save
This commit is contained in:
+78
-59
@@ -1,7 +1,8 @@
|
||||
import os
|
||||
import sys
|
||||
from configs.iccv.config_abl_pyr_32_64_128_256_512 import config
|
||||
from configs.iccv.config_abl_pyr_32_tshirt import config
|
||||
|
||||
os.environ["CUDA_DEVICE_ORDER"]="PCI_BUS_ID"
|
||||
os.environ["CUDA_VISIBLE_DEVICES"] = config['gpu']
|
||||
import tensorflow as tf
|
||||
|
||||
@@ -417,6 +418,7 @@ def test_step(batch):
|
||||
|
||||
|
||||
def plot_everything(batch: dict, type: str):
|
||||
nr_to_plot = 4
|
||||
if config['use_skinning']:
|
||||
mesh_orig, uv_orig = batch['mesh_unposed'].numpy(), batch['uv_unposed'].numpy()
|
||||
else:
|
||||
@@ -426,8 +428,8 @@ def plot_everything(batch: dict, type: str):
|
||||
output_levels = model(levels_input, training=False)
|
||||
|
||||
mesh_pred_levels = [uv2mesh(output_levels[i][0]).numpy() for i in range(len(output_levels))]
|
||||
mesh_orig_posed = batch['mesh_posed'].numpy()
|
||||
body = batch['mesh_body_posed'].numpy()
|
||||
mesh_orig_posed = batch['mesh_posed'].numpy()[:nr_to_plot]
|
||||
body = batch['mesh_body_posed'].numpy()[:nr_to_plot]
|
||||
|
||||
orig_img = (uv_orig + 1) / 2
|
||||
uv_static_img = (batch['uv_static'].numpy() + 1) / 2
|
||||
@@ -436,10 +438,10 @@ def plot_everything(batch: dict, type: str):
|
||||
pred_img = np.concatenate([tf.image.resize(output_levels[i][0], (512, 256), tf.image.ResizeMethod.LANCZOS5) for i in range(len(output_levels))], axis=2)
|
||||
pred_img = (pred_img - pred_img.min(axis=(1,2,3))[:,np.newaxis,np.newaxis,np.newaxis]) / (pred_img.max(axis=(1,2,3))[:,np.newaxis,np.newaxis,np.newaxis] - pred_img.min(axis=(1,2,3))[:,np.newaxis,np.newaxis,np.newaxis])
|
||||
img = tf.concat((uv_static_img, uv_body_posed_img, uv_curvature, orig_img, pred_img), axis=2)
|
||||
imgs = [wandb.Image(img[i]) for i in range(4)]
|
||||
imgs = [wandb.Image(img[i]) for i in range(nr_to_plot)]
|
||||
|
||||
F = np.load("custom_utils/smpl_faces" + skirt_modifier + ("_orig" if config['is_cape'] else "") + ".npy")
|
||||
mask = batch['mesh_mask'].numpy().astype(bool)
|
||||
mask = batch['mesh_mask'].numpy().astype(bool)[:nr_to_plot]
|
||||
_, faces = maskFacesBatch(np.ones_like(mesh_orig), F, mask)
|
||||
|
||||
if config['use_skinning']:
|
||||
@@ -448,30 +450,30 @@ def plot_everything(batch: dict, type: str):
|
||||
shape_offsets = tf.stack([tf.gather(tf.tensordot(shape[i,:10][tf.newaxis, :], smpl.shapedirs[gender[i]], axes=[[1], [2]])[0,:,:], smpl.to_keep, axis=0) for i in range(shape.shape[0])]).numpy()
|
||||
|
||||
|
||||
gt = [mesh_orig[i][mask[i], :] + shape_offsets[i, mask[i], :] for i in range(mask.shape[0])]
|
||||
gt = [mesh_orig[i][mask[i], :] + shape_offsets[i, mask[i], :] for i in range(nr_to_plot)]
|
||||
|
||||
mesh_unposed_shaped_levels = [[] for _ in range(mask.shape[0])]
|
||||
mesh_unposed_shaped_levels = [[] for _ in range(nr_to_plot)]
|
||||
for i in range(len(output_levels)):
|
||||
mesh_unposed_shaped = [mesh_pred_levels[i][j, mask[j], :] + shape_offsets[j, mask[j]] for j in range(mask.shape[0])]
|
||||
for k in range(mask.shape[0]):
|
||||
