def test_3rd_snapshot():
te_set = PosterSet(POSTER_PATH, split, 'test', gen_d=gen_d, augment=False, resize=None, ten_crop=None)#, debug=True)
te_load = DataLoader(te_set, batch_size=64, shuffle=False, num_workers=3, drop_last=True)
model = SmallerNetwork(INP_SIZE, 23)
state = torch.load(SNAP_PATH + "snap3rd.nn")
model.load_state_dict(state['state_dict'])
if CUDA_ON: model.cuda()
model.eval()
loss = 0
skipped = 0
for X, y in tqdm(te_load, desc='3rd'):
X, y = Var(X, volatile=True), Var(y)
if CUDA_ON:
X, y = X.cuda(), y.cuda()
out = model(X)
for i in range(out.size(0)):
try:
loss += accuracy(out.data[i], y.data[i])
except:
skipped += 1
return loss / (len(te_set) - skipped)
def test_5th_snapshot():
te_set = PosterSet(POSTER_PATH, split, 'test', gen_d=gen_d, augment=False, resize=None, ten_crop=CROP_SIZE)#, debug=True)
te_load = DataLoader(te_set, batch_size=64, shuffle=False, num_workers=3, drop_last=True)
model = MidrangeNetwork(CROP_SIZE, 23)
state = torch.load(SNAP_PATH + "snap5th.nn")
model.load_state_dict(state['state_dict'])
if CUDA_ON: model.cuda()
model.eval()
loss = 0
skipped = 0
for X, y in tqdm(te_load, desc='5th'):
X, y = Var(X, volatile=True), Var(y)
bs, ncrops, c, h, w = X.size()
if CUDA_ON:
X, y = X.cuda(), y.cuda()
out = model(X.view(-1, c, h, w))
out = out.view(bs, ncrops, -1).mean(1)
for i in range(out.size(0)):
try:
loss += accuracy(out.data[i], y.data[i])
except:
skipped += 1
return loss / (len(te_set) - skipped)
def test_distribution():
te_set = PosterSet(POSTER_PATH, split, 'test', gen_d=gen_d, augment=False, resize=None, ten_crop=None)#, debug=True)
te_load = DataLoader(te_set, batch_size=64, shuffle=False, num_workers=3, drop_last=True)
pred = torch.Tensor([0.28188283, 0.23829031, 0.27430824, 0.39426496, 0.38900912, 0.22754676, 0.24563302, 0.11153192, 0.29865512, 0.14260318, 0.17011903, 0.0307621 , 0.20026279, 0.2485701 , 0.24037718, 0.17645695, 1. , 0.42100788, 0.11593755, 0.31264492, 0.62699026, 0.1946205 , 0.27446282])
loss = 0
skipped = 0
for X, y in tqdm(te_load, desc='dis'):
for i in range(X.size(0)):
try:
loss += accuracy(pred, y[i])
except:
skipped += 1
return loss / (len(te_set) - skipped)
DEBUG_MODE = False
num_epochs = 101
batch_s = 128
log_percent = 0.25
s_factor = 0.5
learn_r = 0.0001
input_size = (268, 182) #posters are all 182 width, 268 heigth
p = pickle.load(open(DATA_PATH, 'rb'))
gen_d = pickle.load(open(DICT_PATH, 'rb'))
train_set = PosterSet(POSTER_PATH,
p,
'train',
gen_d=gen_d,
tv_norm=True,
augment=True,
resize=input_size,
debug=DEBUG_MODE)
train_loader = torch.utils.data.DataLoader(train_set,
batch_size=batch_s,
shuffle=True,
num_workers=4)
log_interval = np.ceil((len(train_loader.dataset) * log_percent) / batch_s)
val_set = PosterSet(POSTER_PATH,
p,
'val',
gen_d=gen_d,
tv_norm=True,
augment=True,
import sys, os
from extractors import *
from core import PosterSet
DATASET_PATH = "../sets/set_splits.p"
POSTER_PATH = "../posters/"
GENRE_DICT_PATH = "../sets/gen_d.p"
SETS_PATH = "../sets/"
FEATURE_PATH = "../sets/features_all.h5"
CUDA_ON = True
p = pickle.load(open(DATASET_PATH, 'rb'))
gen_d = pickle.load(open(GENRE_DICT_PATH, 'rb'))
h5 = h5py.File(FEATURE_PATH, 'a')
dataset = PosterSet(POSTER_PATH, p, 'all', gen_d=gen_d, normalize=True)
dataloader = torch.utils.data.DataLoader(dataset,
batch_size=32,
shuffle=False,
num_workers=1)
h5.create_dataset("labels", data=np.stack([y for __, y in tqdm(dataset)]))
h5.create_dataset("ids",
data=np.array([s.encode('utf8') for s in dataset.ids]))
cuda_device = int(sys.argv[1]) if len(sys.argv) > 1 else 0
if cuda_device == -1:
cuda_device = 0
CUDA_ON = False
with torch.cuda.device(cuda_device):
print(TERM['y'])
print("mean: ", m.running_mean)
print("var: ", m.running_var)
print("bias:", m.bias.data)
print("weight:", m.weight.data)
print(TERM['clr'])
# prepare data
gen_d = pickle.load(open(GENRE_DICT_PATH, 'rb'))
split = pickle.load(open(DATASET_PATH, 'rb'))
tr_set = PosterSet(POSTER_PATH,
split,
'train',
gen_d=gen_d,
augment=True,
resize=None,
rnd_crop=CROP_SIZE) #, debug=True)
tr_load = DataLoader(tr_set,
batch_size=128,
shuffle=True,
num_workers=3,
drop_last=True)
va_set = PosterSet(POSTER_PATH,
split,
'val',
gen_d=gen_d,
augment=False,
resize=None,