114
+ "loss": 0.6352976989746094,
115
+ "step": 750
116
+ },
117
+ {
118
+ "epoch": 0.6274509803921569,
119
+ "grad_norm": 6.875,
120
+ "learning_rate": 1.7493333333333334e-05,
121
+ "loss": 0.6825536346435547,
122
+ "step": 800
123
+ },
124
+ {
125
+ "epoch": 0.6666666666666666,
126
+ "grad_norm": 8.1875,
127
+ "learning_rate": 1.7336470588235295e-05,
128
+ "loss": 0.6208578491210938,
129
+ "step": 850
130
+ },
131
+ {
132
+ "epoch": 0.7058823529411765,
133
+ "grad_norm": 6.90625,
134
+ "learning_rate": 1.7179607843137256e-05,
135
+ "loss": 0.6486013793945312,
136
+ "step": 900
137
+ },
138
+ {
139
+ "epoch": 0.7450980392156863,
140
+ "grad_norm": 8.375,
141
+ "learning_rate": 1.7022745098039217e-05,
142
+ "loss": 0.6393245697021485,
143
+ "step": 950
144
+ },
145
+ {
146
+ "epoch": 0.7843137254901961,
147
+ "grad_norm": 9.0,
148
+ "learning_rate": 1.6865882352941178e-05,
149
+ "loss": 0.6479877471923828,
150
+ "step": 1000
151
+ },
152
+ {
153
+ "epoch": 0.8235294117647058,
154
+ "grad_norm": 8.625,
155
+ "learning_rate": 1.670901960784314e-05,
156
+ "loss": 0.6383560562133789,
157
+ "step": 1050
158
+ },
159
+ {
160
+ "epoch": 0.8627450980392157,
161
+ "grad_norm": 8.75,
162
+ "learning_rate": 1.65521568627451e-05,
163
+ "loss": 0.6760401916503906,
164
+ "step": 1100
165
+ },
166
+ {
167
+ "epoch": 0.9019607843137255,
168
+ "grad_norm": 10.375,
169
+ "learning_rate": 1.639529411764706e-05,
170
+ "loss": 0.6304359054565429,
171
+ "step": 1150
172
+ },
173
+ {
174
+ "epoch": 0.9411764705882353,
175
+ "grad_norm": 9.4375,
176
+ "learning_rate": 1.6238431372549022e-05,
177
+ "loss": 0.6240536499023438,
178
+ "step": 1200
179
+ },
180
+ {
181
+ "epoch": 0.9803921568627451,
182
+ "grad_norm": 7.46875,
183
+ "learning_rate": 1.608156862745098e-05,
184
+ "loss": 0.6392737579345703,
185
+ "step": 1250
186
+ },
187
+ {
188
+ "epoch": 1.0,
189
+ "eval_loss": 0.674563467502594,
190
+ "eval_runtime": 0.7925,
191
+ "eval_samples_per_second": 211.975,
192
+ "eval_steps_per_second": 3.785,
193
+ "step": 1275
194
+ },
195
+ {
196
+ "epoch": 1.0196078431372548,
197
+ "grad_norm": 8.3125,
198
+ "learning_rate": 1.592470588235294e-05,
199
+ "loss": 0.6577691650390625,
200
+ "step": 1300
201
+ },
202
+ {
203
+ "epoch": 1.0588235294117647,
204
+ "grad_norm": 7.96875,
205
+ "learning_rate": 1.5767843137254902e-05,
206
+ "loss": 0.6757988739013672,
207
+ "step": 1350
208
+ },
209
+ {
210
+ "epoch": 1.0980392156862746,
211
+ "grad_norm": 8.0625,
212
+ "learning_rate": 1.5610980392156866e-05,
213
+ "loss": 0.6209789657592774,
214
+ "step": 1400
215
+ },
216
+ {
217
+ "epoch": 1.1372549019607843,
218
+ "grad_norm": 7.125,
219
+ "learning_rate": 1.5454117647058824e-05,
220
