OAR.2068279.stdout

Namespace(batch_size=64, beta=None, ckpt_dir='checkpoints', ckpt_name='last', cont_capacity=None, dataset='rendered_chairs', disc_capacity=None, epochs=400, experiment_name='VAE_bs_64_ls_15', gpu_devices=[0, 1], is_beta_VAE=False, latent_name='', latent_spec_cont=15, latent_spec_disc=None, load_expe_name='', load_model_checkpoint=False, lr=0.0001, num_worker=4, print_loss_every=50, record_loss_every=50, save_model=True, save_reconstruction_image=False, save_step=1, verbose=True)
creare new diretory experiment: rendered_chairs/VAE_bs_64_ls_15
load dataset: rendered_chairs, with: 69120 train images of shape: (3, 64, 64)
use 2 gpu who named:
GeForce RTX 2080 Ti
GeForce RTX 2080 Ti
DataParallel(
  (module): VAE(
    (img_to_last_conv): Sequential(
      (0): Conv2d(3, 32, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
      (1): ReLU()
      (2): Conv2d(32, 32, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
      (3): ReLU()
      (4): Conv2d(32, 64, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
      (5): ReLU()
      (6): Conv2d(64, 64, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
      (7): ReLU()
    )
    (last_conv_to_continuous_features): Sequential(
      (0): Conv2d(64, 256, kernel_size=(4, 4), stride=(1, 1))
      (1): ReLU()
    )
    (features_to_hidden_continue): Sequential(
      (0): Linear(in_features=256, out_features=30, bias=True)
      (1): ReLU()
    )
    (latent_to_features): Sequential(
      (0): Linear(in_features=15, out_features=256, bias=True)
      (1): ReLU()
    )
    (features_to_img): Sequential(
      (0): ConvTranspose2d(256, 64, kernel_size=(4, 4), stride=(1, 1))
      (1): ReLU()
      (2): ConvTranspose2d(64, 64, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
      (3): ReLU()
      (4): ConvTranspose2d(64, 32, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
      (5): ReLU()
      (6): ConvTranspose2d(32, 32, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
      (7): ReLU()
      (8): ConvTranspose2d(32, 3, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
      (9): Sigmoid()
    )
  )
)
The number of parameters of model is 769185
don't use continuous capacity
=> no checkpoint found at 'trained_models/rendered_chairs/VAE_bs_64_ls_15/checkpoints/last'
0/69092	Loss: 3015.823
3200/69092	Loss: 2867.494
6400/69092	Loss: 899.528
9600/69092	Loss: 536.005
12800/69092	Loss: 478.343
16000/69092	Loss: 455.459
19200/69092	Loss: 457.437
22400/69092	Loss: 370.021
25600/69092	Loss: 263.633
28800/69092	Loss: 232.440
32000/69092	Loss: 210.459
35200/69092	Loss: 217.661
38400/69092	Loss: 216.086
41600/69092	Loss: 215.461
44800/69092	Loss: 208.221
48000/69092	Loss: 205.981
51200/69092	Loss: 208.176
54400/69092	Loss: 204.615
57600/69092	Loss: 205.887
60800/69092	Loss: 203.826
64000/69092	Loss: 202.141
67200/69092	Loss: 198.343
Training time 0:03:51.208936
Epoch: 1 Average loss: 427.71
=> saved checkpoint 'trained_models/rendered_chairs/VAE_bs_64_ls_15/checkpoints/last' (iter 1)
0/69092	Loss: 187.281
3200/69092	Loss: 200.427
6400/69092	Loss: 196.083
9600/69092	Loss: 202.029
12800/69092	Loss: 196.254
16000/69092	Loss: 196.466
19200/69092	Loss: 195.587
22400/69092	Loss: 192.062
25600/69092	Loss: 197.137
28800/69092	Loss: 196.870
32000/69092	Loss: 193.763
35200/69092	Loss: 196.194
38400/69092	Loss: 193.444
41600/69092	Loss: 186.353
44800/69092	Loss: 184.125
48000/69092	Loss: 179.607
51200/69092	Loss: 181.214
54400/69092	Loss: 179.105
57600/69092	Loss: 173.470
60800/69092	Loss: 163.793
64000/69092	Loss: 163.068
67200/69092	Loss: 164.580
Training time 0:03:46.574199
Epoch: 2 Average loss: 186.57
=> saved checkpoint 'trained_models/rendered_chairs/VAE_bs_64_ls_15/checkpoints/last' (iter 2)
0/69092	Loss: 164.363
3200/69092	Loss: 157.112
6400/69092	Loss: 152.674
9600/69092	Loss: 154.297
12800/69092	Loss: 155.036
16000/69092	Loss: 151.871
19200/69092	Loss: 151.537
22400/69092	Loss: 152.374
25600/69092	Loss: 150.578
28800/69092	Loss: 152.244
32000/69092	Loss: 150.801
35200/69092	Loss: 147.880
38400/69092	Loss: 148.003
41600/69092	Loss: 147.567