Hugging Face's logo

huyydangg
/
DEk21_hcmute_embedding

Sentence Similarity
sentence-transformers
Safetensors
Vietnamese
roberta
feature-extraction
Generated from Trainer
dataset_size:80000
loss:MatryoshkaLoss
loss:MultipleNegativesRankingLoss
Eval Results
Inference Endpoints
Model card Files Community
Huy Dang commited on
Commit
a8e3e9e
·
verified ·
1 Parent(s): 71c33a3

Initial model upload

Browse files
Files changed (6) hide show
  1. model.safetensors +1 -1
  2. optimizer.pt +3 -0
  3. rng_state.pth +3 -0
  4. scheduler.pt +3 -0
  5. trainer_state.json +406 -0
  6. training_args.bin +1 -1
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:b9ab10dcf19985c72dbe1bff287d353f47fd2fda3ce267b8edbc4b4c5c351d30
3
  size 540015464
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1d71b98c7810f32cd761119e72b94f7efaebb9b9cdef0c12959c166256cdbd93
3
  size 540015464
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c4ebc28c2ebca4862776a9f89c3c8de972438cd172a22813b89b57db18dc44d2
3
+ size 1080155450
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:19059bd5ccf01b60d1ddae0c675995666bd984bcf457cfdcaf9516d7d8ff43fe
3
+ size 14244
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:37d2c3099ed2a7c9bdcdc0c3d1058ad81e0894112175f8523a80f541d102e381
3
+ size 1064
trainer_state.json ADDED
@@ -0,0 +1,406 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": 0.513288052386553,
3
+ "best_model_checkpoint": "bkai-fine-tuned-legal/checkpoint-240",
4
+ "epoch": 4.0,
5
+ "eval_steps": 500,
6
+ "global_step": 240,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 1.0,
13
+ "grad_norm": 3051630.25,
14
+ "learning_rate": 2.9970400926424075e-05,
15
+ "loss": 0.9663,
16
+ "step": 60
17
+ },
18
+ {
19
+ "epoch": 1.0,
20
+ "eval_dim_128_cosine_accuracy@1": 0.2080602426439827,
21
+ "eval_dim_128_cosine_accuracy@10": 0.6729884256031237,
22
+ "eval_dim_128_cosine_accuracy@3": 0.4049644401059824,
23
+ "eval_dim_128_cosine_accuracy@5": 0.49504950495049505,
24
+ "eval_dim_128_cosine_map@100": 0.34660918730461937,
25
+ "eval_dim_128_cosine_mrr@10": 0.3358910891089097,
26
+ "eval_dim_128_cosine_ndcg@10": 0.41500734332231626,
27
+ "eval_dim_128_cosine_precision@1": 0.2080602426439827,
28
+ "eval_dim_128_cosine_precision@10": 0.06729884256031238,
29
+ "eval_dim_128_cosine_precision@3": 0.13498814670199413,
30
+ "eval_dim_128_cosine_precision@5": 0.09900990099009901,
31
+ "eval_dim_128_cosine_recall@1": 0.2080602426439827,
32
+ "eval_dim_128_cosine_recall@10": 0.6729884256031237,
33
+ "eval_dim_128_cosine_recall@3": 0.4049644401059824,
34
+ "eval_dim_128_cosine_recall@5": 0.49504950495049505,
35
+ "eval_dim_256_cosine_accuracy@1": 0.21321991354064984,
36
+ "eval_dim_256_cosine_accuracy@10": 0.6862362292567284,
37
+ "eval_dim_256_cosine_accuracy@3": 0.41598103472319065,
38
+ "eval_dim_256_cosine_accuracy@5": 0.5063450006972529,
39
+ "eval_dim_256_cosine_map@100": 0.3546268888431031,
40
+ "eval_dim_256_cosine_mrr@10": 0.34420854057236694,
41
+ "eval_dim_256_cosine_ndcg@10": 0.424598027833536,
42
+ "eval_dim_256_cosine_precision@1": 0.21321991354064984,
43
+ "eval_dim_256_cosine_precision@10": 0.06862362292567285,