+ "loss": 0.6348079299926758,
221
+ "step": 1450
222
+ },
223
+ {
224
+ "epoch": 1.1764705882352942,
225
+ "grad_norm": 7.78125,
226
+ "learning_rate": 1.5297254901960785e-05,
227
+ "loss": 0.647025146484375,
228
+ "step": 1500
229
+ },
230
+ {
231
+ "epoch": 1.215686274509804,
232
+ "grad_norm": 7.8125,
233
+ "learning_rate": 1.5140392156862746e-05,
234
+ "loss": 0.6091714477539063,
235
+ "step": 1550
236
+ },
237
+ {
238
+ "epoch": 1.2549019607843137,
239
+ "grad_norm": 8.0625,
240
+ "learning_rate": 1.4983529411764707e-05,
241
+ "loss": 0.6313312149047852,
242
+ "step": 1600
243
+ },
244
+ {
245
+ "epoch": 1.2941176470588236,
246
+ "grad_norm": 9.875,
247
+ "learning_rate": 1.4826666666666666e-05,
248
+ "loss": 0.6484039306640625,
249
+ "step": 1650
250
+ },
251
+ {
252
+ "epoch": 1.3333333333333333,
253
+ "grad_norm": 8.5,
254
+ "learning_rate": 1.4669803921568627e-05,
255
+ "loss": 0.5883905410766601,
256
+ "step": 1700
257
+ },
258
+ {
259
+ "epoch": 1.3725490196078431,
260
+ "grad_norm": 8.25,
261
+ "learning_rate": 1.451294117647059e-05,
262
+ "loss": 0.5900686645507812,
263
+ "step": 1750
264
+ },
265
+ {
266
+ "epoch": 1.4117647058823528,
267
+ "grad_norm": 9.375,
268
+ "learning_rate": 1.4356078431372551e-05,
269
+ "loss": 0.6133530044555664,
270
+ "step": 1800
271
+ },
272
+ {
273
+ "epoch": 1.4509803921568627,
274
+ "grad_norm": 7.40625,
275
+ "learning_rate": 1.419921568627451e-05,
276
+ "loss": 0.5745703887939453,
277
+ "step": 1850
278
+ },
279
+ {
280
+ "epoch": 1.4901960784313726,
281
+ "grad_norm": 8.5625,
282
+ "learning_rate": 1.4042352941176472e-05,
283
+ "loss": 0.6304582214355469,
284
+ "step": 1900
285
+ },
286
+ {
287
+ "epoch": 1.5294117647058822,
288
+ "grad_norm": 8.3125,
289
+ "learning_rate": 1.3885490196078433e-05,
290
+ "loss": 0.6089314651489258,
291
+ "step": 1950
292
+ },
293
+ {
294
+ "epoch": 1.5686274509803921,
295
+ "grad_norm": 8.625,
296
+ "learning_rate": 1.3728627450980394e-05,
297
+ "loss": 0.6195945739746094,
298
+ "step": 2000
299
+ },
300
+ {
301
+ "epoch": 1.607843137254902,
302
+ "grad_norm": 7.34375,
303
+ "learning_rate": 1.3571764705882353e-05,
304
+ "loss": 0.6149443054199218,
305
+ "step": 2050
306
+ },
307
+ {
308
+ "epoch": 1.6470588235294117,
309
+ "grad_norm": 7.03125,
310
+ "learning_rate": 1.3414901960784314e-05,
311
+ "loss": 0.5880471038818359,
312
+ "step": 2100
313
+ },
314
+ {
315
+ "epoch": 1.6862745098039216,
316
+ "grad_norm": 8.25,
317
+ "learning_rate": 1.3258039215686275e-05,
318
+ "loss": 0.5851007461547851,
319
+ "step": 2150
320
+ },
321
+ {
322
+ "epoch": 1.7254901960784315,
323
+ "grad_norm": 7.59375,
324
+ "learning_rate": 1.3101176470588238e-05,
325
+ "loss": 0.6281360626220703,