mesh_unposed_shaped = [mesh_pred_levels[i][j, mask[j], :] + shape_offsets[j, mask[j]] for j in range(nr_to_plot)]
|
||||
for k in range(nr_to_plot):
|
||||
mesh_unposed_shaped_levels[k].append(mesh_unposed_shaped[k])
|
||||
|
||||
# full res version
|
||||
mesh_unposed_shaped_full = np.zeros((mask.shape[0], config['smpl_res'], 3), dtype=np.float32)
|
||||
for i in range(mask.shape[0]):
|
||||
mesh_unposed_shaped_full = np.zeros((nr_to_plot, config['smpl_res'], 3), dtype=np.float32)
|
||||
for i in range(nr_to_plot):
|
||||
mesh_unposed_shaped_full[i, mask[i], :] = mesh_unposed_shaped[i]
|
||||
|
||||
mesh_posed = smpl.run_garment(batch['pose_shape'][:, :72], batch['pose_shape'][:, 72:], mesh_unposed_shaped_full).numpy()
|
||||
mesh_posed = smpl.run_garment(batch['pose_shape'][:nr_to_plot, :72], batch['pose_shape'][:nr_to_plot, 72:], mesh_unposed_shaped_full).numpy()
|
||||
else:
|
||||
mesh_posed = mesh_pred_levels[-1]
|
||||
gt = [mesh_orig[i][mask[i], :] for i in range(mesh_posed.shape[0])]
|
||||
mesh_unposed_shaped_levels = [[mesh_pred_levels[i][j, mask[j], :] for i in range(len(output_levels))] for j in range(mask.shape[0])]
|
||||
gt = [mesh_orig[i][mask[i], :] for i in range(nr_to_plot)]
|
||||
mesh_unposed_shaped_levels = [[mesh_pred_levels[i][j, mask[j], :] for i in range(len(output_levels))] for j in range(nr_to_plot)]
|
||||
|
||||
|
||||
gt_posed = [mesh_orig_posed[i, mask[i], :] for i in range(mask.shape[0])]
|
||||
gt_posed = [mesh_orig_posed[i, mask[i], :] for i in range(nr_to_plot)]
|
||||
mesh_posed_fixed = fix_collision(mesh_posed, body, mask).numpy()
|
||||
|
||||
mesh_posed = [mesh_posed[j, mask[j], :] for j in range(mask.shape[0])]
|
||||
mesh_posed = [mesh_posed[j, mask[j], :] for j in range(nr_to_plot)]
|
||||
|
||||
if config['is_cape']:
|
||||
collision = mesh_posed
|
||||
@@ -487,11 +489,11 @@ def plot_everything(batch: dict, type: str):
|
||||
if config['is_cape']:
|
||||
figs_coll = dict()
|
||||
else:
|
||||
garment_posed_fixed = [mesh_posed_fixed[i, mask[i], :] for i in range(body.shape[0])]
|
||||
garment_posed_fixed = [mesh_posed_fixed[i, mask[i], :] for i in range(nr_to_plot)]
|
||||
collision_fixed = [np.concatenate((body[i], garment_posed_fixed[i]), axis=0) for i in range(len(mesh_posed))]
|
||||
gt_posed = [batch['mesh_posed'].numpy()[i, mask[i], :] for i in range(mask.shape[0])]
|
||||
gt_posed = [batch['mesh_posed'].numpy()[i, mask[i], :] for i in range(nr_to_plot)]
|
||||
collision_gt = [np.concatenate((body[i], gt_posed[i]), axis=0) for i in range(len(gt_posed))]
|
||||
figs_coll = { type + "/garment_" + str(i) + '_collision': plot_3d_collision(collision_gt[i], collision[i], collision_fixed[i], collision_faces[i], collision_color[i]) for i in range(4) }
|
||||
figs_coll = { type + "/garment_" + str(i) + '_collision': plot_3d_collision(collision_gt[i], collision[i], collision_fixed[i], collision_faces[i], collision_color[i]) for i in range(nr_to_plot) }
|
||||
|
||||
return { **figs, **figs_coll, **{ type + '/images': imgs, } }
|
||||
|
||||
@@ -728,14 +730,23 @@ def generate_results():
|
||||
pass
|
||||
|
||||
@tf.function
|
||||
def speed_step(levels_input, conds):
