loaded library: /usr/lib/x86_64-linux-gnu/libibverbs.so.1
loaded library: /usr/lib/x86_64-linux-gnu/libibverbs.so.1
------------------------ arguments ------------------------
  batch_size ...................................... 512
  batch_size_per_proc ............................. 512
  data_dir ........................................ /dataset/f9f659c5/wdl_ofrecord
  data_part_name_suffix_length .................... 5
  data_part_num ................................... 256
  dataset_format .................................. ofrecord
  ddp ............................................. True
  deep_dropout_rate ............................... 0.5
  deep_embedding_vec_size ......................... 16
  deep_vocab_size ................................. 2322444
  eval_after_training ............................. False
  eval_batchs ..................................... 20
  eval_interval ................................... 0
  execution_mode .................................. eager
  hidden_size ..................................... 1024
  hidden_units_num ................................ 2
  learning_rate ................................... 0.001
  loss_print_every_n_iter ......................... 100
  max_iter ........................................ 1100
  model_load_dir .................................. 
  model_save_dir .................................. ./checkpoint
  num_deep_sparse_fields .......................... 26
  num_dense_fields ................................ 13
  num_wide_sparse_fields .......................... 2
  save_initial_model .............................. False
  save_model_after_each_eval ...................... False
  test_name ....................................... noname_test
  wide_vocab_size ................................. 2322444
-------------------- end of arguments ---------------------
[rank:0] iter: 100/1100, loss: 0.5269403457641602, latency(ms): 44.1529837995767593 | 2022-04-01 01:46:02.766
timestamp, name, driver_version, utilization.gpu [%], utilization.memory [%], memory.total [MiB], memory.free [MiB], memory.used [MiB]
2022/04/01 01:46:02.785, Tesla V100-SXM2-32GB, 470.57.02, 87 %, 81 %, 32510 MiB, 30708 MiB, 1802 MiB
[rank:0] iter: 200/1100, loss: 0.5004009008407593, latency(ms): 4.2022902891039848 | 2022-04-01 01:46:03.186
[rank:0] iter: 300/1100, loss: 0.4958913922309875, latency(ms): 4.0490614622831345 | 2022-04-01 01:46:03.591
[rank:0] iter: 400/1100, loss: 0.4919652938842773, latency(ms): 3.9613037556409836 | 2022-04-01 01:46:03.987
[rank:0] iter: 500/1100, loss: 0.4932500422000885, latency(ms): 4.1327982395887375 | 2022-04-01 01:46:04.400
[rank:0] iter: 600/1100, loss: 0.4817394912242889, latency(ms): 3.9844007417559624 | 2022-04-01 01:46:04.799
[rank:0] iter: 700/1100, loss: 0.4859778285026550, latency(ms): 4.0923856571316719 | 2022-04-01 01:46:05.208
[rank:0] iter: 800/1100, loss: 0.4860964119434357, latency(ms): 4.1314920410513878 | 2022-04-01 01:46:05.621
[rank:0] iter: 900/1100, loss: 0.4846715927124023, latency(ms): 4.1203279793262482 | 2022-04-01 01:46:06.033
[rank:0] iter: 1000/1100, loss: 0.4819317758083344, latency(ms): 4.0466975048184395 | 2022-04-01 01:46:06.438
[rank:0] iter: 1100/1100, loss: 0.4841487407684326, latency(ms): 4.0118299797177315 | 2022-04-01 01:46:06.839