326
+ "step": 2200
327
+ },
328
+ {
329
+ "epoch": 1.7647058823529411,
330
+ "grad_norm": 12.0,
331
+ "learning_rate": 1.2944313725490197e-05,
332
+ "loss": 0.5828293228149414,
333
+ "step": 2250
334
+ },
335
+ {
336
+ "epoch": 1.803921568627451,
337
+ "grad_norm": 7.78125,
338
+ "learning_rate": 1.2787450980392158e-05,
339
+ "loss": 0.632794418334961,
340
+ "step": 2300
341
+ },
342
+ {
343
+ "epoch": 1.843137254901961,
344
+ "grad_norm": 9.25,
345
+ "learning_rate": 1.2630588235294119e-05,
346
+ "loss": 0.629760627746582,
347
+ "step": 2350
348
+ },
349
+ {
350
+ "epoch": 1.8823529411764706,
351
+ "grad_norm": 9.4375,
352
+ "learning_rate": 1.247372549019608e-05,
353
+ "loss": 0.562823371887207,
354
+ "step": 2400
355
+ },
356
+ {
357
+ "epoch": 1.9215686274509802,
358
+ "grad_norm": 7.8125,
359
+ "learning_rate": 1.231686274509804e-05,
360
+ "loss": 0.6019530868530274,
361
+ "step": 2450
362
+ },
363
+ {
364
+ "epoch": 1.9607843137254903,
365
+ "grad_norm": 8.4375,
366
+ "learning_rate": 1.216e-05,
367
+ "loss": 0.611793441772461,
368
+ "step": 2500
369
+ },
370
+ {
371
+ "epoch": 2.0,
372
+ "grad_norm": 21.75,
373
+ "learning_rate": 1.2003137254901961e-05,
374
+ "loss": 0.6110140609741211,
375
+ "step": 2550
376
+ },
377
+ {
378
+ "epoch": 2.0,
379
+ "eval_loss": 0.6515551805496216,
380
+ "eval_runtime": 0.7445,
381
+ "eval_samples_per_second": 225.662,
382
+ "eval_steps_per_second": 4.03,
383
+ "step": 2550
384
+ },
385
+ {
386
+ "epoch": 2.0392156862745097,
387
+ "grad_norm": 9.5,
388
+ "learning_rate": 1.1846274509803923e-05,
389
+ "loss": 0.5883094024658203,
390
+ "step": 2600
391
+ },
392
+ {
393
+ "epoch": 2.0784313725490198,
394
+ "grad_norm": 10.4375,
395
+ "learning_rate": 1.1689411764705882e-05,
396
+ "loss": 0.6096354293823242,
397
+ "step": 2650
398
+ },
399
+ {
400
+ "epoch": 2.1176470588235294,
401
+ "grad_norm": 9.5625,
402
+ "learning_rate": 1.1532549019607845e-05,
403
+ "loss": 0.5969795227050781,
404
+ "step": 2700
405
+ },
406
+ {
407
+ "epoch": 2.156862745098039,
408
+ "grad_norm": 9.8125,
409
+ "learning_rate": 1.1375686274509806e-05,
410
+ "loss": 0.6177022171020508,
411
+ "step": 2750
412
+ },
413
+ {
414
+ "epoch": 2.196078431372549,
415
+ "grad_norm": 10.6875,
416
+ "learning_rate": 1.1218823529411767e-05,
417
+ "loss": 0.6016413116455078,
418
+ "step": 2800
419
+ },
420
+ {
421
+ "epoch": 2.235294117647059,
422
+ "grad_norm": 10.875,
423
+ "learning_rate": 1.1061960784313726e-05,
424
+ "loss": 0.615409164428711,
425
+ "step": 2850
426
+ },
427
+ {
428
+ "epoch": 2.2745098039215685,
429
+ "grad_norm": 9.0,
430
+ "learning_rate": 1.0905098039215687e-05,
431
+ "loss": 0.6393753814697266,
432
+ "step": 2900
433
+ },
434
+ {
