luni/cvpr_2023
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

save

Browse Source
This commit is contained in:
luni
2023-07-14 10:12:00 +02:00
parent bb16aaf868
commit 88531da784
16 changed files with 411 additions and 299 deletions
Download Patch File Download Diff File Expand all files Collapse all files
scripts/cvae_pyramid_e2e.py
+87 -41
View File
@@ -11,7 +11,7 @@ parser.add_argument('--debug', action='store_true')
args = vars(parser.parse_args())
if args['debug']:
from configs.iccvw.pyr_64_128_256_512_cape_m_e2e import config
from configs.iccvw.pyr_64_128_256_512_reg_e2e import config
else:
config_module = importlib.import_module(args['config'])
config = config_module.config
@@ -246,34 +246,49 @@ def train_step(batch):
if res != 512:
pred = resizer(pred) * batch['uv_mask'][:,None,:,:]
l2_loss = torch.mean(1e3 * torch.mean(torch.sum(((uv_orig - pred) ** 2), dim=-1), dim=[1, 2])) # levels_input[i][0] = uv_orig
l2_loss = torch.mean(1e3 * torch.mean(torch.sum(((uv_orig - pred) ** 2), dim=-1), dim=[1, 2]))
kl = kl_loss(mean, logvar)
mesh_pred = uv2mesh(pred)
if config['use_extra_losses']:
mesh_pred = uv2mesh(pred)
if config['use_skinning']:
shape = batch['pose_shape'][:, 72:]
gender = torch.greater_equal(shape[:, 10] , 0.5).to(dtype=torch.int64)
shape_offsets = torch.stack([torch.tensordot(shape[i,:10][None, :], smpl.shapedirs[gender[i]], dims=[[1], [2]])[0,smpl.to_keep,:] for i in range(shape.shape[0])])
if config['use_skinning']:
shape = batch['pose_shape'][:, 72:]
gender = torch.greater_equal(shape[:, 10] , 0.5).to(dtype=torch.int64)
shape_offsets = torch.stack([torch.tensordot(shape[i,:10][None, :], smpl.shapedirs[gender[i]], dims=[[1], [2]])[0,smpl.to_keep,:] for i in range(shape.shape[0])])
mesh_pred_shaped = mesh_pred + shape_offsets * batch['mesh_fix']['mesh_mask'][:,:,None]
mesh_pred_posed = smpl(batch['pose_shape'][:, :72], batch['pose_shape'][:, 72:], mesh_pred_shaped)[0]
mesh_pred_shaped = mesh_pred + shape_offsets * batch['mesh_fix']['mesh_mask'][:,:,None]
mesh_pred_posed = smpl(batch['pose_shape'][:, :72], batch['pose_shape'][:, 72:], mesh_pred_shaped)[0]
else:
mesh_pred_posed = mesh_pred
e_loss = torch.mean(edge_loss(mesh_pred, mesh_orig, batch['mesh_fix']['mesh_mask']))
c_loss, c_ratio = collision_loss(mesh_pred_posed, batch['mesh_body_posed'], batch['mesh_fix']['mesh_mask'])
c_loss, c_ratio = torch.mean(c_loss), torch.mean(c_ratio)
n_loss = normal_loss(mesh_pred_posed, batch['mesh_posed'], batch['mesh_fix']['mesh_mask'], is_skirt=config['is_skirt'], is_smpl_orig=config['is_cape'])
l2_mesh_loss = torch.mean(1e3 * torch.mean(torch.sum(batch['mesh_fix']['mesh_mask'][:,:,None] * (batch['mesh_posed'] - mesh_pred_posed) ** 2, dim=-1), dim=-1))
reconstr_loss_total = reconstr_loss_total + config['pyramid']['loss_weights'][i] * (l2_loss + l2_mesh_loss)
if not config['is_cape']:
reconstr_loss_total = reconstr_loss_total + config['pyramid']['loss_weights'][i] * (e_loss + c_loss) * config['reg_loss_coeff']
if config['use_norm_loss'] and step > 2000:
reconstr_loss_total = reconstr_loss_total + n_loss * config['reg_loss_coeff']
output = { **output, **{ 'train/edge_loss' + res_modifier: e_loss, 'train/collision_loss' + res_modifier: c_loss,
