experiment_parameters['n_col'] = 6
experiment_parameters['network_type'] = 'DVAE'
experiment_parameters['save_interval'] = 5
experiment_parameters['fps'] = 25
experiment_parameters['n_clayers'] = 5
experiment_parameters['kernel1_size'] = 4
# Setting up data folder
resultdir = './results/cows/'
if not (path.exists(resultdir)):
os.mkdir(resultdir)
datestring = datetime.datetime.now().strftime("%Y_%m_%d_%H_%M_%S")
os.mkdir(resultdir + datestring)
# Saving parameter file
helper.print_params('./results/cows/' + datestring + '/PARAMETERS.txt',
experiment_parameters)
# Running the experiments
files = glob.glob('./data/cows/*mp4')
files.sort()
vids = []
for f_i, file in enumerate(files):
cv, _ = helper.readvid(file, scale=(224, 224))
vids.append(cv)
dpref = resultdir + datestring
fpref = dpref + '/' + 'cows'
print('Processing now:', fpref)
print()
vid_pairs = []
for V in vids:
multiproc=True,
ncores=12,
filename=os.path.join(save_path, 'explore_perf.hdf5'))
traj = env.v_trajectory
pp.add_parameters(traj, parameter_dict)
explore_dict = pypet.cartesian_product(
explore_dict, tuple(explore_dict.keys())
) #if not all entry of dict need be explored through cartesian product replace tuple(.) only with relevant dict keys in tuple
explore_dict['name'] = pp.set_run_names(explore_dict, parameter_dict['name'])
traj.f_explore(explore_dict)
""" launch simulation with pypet for parameter exploration """
tic = time.time()
env.f_run(pp.launch_exploration, images_train, labels_train, images_test,
labels_test, save_path)
toc = time.time()
""" save parameters to file """
helper.print_params(print_dict, save_path, runtime=toc - tic)
""" plot results """
name_best = pp.plot_results(folder_path=save_path)
pp.faceting(save_path)
print '\nrun name:\t' + parameter_dict['name']
print 'start time:\t' + time.strftime("%a, %d %b %Y %H:%M:%S",
time.localtime(tic))
print 'end time:\t' + time.strftime("%a, %d %b %Y %H:%M:%S",
time.localtime(toc))
print 'train time:\t' + str(datetime.timedelta(seconds=toc - tic))
experiment_parameters['n_col'] = 6
experiment_parameters['network_type'] = 'incomplete'
experiment_parameters['save_interval'] = 5
experiment_parameters['fps'] = 25
experiment_parameters['n_clayers'] = 5
experiment_parameters['kernel1_size'] = 8
# Setting up data folder
resultdir = './results/playing_incomplete/'
if not(path.exists(resultdir)):
os.mkdir(resultdir)
datestring = datetime.datetime.now().strftime("%Y_%m_%d_%H_%M_%S")
os.mkdir(resultdir+datestring)
# Saving parameter file
helper.print_params('./results/playing_incomplete/'+datestring+'/PARAMETERS.txt',
experiment_parameters)
# Running the experiments
V, _ = helper.readvid('./data/playing_incomplete/playing_incomplete.avi')
M, _ = helper.readvid('./data/playing_incomplete/playing_mask.avi')
dpref = resultdir+datestring
fpref = dpref+'/'+'playing_incomplete'
print('Processing now:', fpref)
print()
vid_pairs = []
V_ = np.zeros((V.shape[0]-1, 6, V.shape[2], V.shape[3]))
V_[:, 0, :, :] = V[:-1, 0]
V_[:, 1, :, :] = V[:-1, 1]
V_[:, 2, :, :] = V[:-1, 2]
V_[:, 3, :, :] = V[1:, 0]
os.mkdir(resultdir + datestring)
# Running the experiments
files = glob.glob('./data/stgconv_data/*.avi')
files.sort()
for f_i, file in enumerate(files):
V, fr = helper.readvid(file, scale=(128, 128))
os.mkdir(resultdir + datestring + '/v' + str(f_i))
dpref = resultdir + datestring + '/v' + str(f_i)
fpref = dpref + '/vid' + str(f_i)
print('Processing now:', fpref)
print()
V_ = np.zeros((V.shape[0] - 1, 6, V.shape[2], V.shape[3]))
V_[:, 0, :, :] = V[:-1, 0]
V_[:, 1, :, :] = V[:-1, 1]
V_[:, 2, :, :] = V[:-1, 2]
V_[:, 3, :, :] = V[1:, 0]
V_[:, 4, :, :] = V[1:, 1]
V_[:, 5, :, :] = V[1:, 2]
experiment_parameters['fps'] = fr
experiment_parameters['n_epochs'] = 62500 // V.shape[0]
experiment_parameters[
'save_interval'] = experiment_parameters['n_epochs'] // 20
helper.print_params(dpref + '/PARAMETERS.txt', experiment_parameters)
npdata = V_ * 2 - 1
traindata = torch.from_numpy(npdata).float()
ret = helper.run_experiment(traindata, fpref, experiment_parameters)
if not ret:
print('Could not synthesize', fpref, 'due to runtime error')
total_err += np.abs(err)
print(
'(' + "({0:2d},{1:2d}) -> ({2:2d},{3:2d})".format(
d[n][0], d[n][1], d[n][2], d[n][3]) + ') = ' +
"{0:8.2f} / {1:8.2f} diff: {2:8.2f} 1/cm".format(d[n][4], fit, err))
print("\nTotal error/no of lines = {0:6.5f}\n".format(total_err / len(d)))
###############################################
# calculate predictions
###############################################
print()
print_params(result)
(Yg, Ye) = make_params_dunham(result)
#print_matrix(Yg)
#print_matrix(Ye)
(Ug, Ue) = reduce_dunham(Yg, Ye)
print_matrix(Ug)
print_matrix(Ue)
make_report(Ug, Ue, vmax=1, Jmax=1, save_filename='lines_dunham.txt')
#print_matrix(Ug)
#print_matrix(Ue)