435
+ "epoch": 2.313725490196078,
436
+ "grad_norm": 8.125,
437
+ "learning_rate": 1.0748235294117648e-05,
438
+ "loss": 0.5983933258056641,
439
+ "step": 2950
440
+ },
441
+ {
442
+ "epoch": 2.3529411764705883,
443
+ "grad_norm": 8.0625,
444
+ "learning_rate": 1.0591372549019609e-05,
445
+ "loss": 0.6090192031860352,
446
+ "step": 3000
447
+ },
448
+ {
449
+ "epoch": 2.392156862745098,
450
+ "grad_norm": 11.8125,
451
+ "learning_rate": 1.0434509803921568e-05,
452
+ "loss": 0.612490005493164,
453
+ "step": 3050
454
+ },
455
+ {
456
+ "epoch": 2.431372549019608,
457
+ "grad_norm": 8.0625,
458
+ "learning_rate": 1.027764705882353e-05,
459
+ "loss": 0.5786522674560547,
460
+ "step": 3100
461
+ },
462
+ {
463
+ "epoch": 2.4705882352941178,
464
+ "grad_norm": 11.0625,
465
+ "learning_rate": 1.0120784313725492e-05,
466
+ "loss": 0.5843408966064453,
467
+ "step": 3150
468
+ },
469
+ {
470
+ "epoch": 2.5098039215686274,
471
+ "grad_norm": 9.25,
472
+ "learning_rate": 9.963921568627451e-06,
473
+ "loss": 0.5953018951416016,
474
+ "step": 3200
475
+ },
476
+ {
477
+ "epoch": 2.549019607843137,
478
+ "grad_norm": 9.5,
479
+ "learning_rate": 9.807058823529412e-06,
480
+ "loss": 0.6094627380371094,
481
+ "step": 3250
482
+ },
483
+ {
484
+ "epoch": 2.588235294117647,
485
+ "grad_norm": 9.25,
486
+ "learning_rate": 9.650196078431373e-06,
487
+ "loss": 0.5818396759033203,
488
+ "step": 3300
489
+ },
490
+ {
491
+ "epoch": 2.627450980392157,
492
+ "grad_norm": 10.25,
493
+ "learning_rate": 9.493333333333334e-06,
494
+ "loss": 0.6009473419189453,
495
+ "step": 3350
496
+ },
497
+ {
498
+ "epoch": 2.6666666666666665,
499
+ "grad_norm": 10.75,
500
+ "learning_rate": 9.336470588235294e-06,
501
+ "loss": 0.5970331954956055,
502
+ "step": 3400
503
+ },
504
+ {
505
+ "epoch": 2.7058823529411766,
506
+ "grad_norm": 8.75,
507
+ "learning_rate": 9.179607843137257e-06,
508
+ "loss": 0.5867655181884766,
509
+ "step": 3450
510
+ },
511
+ {
512
+ "epoch": 2.7450980392156863,
513
+ "grad_norm": 9.5,
514
+ "learning_rate": 9.022745098039216e-06,
515
+ "loss": 0.6068114471435547,
516
+ "step": 3500
517
+ },
518
+ {
519
+ "epoch": 2.784313725490196,
520
+ "grad_norm": 8.75,
521
+ "learning_rate": 8.865882352941177e-06,
522
+ "loss": 0.5716716003417969,
523
+ "step": 3550
524
+ },
525
+ {
526
+ "epoch": 2.8235294117647056,
527
+ "grad_norm": 9.8125,
528
+ "learning_rate": 8.709019607843138e-06,
529
+ "loss": 0.5989672470092774,
530
+ "step": 3600
531
+ },
532
+ {
533
+ "epoch": 2.8627450980392157,
534
+ "grad_norm": 10.0,
535
+ "learning_rate": 8.552156862745099e-06,
536
+ "loss": 0.6049938201904297,
537
+ "step": 3650
538
+ },
539
+ {
540
+ "epoch": 2.9019607843137254,
541
