geographconv_cmu.log

Can not use cuDNN on context None: Disabled by dnn.enabled flag
Mapped name None to device cuda0: GeForce GTX 1080 Ti (0000:61:00.0)
11/16/2018 02:06:13 PM In order to work for big datasets fix https://github.com/Theano/Theano/pull/5721 should be applied to theano.
11/16/2018 02:06:13 PM loading the dataset from ./data/cmu/
11/16/2018 02:06:14 PM #labels: 129
11/16/2018 02:06:16 PM TfidfVectorizer(analyzer=u'word', binary=True, decode_error=u'strict',
        dtype='float32', encoding='latin1', input=u'content',
        lowercase=True, max_df=0.2, max_features=None, min_df=10,
        ngram_range=(1, 1), norm='l2', preprocessor=None, smooth_idf=True,
        stop_words='english', strip_accents=None, sublinear_tf=False,
        token_pattern='(?u)(?<![@])#?\\b\\w\\w+\\b', tokenizer=None,
        use_idf=True, vocabulary=None)
/home/felixgwu/.local/lib/python2.7/site-packages/sklearn/feature_extraction/text.py:1089: 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`.
  if hasattr(X, 'dtype') and np.issubdtype(X.dtype, np.float):
11/16/2018 02:06:20 PM training    n_samples: 5685, n_features: 9467
11/16/2018 02:06:20 PM development n_samples: 1895, n_features: 9467
11/16/2018 02:06:20 PM test        n_samples: 1895, n_features: 9467
11/16/2018 02:06:20 PM saving vocab in ./data/cmu/vocab.pkl
11/16/2018 02:06:20 PM vocab dumped successfully!
11/16/2018 02:06:20 PM adding the train graph
11/16/2018 02:06:22 PM adding the dev graph
11/16/2018 02:06:22 PM adding the test graph
11/16/2018 02:06:23 PM removing 91477 celebrity nodes with degree higher than 5
11/16/2018 02:06:23 PM projecting the graph
11/16/2018 02:06:23 PM 0%
11/16/2018 02:06:23 PM 10%
11/16/2018 02:06:23 PM 20%
11/16/2018 02:06:23 PM 30%
11/16/2018 02:06:23 PM 40%
11/16/2018 02:06:23 PM 50%
11/16/2018 02:06:23 PM 60%
11/16/2018 02:06:23 PM 70%
11/16/2018 02:06:23 PM 80%
11/16/2018 02:06:23 PM 90%
11/16/2018 02:06:23 PM 100%
11/16/2018 02:06:23 PM #nodes: 9475, #edges: 77155
11/16/2018 02:06:23 PM creating adjacency matrix...
/home/felixgwu/.local/lib/python2.7/site-packages/scipy/sparse/compressed.py:730: SparseEfficiencyWarning: Changing the sparsity structure of a csr_matrix is expensive. lil_matrix is more efficient.
  SparseEfficiencyWarning)
11/16/2018 02:06:24 PM adjacency matrix created.
11/16/2018 02:06:24 PM dumping data in ./data/cmu/dump.pkl ...
11/16/2018 02:06:26 PM data dump finished!
11/16/2018 02:06:26 PM stacking training, dev and test features and creating indices...
11/16/2018 02:06:26 PM running mlp with graph conv...
11/16/2018 02:06:26 PM highway is True
11/16/2018 02:06:26 PM Graphconv model input size 9467, output size 129 and hidden layers [300, 300, 300] regul 0.0 dropout 0.5.
11/16/2018 02:06:27 PM 3 gconv layers
/home/felixgwu/.conda/envs/geo/lib/python2.7/site-packages/lasagne/layers/helper.py:216: UserWarning: get_output() was called with unused kwargs:
	A
  % "\n\t".join(suggestions))
/share/felixgwu/projects/gcn_project/geographconv/gcnmodel.py:402: UserWarning: theano.function was asked to create a function computing outputs given certain inputs, but the provided input variable at index 1 is not part of the computational graph needed to compute the outputs: SparseVariable{csr,float32}.
