def pre(self):
json_file = 'results/result_' + time.strftime("%Y%m%d%H%M%S") + '.json'
json_file_rk = 'results/result_' + time.strftime(
"%Y%m%d%H%M%S") + '_rk.json'
self.model.eval()
t1 = time.time()
print('extract features, this may take a few minutes')
feats = extract_feature(self.model, tqdm(self.test_loader))
print('feats:', feats.shape)
qf = feats[:self.num_query]
gf = feats[self.num_query:]
print('qf,gf:', qf.shape, gf.shape)
m, n = qf.shape[0], gf.shape[0]
distmat = torch.pow(qf, 2).sum(dim=1, keepdim=True).expand(m, n) + \
torch.pow(gf, 2).sum(dim=1, keepdim=True).expand(n, m).t()
distmat.addmm_(1, -2, qf, gf.t())
distmat = distmat.cpu().numpy()
self.writeResult(distmat, json_file)
print('re_ranking ...')
distmat_rk = reRanking(qf, gf, 7, 3, 0.85)
self.writeResult(distmat_rk, json_file_rk)
print('Time cost is: {:.2f} s'.format(time.time() - t1))
print('over!')
def ensemble(models):
feats = []
for index, model in enumerate(models):
feat = get_feat(model)
if index == 0:
feats = feat
else:
feats = torch.cat((feats, feat), 1)
print('feats:', feats.shape)
# query_paths = data.query_paths
num_query = len(data.query_paths)
qf = feats[:num_query]
gf = feats[num_query:]
print('qf,gf:', qf.shape, gf.shape)
m, n = qf.shape[0], gf.shape[0]
distmat = torch.pow(qf, 2).sum(dim=1, keepdim=True).expand(m, n) + \
torch.pow(gf, 2).sum(dim=1, keepdim=True).expand(n, m).t()
distmat.addmm_(1, -2, qf, gf.t())
# distmat = distmat.cpu().numpy()
print('re_ranking ...')
distmat_rk = reRanking(qf, gf, 7, 3, 0.85)
writeResult(data, distmat_rk, 'results/result_{}_200.json'.format(time_str), 200)
writeResult(data, distmat_rk, 'results/result_{}_300.json'.format(time_str), 300)
def pre(self):
json_file = 'results/result_' + time.strftime("%Y%m%d%H%M%S") + '.json'
json_file_rk = 'results/result_' + time.strftime(
"%Y%m%d%H%M%S") + '_rk.json'
self.model.eval()
t1 = time.time()
print('extract features, this may take a few minutes')
feats = extract_feature(self.model, tqdm(self.test_loader))
print('feats:', feats.shape)
qf = feats[:self.num_query]
gf = feats[self.num_query:]
print('qf,gf:', qf.shape, gf.shape)
m, n = qf.shape[0], gf.shape[0]
distmat = torch.pow(qf, 2).sum(dim=1, keepdim=True).expand(m, n) + \
torch.pow(gf, 2).sum(dim=1, keepdim=True).expand(n, m).t()
distmat.addmm_(1, -2, qf, gf.t())
distmat = distmat.cpu().numpy()
self.writeResult(distmat, json_file)
print('re_ranking ...')
distmat_rk = reRanking(qf, gf, 7, 3, 0.85)
self.writeResult(distmat_rk, json_file_rk)
# print('distmat:', distmat.shape)
# index = np.argsort(distmat) # from small to large
# max_index = index[:, :200]
# print(max_index.shape)
#
# # Visualize the rank result
# results = {}
# json_file = 'results/result_' + time.strftime("%Y%m%d%H%M%S") + '.json'
# json_file_rk = 'results/result_' + time.strftime("%Y%m%d%H%M%S") + '_rk.json'
# for i in range(len(self.query_paths)):
# query_name = self.query_paths[i].split('/')[-1]
# index_mask = max_index[i]
# gallery_name = [self.gallery_paths[k].split('/')[-1] for k in index_mask]
# #print(index_mask)
#
# results[query_name] = gallery_name
# #print(res)
#
# with open(json_file, 'w', encoding='utf-8') as fp:
# json.dump(results, fp)
print('Time cost is: {:.2f} s'.format(time.time() - t1))
print('over!')
def writeResult_rk(distmat, qf, gf, k1, k2, lambda_value, num=100, part=5, dis_th=0.5,
json_file='part_res.json'):
results = {}
st0 = time.time()
print('start part reranking, ini distmat:', distmat.shape)
print('distmat mean:', np.mean(distmat))
print('selet num:', num)
print('selet part:', part)
print('selet dis_th:', dis_th)
index = np.argsort(distmat) # from small to large
print('argsort cost time:{} s'.format(time.time() - st0))
st0 = time.time()
max_index = index[:, :num]
print(max_index.shape)
for ii in range(part):
num_ = qf.size(0) // part
start = ii * num_
end = (ii + 1) * num_
gf_indexs = []
for jj in range(start, end):
gf_indexs.extend(max_index[jj, :].tolist())
# temp_inds = [dis_ind for dis_ind in max_index[jj, :] if distmat[jj, dis_ind] < dis_th]
# gf_indexs.extend(temp_inds)
gf_indexs = list(set(gf_indexs))
gf_indexs.sort()
gf_indexs = np.array(gf_indexs)
if ii != part - 1:
qf_ = qf[start:end, :]
query_paths_ = data.query_paths[start:end]
gallery_paths_ = [data.gallery_paths[k] for k in gf_indexs]
else:
qf_ = qf[start:, :]
query_paths_ = data.query_paths[start:]
gallery_paths_ = [data.gallery_paths[k] for k in gf_indexs]
gf_ = gf[gf_indexs, :]
print('part-({}/{}), qf gf size:{} {}'.format(ii + 1, part, qf_.size(), gf_.size()))
distmat_rk_ = reRanking(qf_, gf_, k1, k2, lambda_value)
writeResult_part(results, distmat_rk_, query_paths_, gallery_paths_, num=200)
print('part-({}/{}), cost time:{} s'.format(ii + 1, part, time.time() - st0))
st0 = time.time()
with open(json_file, 'w', encoding='utf-8') as fp:
json.dump(results, fp)
print('save result :', json_file)