def start_all_logging_instruments(hyper, results_path, test_images):
writer = tf.summary.create_file_writer(logdir=results_path)
logger = setup_logger(log_file_name=append_timestamp_to_file(file_name=results_path + '/loss.log',
termination='.log'),
logger_name=append_timestamp_to_file('logger', termination=''))
log_all_hyperparameters(hyper=hyper, logger=logger)
plot_originals(test_images=test_images, results_path=results_path)
return writer, logger
def start_all_logging_instruments(hyper, test_images):
results_path = determine_path_to_save_results(
model_type=hyper['model_type'], dataset_name=hyper['dataset_name'])
if not os.path.exists(results_path):
os.mkdir(results_path)
logger = setup_logger(log_file_name=append_timestamp_to_file(
file_name=results_path + '/loss.log', termination='.log'),
logger_name=append_timestamp_to_file('logger',
termination=''))
log_all_hyperparameters(hyper=hyper, logger=logger)
plot_originals(test_images=test_images, results_path=results_path)
return logger, results_path
def run_sop(hyper, results_path, data):
train_dataset, test_dataset = data
sop_optimizer = setup_sop_optimizer(hyper=hyper)
logger = setup_logger(log_file_name=append_timestamp_to_file(
file_name=results_path + f'/loss_{sop_optimizer.model.model_type}.log',
termination='.log'))
log_all_hyperparameters(hyper=hyper, logger=logger)
train_sop(sop_optimizer=sop_optimizer,
hyper=hyper,
train_dataset=train_dataset,
test_dataset=test_dataset,
logger=logger)
def run_sop(hyper, results_path):
tf.random.set_seed(seed=hyper['seed'])
data = load_mnist_sop_data(batch_n=hyper['batch_size'])
train_dataset, test_dataset = data
sop_optimizer = setup_sop_optimizer(hyper=hyper)
model_type = sop_optimizer.model.model_type
log_path = results_path + f'/loss_{model_type}.log'
logger = setup_logger(log_file_name=append_timestamp_to_file(
file_name=log_path, termination='.log'),
logger_name=model_type + str(hyper['seed']))
log_all_hyperparameters(hyper=hyper, logger=logger)
save_hyper(hyper)
train_sop(sop_optimizer=sop_optimizer,
hyper=hyper,
train_dataset=train_dataset,
test_dataset=test_dataset,
logger=logger)
import time
import numpy as np
import tensorflow as tf
from Utils.Distributions import compute_gradients, apply_gradients
from Utils.general import initialize_mu_and_xi_for_logistic, initialize_mu_and_xi_equally, setup_logger
logger = setup_logger(log_file_name='./Log/discrete.log')
class MinimizeEmpiricalLoss:
def __init__(self,
params,
learning_rate,
temp,
sample_size=int(1.e3),
max_iterations=int(1.e4),
run_kl=True,
tolerance=1.e-5,
model_type='IGR_I',
threshold=0.9,
planar_flow=None):
self.params = params
self.learning_rate = learning_rate
self.temp = tf.constant(value=temp, dtype=tf.float32)
self.sample_size = sample_size
self.max_iterations = max_iterations
self.run_kl = run_kl
self.tolerance = tolerance
self.model_type = model_type
self.threshold = threshold