StevenLimcorn
commited on
Commit
·
f13ec5c
1
Parent(s):
4457efe
Upload 9 files
Browse files- config.json +3 -4
- optimizer.pt +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +1 -0
- tokenizer_config.json +1 -0
- train_results.txt +3 -0
- trainer_state.json +206 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
config.json
CHANGED
@@ -1,10 +1,9 @@
|
|
1 |
{
|
2 |
-
"_name_or_path": "/content/
|
3 |
"architectures": [
|
4 |
"BertModel"
|
5 |
],
|
6 |
"attention_probs_dropout_prob": 0.1,
|
7 |
-
"classifier_dropout": null,
|
8 |
"gradient_checkpointing": false,
|
9 |
"hidden_act": "gelu",
|
10 |
"hidden_dropout_prob": 0.1,
|
@@ -18,8 +17,8 @@
|
|
18 |
"num_hidden_layers": 12,
|
19 |
"pad_token_id": 0,
|
20 |
"position_embedding_type": "absolute",
|
21 |
-
"transformers_version": "4.
|
22 |
"type_vocab_size": 2,
|
23 |
"use_cache": true,
|
24 |
"vocab_size": 30522
|
25 |
-
}
|
|
|
1 |
{
|
2 |
+
"_name_or_path": "/content/result/semeval-unsup-promcse-bert-base-uncased-facebook-election-ads",
|
3 |
"architectures": [
|
4 |
"BertModel"
|
5 |
],
|
6 |
"attention_probs_dropout_prob": 0.1,
|
|
|
7 |
"gradient_checkpointing": false,
|
8 |
"hidden_act": "gelu",
|
9 |
"hidden_dropout_prob": 0.1,
|
|
|
17 |
"num_hidden_layers": 12,
|
18 |
"pad_token_id": 0,
|
19 |
"position_embedding_type": "absolute",
|
20 |
+
"transformers_version": "4.2.1",
|
21 |
"type_vocab_size": 2,
|
22 |
"use_cache": true,
|
23 |
"vocab_size": 30522
|
24 |
+
}
|
optimizer.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:a1911fd6492a6026302f75b88ec98c15cdaa2dff8298163aa147c822bb25c874
|
3 |
+
size 7086823
|
scheduler.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:373effb463578c5eb8cd7c44b95a93c749d62c339de65f01c9ff11917f09e779
|
3 |
+
size 627
|
special_tokens_map.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]"}
|
tokenizer_config.json
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
{"do_lower_case": true, "unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "tokenize_chinese_chars": true, "strip_accents": null, "special_tokens_map_file": null, "name_or_path": "YuxinJiang/unsup-promcse-bert-base-uncased", "do_basic_tokenize": true, "never_split": null}
|
train_results.txt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
epoch = 5.0
|
2 |
+
train_runtime = 1423.6449
|
3 |
+
train_samples_per_second = 2.002
|
trainer_state.json
ADDED
@@ -0,0 +1,206 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": 0.8241261505887862,
|
3 |
+
"best_model_checkpoint": "/content/result/semeval-unsup-promcse-bert-base-uncased-facebook-election-ads",
|
4 |
+
"epoch": 5.0,
|
5 |
+
"global_step": 2850,
|
6 |
+
"is_hyper_param_search": false,
|
7 |
+
"is_local_process_zero": true,
|
8 |
+
"is_world_process_zero": true,
|
9 |
+
"log_history": [
|
10 |
+
{
|
11 |
+
"epoch": 0.22,
|
12 |
+
"eval_avg_sts": 0.7606087399209226,
|
13 |
+
"eval_sickr_spearman": 0.7073388050199292,
|
14 |
+
"eval_stsb_spearman": 0.813878674821916,
|
15 |
+
