def extend_mlp_exp():
for s in [5]:
for i in [1, 3, 4, 5]:
for a in ['relu']:
for e in [1, 2, 3, 4, 5, 10, 15]:
exp_data = {
'sparseness': s,
'index': i,
'epochs': 50,
'batch_size': 256,
'layers': [64, 32, 16, 8],
'reg_layers': [0, 0, 0, 0],
'fake_layers': [64, 32, 16, 8],
'fake_reg_layers': [0, 0, 0, 0],
'last_activation': a,
'fake_last_activation': a,
'learning_rate': 0.001,
'extend_near_num': e,
'learner': 'adagrad'
}
experiment.extend_mlp_model_exp.experiment(**exp_data)
send_email(receiver='*****@*****.**',
title='EXMLP实验结束',
text="实验结束时间：{}".format(datetime.now()),
**email_config)
def svd():
for s in [1, 3]:
for i in [1]:
for e in [0]:
experiment.svd_exp.experiment(s, i, 0.1, e, matrix_type='rt')
send_email(receiver='*****@*****.**',
title='SVD实验结束',
text="",
**email_config)
def main_fork():
with Pool() as pool:
extends = (1, 2, 3, 4, 5, 10, 15, 20)
# extends = (1, 2)
# pool.map_async(save_extend_array, [(s, i, extends, pool) for s in [5, 10, 15, 20] for i in range(1, 6)])
pool.starmap(save_extend_array, [(s, i, extends)
for s in [5, 10, 15, 20]
for i in range(1, 6)])
send_email(receiver='*****@*****.**',
title='实验结束',
text="",
**email_config)
def ncf_exp():
for s in [1, 3]:
for i in [1]:
for e in [0, 1, 2, 3, 4, 5, 10, 15, 20]:
for d in [8]:
data = ExperimentData()
data.sparseness = s
data.data_index = i
data.mf_dim = d
data.epochs = 30
data.batch_size = 128
data.layers = [64, 32, 16]
data.reg_layers = [0, 0, 0]
data.learning_rate = 0.007
data.extend_near_num = e
data.learner = 'adam'
experiment.ncf_exp.experiment(data, 'relu', matrix_type='tp')
send_email(receiver='*****@*****.**',
title='NCF实验结束',
text="",
**email_config)
def mlp_exp():
for s in [5, 10, 15, 20]:
for i in [1]:
for a in ['relu']:
for e in [0, 1, 2, 3, 4, 5, 10, 15, 20]:
exp_data = {
'sparseness': s,
'index': i,
'epochs': 30,
'batch_size': 128,
'layers': [64, 32, 16],
'reg_layers': [0, 0, 0],
'last_activation': a,
'learning_rate': 0.007,
'extend_near_num': e,
'learner': 'adam',
'matrix_type': 'tp'
}
experiment.mlp_exp.experiment(**exp_data)
send_email(receiver='*****@*****.**',
title='MLP实验结束',
text="",
**email_config)
def gmf_exp():
for s in [1, 3]:
for i in [1, 2, 3, 4, 5]:
for a in ['relu']:
for d in [128]:
for e in [0, 1, 2, 3, 4, 5, 10, 15, 20]:
exp_data = {
'sparseness': s,
'index': i,
'epochs': 30,
'batch_size': 128,
'mf_dim': d,
'regs': [0, 0],
'last_activation': a,
'learning_rate': 0.007,
'extend_near_num': e,
'learner': 'adagrad'
}
experiment.gmf_exp.experiment(**exp_data)
send_email(receiver='*****@*****.**',
title='GMF实验结束',
text="实验结束时间：{}".format(datetime.now()),
**email_config)
if __name__ == '__main__':
os.environ["CUDA_VISIBLE_DEVICES"] = "1"
all_exp_data = []
for s in [5]:
for i in [2, 3, 4, 5]:
for a in ['relu', 'sigmoid']:
for e in [0, 5]:
if i == 2 and a == 'relu' and e == 0:
continue
exp_data = {
'sparseness': s,
'index': i,
'epochs': 30,
'batch_size': 128,
'layers': [64, 32, 16],
'reg_layers': [0, 0, 0],
'last_activation': a,
'learning_rate': 0.007,
'extend_near_num': e,
'learner': 'adagrad'
}
all_exp_data.append(experiment(**exp_data))
text = '\n'.join(map(lambda d: str(d), all_exp_data))
send_email(receiver='*****@*****.**',
title='MLP实验结束',
text=text,
**email_config)
VT1 = VT[:K, :]
Sigma1 = np.eye(K) * Sigma[:K]
R = np.matmul(np.matmul(U1, Sigma1), VT1)
mae, rmse = evaluate(sparseness, index, R, matrix_type=matrix_type)
exp_data["mae"] = float(mae)
exp_data["rmse"] = float(rmse)
exp_data['datetime'] = datetime.now()
print(exp_data)
auto_insert_database(database_remote_config, exp_data,
f'svd_{matrix_type}')
# insert_database('experiment_data.db', "experiment_svd_rt", exp_data)
if __name__ == '__main__':
args = []
for s in [1, 3]:
for i in [1]:
for e in [0]:
args.append((s, i, 0.1, e))
from concurrent.futures import ThreadPoolExecutor
with ThreadPoolExecutor() as executor:
executor.map(lambda x: experiment(*x), args)
executor.shutdown(wait=True)
send_email(receiver='*****@*****.**',
title='SVD实验结束',
text="",
**email_config)