def get_fid(fakes, model, npz, device, batch_size=1, use_tqdm=True):
m1, s1 = npz['mu'], npz['sigma']
fakes = torch.cat(fakes, dim=0)
fakes = util.tensor2im(fakes).astype(float)
m2, s2 = _compute_statistics_of_ims(fakes, model, batch_size, 2048,
device, use_tqdm=use_tqdm)
return float(calculate_frechet_distance(m1, s1, m2, s2))
def main(opt):
dataloader = create_dataloader(opt)
device = torch.device('cuda:{}'.format(opt.gpu_ids[0])) if opt.gpu_ids \
else torch.device('cpu')
block_idx = InceptionV3.BLOCK_INDEX_BY_DIM[2048]
inception_model = InceptionV3([block_idx])
inception_model.to(device)
inception_model.eval()
tensors = []
for i, data_i in enumerate(tqdm.tqdm(dataloader)):
if opt.dataset_mode in ['single', 'aligned']:
tensor = data_i[opt.direction[-1]]
else:
tensor = data_i['image']
tensors.append(tensor)
tensors = torch.cat(tensors, dim=0)
tensors = util.tensor2im(tensors).astype(float)
mu, sigma = _compute_statistics_of_ims(tensors,
inception_model,
32,
2048,
device,
use_tqdm=True)
np.savez(opt.output_path, mu=mu, sigma=sigma)
def get_fid_new(reals, fakes, model, device, batch_size=1, use_tqdm=True):
reals = torch.cat(reals, dim=0)
reals = util.tensor2im(reals).astype(float)
fakes = torch.cat(fakes, dim=0)
fakes = util.tensor2im(fakes).astype(float)
m1, s1 = _compute_statistics_of_ims(reals,
model,
batch_size,
2048,
device,
use_tqdm=use_tqdm,
median=False)
m2, s2 = _compute_statistics_of_ims(fakes,
model,
batch_size,
2048,
device,
use_tqdm=use_tqdm,
median=False)
fid_mean = float(calculate_frechet_distance(m1, s1, m2, s2, median=False))
m1, s1 = _compute_statistics_of_ims(reals,
model,
batch_size,
2048,
device,
use_tqdm=use_tqdm,
median=True)
m2, s2 = _compute_statistics_of_ims(fakes,
model,
batch_size,
2048,
device,
use_tqdm=use_tqdm,
median=True)
fid_median = float(calculate_frechet_distance(m1, s1, m2, s2, median=True))
return fid_mean, fid_median
def main(opt):
dataloader = create_dataset(opt)
device = torch.device('cuda:{}'.format(opt.gpu_ids[0])) if opt.gpu_ids \
else torch.device('cpu')
block_idx = InceptionV3.BLOCK_INDEX_BY_DIM[2048]
inception_model = InceptionV3([block_idx])
inception_model.to(device)
inception_model.eval()
tensors = {}
for i, data_i in enumerate(dataloader):
tensor = data_i['B' if opt.direction == 'AtoB' else 'A']
tensors[data_i['B_paths' if opt.direction == 'AtoB' else 'A_paths']
[0]] = tensor
tensors = torch.cat(list(tensors.values()), dim=0)
tensors = util.tensor2imgs(tensors).astype(float)
mu, sigma = _compute_statistics_of_ims(tensors,
inception_model,
32,
2048,
device,
use_tqdm=True)
np.savez(opt.output_path, mu=mu, sigma=sigma)