Hinglish/sweep.yaml at master · NirantK/Hinglish

History

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

method: bayes

metric:

goal: minimize

name: Valid F1

parameters:

adam_epsilon:

distribution: uniform

max: 0.0001944

min: 2.045e-05

attention_probs_dropout_prob:

distribution: uniform

max: 1.6

min: 0.2

batch_size:

distribution: int_uniform

max: 32

min: 2

epochs:

distribution: int_uniform

max: 6

min: 2

hidden_dropout_prob:

distribution: uniform

max: 1.2

min: 0.1

learning_rate:

distribution: uniform

max: 0.0001026

min: 1.51e-05

model_name:

distribution: categorical

values:

- distilbert

- bert

program: train.py

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

sweep.yaml

sweep.yaml

Files

sweep.yaml

Latest commit

History

sweep.yaml

File metadata and controls