This repository has been archived by the owner on May 16, 2022. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 20
/
config.py
64 lines (53 loc) · 1.76 KB
/
config.py
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
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
from typing import List, Tuple
from dataclasses import dataclass, field
from omegaconf import OmegaConf
@dataclass
class DatasetConfig:
data_path:str='datasets/ml-100k/u.data'
dataname:str='ml-1m'
separator:str='\t'
binarize_threshold:float=0.0
implicit:bool=True
min_item_per_user:int=10
min_user_per_item:int=1
protocol:str='holdout' # holdout, leave_one_out
generalization:str='weak' # weak/strong
holdout_users:int=600
valid_ratio:float=0.1
test_ratio:float=0.2
leave_k:int=1
split_random:bool=True
@dataclass
class EvaluatorConfig:
ks:List[int] = field(default_factory=lambda: [5])
@dataclass
class EarlyStopConfig:
early_stop:int=25
early_stop_measure:str='NDCG@10'
@dataclass
class ExperimentConfig:
debug:bool=False
save_dir:str='saves'
num_epochs:int=10
batch_size:int=256
verbose:int=0
print_step:int=1
test_step:int=1
test_from:int=1
model_name:str='EASE'
num_exp:int=5
seed:int=2020
gpu:int=0
def load_config():
dataset_config = OmegaConf.structured({'dataset' :DatasetConfig})
evaluator_config = OmegaConf.structured({'evaluator': EvaluatorConfig})
early_stop_config = OmegaConf.structured({'early_stop': EarlyStopConfig})
experiment_config = OmegaConf.structured({'experiment': ExperimentConfig})
model_name = experiment_config.experiment.model_name
# model_config = OmegaConf.structured({'hparams': OmegaConf.load(f"conf/{model_name}.yaml")})
model_config = OmegaConf.structured(OmegaConf.load(f"conf/{model_name}.yaml"))
config = OmegaConf.merge(dataset_config, evaluator_config, early_stop_config, experiment_config, model_config)
return config
if __name__ == '__main__':
config = load_config()
print(config)