44
+ "eval_dim_256_cosine_precision@3": 0.1386603449077302,
45
+ "eval_dim_256_cosine_precision@5": 0.10126900013945056,
46
+ "eval_dim_256_cosine_recall@1": 0.21321991354064984,
47
+ "eval_dim_256_cosine_recall@10": 0.6862362292567284,
48
+ "eval_dim_256_cosine_recall@3": 0.41598103472319065,
49
+ "eval_dim_256_cosine_recall@5": 0.5063450006972529,
50
+ "eval_dim_512_cosine_accuracy@1": 0.21433551805884815,
51
+ "eval_dim_512_cosine_accuracy@10": 0.6919537024124948,
52
+ "eval_dim_512_cosine_accuracy@3": 0.4211407056198578,
53
+ "eval_dim_512_cosine_accuracy@5": 0.508855110863199,
54
+ "eval_dim_512_cosine_map@100": 0.3564257538175301,
55
+ "eval_dim_512_cosine_mrr@10": 0.34621192501543907,
56
+ "eval_dim_512_cosine_ndcg@10": 0.42739973393443165,
57
+ "eval_dim_512_cosine_precision@1": 0.21433551805884815,
58
+ "eval_dim_512_cosine_precision@10": 0.06919537024124948,
59
+ "eval_dim_512_cosine_precision@3": 0.14038023520661924,
60
+ "eval_dim_512_cosine_precision@5": 0.1017710221726398,
61
+ "eval_dim_512_cosine_recall@1": 0.21433551805884815,
62
+ "eval_dim_512_cosine_recall@10": 0.6919537024124948,
63
+ "eval_dim_512_cosine_recall@3": 0.4211407056198578,
64
+ "eval_dim_512_cosine_recall@5": 0.508855110863199,
65
+ "eval_dim_64_cosine_accuracy@1": 0.19704364802677451,
66
+ "eval_dim_64_cosine_accuracy@10": 0.6503974341096082,
67
+ "eval_dim_64_cosine_accuracy@3": 0.3846046576488635,
68
+ "eval_dim_64_cosine_accuracy@5": 0.48096499790824154,
69
+ "eval_dim_64_cosine_map@100": 0.3332997375022508,
70
+ "eval_dim_64_cosine_mrr@10": 0.3223062799237654,
71
+ "eval_dim_64_cosine_ndcg@10": 0.399385228043632,
72
+ "eval_dim_64_cosine_precision@1": 0.19704364802677451,
73
+ "eval_dim_64_cosine_precision@10": 0.06503974341096082,
74
+ "eval_dim_64_cosine_precision@3": 0.12820155254962115,
75
+ "eval_dim_64_cosine_precision@5": 0.0961929995816483,
76
+ "eval_dim_64_cosine_recall@1": 0.19704364802677451,
77
+ "eval_dim_64_cosine_recall@10": 0.6503974341096082,
78
+ "eval_dim_64_cosine_recall@3": 0.3846046576488635,
79
+ "eval_dim_64_cosine_recall@5": 0.48096499790824154,
80
+ "eval_dim_768_cosine_accuracy@1": 0.22116859573281272,
81
+ "eval_dim_768_cosine_accuracy@10": 0.6947427137079906,
82
+ "eval_dim_768_cosine_accuracy@3": 0.4196067494073351,
83
+ "eval_dim_768_cosine_accuracy@5": 0.5128991772416678,
84
+ "eval_dim_768_cosine_map@100": 0.3609325312259912,
85
+ "eval_dim_768_cosine_mrr@10": 0.3508835189353944,
86
+ "eval_dim_768_cosine_ndcg@10": 0.4315716703139609,
87
+ "eval_dim_768_cosine_precision@1": 0.22116859573281272,
88
+ "eval_dim_768_cosine_precision@10": 0.06947427137079905,
89
+ "eval_dim_768_cosine_precision@3": 0.13986891646911168,
90
+ "eval_dim_768_cosine_precision@5": 0.10257983544833356,
91
+ "eval_dim_768_cosine_recall@1": 0.22116859573281272,
92
+ "eval_dim_768_cosine_recall@10": 0.6947427137079906,
93
+ "eval_dim_768_cosine_recall@3": 0.4196067494073351,
94
+ "eval_dim_768_cosine_recall@5": 0.5128991772416678,
95
+ "eval_loss": 0.30127617716789246,
96
+ "eval_runtime": 4760.4689,
97
+ "eval_samples_per_second": 1.507,
98
+ "eval_sequential_score": 0.399385228043632,