To make this warning into an error, you can pass the parameter on_unused_input='raise' to theano.function. To disable it completely, use on_unused_input='ignore'.
  self.f_gates.append(theano.function([self.X_sym, self.A_sym], self.gate_outputs[i], on_unused_input='warn'))
11/16/2018 02:06:48 PM ***********percentile 1.000000 ******************
11/16/2018 02:06:48 PM 5685 training samples
11/16/2018 02:06:48 PM training for 10000 epochs with batch size 500
/home/felixgwu/.conda/envs/geo/lib/python2.7/site-packages/theano/tensor/subtensor.py:2339: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.
  out[0][inputs[2:]] = inputs[1]
11/16/2018 02:06:48 PM epoch 0 train loss 4.86 train acc 0.01 val loss 4.86 val acc 0.01 best val acc 0.01 maxdown 0
11/16/2018 02:06:49 PM epoch 1 train loss 4.85 train acc 0.08 val loss 4.85 val acc 0.03 best val acc 0.03 maxdown 0
11/16/2018 02:06:50 PM epoch 2 train loss 4.83 train acc 0.14 val loss 4.85 val acc 0.04 best val acc 0.04 maxdown 0
11/16/2018 02:06:51 PM epoch 3 train loss 4.82 train acc 0.15 val loss 4.84 val acc 0.04 best val acc 0.04 maxdown 0
11/16/2018 02:06:51 PM epoch 4 train loss 4.80 train acc 0.16 val loss 4.84 val acc 0.03 best val acc 0.03 maxdown 0
11/16/2018 02:06:52 PM epoch 5 train loss 4.79 train acc 0.16 val loss 4.83 val acc 0.04 best val acc 0.04 maxdown 0
11/16/2018 02:06:53 PM epoch 6 train loss 4.77 train acc 0.17 val loss 4.82 val acc 0.04 best val acc 0.04 maxdown 0
11/16/2018 02:06:53 PM epoch 7 train loss 4.75 train acc 0.18 val loss 4.81 val acc 0.04 best val acc 0.04 maxdown 0
11/16/2018 02:06:54 PM epoch 8 train loss 4.73 train acc 0.20 val loss 4.80 val acc 0.04 best val acc 0.04 maxdown 0
11/16/2018 02:06:55 PM epoch 9 train loss 4.71 train acc 0.23 val loss 4.79 val acc 0.06 best val acc 0.06 maxdown 0
11/16/2018 02:06:56 PM epoch 10 train loss 4.69 train acc 0.25 val loss 4.78 val acc 0.08 best val acc 0.08 maxdown 0
11/16/2018 02:06:56 PM epoch 11 train loss 4.67 train acc 0.27 val loss 4.76 val acc 0.10 best val acc 0.10 maxdown 0
11/16/2018 02:06:57 PM epoch 12 train loss 4.65 train acc 0.30 val loss 4.75 val acc 0.11 best val acc 0.11 maxdown 0
11/16/2018 02:06:58 PM epoch 13 train loss 4.62 train acc 0.32 val loss 4.74 val acc 0.12 best val acc 0.12 maxdown 0
11/16/2018 02:06:59 PM epoch 14 train loss 4.59 train acc 0.33 val loss 4.72 val acc 0.13 best val acc 0.13 maxdown 0
11/16/2018 02:06:59 PM epoch 15 train loss 4.56 train acc 0.34 val loss 4.70 val acc 0.14 best val acc 0.14 maxdown 0
11/16/2018 02:07:00 PM epoch 16 train loss 4.53 train acc 0.36 val loss 4.68 val acc 0.15 best val acc 0.15 maxdown 0
11/16/2018 02:07:01 PM epoch 17 train loss 4.50 train acc 0.38 val loss 4.66 val acc 0.15 best val acc 0.15 maxdown 0
11/16/2018 02:07:02 PM epoch 18 train loss 4.47 train acc 0.39 val loss 4.64 val acc 0.16 best val acc 0.16 maxdown 0
11/16/2018 02:07:02 PM epoch 19 train loss 4.44 train acc 0.40 val loss 4.62 val acc 0.17 best val acc 0.17 maxdown 0