+ "grad_norm": 9.375,
542
+ "learning_rate": 8.39529411764706e-06,
543
+ "loss": 0.5805830383300781,
544
+ "step": 3700
545
+ },
546
+ {
547
+ "epoch": 2.9411764705882355,
548
+ "grad_norm": 8.6875,
549
+ "learning_rate": 8.238431372549021e-06,
550
+ "loss": 0.5986671447753906,
551
+ "step": 3750
552
+ },
553
+ {
554
+ "epoch": 2.980392156862745,
555
+ "grad_norm": 9.875,
556
+ "learning_rate": 8.08156862745098e-06,
557
+ "loss": 0.5819197463989257,
558
+ "step": 3800
559
+ },
560
+ {
561
+ "epoch": 3.0,
562
+ "eval_loss": 0.6459775567054749,
563
+ "eval_runtime": 0.7468,
564
+ "eval_samples_per_second": 224.958,
565
+ "eval_steps_per_second": 4.017,
566
+ "step": 3825
567
+ },
568
+ {
569
+ "epoch": 3.019607843137255,
570
+ "grad_norm": 8.375,
571
+ "learning_rate": 7.924705882352941e-06,
572
+ "loss": 0.5851152420043946,
573
+ "step": 3850
574
+ },
575
+ {
576
+ "epoch": 3.0588235294117645,
577
+ "grad_norm": 8.25,
578
+ "learning_rate": 7.767843137254902e-06,
579
+ "loss": 0.609749755859375,
580
+ "step": 3900
581
+ },
582
+ {
583
+ "epoch": 3.0980392156862746,
584
+ "grad_norm": 9.5,
585
+ "learning_rate": 7.610980392156863e-06,
586
+ "loss": 0.6063669204711915,
587
+ "step": 3950
588
+ },
589
+ {
590
+ "epoch": 3.1372549019607843,
591
+ "grad_norm": 8.25,
592
+ "learning_rate": 7.4541176470588236e-06,
593
+ "loss": 0.6049988555908203,
594
+ "step": 4000
595
+ },
596
+ {
597
+ "epoch": 3.176470588235294,
598
+ "grad_norm": 10.125,
599
+ "learning_rate": 7.2972549019607854e-06,
600
+ "loss": 0.6084383010864258,
601
+ "step": 4050
602
+ },
603
+ {
604
+ "epoch": 3.215686274509804,
605
+ "grad_norm": 10.125,
606
+ "learning_rate": 7.140392156862746e-06,
607
+ "loss": 0.6269755172729492,
608
+ "step": 4100
609
+ },
610
+ {
611
+ "epoch": 3.2549019607843137,
612
+ "grad_norm": 9.375,
613
+ "learning_rate": 6.983529411764707e-06,
614
+ "loss": 0.5867807006835938,
615
+ "step": 4150
616
+ },
617
+ {
618
+ "epoch": 3.2941176470588234,
619
+ "grad_norm": 9.1875,
620
+ "learning_rate": 6.826666666666667e-06,
621
+ "loss": 0.6203334045410156,
622
+ "step": 4200
623
+ },
624
+ {
625
+ "epoch": 3.3333333333333335,
626
+ "grad_norm": 12.4375,
627
+ "learning_rate": 6.669803921568629e-06,
628
+ "loss": 0.6299712371826172,
629
+ "step": 4250
630
+ },
631
+ {
632
+ "epoch": 3.372549019607843,
633
+ "grad_norm": 8.375,
634
+ "learning_rate": 6.512941176470589e-06,
635
+ "loss": 0.6253769683837891,
636
+ "step": 4300
637
+ },
638
+ {
639
+ "epoch": 3.411764705882353,
640
+ "grad_norm": 7.90625,
641
+ "learning_rate": 6.35607843137255e-06,
642
+ "loss": 0.629886245727539,
643
+ "step": 4350
644
+ },
645
+ {
646
+ "epoch": 3.450980392156863,
647
+ "grad_norm": 9.6875,
648