99
+ "eval_steps_per_second": 0.019,
100
+ "step": 60
101
+ },
102
+ {
103
+ "epoch": 2.0,
104
+ "grad_norm": 1582530.375,
105
+ "learning_rate": 2.655769864163684e-05,
106
+ "loss": 0.2399,
107
+ "step": 120
108
+ },
109
+ {
110
+ "epoch": 2.0,
111
+ "eval_dim_128_cosine_accuracy@1": 0.3225491563240831,
112
+ "eval_dim_128_cosine_accuracy@10": 0.7071538139729466,
113
+ "eval_dim_128_cosine_accuracy@3": 0.46102356714544696,
114
+ "eval_dim_128_cosine_accuracy@5": 0.53772137777158,
115
+ "eval_dim_128_cosine_map@100": 0.43211454127792864,
116
+ "eval_dim_128_cosine_mrr@10": 0.42227551225947657,
117
+ "eval_dim_128_cosine_ndcg@10": 0.488553211966058,
118
+ "eval_dim_128_cosine_precision@1": 0.3225491563240831,
119
+ "eval_dim_128_cosine_precision@10": 0.07071538139729466,
120
+ "eval_dim_128_cosine_precision@3": 0.15367452238181564,
121
+ "eval_dim_128_cosine_precision@5": 0.107544275554316,
122
+ "eval_dim_128_cosine_recall@1": 0.3225491563240831,
123
+ "eval_dim_128_cosine_recall@10": 0.7071538139729466,
124
+ "eval_dim_128_cosine_recall@3": 0.46102356714544696,
125
+ "eval_dim_128_cosine_recall@5": 0.53772137777158,
126
+ "eval_dim_256_cosine_accuracy@1": 0.32854553060939895,
127
+ "eval_dim_256_cosine_accuracy@10": 0.7131501882582625,
128
+ "eval_dim_256_cosine_accuracy@3": 0.464928182959141,
129
+ "eval_dim_256_cosine_accuracy@5": 0.5421837958443732,
130
+ "eval_dim_256_cosine_map@100": 0.4374891918057777,
131
+ "eval_dim_256_cosine_mrr@10": 0.42779830799981505,
132
+ "eval_dim_256_cosine_ndcg@10": 0.49410729071963905,
133
+ "eval_dim_256_cosine_precision@1": 0.32854553060939895,
134
+ "eval_dim_256_cosine_precision@10": 0.07131501882582625,
135
+ "eval_dim_256_cosine_precision@3": 0.1549760609863803,
136
+ "eval_dim_256_cosine_precision@5": 0.10843675916887462,
137
+ "eval_dim_256_cosine_recall@1": 0.32854553060939895,
138
+ "eval_dim_256_cosine_recall@10": 0.7131501882582625,
139
+ "eval_dim_256_cosine_recall@3": 0.464928182959141,
140
+ "eval_dim_256_cosine_recall@5": 0.5421837958443732,
141
+ "eval_dim_512_cosine_accuracy@1": 0.33286849811741737,
142
+ "eval_dim_512_cosine_accuracy@10": 0.7142657927764607,
143
+ "eval_dim_512_cosine_accuracy@3": 0.46966950216148373,
144
+ "eval_dim_512_cosine_accuracy@5": 0.5445544554455446,
145
+ "eval_dim_512_cosine_map@100": 0.4410220731771345,
146
+ "eval_dim_512_cosine_mrr@10": 0.4314777000838923,
147
+ "eval_dim_512_cosine_ndcg@10": 0.4972638536904392,
148
+ "eval_dim_512_cosine_precision@1": 0.33286849811741737,
149
+ "eval_dim_512_cosine_precision@10": 0.07142657927764608,
150
+ "eval_dim_512_cosine_precision@3": 0.15655650072049457,
151
+ "eval_dim_512_cosine_precision@5": 0.10891089108910891,
152
+ "eval_dim_512_cosine_recall@1": 0.33286849811741737,
153
+ "eval_dim_512_cosine_recall@10": 0.7142657927764607,
154
+ "eval_dim_512_cosine_recall@3": 0.46966950216148373,
155
+ "eval_dim_512_cosine_recall@5": 0.5445544554455446,
156
+ "eval_dim_64_cosine_accuracy@1": 0.306233440245433,
157
+ "eval_dim_64_cosine_accuracy@10": 0.6906986473295217,
158
+ "eval_dim_64_cosine_accuracy@3": 0.44205829033607585,
159
