def main():
generator = sample_generator(train_batch_size, mod_n, NR)
device = 'cuda'
model = iterative_classifier(d_model, n_head, nhid, nlayers, mod_n, NR,
d_transmitter_encoding,
generator.real_QAM_const,
generator.imag_QAM_const,
generator.constellation, device, dropout)
model = model.to(device=device)
optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate)
if (load_pretrained_model):
checkpoint = torch.load(model_filename)
model.load_state_dict(checkpoint['model_state_dict'])
optimizer.load_state_dict(checkpoint['optimizer_state_dict'])
lr_scheduler = torch.optim.lr_scheduler.ReduceLROnPlateau(
optimizer, 'min', 0.91, 0, True, 0.0001, 'rel', 0, 0, 1e-08)
print('*******Successfully loaded pre-trained model***********')
else:
lr_scheduler = torch.optim.lr_scheduler.ReduceLROnPlateau(
optimizer, 'min', 0.91, 0, True, 0.0001, 'rel', 0, 0, 1e-08)
train(model, optimizer, lr_scheduler, generator, device)
print(
'******************************** Now Testing **********************************************'
)
def main():
generator = sample_generator(validtn_batch_size, mod_n, NR)
device = 'cuda'
model = iterative_classifier(d_model, n_head, nhid, nlayers, mod_n, NR, d_transmitter_encoding, generator.real_QAM_const, generator.imag_QAM_const, generator.constellation, device, dropout)
model = model.to(device=device)
checkpoint = torch.load(model_filename)
model.load_state_dict(checkpoint['model_state_dict'])
print('*******Successfully loaded pre-trained model*********** from direcotry : ', model_filename)
test(model, generator, device)
print('******************************** Now Testing **********************************************')
def main():
generator = sample_generator(time_batch_size, mod_n, NR)
device = 'cuda'
model_network = iterative_classifier(d_model, n_head, nhid, nlayers, mod_n, NR, d_transmitter_encoding, generator.real_QAM_const, generator.imag_QAM_const, generator.constellation, device, dropout)
model_oampnet = oampnet(num_layers, generator.constellation, generator.real_QAM_const, generator.imag_QAM_const, device=device)
model_network = model_network.to(device=device)
model_oampnet = model_oampnet.to(device=device)
network_checkpoint = torch.load(model_network_filename)
model_network.load_state_dict(network_checkpoint['model_state_dict'])
model_oampnet.load_state_dict(torch.load(model_oampnet_filename))
print('*******Successfully loaded pre-trained model***********')
test(model_network, model_oampnet, generator, device)
print('******************************** Now Testing **********************************************')