+ "learning_rate": 6.19921568627451e-06,
649
+ "loss": 0.6090424728393554,
650
+ "step": 4400
651
+ },
652
+ {
653
+ "epoch": 3.4901960784313726,
654
+ "grad_norm": 8.375,
655
+ "learning_rate": 6.042352941176471e-06,
656
+ "loss": 0.6086097717285156,
657
+ "step": 4450
658
+ },
659
+ {
660
+ "epoch": 3.5294117647058822,
661
+ "grad_norm": 8.9375,
662
+ "learning_rate": 5.885490196078431e-06,
663
+ "loss": 0.6270752716064453,
664
+ "step": 4500
665
+ },
666
+ {
667
+ "epoch": 3.568627450980392,
668
+ "grad_norm": 9.8125,
669
+ "learning_rate": 5.728627450980393e-06,
670
+ "loss": 0.6130237197875976,
671
+ "step": 4550
672
+ },
673
+ {
674
+ "epoch": 3.607843137254902,
675
+ "grad_norm": 9.1875,
676
+ "learning_rate": 5.571764705882353e-06,
677
+ "loss": 0.5669650650024414,
678
+ "step": 4600
679
+ },
680
+ {
681
+ "epoch": 3.6470588235294117,
682
+ "grad_norm": 9.625,
683
+ "learning_rate": 5.414901960784314e-06,
684
+ "loss": 0.5876271057128907,
685
+ "step": 4650
686
+ },
687
+ {
688
+ "epoch": 3.686274509803922,
689
+ "grad_norm": 9.875,
690
+ "learning_rate": 5.2580392156862745e-06,
691
+ "loss": 0.5910399627685546,
692
+ "step": 4700
693
+ },
694
+ {
695
+ "epoch": 3.7254901960784315,
696
+ "grad_norm": 9.375,
697
+ "learning_rate": 5.101176470588236e-06,
698
+ "loss": 0.5917201232910156,
699
+ "step": 4750
700
+ },
701
+ {
702
+ "epoch": 3.764705882352941,
703
+ "grad_norm": 8.375,
704
+ "learning_rate": 4.9443137254901966e-06,
705
+ "loss": 0.6246376419067383,
706
+ "step": 4800
707
+ },
708
+ {
709
+ "epoch": 3.803921568627451,
710
+ "grad_norm": 8.5625,
711
+ "learning_rate": 4.787450980392157e-06,
712
+ "loss": 0.6015016937255859,
713
+ "step": 4850
714
+ },
715
+ {
716
+ "epoch": 3.843137254901961,
717
+ "grad_norm": 7.40625,
718
+ "learning_rate": 4.630588235294118e-06,
719
+ "loss": 0.5997852325439453,
720
+ "step": 4900
721
+ },
722
+ {
723
+ "epoch": 3.8823529411764706,
724
+ "grad_norm": 10.5,
725
+ "learning_rate": 4.473725490196079e-06,
726
+ "loss": 0.6182506561279297,
727
+ "step": 4950
728
+ },
729
+ {
730
+ "epoch": 3.9215686274509802,
731
+ "grad_norm": 9.6875,
732
+ "learning_rate": 4.316862745098039e-06,
733
+ "loss": 0.6194680404663085,
734
+ "step": 5000
735
+ },
736
+ {
737
+ "epoch": 3.9607843137254903,
738
+ "grad_norm": 9.5,
739
+ "learning_rate": 4.16e-06,
740
+ "loss": 0.6190138244628907,
741
+ "step": 5050
742
+ },
743
+ {
744
+ "epoch": 4.0,
745
+ "grad_norm": 19.0,
746
+ "learning_rate": 4.003137254901961e-06,
747
+ "loss": 0.6149836730957031,
748
+ "step": 5100
749
+ },
750
+ {
751
+ "epoch": 4.0,
752
+ "eval_loss": 0.6415771245956421,
753
+ "eval_runtime": 0.7496,
754
+ "eval_samples_per_second": 224.122,
755