+ "eval_dim_64_cosine_accuracy@5": 0.5208478594338307,
160
+ "eval_dim_64_cosine_map@100": 0.416048690892794,
161
+ "eval_dim_64_cosine_mrr@10": 0.4057003738603239,
162
+ "eval_dim_64_cosine_ndcg@10": 0.47195343698452114,
163
+ "eval_dim_64_cosine_precision@1": 0.306233440245433,
164
+ "eval_dim_64_cosine_precision@10": 0.06906986473295217,
165
+ "eval_dim_64_cosine_precision@3": 0.1473527634453586,
166
+ "eval_dim_64_cosine_precision@5": 0.10416957188676614,
167
+ "eval_dim_64_cosine_recall@1": 0.306233440245433,
168
+ "eval_dim_64_cosine_recall@10": 0.6906986473295217,
169
+ "eval_dim_64_cosine_recall@3": 0.44205829033607585,
170
+ "eval_dim_64_cosine_recall@5": 0.5208478594338307,
171
+ "eval_dim_768_cosine_accuracy@1": 0.33412355320039044,
172
+ "eval_dim_768_cosine_accuracy@10": 0.7152419467298843,
173
+ "eval_dim_768_cosine_accuracy@3": 0.47106400780923163,
174
+ "eval_dim_768_cosine_accuracy@5": 0.5448333565750941,
175
+ "eval_dim_768_cosine_map@100": 0.44207831994901653,
176
+ "eval_dim_768_cosine_mrr@10": 0.4325288363846451,
177
+ "eval_dim_768_cosine_ndcg@10": 0.49828039957821635,
178
+ "eval_dim_768_cosine_precision@1": 0.33412355320039044,
179
+ "eval_dim_768_cosine_precision@10": 0.07152419467298843,
180
+ "eval_dim_768_cosine_precision@3": 0.15702133593641054,
181
+ "eval_dim_768_cosine_precision@5": 0.10896667131501885,
182
+ "eval_dim_768_cosine_recall@1": 0.33412355320039044,
183
+ "eval_dim_768_cosine_recall@10": 0.7152419467298843,
184
+ "eval_dim_768_cosine_recall@3": 0.47106400780923163,
185
+ "eval_dim_768_cosine_recall@5": 0.5448333565750941,
186
+ "eval_loss": 0.18898583948612213,
187
+ "eval_runtime": 4757.4,
188
+ "eval_samples_per_second": 1.508,
189
+ "eval_sequential_score": 0.47195343698452114,
190
+ "eval_steps_per_second": 0.019,
191
+ "step": 120
192
+ },
193
+ {
194
+ "epoch": 3.0,
195
+ "grad_norm": 792357.3125,
196
+ "learning_rate": 1.8730348307472828e-05,
197
+ "loss": 0.1032,
198
+ "step": 180
199
+ },
200
+ {
201
+ "epoch": 3.0,
202
+ "eval_dim_128_cosine_accuracy@1": 0.3468135545948961,
203
+ "eval_dim_128_cosine_accuracy@10": 0.7173337052015061,
204
+ "eval_dim_128_cosine_accuracy@3": 0.46520708408869055,
205
+ "eval_dim_128_cosine_accuracy@5": 0.5407892901966252,
206
+ "eval_dim_128_cosine_map@100": 0.447677926706174,
207
+ "eval_dim_128_cosine_mrr@10": 0.43798207285517343,
208
+ "eval_dim_128_cosine_ndcg@10": 0.5026334575249692,
209
+ "eval_dim_128_cosine_precision@1": 0.3468135545948961,
210
+ "eval_dim_128_cosine_precision@10": 0.07173337052015061,
211
+ "eval_dim_128_cosine_precision@3": 0.15506902802956352,
212
+ "eval_dim_128_cosine_precision@5": 0.10815785803932505,
213
+ "eval_dim_128_cosine_recall@1": 0.3468135545948961,
214
+ "eval_dim_128_cosine_recall@10": 0.7173337052015061,
215
+ "eval_dim_128_cosine_recall@3": 0.46520708408869055,
216
+ "eval_dim_128_cosine_recall@5": 0.5407892901966252,
217
+ "eval_dim_256_cosine_accuracy@1": 0.35043926927904057,
218
+ "eval_dim_256_cosine_accuracy@10": 0.7156602984242086,
219
+ "eval_dim_256_cosine_accuracy@3": 0.4700878538558081,
220
+ "eval_dim_256_cosine_accuracy@5": 0.544972807139869,
221
