weights_path = os.path.join(logdir, 'weights_posebaseline_{epoch:03d}.hdf5')
batch_size_mpii = 14
batch_size_ar = 2
"""Load datasets"""
mpii = MpiiSinglePerson(datasetpath('MPII'), dataconf=mpii_dataconf,
poselayout=pa17j3d)
# h36m = Human36M(datasetpath('Human3.6M'), dataconf=human36m_dataconf,
# poselayout=pa17j3d, topology='frames')f
penn_sf = PennAction(datasetpath('Penn_Action'), pennaction_dataconf,
poselayout=pa17j3d, topology='frames', use_gt_bbox=True)
ntu_sf = Ntu(datasetpath('NTU'), ntu_pe_dataconf, poselayout=pa17j3d,
topology='frames', use_gt_bbox=True)
"""Create an object to load data from all datasets."""
data_tr = BatchLoader([mpii, penn_sf, ntu_sf], ['frame'], ['pose'],
TRAIN_MODE, batch_size=[batch_size_mpii, batch_size_ar,
batch_size_ar], num_predictions=num_predictions, shuffle=True)
"""MPII validation samples."""
mpii_val = BatchLoader(mpii, ['frame'], ['pose', 'afmat', 'headsize'],
VALID_MODE, batch_size=mpii.get_length(VALID_MODE), shuffle=True)
printcn(OKBLUE, 'Pre-loading MPII validation data...')
[x_val], [p_val, afmat_val, head_val] = mpii_val[0]
mpii_callback = MpiiEvalCallback(x_val, p_val, afmat_val, head_val,
map_to_pa16j=pa17j3d.map_to_pa16j, logdir=logdir)
# """Human3.6H validation samples."""
num_pose_features=192, num_visual_features=192)
logger.debug(cfg)
num_predictions = spnet.get_num_predictions(cfg.num_pyramids, cfg.num_levels)
num_action_predictions = spnet.get_num_predictions(len(cfg.action_pyramids), cfg.num_levels)
logger.debug("NUM PREDICTIONS")
logger.debug(num_action_predictions)
"""Load datasets"""
# h36m = Human36M(datasetpath('Human3.6M'), dataconf=human36m_dataconf,
# poselayout=pa17j3d, topology='frames')
ntu_data_path = os.getcwd() + '/datasets/NTU'
logger.debug(ntu_data_path)
ntu = Ntu(ntu_data_path, ntu_dataconf, poselayout=pa17j3d,
topology='sequences', use_gt_bbox=True, clip_size=num_frames, num_S=5)
#logger.debug ('WARNING!! USING ONLY S1 FOR EVALUATION!')
"""Build the full model"""
full_model = spnet.build(cfg)
weights_file = os.getcwd() + '/weights/weights_3dp+ntu_ar_062.hdf5'
logger.debug(weights_file)
if os.path.isfile(weights_file) == False:
logger.debug (f'Error: file {weights_file} not found!')
logger.debug (f'\nPlease download it from https://drive.google.com/file/d/1I6GftXEkL5nohLA60Vi6faW0rvTZg6Kx/view?usp=sharing')
sys.stdout.flush()
sys.exit()
"""Load pre-trained weights from pose estimation and copy replica layers."""
sys.stdout = open(str(logdir) + '/log.txt', 'w')
num_frames = 8
cfg = ModelConfig((num_frames,) + ntu_dataconf.input_shape, pa17j3d,
num_actions=[60], num_pyramids=2, action_pyramids=[1, 2],
num_levels=4, pose_replica=False,
num_pose_features=192, num_visual_features=192)
num_predictions = spnet.get_num_predictions(cfg.num_pyramids, cfg.num_levels)
num_action_predictions = spnet.get_num_predictions(len(cfg.action_pyramids), cfg.num_levels)
"""Load datasets"""
# h36m = Human36M(datasetpath('Human3.6M'), dataconf=human36m_dataconf,
# poselayout=pa17j3d, topology='frames')
ntu = Ntu(datasetpath('NTU'), ntu_dataconf, poselayout=pa17j3d,
topology='sequences', use_gt_bbox=True, clip_size=num_frames)#, num_S=1)
# print ('WARNING!! USING ONLY S1 FOR EVALUATION!')
"""Build the full model"""
full_model = spnet.build(cfg)
"""Load pre-trained weights from pose estimation and copy replica layers."""
full_model.load_weights(
# 'output/ntu_spnet_trial-03-ft_replica_0ae2bf7/weights_3dp+ntu_ar_062.hdf5',
'output/ntu_spnet_trial_06_nopose_g_512a239/weights_3dp+ntu_ar_030.hdf5',
by_name=True)
"""Split model to simplify evaluation."""
