save

aaa_helpers.txt

@@ -127,13 +127,14 @@ nohup python cvae_pyramid.py --name="lhunor_iccv_pyr_32_64_128_256_skin" > nohup
 nohup python cvae_pyramid.py --name="lhunor_iccv_pyr_32_64_128_256_curv" > nohup/iccv/pyr/iccv_pyr_32_64_128_256_curv.out 2>&1 &
 nohup python cvae_pyramid.py --name="lhunor_iccv_pyr_32_64_128_256_curv_skirt" > nohup/iccv/pyr/iccv_pyr_32_64_128_256_curv_skirt.out 2>&1 &
 nohup python cvae_pyramid.py --name="lhunor_iccv_pyr_32_64_128_256_reg" > nohup/iccv/pyr/iccv_pyr_32_64_128_256_reg.out 2>&1 &
-nohup python cvae_pyramid.py --name="lhunor_iccv_pyr_32_64_128_256_seq" > nohup/iccv/pyr/iccv_pyr_32_64_128_256_seq.out 2>&1 &
+nohup python cvae_pyramid.py --name="lhunor_iccv_pyr_32_64_128_256_seq_2_fine" > nohup/iccv/pyr/iccv_pyr_32_64_128_256_seq_2_fine.out 2>&1 &
 
 nohup python cvae_pyramid.py --name="lhunor_iccv_pyr_32_64_128_256_512" > nohup/iccv/pyr/iccv_pyr_32_64_128_256_512.out 2>&1 &
 nohup python cvae_pyramid.py --name="lhunor_iccv_pyr_32_64_128_256_512_skirt" > nohup/iccv/pyr/iccv_pyr_32_64_128_256_512_skirt.out 2>&1 &
 nohup python cvae_pyramid.py --name="lhunor_iccv_pyr_32_64_128_256_512_skin" > nohup/iccv/pyr/iccv_pyr_32_64_128_256_512_skin.out 2>&1 &
 nohup python cvae_pyramid.py --name="lhunor_iccv_pyr_32_64_128_256_512_curv" > nohup/iccv/pyr/iccv_pyr_32_64_128_256_512_curv.out 2>&1 &
 nohup python cvae_pyramid.py --name="lhunor_iccv_pyr_32_64_128_256_512_reg" > nohup/iccv/pyr/iccv_pyr_32_64_128_256_512_reg.out 2>&1 &
+nohup python cvae_pyramid.py --name="lhunor_iccv_pyr_32_64_128_256_512_fine" > nohup/iccv/pyr/iccv_pyr_32_64_128_256_512_fine.out 2>&1 &
 
 nohup python evaluate_ae_pyramid.py --name evaluate_ae_iccv_pyr_32 > nohup/iccv/evaluate/iccv_pyr_32.out 2>&1 &
 nohup python evaluate_ae_pyramid.py --name evaluate_ae_iccv_pyr_32_skirt > nohup/iccv/evaluate/iccv_pyr_32_skirt.out 2>&1 &

configs/iccv/config_abl_pyr_32_64_128_256_seq.py

@@ -1,7 +1,7 @@
 config = dict()
 
-config['gpu'] = '3'
-config['name'] = 'iccv_pyr_32_64_128_256_seq_2'
+config['gpu'] = '1'
+config['name'] = 'iccv_pyr_32_64_128_256_seq_2_fine'
 config['notes'] = ''
 config['tags'] = ['pyr', 'iccv']
 
@@ -22,12 +22,12 @@ config['use_offsets'] =               False
 config['use_norm_loss'] =             True
 config['use_curvature'] =             True
 
-config['pretrained_model'] = ''
+config['pretrained_model'] = 'checkpoints/iccv/iccv_pyr_32_64_128_256_seq_2'
 config['pretrained_model_iteration'] = ''
 
 config['pyramid'] = dict()
 config['pyramid']['resolutions'] = [32, 64, 128, 256]
-config['pyramid']['pretrained_pyramid'] = 'checkpoints/iccv/iccv_pyr_32_64_128_seq_2'
+config['pyramid']['pretrained_pyramid'] = ''
 config['pyramid']['final_upscale_enhancer'] = ''
 
 config['train_dir'] = '/home/workspace/local-data/train'

