def var_coat(alg_kwargs, var_all):
alg_kwargs = config.dictify(alg_kwargs)
alg_kwargs['n_epochs'] = n_epochs
alg_kwargs['var_all'] = var_all
kwargs_str = config.stringify(alg_kwargs)
kwargs_file = path.join(curve_dir, 'COAT_%s.p' % kwargs_str)
alg_kwargs['eval_space'] = eval_space
alg_kwargs['kwargs_file'] = kwargs_file
algo = VARREC(**alg_kwargs)
algo.fit(trainset, testset)
predictions = algo.test(testset)
mae = accuracy.mae(predictions, **{'verbose': False})
mse = pow(accuracy.rmse(predictions, **{'verbose': False}), 2.0)
print('%.4f %.4f %s' % (mae, mse, kwargs_str))
stdout.flush()
from os import path
from sys import stdout
import config
import itertools
import matplotlib.pyplot as plt
import numpy as np
import operator
import time
gsearch_file = tune_coat_file
err_kwargs, kwargs_set = config.read_gsearch(gsearch_file)
lr_all_opt, reg_all_opt = set(), set()
for kwargs_str in kwargs_set:
alg_kwargs = config.dictify(kwargs_str)
lr_all_opt.add(alg_kwargs['lr_all'])
reg_all_opt.add(alg_kwargs['reg_all'])
n_epochs = coat_n_epochs
epochs = np.arange(1, 1 + n_epochs)
lr_all_opt = sorted(list(lr_all_opt))
reg_all_opt = sorted(list(reg_all_opt))
lr_part_opt = [lr_all_opt[i] for i in [
2,
]]
lr_part_opt = [lr_all_opt[i] for i in [
0,
1,
2,