+ "eval_dim_256_cosine_map@100": 0.4507378853512356,
222
+ "eval_dim_256_cosine_mrr@10": 0.44098468699988747,
223
+ "eval_dim_256_cosine_ndcg@10": 0.5046872072161374,
224
+ "eval_dim_256_cosine_precision@1": 0.35043926927904057,
225
+ "eval_dim_256_cosine_precision@10": 0.07156602984242087,
226
+ "eval_dim_256_cosine_precision@3": 0.15669595128526936,
227
+ "eval_dim_256_cosine_precision@5": 0.1089945614279738,
228
+ "eval_dim_256_cosine_recall@1": 0.35043926927904057,
229
+ "eval_dim_256_cosine_recall@10": 0.7156602984242086,
230
+ "eval_dim_256_cosine_recall@3": 0.4700878538558081,
231
+ "eval_dim_256_cosine_recall@5": 0.544972807139869,
232
+ "eval_dim_512_cosine_accuracy@1": 0.35169432436201364,
233
+ "eval_dim_512_cosine_accuracy@10": 0.7231906289220471,
234
+ "eval_dim_512_cosine_accuracy@3": 0.4700878538558081,
235
+ "eval_dim_512_cosine_accuracy@5": 0.5448333565750941,
236
+ "eval_dim_512_cosine_map@100": 0.45212818188900633,
237
+ "eval_dim_512_cosine_mrr@10": 0.44279272997722396,
238
+ "eval_dim_512_cosine_ndcg@10": 0.5076751608428753,
239
+ "eval_dim_512_cosine_precision@1": 0.35169432436201364,
240
+ "eval_dim_512_cosine_precision@10": 0.07231906289220472,
241
+ "eval_dim_512_cosine_precision@3": 0.15669595128526936,
242
+ "eval_dim_512_cosine_precision@5": 0.10896667131501885,
243
+ "eval_dim_512_cosine_recall@1": 0.35169432436201364,
244
+ "eval_dim_512_cosine_recall@10": 0.7231906289220471,
245
+ "eval_dim_512_cosine_recall@3": 0.4700878538558081,
246
+ "eval_dim_512_cosine_recall@5": 0.5448333565750941,
247
+ "eval_dim_64_cosine_accuracy@1": 0.3356575094129131,
248
+ "eval_dim_64_cosine_accuracy@10": 0.6976711755682611,
249
+ "eval_dim_64_cosine_accuracy@3": 0.4523776321294101,
250
+ "eval_dim_64_cosine_accuracy@5": 0.5307488495328406,
251
+ "eval_dim_64_cosine_map@100": 0.43559811808588345,
252
+ "eval_dim_64_cosine_mrr@10": 0.42538852476797007,
253
+ "eval_dim_64_cosine_ndcg@10": 0.48852868771634256,
254
+ "eval_dim_64_cosine_precision@1": 0.3356575094129131,
255
+ "eval_dim_64_cosine_precision@10": 0.0697671175568261,
256
+ "eval_dim_64_cosine_precision@3": 0.1507925440431367,
257
+ "eval_dim_64_cosine_precision@5": 0.10614976990656813,
258
+ "eval_dim_64_cosine_recall@1": 0.3356575094129131,
259
+ "eval_dim_64_cosine_recall@10": 0.6976711755682611,
260
+ "eval_dim_64_cosine_recall@3": 0.4523776321294101,
261
+ "eval_dim_64_cosine_recall@5": 0.5307488495328406,
262
+ "eval_dim_768_cosine_accuracy@1": 0.3568539952586808,
263
+ "eval_dim_768_cosine_accuracy@10": 0.7237484311811463,
264
+ "eval_dim_768_cosine_accuracy@3": 0.473434667410403,
265
+ "eval_dim_768_cosine_accuracy@5": 0.5530609398968066,
266
+ "eval_dim_768_cosine_map@100": 0.456489701012935,
267
+ "eval_dim_768_cosine_mrr@10": 0.44698465822437466,
268
+ "eval_dim_768_cosine_ndcg@10": 0.5110884733352402,
269
+ "eval_dim_768_cosine_precision@1": 0.3568539952586808,
270
+ "eval_dim_768_cosine_precision@10": 0.07237484311811462,
271
+ "eval_dim_768_cosine_precision@3": 0.15781155580346765,
272
+ "eval_dim_768_cosine_precision@5": 0.11061218797936133,
273
+ "eval_dim_768_cosine_recall@1": 0.3568539952586808,
274
