➜  ProxylessNAS git:(master) CUDA_VISIBLE_DEVICES=0 python eval_tf.py -a proxyless_gpu
~/anaconda3/lib/python3.6/site-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.
  from ._conv import register_converters as _register_converters
Downloading: "http://hanlab.mit.edu/files/proxylessNAS/proxyless_gpu.tfinit" to ~/.torch/proxyless_nas/proxyless_gpu.tfinit
WARNING:tensorflow:From ~/Workspace/ProxylessNAS/proxyless_nas_tensorflow/tf_modules.py:35: softmax_cross_entropy_with_logits (from tensorflow.python.ops.nn_ops) is deprecated and will be removed in a future version.
Instructions for updating:

Future major versions of TensorFlow will allow gradients to flow
into the labels input on backprop by default.

See @{tf.nn.softmax_cross_entropy_with_logits_v2}.

2018-12-19 18:38:36.514256: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA
2018-12-19 18:38:36.684471: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1392] Found device 0 with properties:
name: GeForce GTX 1080 major: 6 minor: 1 memoryClockRate(GHz): 1.7335
pciBusID: 0000:05:00.0
totalMemory: 7.92GiB freeMemory: 7.75GiB
2018-12-19 18:38:36.684501: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1471] Adding visible gpu devices: 0
2018-12-19 18:38:37.074122: I tensorflow/core/common_runtime/gpu/gpu_device.cc:952] Device interconnect StreamExecutor with strength 1 edge matrix:
2018-12-19 18:38:37.074155: I tensorflow/core/common_runtime/gpu/gpu_device.cc:958]      0
2018-12-19 18:38:37.074162: I tensorflow/core/common_runtime/gpu/gpu_device.cc:971] 0:   N
2018-12-19 18:38:37.074364: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1084] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 7478 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1080, pci bus id: 0000:05:00.0, compute capability: 6.1)
0 	Loss 1.0945 (1.0945) 	Top 1-acc 79.688 (79.688)
50 	Loss 0.7295 (1.0897) 	Top 1-acc 79.688 (75.429)
100 	Loss 0.9993 (1.0968) 	Top 1-acc 75.000 (74.381)
150 	Loss 1.2386 (1.0955) 	Top 1-acc 71.875 (74.721)
200 	Loss 1.0635 (1.0977) 	Top 1-acc 73.438 (74.495)
250 	Loss 1.2828 (1.0830) 	Top 1-acc 68.750 (74.869)
300 	Loss 1.3761 (1.0883) 	Top 1-acc 65.625 (74.720)
350 	Loss 0.8165 (1.0868) 	Top 1-acc 87.500 (74.666)
400 	Loss 0.8985 (1.0851) 	Top 1-acc 81.250 (74.700)
450 	Loss 1.1681 (1.0837) 	Top 1-acc 75.000 (74.699)
500 	Loss 0.6891 (1.0811) 	Top 1-acc 82.812 (74.797)
550 	Loss 1.0987 (1.0782) 	Top 1-acc 78.125 (74.821)
600 	Loss 1.0207 (1.0788) 	Top 1-acc 76.562 (74.914)
650 	Loss 1.7205 (1.0787) 	Top 1-acc 67.188 (74.940)
700 	Loss 0.9781 (1.0756) 	Top 1-acc 75.000 (74.978)
750 	Loss 1.2338 (1.0745) 	Top 1-acc 75.000 (75.073)
Loss: 1.0737 	Top-1: 75.084