phanerozoic
commited on
Commit
•
cf88411
1
Parent(s):
86bd163
Upload 2 files
Browse files- config.json +27 -0
- runinfo.json +200 -0
config.json
ADDED
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "bert-base-uncased",
|
3 |
+
"architectures": [
|
4 |
+
"BertForSequenceClassification"
|
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,
|
11 |
+
"hidden_size": 768,
|
12 |
+
"initializer_range": 0.02,
|
13 |
+
"intermediate_size": 3072,
|
14 |
+
"layer_norm_eps": 1e-12,
|
15 |
+
"max_position_embeddings": 512,
|
16 |
+
"model_type": "bert",
|
17 |
+
"num_attention_heads": 12,
|
18 |
+
"num_hidden_layers": 12,
|
19 |
+
"pad_token_id": 0,
|
20 |
+
"position_embedding_type": "absolute",
|
21 |
+
"problem_type": "single_label_classification",
|
22 |
+
"torch_dtype": "float32",
|
23 |
+
"transformers_version": "4.39.3",
|
24 |
+
"type_vocab_size": 2,
|
25 |
+
"use_cache": true,
|
26 |
+
"vocab_size": 30522
|
27 |
+
}
|
runinfo.json
ADDED
@@ -0,0 +1,200 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_params": {
|
3 |
+
"epochs": 2,
|
4 |
+
"batch_size": 32,
|
5 |
+
"warmup_steps": 0,
|
6 |
+
"weight_decay": 0.01
|
7 |
+
},
|
8 |
+
"all_results": [
|
9 |
+
{
|
10 |
+
"params": {
|
11 |
+
"epochs": 2,
|
12 |
+
"batch_size": 32,
|
13 |
+
"warmup_steps": 0,
|
14 |
+
"weight_decay": 0.01
|
15 |
+
},
|
16 |
+
"eval_result": {
|
17 |
+
"eval_loss": 0.31092265248298645,
|
18 |
+
"eval_accuracy": 0.89128,
|
19 |
+
"eval_f1": 0.8922112944162437,
|
20 |
+
"eval_precision": 0.884633532557408,
|
21 |
+
"eval_recall": 0.89992,
|
22 |
+
"eval_runtime": 32.9216,
|
23 |
+
"eval_samples_per_second": 759.379,
|
24 |
+
"eval_steps_per_second": 23.753,
|
25 |
+
"epoch": 2.0
|
26 |
+
}
|
27 |
+
},
|
28 |
+
{
|
29 |
+
"params": {
|
30 |
+
"epochs": 2,
|
31 |
+
"batch_size": 16,
|
32 |
+
"warmup_steps": 100,
|
33 |
+
"weight_decay": 0.01
|
34 |
+
},
|
35 |
+
"eval_result": {
|
36 |
+
"eval_loss": 0.3510156273841858,
|
37 |
+
"eval_accuracy": 0.8912,
|
38 |
+
"eval_f1": 0.8914431673052363,
|
39 |
+
"eval_precision": 0.8894552405224594,
|
40 |
+
"eval_recall": 0.89344,
|
41 |
+
"eval_runtime": 34.3301,
|
42 |
+
"eval_samples_per_second": 728.223,
|
43 |
+
"eval_steps_per_second": 45.529,
|
44 |
+
"epoch": 2.0
|
45 |
+
}
|
46 |
+
},
|
47 |
+
{
|
48 |
+
"params": {
|
49 |
+
"epochs": 2,
|
50 |
+
"batch_size": 16,
|
51 |
+
"warmup_steps": 100,
|
52 |
+
"weight_decay": 0.01
|
53 |
+
},
|
54 |
+
"eval_result": {
|
55 |
+
"eval_loss": 0.3510156273841858,
|
56 |
+
"eval_accuracy": 0.8912,
|
57 |
+
"eval_f1": 0.8914431673052363,
|
58 |
+
"eval_precision": 0.8894552405224594,
|
59 |
+
"eval_recall": 0.89344,
|
60 |
+
"eval_runtime": 34.4751,
|
61 |
+
"eval_samples_per_second": 725.161,
|
62 |
+
"eval_steps_per_second": 45.337,
|
63 |
+
"epoch": 2.0
|
64 |
+
}
|
65 |
+
},
|
66 |
+
{
|
67 |
+
"params": {
|
68 |
+
"epochs": 2,
|
69 |
+
"batch_size": 16,
|
70 |
+
"warmup_steps": 100,
|
71 |
+
"weight_decay": 0.01
|
72 |
+
},
|
73 |
+
"eval_result": {
|
74 |
+
"eval_loss": 0.3510156273841858,
|
75 |
+
"eval_accuracy": 0.8912,
|
76 |
+
"eval_f1": 0.8914431673052363,
|
77 |
+
"eval_precision": 0.8894552405224594,
|
78 |
+
"eval_recall": 0.89344,
|
79 |
+
"eval_runtime": 34.5714,
|
80 |
+
"eval_samples_per_second": 723.14,
|
81 |
+
"eval_steps_per_second": 45.211,
|
82 |
+
"epoch": 2.0
|
83 |
+
}
|
84 |
+
},
|
85 |
+
{
|
86 |
