File size: 2,680 Bytes
c090ba2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
Writing logs to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/bert-base-uncased-glue:rte-2020-06-29-14:24/log.txt.
Loading nlp dataset glue, subset rte, split train.
Loading nlp dataset glue, subset rte, split validation.
Loaded dataset. Found: 2 labels: ([0, 1])
Loading transformers AutoModelForSequenceClassification: bert-base-uncased
Tokenizing training data. (len: 2490)
Tokenizing eval data (len: 277)
Loaded data and tokenized in 14.295648097991943s
Training model across 1 GPUs
***** Running training *****
	Num examples = 2490
	Batch size = 128
	Max sequence length = 128
	Num steps = 95
	Num epochs = 5
	Learning rate = 3e-05
Failed to predict with model <class 'transformers.modeling_bert.BertForSequenceClassification'>. Check tokenizer configuration.
Writing logs to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/bert-base-uncased-glue:rte-2020-06-29-14:24/log.txt.
Loading nlp dataset glue, subset rte, split train.
Loading nlp dataset glue, subset rte, split validation.
Loaded dataset. Found: 2 labels: ([0, 1])
Loading transformers AutoModelForSequenceClassification: bert-base-uncased
Tokenizing training data. (len: 2490)
Tokenizing eval data (len: 277)
Loaded data and tokenized in 13.395596742630005s
Training model across 1 GPUs
***** Running training *****
	Num examples = 2490
	Batch size = 8
	Max sequence length = 128
	Num steps = 1555
	Num epochs = 5
	Learning rate = 2e-05
Eval accuracy: 68.23104693140795%
Best acc found. Saved model to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/bert-base-uncased-glue:rte-2020-06-29-14:24/.
Eval accuracy: 70.03610108303249%
Best acc found. Saved model to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/bert-base-uncased-glue:rte-2020-06-29-14:24/.
Eval accuracy: 72.56317689530685%
Best acc found. Saved model to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/bert-base-uncased-glue:rte-2020-06-29-14:24/.
Eval accuracy: 69.67509025270758%
Eval accuracy: 69.31407942238266%
Saved tokenizer <textattack.models.tokenizers.auto_tokenizer.AutoTokenizer object at 0x7fc688911c10> to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/bert-base-uncased-glue:rte-2020-06-29-14:24/.
Wrote README to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/bert-base-uncased-glue:rte-2020-06-29-14:24/README.md.
Wrote training args to /p/qdata/jm8wx/research/text_attacks/textattack/outputs/training/bert-base-uncased-glue:rte-2020-06-29-14:24/train_args.json.