cvae_pyramid.py

@@ -1,6 +1,6 @@
 import os
 import sys
-from configs.iccv.config_abl_pyr_32_tshirt import config
+from configs.iccv.config_abl_pyr_32_64_128_256_seq import config
 
 os.environ["CUDA_DEVICE_ORDER"]="PCI_BUS_ID"
 os.environ["CUDA_VISIBLE_DEVICES"] = config['gpu']
@@ -221,20 +221,25 @@ def train_step(batch):
         out_levels = model([levels_input, conds], training=True)
 
         if config['use_skinning']:
-            mesh_orig, uv_orig = batch['mesh_unposed'], batch['uv_unposed']
+            mesh_orig = batch['mesh_unposed']
         else:
-            mesh_orig, uv_orig = batch['mesh_posed'], batch['uv_posed']
+            mesh_orig = batch['mesh_posed']
 
         output = dict()
         reconstr_loss_total = 0
         kl_total = 0
 
-        #for i, res in enumerate(config['pyramid']['resolutions']):
-        i = -1
-        res = config['pyramid']['resolutions'][-1]
+        for i, res in enumerate(config['pyramid']['resolutions']):
+        # i = -1
+            res = config['pyramid']['resolutions'][i] if i != len(config['pyramid']['resolutions']) - 1 else 512
             res_modifier = '_' + str(res) if res != 512 else ''
             pred, mean, logvar = out_levels[i]
 
+            if config['use_skinning']:
+                uv_orig = batch['uv_unposed' + res_modifier]
+            else:
+                uv_orig = batch['uv_posed' + res_modifier]
+
             mesh_pred = uv2mesh(pred)
 
             if config['use_skinning']:
@@ -252,7 +257,7 @@ def train_step(batch):
             e_loss = tf.reduce_mean(edge_loss(mesh_pred, mesh_orig, batch['mesh_mask']))
             c_loss, c_ratio = collision_penalty(mesh_pred_posed, batch['mesh_body_posed'], batch['mesh_mask'])
             c_loss, c_ratio = tf.reduce_mean(c_loss), tf.reduce_mean(c_ratio)
-        l2_loss = tf.reduce_mean(1e3 * tf.reduce_mean(tf.reduce_sum(((uv_orig - pred) ** 2), axis=-1), axis=[1, 2]))
+            l2_loss = tf.reduce_mean(1e3 * tf.reduce_mean(tf.reduce_sum(((uv_orig - pred) ** 2), axis=-1) * batch['uv_mask' + res_modifier], axis=[1, 2]))
             # ms_ssim = tf.reduce_mean(tf.image.ssim_multiscale(pred + 5, batch['uv_unposed'] + 5, max_val=10, power_factors=(0.5, 0.2, 0.1, 0.1, 0.1)))
             # ms_ssim = tf.reduce_mean(tf.image.ssim(pred + 1, batch['uv_unposed'] + 1, max_val=3))
             l2_mesh_loss = tf.reduce_mean(1e3 * tf.reduce_mean(tf.reduce_sum(batch['mesh_mask'][:,:,tf.newaxis] * (batch['mesh_posed'] - mesh_pred_posed) ** 2, axis=-1), axis=-1))
@@ -278,8 +283,8 @@ def train_step(batch):
 
             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/var' + res_modifier: np.e ** logvar, 'train/l2_loss' + res_modifier: l2_loss, 'train/normal_loss' + res_modifier: n_loss,
+                                    'train/euclid_error' + res_modifier: metric, 'train/l2_mesh_loss' + res_modifier: l2_mesh_loss } } # 'train/ms_ssim' + res_modifier: ms_ssim,
 
         loss = reconstr_loss_total + balancing_coeff * kl_total
 
@@ -295,20 +300,25 @@ def valid_step(batch):
 
 
     if config['use_skinning']:
-        mesh_orig, uv_orig = batch['mesh_unposed'], batch['uv_unposed']
+        mesh_orig = batch['mesh_unposed']
     else:
-        mesh_orig, uv_orig = batch['mesh_posed'], batch['uv_posed']
+        mesh_orig = batch['mesh_posed']
 
     output = dict()
     reconstr_loss_total = 0
     kl_total = 0
 
-    #for i, res in enumerate(config['pyramid']['resolutions']):
-    i = -1
-    res = config['pyramid']['resolutions'][-1]
+    for i, res in enumerate(config['pyramid']['resolutions']):
+    # i = -1
+        res = config['pyramid']['resolutions'][i] if i != len(config['pyramid']['resolutions']) - 1 else 512
         res_modifier = '_' + str(res) if res != 512 else ''
         pred, mean, logvar = out_levels[i]
 