models = split_model(full_model, cfg, interlaced=False,
model_names=['3DPose', '3DAction'])
mpii = MpiiSinglePerson("D:\\MPII", dataconf=mpii_dataconf, poselayout=pa17j3d)
h36m = Human36M("B:\\Human3.6M",
dataconf=human36m_dataconf,
poselayout=pa17j3d,
topology='frames')
penn_sf = PennAction("D:\\PennAction",
pennaction_dataconf,
poselayout=pa17j3d,
topology='frames',
use_gt_bbox=True)
ntu_sf = Ntu("E:\\Bachelorarbeit-SS20\\datasets\\NTU",
ntu_pe_dataconf,
poselayout=pa17j3d,
topology='frames',
use_gt_bbox=True)
"""Create an object to load data from all datasets."""
data_tr = BatchLoader([mpii, h36m, penn_sf, ntu_sf], ['frame'], ['pose'],
TRAIN_MODE,
batch_size=[
batch_size_mpii, batch_size_mpii, batch_size_ar,
batch_size_ar
],
num_predictions=num_predictions,
shuffle=True)
"""MPII validation samples."""
mpii_val = BatchLoader(mpii, ['frame'], ['pose', 'afmat', 'headsize'],
VALID_MODE,
batch_size=mpii.get_length(VALID_MODE),
action_weight = 0.1
batch_size_mpii = 3
batch_size_h36m = 4
batch_size_ntu = 8 #1
batch_clips = 4 # 8/4
"""Load datasets"""
mpii = MpiiSinglePerson(datasetpath('MPII'),
dataconf=mpii_dataconf,
poselayout=pa17j3d)
# h36m = Human36M(datasetpath('Human3.6M'), dataconf=human36m_dataconf,
# poselayout=pa17j3d, topology='frames')
ntu_sf = Ntu(datasetpath('NTU'),
ntu_pe_dataconf,
poselayout=pa17j3d,
topology='frames',
use_gt_bbox=True)
ntu = Ntu(datasetpath('NTU'),
ntu_dataconf,
poselayout=pa17j3d,
topology='sequences',
use_gt_bbox=True,
clip_size=num_frames)
ntu_s1 = Ntu(datasetpath('NTU'),
ntu_dataconf,
poselayout=pa17j3d,
topology='sequences',
use_gt_bbox=True,
model = action.build_merge_model(model_pe,
num_actions,
input_shape,
num_frames,
num_joints,
num_blocks,
pose_dim=pose_dim,
num_context_per_joint=0,
pose_net_version='v2')
"""Load pre-trained model."""
# weights_path = get_file(weights_file, TF_WEIGHTS_PATH, md5_hash=md5_hash,
# cache_subdir='models')
weights_path = os.path.join(TF_WEIGHTS_PATH, weights_file)
model.load_weights(weights_path)
"""Load NTU dataset."""
ntu = Ntu('../../datasets/NTU',
ntu_dataconf,
poselayout=pa20j3d,
topology='sequences',
use_gt_bbox=True,
clip_size=num_frames)
ntu_te = BatchLoader(ntu, ['frame'], ['ntuaction'],
TEST_MODE,
batch_size=1,
shuffle=False)
printcn(OKGREEN, 'Evaluation on NTU single-clip using GT bboxes')
eval_singleclip_generator(model, ntu_te)
cfg = ModelConfig(dconf.input_shape,
pa17j3d,
num_pyramids=8,
action_pyramids=[],
num_levels=4)
full_model = spnet.build(cfg)
full_model.load_weights(
'output/pose_baseline_3dp_02_94226e0/weights_posebaseline_060.hdf5')
"""Squeeze the model for only one output."""