'train/mean' + res_modifier: mean, 'train/logvar' + res_modifier: logvar, 'train/kl_loss' + res_modifier: kl,
'train/var' + res_modifier: np.e ** logvar, 'train/l2_loss' + res_modifier: l2_loss, 'train/normal_loss': n_loss,
'train/l2_mesh_loss': l2_mesh_loss, 'train/collision_ratio': c_ratio } }
else:
mesh_pred_posed = mesh_pred
reconstr_loss_total = reconstr_loss_total + config['pyramid']['loss_weights'][i] * (l2_loss)
e_loss = torch.mean(edge_loss(mesh_pred, mesh_orig, batch['mesh_fix']['mesh_mask']))
c_loss, c_ratio = collision_loss(mesh_pred_posed, batch['mesh_body_posed'], batch['mesh_fix']['mesh_mask'])
c_loss, c_ratio = torch.mean(c_loss), torch.mean(c_ratio)
n_loss = normal_loss(mesh_pred_posed, batch['mesh_posed'], batch['mesh_fix']['mesh_mask'], is_skirt=config['is_skirt'], is_smpl_orig=config['is_cape'])
l2_mesh_loss = torch.mean(1e3 * torch.mean(torch.sum(batch['mesh_fix']['mesh_mask'][:,:,None] * (batch['mesh_posed'] - mesh_pred_posed) ** 2, dim=-1), dim=-1))
reconstr_loss_total = reconstr_loss_total + config['pyramid']['loss_weights'][i] * (l2_loss + l2_mesh_loss)
if not config['is_cape']:
reconstr_loss_total = reconstr_loss_total + config['pyramid']['loss_weights'][i] * (e_loss + c_loss) * config['reg_loss_coeff']
if config['use_norm_loss'] and step > 2000:
reconstr_loss_total = reconstr_loss_total + config['pyramid']['loss_weights'][i] * n_loss * config['reg_loss_coeff']
kl_total += kl
model.eval()
with torch.inference_mode():
mean = torch.mean(mean)
logvar = torch.mean(logvar)
loss: torch.Tensor = reconstr_loss_total + g_balancing_coeff * kl_total
optimizer.zero_grad()
@@ -282,17 +297,45 @@ def train_step(batch):
model.eval()
with torch.inference_mode():
mean = torch.mean(mean)
logvar = torch.mean(logvar)
for i, res in enumerate(config['pyramid']['resolutions']):
# i = -1
res = config['pyramid']['resolutions'][i]
res_modifier = '_' + str(res) if res != 512 else ''
pred, mean, logvar = out_levels[i]
if res != 512:
pred = resizer(pred) * batch['uv_mask'][:,None,:,:]
ms_ssim = torch.mean(_MSSSIM(pred + 1, uv_orig + 1))
metric = torch.sum(torch.sqrt(torch.sum(batch['mesh_fix']['mesh_mask'][:,:,None] * (mesh_orig - mesh_pred.detach()) ** 2, axis=-1))) / torch.sum(batch['mesh_fix']['mesh_mask'])
mean = torch.mean(mean)
logvar = torch.mean(logvar)
output = { **output, **{ 'train/edge_loss' + res_modifier: e_loss, 'train/collision_loss' + res_modifier: c_loss,'train/euclid_error' + res_modifier: metric,
'train/mean' + res_modifier: mean, 'train/logvar' + res_modifier: logvar, 'train/kl_loss' + res_modifier: kl,
'train/var' + res_modifier: np.e ** logvar, 'train/l2_loss' + res_modifier: l2_loss, 'train/normal_loss': n_loss,
'train/euclid_error': metric, 'train/l2_mesh_loss': l2_mesh_loss, 'train/ms_ssim' + res_modifier: ms_ssim,
'train/collision_ratio': c_ratio } } #
if not config['use_extra_losses']:
mesh_pred = uv2mesh(pred)
if config['use_skinning']:
shape = batch['pose_shape'][:, 72:]
gender = torch.greater_equal(shape[:, 10] , 0.5).to(dtype=torch.int64)
shape_offsets = torch.stack([torch.tensordot(shape[i,:10][None, :], smpl.shapedirs[gender[i]], dims=[[1], [2]])[0,smpl.to_keep,:] for i in range(shape.shape[0])])
mesh_pred_shaped = mesh_pred + shape_offsets * batch['mesh_fix']['mesh_mask'][:,:,None]
