def accuracy(output, target, topk=(1,)):
output, target = to_torch(output), to_torch(target)
maxk = max(topk)
batch_size = target.size(0)
_, pred = output.topk(maxk, 1, True, True)
pred = pred.t()
correct = pred.eq(target.view(1, -1).expand_as(pred))
ret = []
for k in topk:
correct_k = correct[:k].view(-1).float().sum(0)
ret.append(correct_k.mul_(1. / batch_size))
return ret
def extract_cnn_embeddings(model, inputs, modules=None):
model.eval()
for ind, inp in enumerate(inputs):
inputs[ind] = to_torch(inp)
inputs = [Variable(x, volatile=True).cuda() for x in inputs]
assert modules is None
outputs = model(*inputs)
outputs = outputs.data
return outputs
def predict_prob(model, data, data_dir, config):
config.set_training(False)
model.eval()
dataloader = dp.get_dataloader(data, data_dir, config)
probs = []
for i, (imgs, _, _, _, _) in enumerate(dataloader):
inputs = to_torch(imgs)
inputs = Variable(inputs, volatile=True)
output = model(inputs)
prob = nn.functional.softmax(output, dim=1)
probs += [prob.data.cpu().numpy()]
probs = np.concatenate(probs)
return probs
def extract_cnn_feature(model, inputs, modules=None):
model.eval()
inputs = to_torch(inputs)
inputs = Variable(inputs, volatile=True)
if modules is None:
outputs = model(inputs)
outputs = outputs.data.cpu()
return outputs
# Register forward hook for each module
outputs = OrderedDict()
handles = []
for m in modules:
outputs[id(m)] = None
def func(m, i, o):
outputs[id(m)] = o.data.cpu()
handles.append(m.register_forward_hook(func))
model(inputs)
for h in handles:
h.remove()
return list(outputs.values())