+ "eval_steps_per_second": 4.002,
756
+ "step": 5100
757
+ },
758
+ {
759
+ "epoch": 4.03921568627451,
760
+ "grad_norm": 8.875,
761
+ "learning_rate": 3.846274509803922e-06,
762
+ "loss": 0.5928133773803711,
763
+ "step": 5150
764
+ },
765
+ {
766
+ "epoch": 4.078431372549019,
767
+ "grad_norm": 8.3125,
768
+ "learning_rate": 3.6894117647058826e-06,
769
+ "loss": 0.6135052871704102,
770
+ "step": 5200
771
+ },
772
+ {
773
+ "epoch": 4.117647058823529,
774
+ "grad_norm": 9.4375,
775
+ "learning_rate": 3.5325490196078432e-06,
776
+ "loss": 0.6121374893188477,
777
+ "step": 5250
778
+ },
779
+ {
780
+ "epoch": 4.1568627450980395,
781
+ "grad_norm": 10.5625,
782
+ "learning_rate": 3.375686274509804e-06,
783
+ "loss": 0.6306461334228516,
784
+ "step": 5300
785
+ },
786
+ {
787
+ "epoch": 4.196078431372549,
788
+ "grad_norm": 9.0625,
789
+ "learning_rate": 3.218823529411765e-06,
790
+ "loss": 0.6313008880615234,
791
+ "step": 5350
792
+ },
793
+ {
794
+ "epoch": 4.235294117647059,
795
+ "grad_norm": 8.5,
796
+ "learning_rate": 3.0619607843137255e-06,
797
+ "loss": 0.6515373229980469,
798
+ "step": 5400
799
+ },
800
+ {
801
+ "epoch": 4.2745098039215685,
802
+ "grad_norm": 11.4375,
803
+ "learning_rate": 2.9050980392156865e-06,
804
+ "loss": 0.5686242294311523,
805
+ "step": 5450
806
+ },
807
+ {
808
+ "epoch": 4.313725490196078,
809
+ "grad_norm": 9.5625,
810
+ "learning_rate": 2.748235294117647e-06,
811
+ "loss": 0.6116217041015625,
812
+ "step": 5500
813
+ },
814
+ {
815
+ "epoch": 4.352941176470588,
816
+ "grad_norm": 8.375,
817
+ "learning_rate": 2.5913725490196077e-06,
818
+ "loss": 0.5991297912597656,
819
+ "step": 5550
820
+ },
821
+ {
822
+ "epoch": 4.392156862745098,
823
+ "grad_norm": 8.5625,
824
+ "learning_rate": 2.434509803921569e-06,
825
+ "loss": 0.6036840057373047,
826
+ "step": 5600
827
+ },
828
+ {
829
+ "epoch": 4.431372549019608,
830
+ "grad_norm": 8.75,
831
+ "learning_rate": 2.2776470588235297e-06,
832
+ "loss": 0.6462667846679687,
833
+ "step": 5650
834
+ },
835
+ {
836
+ "epoch": 4.470588235294118,
837
+ "grad_norm": 9.5,
838
+ "learning_rate": 2.1207843137254903e-06,
839
+ "loss": 0.6222269058227539,
840
+ "step": 5700
841
+ },
842
+ {
843
+ "epoch": 4.509803921568627,
844
+ "grad_norm": 8.5,
845
+ "learning_rate": 1.9639215686274514e-06,
846
+ "loss": 0.6418203735351562,
847
+ "step": 5750
848
+ },
849
+ {
850
+ "epoch": 4.549019607843137,
851
+ "grad_norm": 9.0,
852
+ "learning_rate": 1.807058823529412e-06,
853
+ "loss": 0.6009382247924805,
854
+ "step": 5800
855
+ },
856
+ {
857
+ "epoch": 4.588235294117647,
858
+ "grad_norm": 9.75,
859
+ "learning_rate": 1.6501960784313728e-06,
860
+ "loss": 0.644178237915039,
861
+ "step": 5850
862
+ },