mesh_pred_posed = smpl(batch['pose_shape'][:, :72], batch['pose_shape'][:, 72:], mesh_pred_shaped)[0]
else:
mesh_pred_posed = mesh_pred
e_loss = torch.mean(edge_loss(mesh_pred, mesh_orig, batch['mesh_fix']['mesh_mask']))
c_loss, c_ratio = collision_loss(mesh_pred_posed, batch['mesh_body_posed'], batch['mesh_fix']['mesh_mask'])
c_loss, c_ratio = torch.mean(c_loss), torch.mean(c_ratio)
l2_mesh_loss = torch.mean(1e3 * torch.mean(torch.sum(batch['mesh_fix']['mesh_mask'][:,:,None] * (batch['mesh_posed'] - mesh_pred_posed) ** 2, dim=-1), dim=-1))
n_loss = normal_loss(mesh_pred_posed, batch['mesh_posed'], batch['mesh_fix']['mesh_mask'], is_skirt=config['is_skirt'], is_smpl_orig=config['is_cape'])
output = { **output, **{ 'train/edge_loss' + res_modifier: e_loss, 'train/collision_loss' + res_modifier: c_loss,
'train/mean' + res_modifier: mean, 'train/logvar' + res_modifier: logvar, 'train/kl_loss' + res_modifier: kl,
'train/var' + res_modifier: np.e ** logvar, 'train/l2_loss' + res_modifier: l2_loss, 'train/normal_loss': n_loss,
'train/l2_mesh_loss': l2_mesh_loss, 'train/collision_ratio': c_ratio } }
ms_ssim = torch.mean(_MSSSIM(pred + 1, uv_orig + 1))
metric = torch.sum(torch.sqrt(torch.sum(batch['mesh_fix']['mesh_mask'][:,:,None] * (mesh_orig - mesh_pred.detach()) ** 2, axis=-1))) / torch.sum(batch['mesh_fix']['mesh_mask'])
output = { **output, **{ 'train/euclid_error' + res_modifier: metric, 'train/ms_ssim' + res_modifier: ms_ssim } }
return { **{ 'train/loss': loss, 'train/reconstr_loss_total': reconstr_loss_total, 'train/kl_total': kl_total }, **output }
@@ -356,18 +399,21 @@ def valid_step(batch):
reconstr_loss_total = reconstr_loss_total + config['pyramid']['loss_weights'][i] * (l2_loss)
kl_total += kl
mean = torch.mean(mean)
logvar = torch.mean(logvar)
ms_ssim = torch.mean(_MSSSIM(pred + 1, uv_orig + 1))
metric = torch.sum(torch.sqrt(torch.sum(batch['mesh_fix']['mesh_mask'][:,:,None] * (mesh_orig - mesh_pred) ** 2, axis=-1))) / torch.sum(batch['mesh_fix']['mesh_mask'])
output = { **output, **{ 'valid/edge_loss' + res_modifier: e_loss, 'valid/collision_loss' + res_modifier: c_loss,
'valid/mean' + res_modifier: mean, 'valid/logvar' + res_modifier: logvar, 'valid/kl_loss' + res_modifier: kl,
'valid/var' + res_modifier: np.e ** logvar, 'valid/l2_loss' + res_modifier: l2_loss, 'valid/normal_loss': n_loss,
'valid/euclid_error' + res_modifier: metric, 'valid/l2_mesh_loss': l2_mesh_loss, 'valid/ms_ssim' + res_modifier: ms_ssim, } } #
loss = reconstr_loss_total + g_balancing_coeff * kl_total
mean = torch.mean(mean)
logvar = torch.mean(logvar)
ms_ssim = torch.mean(_MSSSIM(pred + 1, uv_orig + 1))
metric = torch.sum(torch.sqrt(torch.sum(batch['mesh_fix']['mesh_mask'][:,:,None] * (mesh_orig - mesh_pred) ** 2, axis=-1))) / torch.sum(batch['mesh_fix']['mesh_mask'])
output = { **output, **{ 'valid/edge_loss' + res_modifier: e_loss, 'valid/collision_loss' + res_modifier: c_loss,'valid/euclid_error' + res_modifier: metric,
'valid/mean' + res_modifier: mean, 'valid/logvar' + res_modifier: logvar, 'valid/kl_loss' + res_modifier: kl,
'valid/var' + res_modifier: np.e ** logvar, 'valid/l2_loss' + res_modifier: l2_loss, 'valid/normal_loss': n_loss,
'valid/euclid_error': metric, 'valid/l2_mesh_loss': l2_mesh_loss, 'valid/ms_ssim' + res_modifier: ms_ssim, } } #