11/16/2018 02:07:03 PM epoch 20 train loss 4.40 train acc 0.40 val loss 4.60 val acc 0.17 best val acc 0.17 maxdown 0
11/16/2018 02:07:04 PM epoch 21 train loss 4.37 train acc 0.42 val loss 4.58 val acc 0.17 best val acc 0.17 maxdown 0
11/16/2018 02:07:05 PM epoch 22 train loss 4.33 train acc 0.42 val loss 4.55 val acc 0.18 best val acc 0.18 maxdown 0
11/16/2018 02:07:05 PM epoch 23 train loss 4.30 train acc 0.43 val loss 4.53 val acc 0.18 best val acc 0.18 maxdown 0
11/16/2018 02:07:06 PM epoch 24 train loss 4.26 train acc 0.44 val loss 4.51 val acc 0.18 best val acc 0.18 maxdown 0
11/16/2018 02:07:07 PM epoch 25 train loss 4.22 train acc 0.44 val loss 4.48 val acc 0.18 best val acc 0.18 maxdown 0
11/16/2018 02:07:08 PM epoch 26 train loss 4.18 train acc 0.45 val loss 4.46 val acc 0.20 best val acc 0.20 maxdown 0
11/16/2018 02:07:08 PM epoch 27 train loss 4.15 train acc 0.46 val loss 4.44 val acc 0.20 best val acc 0.20 maxdown 0
11/16/2018 02:07:09 PM epoch 28 train loss 4.11 train acc 0.46 val loss 4.42 val acc 0.20 best val acc 0.20 maxdown 0
11/16/2018 02:07:10 PM epoch 29 train loss 4.07 train acc 0.47 val loss 4.39 val acc 0.20 best val acc 0.20 maxdown 0
11/16/2018 02:07:11 PM epoch 30 train loss 4.04 train acc 0.47 val loss 4.37 val acc 0.19 best val acc 0.19 maxdown 0
11/16/2018 02:07:11 PM epoch 31 train loss 4.00 train acc 0.48 val loss 4.36 val acc 0.20 best val acc 0.20 maxdown 0
11/16/2018 02:07:12 PM epoch 32 train loss 3.97 train acc 0.48 val loss 4.34 val acc 0.20 best val acc 0.20 maxdown 0
11/16/2018 02:07:13 PM epoch 33 train loss 3.93 train acc 0.49 val loss 4.32 val acc 0.20 best val acc 0.20 maxdown 0
11/16/2018 02:07:14 PM epoch 34 train loss 3.90 train acc 0.49 val loss 4.31 val acc 0.20 best val acc 0.20 maxdown 0
11/16/2018 02:07:14 PM epoch 35 train loss 3.87 train acc 0.50 val loss 4.28 val acc 0.21 best val acc 0.21 maxdown 0
11/16/2018 02:07:15 PM epoch 36 train loss 3.83 train acc 0.50 val loss 4.27 val acc 0.21 best val acc 0.21 maxdown 0
11/16/2018 02:07:16 PM epoch 37 train loss 3.80 train acc 0.51 val loss 4.26 val acc 0.21 best val acc 0.21 maxdown 0
11/16/2018 02:07:17 PM epoch 38 train loss 3.76 train acc 0.52 val loss 4.24 val acc 0.22 best val acc 0.22 maxdown 0
11/16/2018 02:07:17 PM epoch 39 train loss 3.73 train acc 0.52 val loss 4.22 val acc 0.21 best val acc 0.21 maxdown 0
11/16/2018 02:07:18 PM epoch 40 train loss 3.69 train acc 0.53 val loss 4.20 val acc 0.22 best val acc 0.22 maxdown 0
11/16/2018 02:07:19 PM epoch 41 train loss 3.65 train acc 0.52 val loss 4.19 val acc 0.22 best val acc 0.22 maxdown 0
11/16/2018 02:07:20 PM epoch 42 train loss 3.61 train acc 0.53 val loss 4.17 val acc 0.23 best val acc 0.23 maxdown 0
11/16/2018 02:07:20 PM epoch 43 train loss 3.58 train acc 0.54 val loss 4.15 val acc 0.23 best val acc 0.23 maxdown 0
11/16/2018 02:07:21 PM epoch 44 train loss 3.54 train acc 0.54 val loss 4.13 val acc 0.23 best val acc 0.23 maxdown 0
11/16/2018 02:07:22 PM epoch 45 train loss 3.50 train acc 0.54 val loss 4.11 val acc 0.23 best val acc 0.23 maxdown 0