|
||||
def speed_step(levels_input, conds, pose_shape, mask):
|
||||
shape = pose_shape[:, 72:]
|
||||
gender = tf.cast(tf.greater_equal(shape[:, 10] , 0.5), tf.int32)
|
||||
shape_offsets = tf.stack([tf.gather(tf.tensordot(shape[i,:10][tf.newaxis, :], smpl.shapedirs[gender[i]], axes=[[1], [2]])[0,:,:], smpl.to_keep, axis=0) for i in range(shape.shape[0])])
|
||||
|
||||
return model.sample([levels_input, conds])
|
||||
# pred = model.sample([levels_input, conds])
|
||||
out_levels = model([levels_input, conds], training=False)
|
||||
pred = out_levels[-1][0]
|
||||
mesh_pred = uv2mesh(pred)
|
||||
mesh_pred_shaped = mesh_pred + shape_offsets * mask[:,:,tf.newaxis]
|
||||
mesh_pred_posed = smpl.run_garment(pose_shape[:, :72], pose_shape[:, 72:], mesh_pred_shaped)
|
||||
return mesh_pred_posed
|
||||
|
||||
|
||||
|
||||
if __name__ == '__main__':
|
||||
# init_wandb()
|
||||
init_wandb()
|
||||
parser = argparse.ArgumentParser(description='')
|
||||
parser.add_argument('--name', type=str, default='lhunor')
|
||||
args = vars(parser.parse_args())
|
||||
@@ -760,8 +771,12 @@ if __name__ == '__main__':
|
||||
resolutions=dataset_resolutions,
|
||||
in_memory=config['dataset_in_memory'],
|
||||
is_cape=config['is_cape'],
|
||||
frames_file=config['frames_file'] if 'frames_file' in config.keys() else None)
|
||||
frames_file=config['frames_file'] if 'frames_file' in config.keys() else None,
|
||||
frames_file_skirt=config['frames_file'] if 'frames_file' in config.keys() else '',
|
||||
folder_skirt=config['train_dir_skirt'] if 'train_dir_skirt' in config.keys() else '',
|
||||
)
|
||||
static_dataset_train = StaticDataset(folder=config['train_dir'],
|
||||
folder_skirt=config['train_dir_skirt'] if 'train_dir_skirt' in config.keys() else '',
|
||||
signature=config['signature_static'],
|
||||
resolutions=dataset_resolutions,
|
||||
in_memory=True,
|
||||
@@ -772,8 +787,12 @@ if __name__ == '__main__':
|
||||
resolutions=dataset_resolutions,
|
||||
in_memory=config['dataset_in_memory'],
|
||||
is_cape=config['is_cape'],
|
||||
frames_file=config['frames_file_valid'] if 'frames_file_valid' in config.keys() else None)
|
||||
frames_file=config['frames_file_valid'] if 'frames_file_valid' in config.keys() else None,
|
||||
frames_file_skirt=config['frames_file_valid'] if 'frames_file_valid' in config.keys() else '',
|
||||
folder_skirt=config['valid_dir_skirt'] if 'valid_dir_skirt' in config.keys() else '',
|
||||
)
|
||||
static_dataset_valid = StaticDataset(folder=config['valid_dir'],
|
||||
folder_skirt=config['valid_dir_skirt'] if 'valid_dir_skirt' in config.keys() else '',
|
||||
signature=config['signature_static'],
|
||||
resolutions=dataset_resolutions,
|
||||
in_memory=True,
|
||||
@@ -848,50 +867,50 @@ if __name__ == '__main__':
|
||||
|
||||
levels_input, conds = group_inputs_for_levels(batch, config)
|
||||
|
||||
start = time.perf_counter()
|
||||
out = speed_step(levels_input, conds)
|
||||
t = time.perf_counter() - start
|
||||
print(t)
|
||||
times.append(t)
|
||||
# start = time.perf_counter()
|
||||
# out = speed_step(levels_input, conds, batch['pose_shape'] ,batch['mesh_mask'])
|
