outputs = netG(inputs)
outputs_cpu = outputs.data.cpu().numpy()
lossL1 = criterionL1(outputs, targets)
L1val_accum += lossL1.item()
if i == 0:
input_ndarray = inputs_cpu.cpu().numpy()[0]
v_norm = (np.max(np.abs(input_ndarray[0, :, :]))**2 +
np.max(np.abs(input_ndarray[1, :, :]))**2)**0.5
outputs_denormalized = data.denormalize(outputs_cpu[0], v_norm)
targets_denormalized = data.denormalize(
targets_cpu.cpu().numpy()[0], v_norm)
utils.makeDirs(["results_train"])
utils.imageOut("results_train/epoch{}_{}".format(epoch, i),
outputs_denormalized,
targets_denormalized,
saveTargets=True)
# data for graph plotting
L1_accum /= len(trainLoader)
L1val_accum /= len(valiLoader)
if saveL1:
if epoch == 0:
utils.resetLog(prefix + "L1.txt")
utils.resetLog(prefix + "L1val.txt")
utils.log(prefix + "L1.txt", "{} ".format(L1_accum), False)
utils.log(prefix + "L1val.txt", "{} ".format(L1val_accum), False)
#fileName = dataDir + str(uuid.uuid4()) # randomized name
fileName = dataDir + "%s_%d_%d" % (basename, int(
freestreamX * 100), int(freestreamY * 100))
print("\tsaving in " + fileName + ".npz")
np.savez_compressed(fileName, a=npOutput)
files = os.listdir(airfoil_database)
files.sort()
if len(files) == 0:
print("error - no airfoils found in %s" % airfoil_database)
exit(1)
utils.makeDirs([
"./data_pictures", "./train", "./OpenFOAM/constant/polyMesh/sets",
"./OpenFOAM/constant/polyMesh"
])
# main
fout = open('train.txt', 'wt')
for n in range(samples):
print("Run {}:".format(n))
print("Run {}:".format(n), file=fout)
#fileNumber = np.random.randint(0, len(files))
#basename = os.path.splitext( os.path.basename(files[fileNumber]) )[0]
#print("\tusing {}".format(files[fileNumber]))
#print("\tusing {}".format(files[fileNumber]), file=fout)
basename = 'cylinder.dat'
targets = Variable(targets)
targets = targets.cuda()
inputs = torch.FloatTensor(1, 3, res, res)
inputs = Variable(inputs)
inputs = inputs.cuda()
targets_dn = torch.FloatTensor(1, 3, res, res)
targets_dn = Variable(targets_dn)
targets_dn = targets_dn.cuda()
outputs_dn = torch.FloatTensor(1, 3, res, res)
outputs_dn = Variable(outputs_dn)
outputs_dn = outputs_dn.cuda()
netG = TurbNetG(channelExponent=expo)
lf = "./" + prefix + "testout{}.txt".format(suffix)
utils.makeDirs(["results_test"])
# loop over different trained models
avgLoss = 0.
losses = []
models = []
loss_p_list = []
loss_v_list = []
accum_list = []
for si in range(25):
s = chr(96 + si)
if (si == 0):
s = "" # check modelG, and modelG + char
modelFn = "./" + prefix + "modelG{}{}".format(suffix, s)
if not os.path.isfile(modelFn):
def init():
'''init repository'''
try:
utils.makeDirs(getPath())
except Exception as e:
raise Exception('could not init repository. reason: {reason}'.format(reason = e.message))