11/16/2018 02:07:22 PM epoch 46 train loss 3.46 train acc 0.54 val loss 4.10 val acc 0.24 best val acc 0.24 maxdown 0
11/16/2018 02:07:23 PM epoch 47 train loss 3.41 train acc 0.55 val loss 4.07 val acc 0.23 best val acc 0.23 maxdown 0
11/16/2018 02:07:24 PM epoch 48 train loss 3.38 train acc 0.55 val loss 4.05 val acc 0.24 best val acc 0.24 maxdown 0
11/16/2018 02:07:25 PM epoch 49 train loss 3.34 train acc 0.55 val loss 4.04 val acc 0.23 best val acc 0.23 maxdown 0
11/16/2018 02:07:25 PM epoch 50 train loss 3.29 train acc 0.56 val loss 4.01 val acc 0.24 best val acc 0.24 maxdown 0
11/16/2018 02:07:26 PM epoch 51 train loss 3.25 train acc 0.56 val loss 3.98 val acc 0.24 best val acc 0.24 maxdown 0
11/16/2018 02:07:27 PM epoch 52 train loss 3.21 train acc 0.57 val loss 3.96 val acc 0.24 best val acc 0.24 maxdown 0
11/16/2018 02:07:28 PM epoch 53 train loss 3.16 train acc 0.57 val loss 3.94 val acc 0.24 best val acc 0.24 maxdown 0
11/16/2018 02:07:28 PM epoch 54 train loss 3.12 train acc 0.58 val loss 3.92 val acc 0.24 best val acc 0.24 maxdown 0
11/16/2018 02:07:29 PM epoch 55 train loss 3.07 train acc 0.57 val loss 3.89 val acc 0.24 best val acc 0.24 maxdown 0
11/16/2018 02:07:30 PM epoch 56 train loss 3.03 train acc 0.58 val loss 3.87 val acc 0.25 best val acc 0.25 maxdown 0
11/16/2018 02:07:31 PM epoch 57 train loss 2.98 train acc 0.58 val loss 3.84 val acc 0.24 best val acc 0.24 maxdown 0
11/16/2018 02:07:31 PM epoch 58 train loss 2.94 train acc 0.58 val loss 3.81 val acc 0.24 best val acc 0.24 maxdown 0
11/16/2018 02:07:32 PM epoch 59 train loss 2.89 train acc 0.59 val loss 3.80 val acc 0.24 best val acc 0.24 maxdown 0
11/16/2018 02:07:33 PM epoch 60 train loss 2.85 train acc 0.59 val loss 3.77 val acc 0.23 best val acc 0.23 maxdown 0
11/16/2018 02:07:34 PM epoch 61 train loss 2.80 train acc 0.58 val loss 3.75 val acc 0.24 best val acc 0.24 maxdown 0
11/16/2018 02:07:34 PM epoch 62 train loss 2.75 train acc 0.59 val loss 3.73 val acc 0.25 best val acc 0.25 maxdown 0
11/16/2018 02:07:35 PM epoch 63 train loss 2.70 train acc 0.59 val loss 3.72 val acc 0.25 best val acc 0.25 maxdown 0
11/16/2018 02:07:36 PM epoch 64 train loss 2.66 train acc 0.58 val loss 3.69 val acc 0.23 best val acc 0.23 maxdown 0
11/16/2018 02:07:37 PM epoch 65 train loss 2.61 train acc 0.58 val loss 3.67 val acc 0.24 best val acc 0.24 maxdown 0
11/16/2018 02:07:37 PM epoch 66 train loss 2.57 train acc 0.58 val loss 3.65 val acc 0.24 best val acc 0.24 maxdown 0
11/16/2018 02:07:38 PM epoch 67 train loss 2.52 train acc 0.58 val loss 3.63 val acc 0.25 best val acc 0.25 maxdown 0
11/16/2018 02:07:39 PM epoch 68 train loss 2.48 train acc 0.59 val loss 3.61 val acc 0.25 best val acc 0.25 maxdown 0
11/16/2018 02:07:40 PM epoch 69 train loss 2.44 train acc 0.59 val loss 3.60 val acc 0.24 best val acc 0.24 maxdown 0
11/16/2018 02:07:40 PM epoch 70 train loss 2.40 train acc 0.59 val loss 3.58 val acc 0.25 best val acc 0.25 maxdown 0
11/16/2018 02:07:41 PM epoch 71 train loss 2.36 train acc 0.59 val loss 3.56 val acc 0.25 best val acc 0.25 maxdown 0