||||
# t = time.perf_counter() - start
|
||||
# print(t)
|
||||
# times.append(t)
|
||||
|
||||
if step > 200:
|
||||
break
|
||||
# if step > 200:
|
||||
# break
|
||||
|
||||
|
||||
# train_loss = train_step(batch)
|
||||
# train_accumulator.update(train_loss)
|
||||
train_loss = train_step(batch)
|
||||
train_accumulator.update(train_loss)
|
||||
|
||||
# if step % 25 == 0:
|
||||
# data_valid = valid_iterator.get_next()
|
||||
# data_valid = static_dataset_valid.expand_batch(batch_to_dict(data_valid, config['signature_dynamic']))
|
||||
# valid_loss = valid_step(data_valid)
|
||||
# valid_accumulator.update(valid_loss)
|
||||
if step % 25 == 0:
|
||||
data_valid = valid_iterator.get_next()
|
||||
data_valid = static_dataset_valid.expand_batch(batch_to_dict(data_valid, config['signature_dynamic']))
|
||||
valid_loss = valid_step(data_valid)
|
||||
valid_accumulator.update(valid_loss)
|
||||
|
||||
# if train_loss['train/euclid_error'] < 0.015:
|
||||
# if balancing_activated_step == 0:
|
||||
# balancing_activated_step = step
|
||||
# balancing_coeff = balancing_coeff + 1 / 20000
|
||||
if train_loss['train/euclid_error'] < 0.015:
|
||||
if balancing_activated_step == 0:
|
||||
balancing_activated_step = step
|
||||
balancing_coeff = balancing_coeff + 1 / 20000
|
||||
|
||||
# if step % 1000 == 0 or (step < 1000 and step % 100 == 0):
|
||||
# to_log = dict({ **train_loss, **valid_loss, **{ 'current_step': step } })
|
||||
# if step % 5000 == 0 or (step < 5000 and step % 1000 == 0):
|
||||
# train_figs = plot_everything(batch, 'train')
|
||||
# valid_figs = plot_everything(data_valid, 'valid')
|
||||
# to_log = { **to_log, **train_figs, **valid_figs }
|
||||
# if step == 0:
|
||||
# to_log = { **to_log, **{ 'model/summary': wandb.Html(get_summary_html(model)) } }
|
||||
if step % 1000 == 0 or (step < 1000 and step % 100 == 0):
|
||||
to_log = dict({ **train_loss, **valid_loss, **{ 'current_step': step } })
|
||||
if step % 5000 == 0 or (step < 5000 and step % 1000 == 0):
|
||||
train_figs = plot_everything(batch, 'train')
|
||||
valid_figs = plot_everything(data_valid, 'valid')
|
||||
to_log = { **to_log, **train_figs, **valid_figs }
|
||||
if step == 0:
|
||||
to_log = { **to_log, **{ 'model/summary': wandb.Html(get_summary_html(model)) } }
|
||||
|
||||
# to_log = { **to_log, **{ 'train/balancing_coeff': balancing_coeff, 'train/gamma': gamma } }
|
||||
to_log = { **to_log, **{ 'train/balancing_coeff': balancing_coeff, 'train/gamma': gamma } }
|
||||
|
||||
# wandb.log(to_log, step=step)
|
||||
# train_accumulator.reset()
|
||||
# valid_accumulator.reset()
|
||||
wandb.log(to_log, step=step)
|
||||
train_accumulator.reset()
|
||||
valid_accumulator.reset()
|
||||
|
||||
# if step % 5000 == 0:
|
||||
# manager.save(checkpoint_number=step)
|
||||
if step % 5000 == 0:
|
||||
manager.save(checkpoint_number=step)
|
||||
|
||||
# print(step)
|
||||
print(step)
|
||||
|
||||
step += 1
|
||||
print('mean:')
|
||||
print(np.mean(times[10:]))
|
||||
# print('mean:')
|
||||
# print(np.mean(times[10:]))
|
||||
Reference in New Issue
Block a user