save_weights(layers, config['weights_dir'], epoch)
np.save(config['weights_dir'] + 'lr_' + str(epoch) + '.npy',
learning_rate.get_value())
save_momentums(vels, config['weights_dir'], epoch)
print('Optimization complete.')
if __name__ == '__main__':
with open('config.yaml', 'r') as f:
config = yaml.load(f)
with open('spec_1gpu.yaml', 'r') as f:
config = dict(config.items() + yaml.load(f).items())
config = proc_configs(config)
if config['para_load']:
from proc_load import fun_load
config['queue_l2t'] = Queue(1)
config['queue_t2l'] = Queue(1)
train_proc = Process(target=train_net, args=(config, ))
load_proc = Process(target=fun_load,
args=(config, config['sock_data']))
train_proc.start()
load_proc.start()
train_proc.join()
load_proc.join()
else:
train_proc = Process(target=train_net, args=(config, ))
print('validation loss %f ' %
(this_validation_loss))
return this_validation_error, this_validation_loss
############################################
if __name__ == '__main__':
with open('config.yaml', 'r') as f:
config = yaml.load(f)
with open('spec_1gpu.yaml', 'r') as f:
config = dict(config.items() + yaml.load(f).items())
config = proc_configs(config)
config['resume_train'] = True
config['load_epoch'] = 60
if config['para_load']:
from proc_load import fun_load
config['queue_l2t'] = Queue(1)
config['queue_t2l'] = Queue(1)
train_proc = Process(target=validate_performance, args=(config,))
load_proc = Process(
target=fun_load, args=(config, config['sock_data']))
train_proc.start()
load_proc.start()
train_proc.join()
load_proc.join()