model = Model(full_model.input, full_model.outputs[-1])
model.summary()
"""Load dataset"""
ntu = Ntu(datasetpath('NTU'),
dconf,
poselayout=pa17j3d,
topology='frames',
use_gt_bbox=False)
def predict_frame_bboxes(mode):
bboxes = {}
num_samples = ntu.get_length(mode)
for i in range(num_samples):
printnl('mode %d: %07d/%07d' % (mode, i + 1, num_samples))
data = ntu.get_data(i, mode)
poses = model.predict(np.expand_dims(data['frame'], axis=0))
bbox = get_bbox_from_poses(poses, data['afmat'], scale=1.5)
seq_idx = data['seq_idx']
action_weight = 0.1
batch_size_mpii = 3
#batch_size_h36m = 4
batch_size_ntu = 6 #1
batch_clips = 3 # 8/4
"""Load datasets"""
mpii = MpiiSinglePerson("E:\\Bachelorarbeit-SS20\\datasets\\MPII",
dataconf=mpii_dataconf,
poselayout=pa17j3d)
# h36m = Human36M(datasetpath('Human3.6M'), dataconf=human36m_dataconf,
# poselayout=pa17j3d, topology='frames')
ntu_sf = Ntu("E:\\Bachelorarbeit-SS20\\datasets\\NTU",
ntu_pe_dataconf,
poselayout=pa17j3d,
topology='frames',
use_gt_bbox=True)
ntu = Ntu("E:\\Bachelorarbeit-SS20\\datasets\\NTU",
ntu_dataconf,
poselayout=pa17j3d,
topology='sequences',
use_gt_bbox=True,
clip_size=num_frames)
ntu_s1 = Ntu("E:\\Bachelorarbeit-SS20\\datasets\\NTU",
ntu_dataconf,
poselayout=pa17j3d,
topology='sequences',
use_gt_bbox=True,
action_pyramids=[1, 2],
num_levels=4,
pose_replica=False,
num_pose_features=192,
num_visual_features=192)
num_predictions = spnet.get_num_predictions(cfg.num_pyramids, cfg.num_levels)
num_action_predictions = spnet.get_num_predictions(len(cfg.action_pyramids),
cfg.num_levels)
"""Load datasets"""
# h36m = Human36M(datasetpath('Human3.6M'), dataconf=human36m_dataconf,
# poselayout=pa17j3d, topology='frames')
ntu = Ntu("E:\\Bachelorarbeit-SS20\\datasets\\NTU",
ntu_dataconf,
poselayout=pa17j3d,
topology='sequences',
use_gt_bbox=True,
clip_size=num_frames) #, num_S=1)
# print ('WARNING!! USING ONLY S1 FOR EVALUATION!')
"""Build the full model"""
full_model = spnet.build(cfg)
"""Load pre-trained weights from pose estimation and copy replica layers."""
full_model.load_weights(
# 'output/ntu_spnet_trial-03-ft_replica_0ae2bf7/weights_3dp+ntu_ar_062.hdf5',
"E:\\Bachelorarbeit-SS20\\weights\\deephar\\output\\spnet\\0429\\weights_3dp+ntu_ar_048.hdf5",
by_name=True)
"""Split model to simplify evaluation."""
models = split_model(full_model,
cfg,
interlaced=False,
model_names=['3DPose', '3DAction'])
start_lr = 0.01
action_weight = 0.1
batch_size_mpii = 3
#batch_size_h36m = 4
batch_size_ntu = 6 #1
batch_clips = 3 # 8/4
"""Load datasets"""
mpii = MpiiSinglePerson("/home/ispl-ex39/Downloads/deephar-master/datasets/MPII", dataconf=mpii_dataconf,
poselayout=pa17j3d)
# h36m = Human36M(datasetpath('Human3.6M'), dataconf=human36m_dataconf,
# poselayout=pa17j3d, topology='frames')
ntu_sf = Ntu("/home/ispl-ex39/hdd_ext/hdd2000/NTU", ntu_pe_dataconf, poselayout=pa17j3d,
topology='frames', use_gt_bbox=True)
ntu = Ntu("/home/ispl-ex39/hdd_ext/hdd2000/NTU", ntu_dataconf, poselayout=pa17j3d,
topology='sequences', use_gt_bbox=True, clip_size=num_frames)
ntu_s1 = Ntu("/home/ispl-ex39/hdd_ext/hdd2000/NTU", ntu_dataconf, poselayout=pa17j3d,
topology='sequences', use_gt_bbox=True, clip_size=num_frames)
# topology='sequences', use_gt_bbox=True, clip_size=num_frames, num_S=1)
pe_data_tr = BatchLoader([ntu_sf], ['frame'], ['pose'], TRAIN_MODE,
batch_size=[batch_size_ntu],
shuffle=True)
pe_data_tr = BatchLoader(pe_data_tr, ['frame'], ['pose'], TRAIN_MODE,
batch_size=batch_clips, num_predictions=num_predictions, shuffle=False)
ar_data_tr = BatchLoader(ntu, ['frame'], ['ntuaction'], TRAIN_MODE,
cfg = ModelConfig(dconf.input_shape,
pa17j3d,
num_pyramids=8,
action_pyramids=[],
num_levels=4)
full_model = spnet.build(cfg)
full_model.load_weights(
'output/pose_baseline_3dp_02_94226e0/weights_posebaseline_060.hdf5')
"""Squeeze the model for only one output."""
model = Model(full_model.input, full_model.outputs[-1])
model.summary()
"""Load dataset"""
ntu = Ntu("E:\\Bachelorarbeit-SS20\\datasets\\NTU",
dconf,
poselayout=pa17j3d,
topology='frames',
use_gt_bbox=False)
def predict_frame_bboxes(mode):
bboxes = {}
num_samples = ntu.get_length(mode)
for i in range(num_samples):
printnl('mode %d: %07d/%07d' % (mode, i + 1, num_samples))
data = ntu.get_data(i, mode)
poses = model.predict(np.expand_dims(data['frame'], axis=0))
bbox = get_bbox_from_poses(poses, data['afmat'], scale=1.5)
seq_idx = data['seq_idx']