"step": 125
|
16 |
+
},
|
17 |
+
{
|
18 |
+
"epoch": 0.44,
|
19 |
+
"eval_avg_sts": 0.7746676184109096,
|
20 |
+
"eval_sickr_spearman": 0.7274445760508217,
|
21 |
+
"eval_stsb_spearman": 0.8218906607709974,
|
22 |
+
"step": 250
|
23 |
+
},
|
24 |
+
{
|
25 |
+
"epoch": 0.66,
|
26 |
+
"eval_avg_sts": 0.7795345076215956,
|
27 |
+
"eval_sickr_spearman": 0.7431172685433821,
|
28 |
+
"eval_stsb_spearman": 0.8159517466998092,
|
29 |
+
"step": 375
|
30 |
+
},
|
31 |
+
{
|
32 |
+
"epoch": 0.88,
|
33 |
+
"learning_rate": 0.024736842105263158,
|
34 |
+
"loss": 0.0019,
|
35 |
+
"step": 500
|
36 |
+
},
|
37 |
+
{
|
38 |
+
"epoch": 0.88,
|
39 |
+
"eval_avg_sts": 0.7790991209780169,
|
40 |
+
"eval_sickr_spearman": 0.7346580309581018,
|
41 |
+
"eval_stsb_spearman": 0.8235402109979321,
|
42 |
+
"step": 500
|
43 |
+
},
|
44 |
+
{
|
45 |
+
"epoch": 1.1,
|
46 |
+
"eval_avg_sts": 0.7768726280130896,
|
47 |
+
"eval_sickr_spearman": 0.7321121424263746,
|
48 |
+
"eval_stsb_spearman": 0.8216331135998047,
|
49 |
+
"step": 625
|
50 |
+
},
|
51 |
+
{
|
52 |
+
"epoch": 1.32,
|
53 |
+
"eval_avg_sts": 0.7580020019953406,
|
54 |
+
"eval_sickr_spearman": 0.7048930613347005,
|
55 |
+
"eval_stsb_spearman": 0.8111109426559808,
|
56 |
+
"step": 750
|
57 |
+
},
|
58 |
+
{
|
59 |
+
"epoch": 1.54,
|
60 |
+
"eval_avg_sts": 0.7643731301566361,
|
61 |
+
"eval_sickr_spearman": 0.7068401941559002,
|
62 |
+
"eval_stsb_spearman": 0.8219060661573719,
|
63 |
+
"step": 875
|
64 |
+
},
|
65 |
+
{
|
66 |
+
"epoch": 1.75,
|
67 |
+
"learning_rate": 0.019473684210526317,
|
68 |
+
"loss": 0.0023,
|
69 |
+
"step": 1000
|
70 |
+
},
|
71 |
+
{
|
72 |
+
"epoch": 1.75,
|
73 |
+
"eval_avg_sts": 0.7592660300891461,
|
74 |
+
"eval_sickr_spearman": 0.7015394818011901,
|
75 |
+
"eval_stsb_spearman": 0.816992578377102,
|
76 |
+
"step": 1000
|
77 |
+
},
|
78 |
+
{
|
79 |
+
"epoch": 1.97,
|
80 |
+
"eval_avg_sts": 0.7638814290387972,
|
81 |
+
"eval_sickr_spearman": 0.7100854635556677,
|
82 |
+
"eval_stsb_spearman": 0.8176773945219268,
|
83 |
+
"step": 1125
|
84 |
+
},
|
85 |
+
{
|
86 |
+
"epoch": 2.19,
|
87 |
+
"eval_avg_sts": 0.761929973436454,
|
88 |
+
"eval_sickr_spearman": 0.7092494822351373,
|
89 |
+
"eval_stsb_spearman": 0.8146104646377706,
|
90 |
+
"step": 1250
|
91 |
+
},
|
92 |
+
{
|
93 |
+
"epoch": 2.41,
|
94 |
+
"eval_avg_sts": 0.7585275538448379,
|
95 |
+
"eval_sickr_spearman": 0.707602783953433,
|
96 |
+
"eval_stsb_spearman": 0.809452323736243,
|
97 |
+
"step": 1375
|
98 |
+
},
|
99 |
+
{
|
100 |
+
"epoch": 2.63,
|
101 |
+
"learning_rate": 0.014210526315789472,
|
102 |
+
"loss": 0.0014,
|
103 |
+
"step": 1500
|
104 |
+
},
|
105 |
+
{
|
106 |
+
"epoch": 2.63,
|
107 |
+
"eval_avg_sts": 0.756410356894333,
|
108 |
+
"eval_sickr_spearman": 0.6931919645578548,
|
109 |
+
"eval_stsb_spearman": 0.8196287492308111,
|
110 |
+
"step": 1500
|
111 |
+
},
|
112 |
+
{
|
113 |
+
"epoch": 2.85,
|
114 |
+
"eval_avg_sts": 0.7620477624576579,
|
115 |
+
"eval_sickr_spearman": 0.7012191143545978,
|
116 |