+        if config['use_skinning']:
+            uv_orig = batch['uv_unposed' + res_modifier]
+        else:
+            uv_orig = batch['uv_posed' + res_modifier]
+
         mesh_pred = uv2mesh(pred)
 
         if config['use_skinning']:
@@ -325,7 +335,7 @@ def valid_step(batch):
         e_loss = tf.reduce_mean(edge_loss(mesh_pred, mesh_orig, batch['mesh_mask']))
         c_loss, c_ratio = collision_penalty(mesh_pred_posed, batch['mesh_body_posed'], batch['mesh_mask'])
         c_loss, c_ratio = tf.reduce_mean(c_loss), tf.reduce_mean(c_ratio)
-    l2_loss = tf.reduce_mean(1e3 * tf.reduce_mean(tf.reduce_sum(((uv_orig - pred) ** 2), axis=-1), axis=[1, 2]))
+        l2_loss = tf.reduce_mean(1e3 * tf.reduce_mean(tf.reduce_sum(((uv_orig - pred) ** 2), axis=-1) * batch['uv_mask' + res_modifier], axis=[1, 2]))
         # ms_ssim = tf.reduce_mean(tf.image.ssim_multiscale(pred + 5, batch['uv_unposed'] + 5, max_val=10, power_factors=(0.5, 0.2, 0.1, 0.1, 0.1)))
         # ms_ssim = tf.reduce_mean(tf.image.ssim(pred + 1, batch['uv_unposed'] + 1, max_val=3))
         l2_mesh_loss = tf.reduce_mean(1e3 * tf.reduce_mean(tf.reduce_sum(batch['mesh_mask'][:,:,tf.newaxis] * (batch['mesh_posed'] - mesh_pred_posed) ** 2, axis=-1), axis=-1))
@@ -351,8 +361,8 @@ def valid_step(batch):
 
         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,
+                                    'valid/var' + res_modifier: np.e ** logvar, 'valid/l2_loss' + res_modifier: l2_loss, 'valid/normal_loss' + res_modifier: n_loss,
+                                    'valid/euclid_error' + res_modifier: metric, 'valid/l2_mesh_loss' + res_modifier: l2_mesh_loss } } # 'valid/ms_ssim' + res_modifier: ms_ssim,
 
     loss = reconstr_loss_total + balancing_coeff * kl_total
 
@@ -852,8 +862,8 @@ if __name__ == '__main__':
     #     load_model(model.upscale_enhancer, config['pyramid']['final_upscale_enhancer'])
     #     model.upscale_enhancer.trainable = False
 
-    for i in range(len(model.levels)-1):
-        model.levels[i].trainable = False
+    # for i in range(len(model.levels)-1):
+    #     model.levels[i].trainable = False
 
     # load_model(model.cond_encoder, config['pyramid']['pretrained_pyramid'])
     # load_model(model.levels[0], config['pyramid']['pretrained_pyramid'])

models/pyramid_model_optimized.py

@@ -163,8 +163,8 @@ class PyramidNew(tf.keras.Model):
             pred = pred + levels_input[0][2] # uv_body_posed_32
         if 1 != len(self.levels):
             pred = self.levels[0].enhance(pred, levels_input[0][-2], training=training)  # levels_input[i][-2]=uv_mask_res
-            pred = tf.image.resize(pred, (self.levels[1].size, self.levels[1].size//2), tf.image.ResizeMethod.BILINEAR)
             out_levels.append([pred, mean, logvar])
+            pred = tf.image.resize(pred, (self.levels[1].size, self.levels[1].size//2), tf.image.ResizeMethod.BILINEAR)
             pred = pred * levels_input[1][-2][:,:,:,tf.newaxis]
         else:
             # pred = self.enhance(pred, levels_input[0][-2], training=training)