+ "eval_dim_768_cosine_recall@10": 0.7237484311811463,
275
+ "eval_dim_768_cosine_recall@3": 0.473434667410403,
276
+ "eval_dim_768_cosine_recall@5": 0.5530609398968066,
277
+ "eval_loss": 0.15947215259075165,
278
+ "eval_runtime": 4772.3881,
279
+ "eval_samples_per_second": 1.503,
280
+ "eval_sequential_score": 0.48852868771634256,
281
+ "eval_steps_per_second": 0.019,
282
+ "step": 180
283
+ },
284
+ {
285
+ "epoch": 4.0,
286
+ "grad_norm": 528732.6875,
287
+ "learning_rate": 9.478131709729831e-06,
288
+ "loss": 0.0562,
289
+ "step": 240
290
+ },
291
+ {
292
+ "epoch": 4.0,
293
+ "eval_dim_128_cosine_accuracy@1": 0.35016036814949103,
294
+ "eval_dim_128_cosine_accuracy@10": 0.71803095802538,
295
+ "eval_dim_128_cosine_accuracy@3": 0.466880490865988,
296
+ "eval_dim_128_cosine_accuracy@5": 0.5453911588341933,
297
+ "eval_dim_128_cosine_map@100": 0.45032612042753656,
298
+ "eval_dim_128_cosine_mrr@10": 0.44050070500007754,
299
+ "eval_dim_128_cosine_ndcg@10": 0.5047593926261945,
300
+ "eval_dim_128_cosine_precision@1": 0.35016036814949103,
301
+ "eval_dim_128_cosine_precision@10": 0.07180309580253802,
302
+ "eval_dim_128_cosine_precision@3": 0.15562683028866267,
303
+ "eval_dim_128_cosine_precision@5": 0.10907823176683866,
304
+ "eval_dim_128_cosine_recall@1": 0.35016036814949103,
305
+ "eval_dim_128_cosine_recall@10": 0.71803095802538,
306
+ "eval_dim_128_cosine_recall@3": 0.466880490865988,
307
+ "eval_dim_128_cosine_recall@5": 0.5453911588341933,
308
+ "eval_dim_256_cosine_accuracy@1": 0.34932366476084226,
309
+ "eval_dim_256_cosine_accuracy@10": 0.7227722772277227,
310
+ "eval_dim_256_cosine_accuracy@3": 0.46660158973643845,
311
+ "eval_dim_256_cosine_accuracy@5": 0.5445544554455446,
312
+ "eval_dim_256_cosine_map@100": 0.4503116262429447,
313
+ "eval_dim_256_cosine_mrr@10": 0.44069499505282583,
314
+ "eval_dim_256_cosine_ndcg@10": 0.5059436181895386,
315
+ "eval_dim_256_cosine_precision@1": 0.34932366476084226,
316
+ "eval_dim_256_cosine_precision@10": 0.07227722772277227,
317
+ "eval_dim_256_cosine_precision@3": 0.15553386324547946,
318
+ "eval_dim_256_cosine_precision@5": 0.10891089108910891,
319
+ "eval_dim_256_cosine_recall@1": 0.34932366476084226,
320
+ "eval_dim_256_cosine_recall@10": 0.7227722772277227,
321
+ "eval_dim_256_cosine_recall@3": 0.46660158973643845,
322
+ "eval_dim_256_cosine_recall@5": 0.5445544554455446,
323
+ "eval_dim_512_cosine_accuracy@1": 0.3550411379166086,
324
+ "eval_dim_512_cosine_accuracy@10": 0.7234695300515968,
325
+ "eval_dim_512_cosine_accuracy@3": 0.4693906010319342,
326
+ "eval_dim_512_cosine_accuracy@5": 0.5484590712592387,
327
+ "eval_dim_512_cosine_map@100": 0.45428241375668854,
328
+ "eval_dim_512_cosine_mrr@10": 0.4447460339595335,
329
+ "eval_dim_512_cosine_ndcg@10": 0.5091968606671677,
330
+ "eval_dim_512_cosine_precision@1": 0.3550411379166086,
331
+ "eval_dim_512_cosine_precision@10": 0.07234695300515968,
332
+ "eval_dim_512_cosine_precision@3": 0.1564635336773114,
333
+ "eval_dim_512_cosine_precision@5": 0.10969181425184772,
334
+ "eval_dim_512_cosine_recall@1": 0.3550411379166086,
335
+ "eval_dim_512_cosine_recall@10": 0.7234695300515968,
336