+
"eval_stsb_spearman": 0.822876410560718,
|
117 |
+
"step": 1625
|
118 |
+
},
|
119 |
+
{
|
120 |
+
"epoch": 3.07,
|
121 |
+
"eval_avg_sts": 0.7608070157867401,
|
122 |
+
"eval_sickr_spearman": 0.7019747876735236,
|
123 |
+
"eval_stsb_spearman": 0.8196392438999567,
|
124 |
+
"step": 1750
|
125 |
+
},
|
126 |
+
{
|
127 |
+
"epoch": 3.29,
|
128 |
+
"eval_avg_sts": 0.7625034782535425,
|
129 |
+
"eval_sickr_spearman": 0.703315191621357,
|
130 |
+
"eval_stsb_spearman": 0.821691764885728,
|
131 |
+
"step": 1875
|
132 |
+
},
|
133 |
+
{
|
134 |
+
"epoch": 3.51,
|
135 |
+
"learning_rate": 0.008947368421052631,
|
136 |
+
"loss": 0.001,
|
137 |
+
"step": 2000
|
138 |
+
},
|
139 |
+
{
|
140 |
+
"epoch": 3.51,
|
141 |
+
"eval_avg_sts": 0.7631369232723644,
|
142 |
+
"eval_sickr_spearman": 0.7082736343472192,
|
143 |
+
"eval_stsb_spearman": 0.8180002121975096,
|
144 |
+
"step": 2000
|
145 |
+
},
|
146 |
+
{
|
147 |
+
"epoch": 3.73,
|
148 |
+
"eval_avg_sts": 0.7620440360304972,
|
149 |
+
"eval_sickr_spearman": 0.7117528082775756,
|
150 |
+
"eval_stsb_spearman": 0.8123352637834187,
|
151 |
+
"step": 2125
|
152 |
+
},
|
153 |
+
{
|
154 |
+
"epoch": 3.95,
|
155 |
+
"eval_avg_sts": 0.7686269781009907,
|
156 |
+
"eval_sickr_spearman": 0.7162074597138681,
|
157 |
+
"eval_stsb_spearman": 0.8210464964881133,
|
158 |
+
"step": 2250
|
159 |
+
},
|
160 |
+
{
|
161 |
+
"epoch": 4.17,
|
162 |
+
"eval_avg_sts": 0.7701487047832861,
|
163 |
+
"eval_sickr_spearman": 0.7184914826805546,
|
164 |
+
"eval_stsb_spearman": 0.8218059268860176,
|
165 |
+
"step": 2375
|
166 |
+
},
|
167 |
+
{
|
168 |
+
"epoch": 4.39,
|
169 |
+
"learning_rate": 0.003684210526315789,
|
170 |
+
"loss": 0.0009,
|
171 |
+
"step": 2500
|
172 |
+
},
|
173 |
+
{
|
174 |
+
"epoch": 4.39,
|
175 |
+
"eval_avg_sts": 0.7700311247392564,
|
176 |
+
"eval_sickr_spearman": 0.7182512311111612,
|
177 |
+
"eval_stsb_spearman": 0.8218110183673516,
|
178 |
+
"step": 2500
|
179 |
+
},
|
180 |
+
{
|
181 |
+
"epoch": 4.61,
|
182 |
+
"eval_avg_sts": 0.7702704869245434,
|
183 |
+
"eval_sickr_spearman": 0.7169341702786267,
|
184 |
+
"eval_stsb_spearman": 0.8236068035704601,
|
185 |
+
"step": 2625
|
186 |
+
},
|
187 |
+
{
|
188 |
+
"epoch": 4.82,
|
189 |
+
"eval_avg_sts": 0.7706318903065524,
|
190 |
+
"eval_sickr_spearman": 0.7171376300243186,
|
191 |
+
"eval_stsb_spearman": 0.8241261505887862,
|
192 |
+
"step": 2750
|
193 |
+
},
|
194 |
+
{
|
195 |
+
"epoch": 5.0,
|
196 |
+
"step": 2850,
|
197 |
+
"train_runtime": 1423.6449,
|
198 |
+
"train_samples_per_second": 2.002
|
199 |
+
}
|
200 |
+
],
|
201 |
+
"max_steps": 2850,
|
202 |
+
"num_train_epochs": 5,
|
203 |
+
"total_flos": 3840268242124800,
|
204 |
+
"trial_name": null,
|
205 |
+
"trial_params": null
|
206 |
+
}
|
training_args.bin
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:11aaa53bf3f00632105b3a88719e9db8fb27de4693d7809b2ee4f962a42ef004
|
3 |
+
size 2107
|
vocab.txt
ADDED
The diff for this file is too large to render.
See raw diff
|
|