-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathedit.py
160 lines (144 loc) · 7.3 KB
/
edit.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
import os
from os.path import join as pjoin
import torch
import numpy as np
import random
from models.AE import AE_models
from models.MARDM import MARDM_models
from utils.motion_process import recover_from_ric, plot_3d_motion, kit_kinematic_chain, t2m_kinematic_chain
import argparse
def main(args):
#################################################################################
# Seed #
#################################################################################
torch.backends.cudnn.benchmark = False
random.seed(args.seed)
np.random.seed(args.seed)
torch.manual_seed(args.seed)
# setting this to true significantly increase training and sampling speed
torch.backends.cuda.matmul.allow_tf32 = False
torch.backends.cudnn.allow_tf32 = False
#################################################################################
# Data #
#################################################################################
dim_pose = 64 if args.dataset_name == 'kit' else 67
nb_joints = 21 if args.dataset_name == 'kit' else 22
data_root = f'{args.dataset_dir}/KIT-ML/' if args.dataset_name == 'kit' else f'{args.dataset_dir}/HumanML3D/'
mean = np.load(pjoin(data_root, 'Mean.npy'))
std = np.load(pjoin(data_root, 'Std.npy'))
motion = np.load(args.source_motion)
m_length = len(motion)
motion = (motion[:, :dim_pose] - mean) / std
#################################################################################
# Models #
#################################################################################
model_dir = pjoin(args.checkpoints_dir, args.dataset_name, args.name, 'model')
result_dir = pjoin('./editing', args.name)
os.makedirs(result_dir, exist_ok=True)
ae = AE_models[args.ae_model](input_width=dim_pose)
ckpt = torch.load(pjoin(args.checkpoints_dir, args.dataset_name, args.ae_name, 'model',
'latest.tar' if args.dataset_name == 't2m' else 'net_best_fid.tar'), map_location='cpu')
model_key = 'ae'
ae.load_state_dict(ckpt[model_key])
ema_mardm = MARDM_models[args.model](ae_dim=ae.output_emb_width, cond_mode='text')
model_dir = pjoin(model_dir, 'latest.tar')
checkpoint = torch.load(model_dir, map_location='cpu')
missing_keys2, unexpected_keys2 = ema_mardm.load_state_dict(checkpoint['ema_mardm'], strict=False)
assert len(unexpected_keys2) == 0
assert all([k.startswith('clip_model.') for k in missing_keys2])
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
#################################################################################
# Sampling #
#################################################################################
prompt_list = []
length_list = []
if args.motion_length == 0:
args.motion_length = m_length
print("Using default motion length.")
prompt_list.append(args.text_prompt)
length_list.append(args.motion_length)
if args.text_prompt == "":
raise "Using an empty text prompt."
ae.to(device)
ema_mardm.to(device)
ae.eval()
ema_mardm.eval()
token_lens = torch.LongTensor(length_list) // 4
token_lens = token_lens.to(device).long()
m_length = token_lens * 4
captions = prompt_list
print_captions = captions[0]
_edit_slice = args.mask_edit_section
edit_slice = []
for eds in _edit_slice:
_start, _end = eds.split(',')
_start = eval(_start)
_end = eval(_end)
edit_slice.append([_start, _end])
kinematic_chain = kit_kinematic_chain if args.dataset_name == 'kit' else t2m_kinematic_chain
motion = torch.from_numpy(motion)[None].to(device)
with torch.no_grad():
latents = ae.encode(motion)
edit_mask = torch.zeros_like(latents)[:, 0, :]
seq_len = latents.shape[-1]
for _start, _end in edit_slice:
if isinstance(_start, float):
_start = int(_start*seq_len)
_end = int(_end*seq_len)
else:
_start //= 4
_end //= 4
edit_mask[:, _start: _end] = 1
print_captions = f'{print_captions} [{_start*4/20.}s - {_end*4/20.}s]'
edit_mask = torch.zeros_like(latents)[:, 0, :]
seq_len = latents.shape[-1]
for _start, _end in edit_slice:
if isinstance(_start, float):
_start = int(_start * seq_len)
_end = int(_end * seq_len)
else:
_start //= 4
_end //= 4
edit_mask[:, _start: _end] = 1
print_captions = f'{print_captions} [{_start * 4 / 20.}s - {_end * 4 / 20.}s]'
edit_mask = edit_mask.bool()
for r in range(args.repeat_times):
print("-->Repeat %d" % r)
with torch.no_grad():
pred_latents = ema_mardm.edit(
captions, latents.clone(), m_length // 4, timesteps=args.time_steps, cond_scale=args.cfg,
temperature=args.temperature, force_mask=False, edit_mask=edit_mask.clone(),
hard_pseudo_reorder=args.hard_pseudo_reorder)
pred_motions = ae.decode(pred_latents)
pred_motions = pred_motions.detach().cpu().numpy()
data = pred_motions * std + mean
for k, (caption, joint_data) in enumerate(zip(captions, data)):
print("---->Sample %d: %s %d" % (k, caption, m_length[k]))
s_path = pjoin(result_dir, str(k))
os.makedirs(s_path, exist_ok=True)
joint_data = joint_data[:m_length[k]]
joint = recover_from_ric(torch.from_numpy(joint_data).float(), nb_joints).numpy()
save_path = pjoin(s_path, "editing_caption:%s_sample%d_repeat%d_len%d.mp4" % (caption, k, r, m_length[k]))
plot_3d_motion(save_path, kinematic_chain, joint, title=caption, fps=20)
np.save(pjoin(s_path, "editing_caption:%s_sample%d_repeat%d_len%d.npy" % (caption, k, r, m_length[k])), joint)
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument('--name', type=str, default='MARDM')
parser.add_argument('--ae_name', type=str, default="AE")
parser.add_argument('--ae_model', type=str, default='AE_Model')
parser.add_argument('--model', type=str, default='MARDM-SiT-XL')
parser.add_argument('--dataset_name', type=str, default='t2m')
parser.add_argument('--dataset_dir', type=str, default='./datasets')
parser.add_argument("--seed", type=int, default=3407)
parser.add_argument('--checkpoints_dir', type=str, default='./checkpoints')
parser.add_argument("--time_steps", default=18, type=int)
parser.add_argument("--cfg", default=4.5, type=float)
parser.add_argument("--temperature", default=1, type=float)
parser.add_argument('--text_prompt', default='', type=str)
parser.add_argument('-msec', '--mask_edit_section', nargs='*', type=str)
parser.add_argument('--source_motion', type=str, default='')
parser.add_argument("--motion_length", default=0, type=int)
parser.add_argument("--repeat_times", default=1, type=int)
parser.add_argument('--hard_pseudo_reorder', action="store_true")
arg = parser.parse_args()
main(arg)