+ "eval_dim_512_cosine_recall@3": 0.4693906010319342,
337
+ "eval_dim_512_cosine_recall@5": 0.5484590712592387,
338
+ "eval_dim_64_cosine_accuracy@1": 0.33760981731976014,
339
+ "eval_dim_64_cosine_accuracy@10": 0.7082694184911449,
340
+ "eval_dim_64_cosine_accuracy@3": 0.45349323664760843,
341
+ "eval_dim_64_cosine_accuracy@5": 0.5308883000976153,
342
+ "eval_dim_64_cosine_map@100": 0.4383018870756752,
343
+ "eval_dim_64_cosine_mrr@10": 0.4285745828103948,
344
+ "eval_dim_64_cosine_ndcg@10": 0.49330694449266793,
345
+ "eval_dim_64_cosine_precision@1": 0.33760981731976014,
346
+ "eval_dim_64_cosine_precision@10": 0.07082694184911449,
347
+ "eval_dim_64_cosine_precision@3": 0.1511644122158695,
348
+ "eval_dim_64_cosine_precision@5": 0.10617766001952307,
349
+ "eval_dim_64_cosine_recall@1": 0.33760981731976014,
350
+ "eval_dim_64_cosine_recall@10": 0.7082694184911449,
351
+ "eval_dim_64_cosine_recall@3": 0.45349323664760843,
352
+ "eval_dim_64_cosine_recall@5": 0.5308883000976153,
353
+ "eval_dim_768_cosine_accuracy@1": 0.36061916050760007,
354
+ "eval_dim_768_cosine_accuracy@10": 0.7238878817459211,
355
+ "eval_dim_768_cosine_accuracy@3": 0.478036536047971,
356
+ "eval_dim_768_cosine_accuracy@5": 0.5529214893320318,
357
+ "eval_dim_768_cosine_map@100": 0.45954079808905995,
358
+ "eval_dim_768_cosine_mrr@10": 0.44983946583793183,
359
+ "eval_dim_768_cosine_ndcg@10": 0.513288052386553,
360
+ "eval_dim_768_cosine_precision@1": 0.36061916050760007,
361
+ "eval_dim_768_cosine_precision@10": 0.07238878817459211,
362
+ "eval_dim_768_cosine_precision@3": 0.15934551201599031,
363
+ "eval_dim_768_cosine_precision@5": 0.11058429786640636,
364
+ "eval_dim_768_cosine_recall@1": 0.36061916050760007,
365
+ "eval_dim_768_cosine_recall@10": 0.7238878817459211,
366
+ "eval_dim_768_cosine_recall@3": 0.478036536047971,
367
+ "eval_dim_768_cosine_recall@5": 0.5529214893320318,
368
+ "eval_loss": 0.14603222906589508,
369
+ "eval_runtime": 4765.1388,
370
+ "eval_samples_per_second": 1.505,
371
+ "eval_sequential_score": 0.49330694449266793,
372
+ "eval_steps_per_second": 0.019,
373
+ "step": 240
374
+ }
375
+ ],
376
+ "logging_steps": 500,
377
+ "max_steps": 354,
378
+ "num_input_tokens_seen": 0,
379
+ "num_train_epochs": 6,
380
+ "save_steps": 500,
381
+ "stateful_callbacks": {
382
+ "EarlyStoppingCallback": {
383
+ "args": {
384
+ "early_stopping_patience": 1,
385
+ "early_stopping_threshold": 0.001
386
+ },
387
+ "attributes": {
388
+ "early_stopping_patience_counter": 0
389
+ }
390
+ },
391
+ "TrainerControl": {
392
+ "args": {
393
+ "should_epoch_stop": false,
394
+ "should_evaluate": false,
395
+ "should_log": false,
396
+ "should_save": true,
397
+ "should_training_stop": false
398
+ },
399
+ "attributes": {}
400
+ }
401
+ },
402
+ "total_flos": 0.0,
403
+ "train_batch_size": 80,
404
+ "trial_name": null,
405
+ "trial_params": null
406
+ }
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d66b1220bbab1fa5f6670d23d913b897cfbc875c08c297bf537703298e10ceba
3
  size 5624
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a1704bcfcfd1965a4c663d038235420f54fec65277e92be8502fddf96743aea5
3
  size 5624