+
"params": {
|
87 |
+
"epochs": 2,
|
88 |
+
"batch_size": 16,
|
89 |
+
"warmup_steps": 100,
|
90 |
+
"weight_decay": 0.01
|
91 |
+
},
|
92 |
+
"eval_result": {
|
93 |
+
"eval_loss": 0.3510156273841858,
|
94 |
+
"eval_accuracy": 0.8912,
|
95 |
+
"eval_f1": 0.8914431673052363,
|
96 |
+
"eval_precision": 0.8894552405224594,
|
97 |
+
"eval_recall": 0.89344,
|
98 |
+
"eval_runtime": 34.3958,
|
99 |
+
"eval_samples_per_second": 726.832,
|
100 |
+
"eval_steps_per_second": 45.442,
|
101 |
+
"epoch": 2.0
|
102 |
+
}
|
103 |
+
},
|
104 |
+
{
|
105 |
+
"params": {
|
106 |
+
"epochs": 2,
|
107 |
+
"batch_size": 16,
|
108 |
+
"warmup_steps": 100,
|
109 |
+
"weight_decay": 0.01
|
110 |
+
},
|
111 |
+
"eval_result": {
|
112 |
+
"eval_loss": 0.3510156273841858,
|
113 |
+
"eval_accuracy": 0.8912,
|
114 |
+
"eval_f1": 0.8914431673052363,
|
115 |
+
"eval_precision": 0.8894552405224594,
|
116 |
+
"eval_recall": 0.89344,
|
117 |
+
"eval_runtime": 34.3952,
|
118 |
+
"eval_samples_per_second": 726.845,
|
119 |
+
"eval_steps_per_second": 45.442,
|
120 |
+
"epoch": 2.0
|
121 |
+
}
|
122 |
+
},
|
123 |
+
{
|
124 |
+
"params": {
|
125 |
+
"epochs": 2,
|
126 |
+
"batch_size": 16,
|
127 |
+
"warmup_steps": 100,
|
128 |
+
"weight_decay": 0.01
|
129 |
+
},
|
130 |
+
"eval_result": {
|
131 |
+
"eval_loss": 0.3510156273841858,
|
132 |
+
"eval_accuracy": 0.8912,
|
133 |
+
"eval_f1": 0.8914431673052363,
|
134 |
+
"eval_precision": 0.8894552405224594,
|
135 |
+
"eval_recall": 0.89344,
|
136 |
+
"eval_runtime": 34.5864,
|
137 |
+
"eval_samples_per_second": 722.827,
|
138 |
+
"eval_steps_per_second": 45.191,
|
139 |
+
"epoch": 2.0
|
140 |
+
}
|
141 |
+
},
|
142 |
+
{
|
143 |
+
"params": {
|
144 |
+
"epochs": 2,
|
145 |
+
"batch_size": 16,
|
146 |
+
"warmup_steps": 100,
|
147 |
+
"weight_decay": 0.01
|
148 |
+
},
|
149 |
+
"eval_result": {
|
150 |
+
"eval_loss": 0.3510156273841858,
|
151 |
+
"eval_accuracy": 0.8912,
|
152 |
+
"eval_f1": 0.8914431673052363,
|
153 |
+
"eval_precision": 0.8894552405224594,
|
154 |
+
"eval_recall": 0.89344,
|
155 |
+
"eval_runtime": 34.4644,
|
156 |
+
"eval_samples_per_second": 725.386,
|
157 |
+
"eval_steps_per_second": 45.351,
|
158 |
+
"epoch": 2.0
|
159 |
+
}
|
160 |
+
},
|
161 |
+
{
|
162 |
+
"params": {
|
163 |
+
"epochs": 2,
|
164 |
+
"batch_size": 16,
|
165 |
+
"warmup_steps": 100,
|
166 |
+
"weight_decay": 0.01
|
167 |
+
},
|
168 |
+
"eval_result": {
|
169 |
+
"eval_loss": 0.3510156273841858,
|
170 |
+
"eval_accuracy": 0.8912,
|
171 |
+
"eval_f1": 0.8914431673052363,
|
172 |
+
"eval_precision": 0.8894552405224594,
|
173 |
+
"eval_recall": 0.89344,
|
174 |
+
"eval_runtime": 34.5881,
|
175 |
+
"eval_samples_per_second": 722.792,
|
176 |
+
"eval_steps_per_second": 45.189,
|
177 |
+
"epoch": 2.0
|
178 |
+
}
|
179 |
+
},
|
180 |
+
{
|
181 |
+
"params": {
|
182 |
+
"epochs": 2,
|
183 |
+
"batch_size": 16,
|
184 |
+
"warmup_steps": 100,
|
185 |
+
"weight_decay": 0.01
|
186 |
+
},
|
187 |
+
"eval_result": {
|
188 |
+
"eval_loss": 0.3510156273841858,
|
189 |
+
"eval_accuracy": 0.8912,
|
190 |
+
"eval_f1": 0.8914431673052363,
|
191 |
+
"eval_precision": 0.8894552405224594,
|
192 |
+
"eval_recall": 0.89344,
|
193 |
+
"eval_runtime": 34.3332,
|
194 |
+
"eval_samples_per_second": 728.157,
|
195 |
+
"eval_steps_per_second": 45.524,
|
196 |
+
"epoch": 2.0
|
197 |
+
}
|
198 |
+
}
|
199 |
+
]
|
200 |
+
}
|