OpeNTF / param.py
MarcoKurepa's picture
Upload param.py
a1b3d87
import random
import torch
import numpy as np
random.seed(0)
torch.manual_seed(0)
torch.cuda.manual_seed_all(0)
np.random.seed(0)
settings = {
'model':{
'baseline': {
'random': {
'b': 128
},
'fnn':{
'l': [128], # list of number of nodes in each layer
'lr': 0.1, # learning rate
'b': 128, # batch size
'e': 20, # epoch
'nns': 3, # number of negative samples
'ns': 'unigram_b', # 'uniform', 'unigram', 'unigram_b'
},
'bnn':{
'l': [128], # list of number of nodes in each layer
'lr': 0.1, # learning rate
'b': 128, # batch size
'e': 20, # epoch
'nns': 3, # number of negative samples
'ns': 'unigram_b', # 'uniform', 'unigram', 'unigram_b'
's': 1 # # sample_elbo for bnn
},
'nmt': {
'base_config': './mdl/nmt_config.yaml'
},
'caser': {},
'rrn': {
'with_zero': True
},
'emb':{
'd': 100,# embedding dimension
'e': 100,# epoch
'dm': 1,# training algorithm. 1: distributed memory (PV-DM), 0: distributed bag of words (PV-DBOW)
'w': 1 #cooccurrence window
}
},
'cmd': ['train', 'test', 'eval', 'fair'], # 'train', 'test', 'eval', 'plot', 'agg', 'adila'
'nfolds': 5,
'train_test_split': 0.85,
'step_ahead': 2,#for now, it means that whatever are in the last [step_ahead] time interval will be the test set!
},
'data':{
'domain': {
'dblp':{},
'uspt':{},
'imdb':{},
},
'location_type': 'country', #should be one of 'city', 'state', 'country' and represents the location of members in teams (not the location of teams)
'filter': {
'min_nteam': 75,
'min_team_size': 3,
},
'parallel': 1,
'ncore': 0,# <= 0 for all
'bucket_size': 500
},
'fair': {'np_ratio': None,
'fairness': ['det_greedy',],
'k_max': None,
'fairness_metrics': {'ndkl'},
'utility_metrics': {'map_cut_2,5,10'},
'eq_op': False,
'mode': 0,
'core': -1,
'attribute': ['gender', 'popularity']},
}