import os from configs.config_pyramid import config os.environ["CUDA_VISIBLE_DEVICES"] = config['gpu'] import tensorflow as tf gpus = tf.config.experimental.list_physical_devices('GPU') if gpus: try: # Currently, memory growth needs to be the same across GPUs for gpu in gpus: tf.config.experimental.set_memory_growth(gpu, True) logical_gpus = tf.config.experimental.list_logical_devices('GPU') print(len(gpus), "Physical GPUs,", len(logical_gpus), "Logical GPUs") except RuntimeError as e: # Memory growth must be set before GPUs have been initialized print(e) import time import argparse import wandb import random import numpy as np import scipy as sp from custom_utils.custom_utils import UV2Mesh, maskFacesBatch, _maskFaces import plotly.graph_objects as go import plotly.express as ex from plotly.subplots import make_subplots from PIL import Image from tf_utils import check_tf_version, allow_memory_growth, split_gpu_for_testing from losses import edge_loss, uv_mask_loss, bend_loss, kl_loss # from losses import collision_penalty as collision_penalty from losses import collision_loss as collision_penalty from custom_utils.custom_utils import batch_to_dict, load_model from dataset import DatasetFactory, StaticDataset from custom_utils.smpl import SMPL from models.pyramid_model import Pyramid def init_wandb(): run = wandb.init(project='vae', name=config['name'], notes=config['notes']) artifact = wandb.Artifact('code', type='code') artifact.add_file('cvae_pyramid.py') artifact.add_file('models/cvae_model.py') artifact.add_file('models/pyramid_model.py') artifact.add_file('models/pyramid_level.py') artifact.add_file('models/upscaler.py') artifact.add_file('losses.py') artifact.add_file('configs/config_base.py') artifact.add_file('configs/config_cvae.py') artifact.add_file('models/base.py') artifact.add_file('models/cond_encoder.py') artifact.add_file('dataset.py') artifact.add_file('convnext/models/convnext_tf.py') run.log_artifact(artifact) uv2mesh = UV2Mesh(res=256) def plot_3d(mesh_gt, mesh_pred_unposed_levels, faces, mesh_posed, collision, collision_faces, collision_color): fig = make_subplots(rows=1, cols=1+len(mesh_pred_unposed_levels)+1+1, start_cell="top-left", specs = [ [{ 'type': 'scene' } for _ in range(1+len(mesh_pred_unposed_levels)+1+1)] ]) fig.add_trace(go.Mesh3d( x=mesh_gt[:,0], y=mesh_gt[:,2], z=mesh_gt[:,1], i = faces[:,0], j = faces[:,1], k = faces[:,2], text="Ground Truth" ), row=1, col=1) for i, mesh in enumerate(mesh_pred_unposed_levels): fig.add_trace(go.Mesh3d( x=mesh[:,0], y=mesh[:,2], z=mesh[:,1], i = faces[:,0], j = faces[:,1], k = faces[:,2], text='Unposed level_' + str(i) ), row=1, col=i+2) fig.add_trace(go.Mesh3d( x=mesh_posed[:,0], y=mesh_posed[:,1], z=mesh_posed[:,2], i = faces[:,0], j = faces[:,1], k = faces[:,2], text='Posed GT' ), row=1, col=1+len(mesh_pred_unposed_levels)+1) fig.add_trace(go.Mesh3d( x=collision[:,0], y=collision[:,2], z=collision[:,1], i = collision_faces[:,0], j = collision_faces[:,1], k = collision_faces[:,2], vertexcolor=collision_color, text='Posed output on body' ), row=1, col=1+len(mesh_pred_unposed_levels)+1+1) if len(mesh_pred_unposed_levels) == 3: fig.update_layout(scene_aspectmode='manual', scene_aspectratio=dict(x=1, y=1, z=1), scene2_aspectmode='manual', scene2_aspectratio=dict(x=1, y=1, z=1), scene3_aspectmode='manual', scene3_aspectratio=dict(x=1, y=1, z=1), scene4_aspectmode='manual', scene4_aspectratio=dict(x=1, y=1, z=1), scene5_aspectmode='manual', scene5_aspectratio=dict(x=1, y=1, z=1), scene6_aspectmode='manual', scene6_aspectratio=dict(x=1, y=1, z=1,) ) helper = dict(xaxis = dict(nticks=4, range=[-1,1],), yaxis = dict(nticks=4, range=[-1,1],), zaxis = dict(nticks=4, range=[-1,1],),) fig.update_layout(scene = helper, scene2 = helper, scene3 = helper, scene4 = helper, scene5 = helper, scene6 = helper, ) if len(mesh_pred_unposed_levels) == 2: fig.update_layout(scene_aspectmode='manual', scene_aspectratio=dict(x=1, y=1, z=1), scene2_aspectmode='manual', scene2_aspectratio=dict(x=1, y=1, z=1), scene3_aspectmode='manual', scene3_aspectratio=dict(x=1, y=1, z=1), scene4_aspectmode='manual', scene4_aspectratio=dict(x=1, y=1, z=1), scene5_aspectmode='manual', scene5_aspectratio=dict(x=1, y=1, z=1), ) helper = dict(xaxis = dict(nticks=4, range=[-1,1],), yaxis = dict(nticks=4, range=[-1,1],), zaxis = dict(nticks=4, range=[-1,1],),) fig.update_layout(scene = helper, scene2 = helper, scene3 = helper, scene4 = helper, scene5 = helper, ) return fig def group_inputs_for_levels(batch): levels_input = [] for res in config['pyramid']['resolutions']: res_modifier = '_' + str(res) if res != 512 else '' levels_input.append([batch['uv_unposed' + res_modifier], batch['uv_static' + res_modifier], batch['uv_body_posed' + res_modifier], batch['uv_curvature' + res_modifier], batch['uv_mask' + res_modifier], batch['uv_mask']]) return levels_input @tf.function def train_step(batch): with tf.GradientTape() as tape: levels_input = group_inputs_for_levels(batch) out_levels = model(levels_input, training=True) shape = batch['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])]) mesh_orig = batch['mesh_unposed'] output = dict() reconstr_loss_total = 0 kl_total = 0 for i, res in enumerate(config['pyramid']['resolutions']): res_modifier = '_' + str(res) if res != 512 else '' pred, mean, logvar = out_levels[i] mesh_pred = uv2mesh(pred) mesh_pred_shaped = mesh_pred + shape_offsets * batch['mesh_mask'][:,:,tf.newaxis] mesh_pred_posed = smpl.run_garment(batch['pose_shape'][:, :72], batch['pose_shape'][:, 72:], mesh_pred_shaped) e_loss = tf.reduce_mean(edge_loss(mesh_pred, mesh_orig, batch['mesh_mask'])) c_loss = tf.reduce_mean(collision_penalty(mesh_pred_posed, batch['mesh_body_posed'], batch['mesh_mask'])) l2_loss = tf.reduce_mean(1e3 * tf.reduce_mean(tf.reduce_sum(((batch['uv_unposed'] - pred) ** 2), axis=-1), axis=[1, 2])) ms_ssim = tf.reduce_mean(tf.image.ssim_multiscale(pred + 1, batch['uv_unposed'] + 1, max_val=3, power_factors=(0.5, 0.2, 0.1, 0.1, 0.1))) kl = kl_loss(mean, logvar) reconstr_loss_total += l2_loss #+ e_loss + c_loss kl_total += kl mean = tf.reduce_mean(mean) logvar = tf.reduce_mean(logvar) metric = tf.reduce_sum(tf.math.sqrt(tf.reduce_sum(batch['mesh_mask'][:,:,tf.newaxis] * (mesh_orig - mesh_pred) ** 2, axis=-1))) / tf.reduce_sum(batch['mesh_mask']) output = { **output, **{ 'train/edge_loss' + res_modifier: e_loss, 'train/collision_loss' + res_modifier: c_loss,'train/euclid_error' + res_modifier: metric, 'train/ms_ssim' + res_modifier: ms_ssim, '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/euclid_error': metric } } loss = reconstr_loss_total + balancing_coeff * kl_total gradients = tape.gradient(loss, model.trainable_variables) optimizer.apply_gradients(zip(gradients, model.trainable_variables)) return { **{ 'train/loss': loss, 'train/reconstr_loss_total': reconstr_loss_total, 'train/kl_total': kl_total }, **output } @tf.function def valid_step(batch): uv_curvature = tf.one_hot(batch['uv_curvature'], 7) pred, mean, logvar, z = model([batch['uv_unposed'], batch['uv_static'], batch['uv_body_posed'], uv_curvature, batch['uv_mask'], batch['uv_shaped_body']]) mesh_pred = uv2mesh(tf.transpose(pred, [0,3,1,2])) mesh_orig = batch['mesh_unposed'] e_loss = 1e-2 * tf.reduce_mean(edge_loss(mesh_pred, mesh_orig, batch['mesh_mask'])) c_loss = 1e-2 * tf.reduce_mean(collision_penalty(mesh_pred, batch['mesh_body_posed'], batch['mesh_mask'])) l1_loss = tf.reduce_sum(tf.reduce_sum((tf.math.abs(tf.transpose(batch['uv_unposed'], [0,2,3,1]) - pred)), axis=-1) * batch['uv_mask'], axis=[1, 2]) b_loss = 0 #1e-3 * bend_loss(mesh_pred, smpl_faces, laplacian_loss, batch['mesh_mask']) # ms_ssim = tf.image.ssim_multiscale(tf.transpose(pred, [0,2,3,1]) + 1, tf.transpose(batch['uv_unposed'], [0,2,3,1]) + 1, max_val=2) ms_ssim = tf.image.ssim_multiscale(pred + 1, tf.transpose(batch['uv_unposed'], [0,2,3,1]) + 1, max_val=3) fft_loss = 0 #tf.reduce_sum(tf.math.abs(tf.signal.fft3d(tf.complex(pred, tf.zeros_like(pred))) - tf.signal.fft3d(tf.complex(tf.transpose(batch['uv_unposed'], [0,2,3,1]), tf.zeros_like(tf.transpose(batch['uv_unposed'], [0,2,3,1]))))), axis=[1,2,3]) # mesh_orig_2 = uv2mesh(batch['uv_unposed']) metric = tf.reduce_sum(tf.math.sqrt(tf.reduce_sum(batch['mesh_mask'][:,:,tf.newaxis] * (mesh_orig - mesh_pred) ** 2, axis=-1))) / tf.reduce_sum(batch['mesh_mask']) metric_2 = 0 #tf.reduce_sum(tf.math.sqrt(tf.reduce_sum(batch['mesh_mask'][:,:,tf.newaxis] * (mesh_orig_2 - mesh_pred) ** 2, axis=-1))) / tf.reduce_sum(batch['mesh_mask']) reconstr_loss = l1_loss # logpx_z = -reconstr_loss # logpz = log_normal_pdf(z[:,:model.latent_dim], 0., 0.) # logqz_x = log_normal_pdf(z[:,:model.latent_dim], mean, logvar) # loss = -tf.reduce_mean(logpx_z + balancing_coeff * (logpz - logqz_x)) kl = kl_loss(mean, logvar) # loss = reconstr_loss / (2 * (gamma ** 2)) + kl loss = reconstr_loss + balancing_coeff * kl ms_ssim = tf.reduce_mean(ms_ssim) l1_loss = tf.reduce_mean(l1_loss) fft_loss = tf.reduce_mean(fft_loss) b_loss = tf.reduce_mean(b_loss) reconstr_loss = tf.reduce_mean(reconstr_loss) mean = tf.reduce_mean(mean) logvar = tf.reduce_mean(logvar) return { 'valid/loss': loss, 'valid/edge_loss': e_loss, 'valid/collision_loss': c_loss, 'valid/l1_loss': l1_loss, 'valid/euclid_error': metric, 'valid/ms_ssim': ms_ssim, 'valid/fft_loss': fft_loss, 'valid/euclid_error_reconst': metric_2, 'train/bend_loss': b_loss, 'train/reconstr_loss': reconstr_loss, 'valid/mean': mean, 'valid/logvar': logvar, 'valid/kl_loss': kl, 'valid/var': np.e ** logvar } def plot_everything(batch: dict, type: str): levels_input = group_inputs_for_levels(batch) 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 = batch['mesh_unposed'].numpy() mesh_orig_posed = batch['mesh_posed'].numpy() body = batch['mesh_body_posed'].numpy() orig_img = batch['uv_unposed'] uv_static_img = batch['uv_static'] uv_body_posed_img = batch['uv_body_posed'] uv_curvature = np.tile(batch['uv_curvature'].numpy().argmax(-1)[:,:,:,np.newaxis], [1,1,1,3]) / 2 - 1 pred_img = np.concatenate([output_levels[i][0] for i in range(len(output_levels))], axis=2) img = tf.concat((uv_static_img, uv_body_posed_img, uv_curvature, orig_img, pred_img), axis=2) imgs = [wandb.Image((img[i] + 1) / 2.0) for i in range(4)] shape = batch['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])]).numpy() F = np.load("custom_utils/smpl_faces.npy") mask = batch['mesh_mask'].numpy().astype(bool) _, faces = maskFacesBatch(np.ones_like(mesh_orig), F, mask) gt = [mesh_orig[i][mask[i], :] + shape_offsets[i, mask[i], :] for i in range(mask.shape[0])] gt_posed = [mesh_orig_posed[i, mask[i], :] for i in range(mask.shape[0])] mesh_unposed_shaped_levels = [[] for _ in range(mask.shape[0])] 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_levels[k].append(mesh_unposed_shaped[k]) # full res version mesh_unposed_shaped_full = np.zeros((mask.shape[0], 14475, 3), dtype=np.float32) for i in range(mask.shape[0]): 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 = [mesh_posed[j, mask[j], :] for j in range(mask.shape[0])] collision = [np.concatenate((body[i], mesh_posed[i]), axis=0) for i in range(len(mesh_posed))] collision_faces = [np.concatenate((F, faces[i] + 14475), axis=0) for i in range(len(mesh_posed))] collision_color = [np.concatenate((np.ones_like(body[i], dtype=np.int32)*100, np.ones_like(mesh_posed[i], dtype=np.int32)*230), axis=0) for i in range(len(mesh_posed))] figs = { type + "/garment_" + str(i): plot_3d(gt[i], mesh_unposed_shaped_levels[i], faces[i], gt_posed[i], collision[i], collision_faces[i], collision_color[i]) for i in range(4) } return { **figs, **{ type + '/images': imgs, } } def sample(folder): # run = wandb.init(project='vae_plots', name=folder.split('/')[-1]) load_model(model,folder) count = 0 for batch in dataset_train: batch = static_dataset_train.expand_batch(batch_to_dict(batch, config['signature_dynamic'])) uv_curvature = tf.one_hot(batch['uv_curvature'], 7) nr_of_samples=2 pred = model.sample(batch['uv_static'][0], batch['uv_body_posed'][0], uv_curvature[0], batch['uv_mask'][0], nr_of_samples=nr_of_samples) pose, shape = batch['pose_shape'][0, :72], batch['pose_shape'][0, 72:] pose, shape = tf.tile(pose[tf.newaxis,:], [nr_of_samples,1]), tf.tile(shape[tf.newaxis,:], [nr_of_samples,1]) mesh_pred = uv2mesh(tf.transpose(pred, [0,3,1,2])) mesh_posed = smpl.run_garment(pose, shape, mesh_pred).numpy() mesh_orig = batch['mesh_posed'][0].numpy() mesh_orig_reposed = smpl.run_garment(pose[:1], shape[:1], batch['mesh_unposed'][:1]).numpy()[0] body = batch['mesh_body_posed'][0].numpy() body_color = np.ones_like(body, dtype=np.int32)*100 F = np.load("custom_utils/smpl_faces.npy") mask = batch['mesh_mask'].numpy().astype(bool)[0] _, faces = _maskFaces(np.ones((14475, 3)), F, mask) pred_zero = model.sample(batch['uv_static'][0], batch['uv_body_posed'][0], uv_curvature[0], batch['uv_mask'][0], nr_of_samples=nr_of_samples, zero_latent=True) mesh_pred_zero = uv2mesh(tf.transpose(pred_zero, [0,3,1,2])) mesh_posed_zero = smpl.run_garment(pose, shape, mesh_pred_zero).numpy() meshes = [] meshes.append(go.Mesh3d( # 8 vertices of a cube x=mesh_orig_reposed[mask,0], y=mesh_orig_reposed[mask,2], z=mesh_orig_reposed[mask,1], # i, j and k give the vertices of triangles i = faces[:,0], j = faces[:,1], k = faces[:,2] )) for i in range(nr_of_samples): x = mesh_posed[i][mask, :].copy() # x[:, 2] += 1 * (i + 1) x[:, 2] += 1 meshes.append(go.Mesh3d( # 8 vertices of a cube x=x[:,0], y=x[:,2], z=x[:,1], # i, j and k give the vertices of triangles i = faces[:,0], j = faces[:,1], k = faces[:,2] )) for i in range(nr_of_samples): x = mesh_posed_zero[i][mask, :].copy() # x[:, 2] += 1 * (i + 1) x[:, 2] += 2 meshes.append(go.Mesh3d( # 8 vertices of a cube x=x[:,0], y=x[:,2], z=x[:,1], # i, j and k give the vertices of triangles i = faces[:,0], j = faces[:,1], k = faces[:,2] )) # for i in range(5): # x = body.copy() # x[:, 2] += 1 * i # meshes.append(go.Mesh3d( # # 8 vertices of a cube # x=x[:,0], # y=x[:,2], # z=x[:,1], # # i, j and k give the vertices of triangles # i = F[:,0], # j = F[:,1], # k = F[:,2], # vertexcolor=body_color # )) fig = go.Figure(data=meshes) fig.update_layout(scene_aspectmode='manual', scene_aspectratio=dict(x=1, y=2, z=1)) # fig.update_layout(scene_aspectmode='manual', scene_aspectratio=dict(x=1, y=nr_of_samples-1, z=1)) fig.update_layout( scene = dict( xaxis = dict(nticks=4, range=[-1,1],), # yaxis = dict(nticks=4, range=[-1,nr_of_samples+1],), yaxis = dict(nticks=4, range=[-1,3],), zaxis = dict(nticks=4, range=[-1,1],),) ) # wandb.log({'plot_' + str(count): fig}) fig.show() pred = pred.numpy() frames = [Image.fromarray(((pred[i] + 1) * 128).astype(np.int8), mode='RGB') for i in range(pred.shape[0])] frames[0].save("temp/garment_1.gif", format="GIF", append_images=frames[1:], save_all=True, duration=500, loop=0) if count > 10: break print(count) count += 1 print() if __name__ == '__main__': init_wandb() allow_memory_growth() parser = argparse.ArgumentParser(description='') parser.add_argument('--name', type=str) args = vars(parser.parse_args()) smpl = SMPL() factory = DatasetFactory() dataset_train = factory.get_dataset(folder=config['train_dir'], batch_size=config['batch_size'], signature=config['signature_dynamic'], use_filled=config['use_filled'], resolutions=config['pyramid']['resolutions']) static_dataset_train = StaticDataset(folder=config['train_dir'], signature=config['signature_static'], use_filled=config['use_filled'], resolutions=config['pyramid']['resolutions']) dataset_valid = factory.get_dataset(folder=config['valid_dir'], batch_size=config['batch_size'], signature=config['signature_dynamic'], use_filled=config['use_filled'], resolutions=config['pyramid']['resolutions']) static_dataset_valid = StaticDataset(folder=config['valid_dir'], signature=config['signature_static']) model: Pyramid = Pyramid(config) optimizer = tf.optimizers.Adam(learning_rate=config['lr']) balancing_coeff = config['starting_balancing_coeff'] reg_coeff = 1 valid_iterator = iter(dataset_valid) gamma = 1 balancing_activated_step = 0 # sample('checkpoints/cvae_13') ckpt = tf.train.Checkpoint(model=model) manager = tf.train.CheckpointManager(ckpt, 'checkpoints/' + config['name'], max_to_keep=3) load_model(model.cond_encoder, 'checkpoints/cvae_19') load_model(model.levels[0], 'checkpoints/cvae_low_4_3_d32') load_model(model.levels[-1], 'checkpoints/cvae_19') step = 0 for batch in dataset_train: batch = static_dataset_train.expand_batch(batch_to_dict(batch, config['signature_dynamic'])) train_loss = train_step(batch) # data_valid = valid_iterator.get_next() # data_valid = static_dataset_valid.expand_batch(batch_to_dict(data_valid, signature_dynamic)) # valid_loss = valid_step(data_valid) valid_loss = {} to_log = { **train_loss, **valid_loss, **{ 'current_step': step} } if step % 2500 == 0: train_figs = plot_everything(batch, 'train') # valid_figs = plot_everything(data_valid, 'valid') valid_figs = {} to_log = { **to_log, **train_figs, **valid_figs } if step % 10000 == 0: manager.save(checkpoint_number=step) if train_loss['train/euclid_error'] < 0.01: if balancing_activated_step == 0: balancing_activated_step = step balancing_coeff = balancing_coeff + 1 / 15000 to_log = { **to_log, **{ 'train/balancing_coeff': balancing_coeff, 'train/gamma': gamma } } wandb.log(to_log, step = step) print(step) step += 1