863
+ {
864
+ "epoch": 4.627450980392156,
865
+ "grad_norm": 9.0625,
866
+ "learning_rate": 1.4933333333333336e-06,
867
+ "loss": 0.6416984558105469,
868
+ "step": 5900
869
+ },
870
+ {
871
+ "epoch": 4.666666666666667,
872
+ "grad_norm": 8.75,
873
+ "learning_rate": 1.3364705882352944e-06,
874
+ "loss": 0.600451431274414,
875
+ "step": 5950
876
+ },
877
+ {
878
+ "epoch": 4.705882352941177,
879
+ "grad_norm": 8.625,
880
+ "learning_rate": 1.179607843137255e-06,
881
+ "loss": 0.5960040283203125,
882
+ "step": 6000
883
+ },
884
+ {
885
+ "epoch": 4.745098039215686,
886
+ "grad_norm": 9.75,
887
+ "learning_rate": 1.0227450980392158e-06,
888
+ "loss": 0.6293835830688477,
889
+ "step": 6050
890
+ },
891
+ {
892
+ "epoch": 4.784313725490196,
893
+ "grad_norm": 11.8125,
894
+ "learning_rate": 8.658823529411765e-07,
895
+ "loss": 0.6368055725097657,
896
+ "step": 6100
897
+ },
898
+ {
899
+ "epoch": 4.823529411764706,
900
+ "grad_norm": 9.375,
901
+ "learning_rate": 7.090196078431373e-07,
902
+ "loss": 0.609964599609375,
903
+ "step": 6150
904
+ },
905
+ {
906
+ "epoch": 4.862745098039216,
907
+ "grad_norm": 9.625,
908
+ "learning_rate": 5.521568627450981e-07,
909
+ "loss": 0.6066595458984375,
910
+ "step": 6200
911
+ },
912
+ {
913
+ "epoch": 4.901960784313726,
914
+ "grad_norm": 10.1875,
915
+ "learning_rate": 3.9529411764705884e-07,
916
+ "loss": 0.610704460144043,
917
+ "step": 6250
918
+ },
919
+ {
920
+ "epoch": 4.9411764705882355,
921
+ "grad_norm": 8.5625,
922
+ "learning_rate": 2.3843137254901965e-07,
923
+ "loss": 0.6032954788208008,
924
+ "step": 6300
925
+ },
926
+ {
927
+ "epoch": 4.980392156862745,
928
+ "grad_norm": 8.5,
929
+ "learning_rate": 8.156862745098041e-08,
930
+ "loss": 0.6181810760498047,
931
+ "step": 6350
932
+ },
933
+ {
934
+ "epoch": 5.0,
935
+ "eval_loss": 0.6410982012748718,
936
+ "eval_runtime": 0.7957,
937
+ "eval_samples_per_second": 211.142,
938
+ "eval_steps_per_second": 3.77,
939
+ "step": 6375
940
+ }
941
+ ],
942
+ "logging_steps": 50,
943
+ "max_steps": 6375,
944
+ "num_input_tokens_seen": 0,
945
+ "num_train_epochs": 5,
946
+ "save_steps": 500,
947
+ "stateful_callbacks": {
948
+ "TrainerControl": {
949
+ "args": {
950
+ "should_epoch_stop": false,
951
+ "should_evaluate": false,
952
+ "should_log": false,
953
+ "should_save": true,
954
+ "should_training_stop": true
955
+ },
956
+ "attributes": {}
957
+ }
958
+ },
959
+ "total_flos": 0.0,
960
+ "train_batch_size": 32,
961
+ "trial_name": null,
962
+ "trial_params": null
963
+ }
cambridgeltl-SapBERT-from-PubMedBERT-fulltext-E5-Pairwise/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b05e02f1846d48c6fd0b0aafc33920646c6d6f4289fba46637a3c4584b077afd
3
+ size 5329