11/16/2018 02:07:42 PM epoch 72 train loss 2.31 train acc 0.61 val loss 3.56 val acc 0.25 best val acc 0.25 maxdown 0
11/16/2018 02:07:42 PM epoch 73 train loss 2.28 train acc 0.61 val loss 3.53 val acc 0.24 best val acc 0.24 maxdown 0
11/16/2018 02:07:43 PM epoch 74 train loss 2.24 train acc 0.61 val loss 3.51 val acc 0.26 best val acc 0.26 maxdown 0
11/16/2018 02:07:44 PM epoch 75 train loss 2.20 train acc 0.61 val loss 3.50 val acc 0.25 best val acc 0.25 maxdown 0
11/16/2018 02:07:45 PM epoch 76 train loss 2.16 train acc 0.62 val loss 3.47 val acc 0.25 best val acc 0.25 maxdown 0
11/16/2018 02:07:45 PM epoch 77 train loss 2.13 train acc 0.63 val loss 3.48 val acc 0.25 best val acc 0.25 maxdown 1
11/16/2018 02:07:46 PM epoch 78 train loss 2.09 train acc 0.63 val loss 3.45 val acc 0.26 best val acc 0.26 maxdown 0
11/16/2018 02:07:47 PM epoch 79 train loss 2.05 train acc 0.64 val loss 3.46 val acc 0.26 best val acc 0.26 maxdown 1
11/16/2018 02:07:48 PM epoch 80 train loss 2.01 train acc 0.64 val loss 3.45 val acc 0.26 best val acc 0.26 maxdown 0
11/16/2018 02:07:48 PM epoch 81 train loss 1.98 train acc 0.64 val loss 3.43 val acc 0.25 best val acc 0.25 maxdown 0
11/16/2018 02:07:49 PM epoch 82 train loss 1.94 train acc 0.65 val loss 3.44 val acc 0.26 best val acc 0.25 maxdown 1
11/16/2018 02:07:50 PM epoch 83 train loss 1.91 train acc 0.66 val loss 3.42 val acc 0.26 best val acc 0.26 maxdown 0
11/16/2018 02:07:51 PM epoch 84 train loss 1.87 train acc 0.66 val loss 3.42 val acc 0.25 best val acc 0.26 maxdown 1
11/16/2018 02:07:51 PM epoch 85 train loss 1.84 train acc 0.66 val loss 3.43 val acc 0.25 best val acc 0.26 maxdown 2
11/16/2018 02:07:52 PM epoch 86 train loss 1.81 train acc 0.67 val loss 3.42 val acc 0.26 best val acc 0.26 maxdown 3
11/16/2018 02:07:53 PM epoch 87 train loss 1.77 train acc 0.67 val loss 3.42 val acc 0.25 best val acc 0.26 maxdown 4
11/16/2018 02:07:54 PM epoch 88 train loss 1.74 train acc 0.67 val loss 3.43 val acc 0.26 best val acc 0.26 maxdown 5
11/16/2018 02:07:54 PM epoch 89 train loss 1.71 train acc 0.68 val loss 3.43 val acc 0.25 best val acc 0.26 maxdown 6
11/16/2018 02:07:55 PM epoch 90 train loss 1.68 train acc 0.68 val loss 3.43 val acc 0.25 best val acc 0.26 maxdown 7
11/16/2018 02:07:56 PM epoch 91 train loss 1.65 train acc 0.69 val loss 3.45 val acc 0.25 best val acc 0.26 maxdown 8
11/16/2018 02:07:57 PM epoch 92 train loss 1.62 train acc 0.69 val loss 3.45 val acc 0.25 best val acc 0.26 maxdown 9
11/16/2018 02:07:57 PM epoch 93 train loss 1.59 train acc 0.70 val loss 3.45 val acc 0.24 best val acc 0.26 maxdown 10
11/16/2018 02:07:58 PM epoch 94 train loss 1.56 train acc 0.71 val loss 3.45 val acc 0.26 best val acc 0.26 maxdown 11
11/16/2018 02:07:58 PM validation results went down. early stopping ...
11/16/2018 02:07:58 PM dev results:
11/16/2018 02:07:58 PM Mean: 520 Median: 46 Acc@161: 60
11/16/2018 02:07:58 PM test results:
11/16/2018 02:07:59 PM Mean: 540 Median: 47 Acc@161: 60