)
print(opt)
random.seed(opt.manualSeed)
torch.manual_seed(opt.manualSeed)
if opt.gpu:
torch.cuda.manual_seed_all(opt.manualSeed)
model = locate("models.%s" % opt.model)(opt)
if opt.load:
model.load(opt.load)
if opt.gpu:
model.gpu()
print("n parameters: %d" % sum([m.numel() for m in model.parameters()]))
viz = utils.Viz(opt)
def process_batch(batch, loss, i, k, set_, t0):
"""Optimization step.
batch = [input, target]: contains data for optim step [input, target]
loss: dict containing statistics about optimization
i: epoch
k: index of the current batch
set_: type of batch (\"train\" or \"dev\")
t0: time of the beginning of epoch
"""
nbatch = vars(opt)["nbatch_" + set_]
res = model.step(batch, set_)
opt = json.loads(data[0])
opt['p_red'] = 0
opt['mask_object'] = opt_test.mask_object
print(opt)
opt = utils.to_namespace(opt)
opt.bsz = opt.m
opt.count = opt_test.count
model = locate('models.%s' %opt.model)(opt, test=True)
model.load(opt_test.load)
if opt_test.gpu:
model.gpu()
if opt_test.eval:
model.eval()
else:
print('WARNING: no call to eval()')
viz = utils.Viz(opt_test)
viz_output = utils.Viz(opt_test)
else:
assert opt_test.load_scores
def process_batch(batch, j, t0):
"""Compute score for every frames in a video (which is also a batch).
batch = [input, target]: all frames in the video
j: index of the video
t0: time when the test started
"""
nbatch = vars(opt)['nbatch_test']
frame_scores = np.zeros((opt.m, 4))
d3, d4 = batch[0].size(3), batch[0].size(4)
for i in range(opt.bsz):