/dports/misc/mxnet/incubator-mxnet-1.9.0/3rdparty/tvm/python/tvm/relay/testing/ |
H A D | lstm.py | 30 def lstm_cell(num_hidden, batch_size=1, dtype="float32", name=""): argument 52 input_type = relay.TensorType((batch_size, num_hidden), dtype) 53 weight_type = relay.TensorType((4 * num_hidden, num_hidden), dtype) 54 bias_type = relay.TensorType((4 * num_hidden,), dtype) 56 dense_type = relay.TensorType((batch_size, 4 * num_hidden), dtype) 79 units=num_hidden * 4, 122 def get_net(iterations, num_hidden, batch_size=1, dtype="float32"): argument 124 input_type = relay.TensorType((batch_size, num_hidden), dtype) 125 weight_type = relay.TensorType((4 * num_hidden, num_hidden), dtype) 126 bias_type = relay.TensorType((4 * num_hidden,), dtype) [all …]
|
/dports/misc/tvm/incubator-tvm-0.6.1/python/tvm/relay/testing/ |
H A D | lstm.py | 29 def lstm_cell(num_hidden, batch_size=1, dtype="float32", name=""): argument 51 input_type = relay.TensorType((batch_size, num_hidden), dtype) 52 weight_type = relay.TensorType((4*num_hidden, num_hidden), dtype) 53 bias_type = relay.TensorType((4*num_hidden,), dtype) 55 dense_type = relay.TensorType((batch_size, 4*num_hidden), dtype) 74 units=num_hidden * 4, 80 units=num_hidden * 4, 118 input_type = relay.TensorType((batch_size, num_hidden), dtype) 119 weight_type = relay.TensorType((4*num_hidden, num_hidden), dtype) 120 bias_type = relay.TensorType((4*num_hidden,), dtype) [all …]
|
/dports/misc/py-tvm/incubator-tvm-0.6.1/python/tvm/relay/testing/ |
H A D | lstm.py | 29 def lstm_cell(num_hidden, batch_size=1, dtype="float32", name=""): argument 51 input_type = relay.TensorType((batch_size, num_hidden), dtype) 52 weight_type = relay.TensorType((4*num_hidden, num_hidden), dtype) 53 bias_type = relay.TensorType((4*num_hidden,), dtype) 55 dense_type = relay.TensorType((batch_size, 4*num_hidden), dtype) 74 units=num_hidden * 4, 80 units=num_hidden * 4, 118 input_type = relay.TensorType((batch_size, num_hidden), dtype) 119 weight_type = relay.TensorType((4*num_hidden, num_hidden), dtype) 120 bias_type = relay.TensorType((4*num_hidden,), dtype) [all …]
|
/dports/misc/mxnet/incubator-mxnet-1.9.0/example/speech_recognition/ |
H A D | stt_layer_fc.py | 24 num_hidden, argument 33 net = mx.sym.FullyConnected(data=net, num_hidden=num_hidden, no_bias=no_bias, name=name) 37 net = mx.sym.FullyConnected(data=net, num_hidden=num_hidden, no_bias=no_bias, name=name) 39 …net = mx.sym.FullyConnected(data=net, num_hidden=num_hidden, bias=bias, no_bias=no_bias, name=name) 42 …net = mx.sym.FullyConnected(data=net, num_hidden=num_hidden, weight=weight, no_bias=no_bias, name=… 46 …net = mx.sym.FullyConnected(data=net, num_hidden=num_hidden, weight=weight, no_bias=no_bias, name=… 48 …net = mx.sym.FullyConnected(data=net, num_hidden=num_hidden, weight=weight, bias=bias, no_bias=no_… 99 num_hidden=num_hidden_list[layer_index], 114 num_hidden=num_hidden_list[layer_index],
|
/dports/misc/py-mxnet/incubator-mxnet-1.9.0/example/speech_recognition/ |
H A D | stt_layer_fc.py | 24 num_hidden, argument 33 net = mx.sym.FullyConnected(data=net, num_hidden=num_hidden, no_bias=no_bias, name=name) 37 net = mx.sym.FullyConnected(data=net, num_hidden=num_hidden, no_bias=no_bias, name=name) 39 …net = mx.sym.FullyConnected(data=net, num_hidden=num_hidden, bias=bias, no_bias=no_bias, name=name) 42 …net = mx.sym.FullyConnected(data=net, num_hidden=num_hidden, weight=weight, no_bias=no_bias, name=… 46 …net = mx.sym.FullyConnected(data=net, num_hidden=num_hidden, weight=weight, no_bias=no_bias, name=… 48 …net = mx.sym.FullyConnected(data=net, num_hidden=num_hidden, weight=weight, bias=bias, no_bias=no_… 99 num_hidden=num_hidden_list[layer_index], 114 num_hidden=num_hidden_list[layer_index],
|
/dports/misc/mxnet/incubator-mxnet-1.9.0/contrib/clojure-package/examples/imclassification/test/ |
H A D | test-symbol.json.ref | 11 "attrs": {"num_hidden": "128"}, 17 "attrs": {"num_hidden": "128"}, 23 "attrs": {"num_hidden": "128"}, 35 "attrs": {"num_hidden": "64"}, 41 "attrs": {"num_hidden": "64"}, 47 "attrs": {"num_hidden": "64"}, 59 "attrs": {"num_hidden": "10"}, 65 "attrs": {"num_hidden": "10"}, 71 "attrs": {"num_hidden": "10"},
|
/dports/misc/py-mxnet/incubator-mxnet-1.9.0/contrib/clojure-package/examples/imclassification/test/ |
H A D | test-symbol.json.ref | 11 "attrs": {"num_hidden": "128"}, 17 "attrs": {"num_hidden": "128"}, 23 "attrs": {"num_hidden": "128"}, 35 "attrs": {"num_hidden": "64"}, 41 "attrs": {"num_hidden": "64"}, 47 "attrs": {"num_hidden": "64"}, 59 "attrs": {"num_hidden": "10"}, 65 "attrs": {"num_hidden": "10"}, 71 "attrs": {"num_hidden": "10"},
|
/dports/misc/mxnet/incubator-mxnet-1.9.0/example/gluon/word_language_model/ |
H A D | model.py | 25 def __init__(self, mode, vocab_size, num_embed, num_hidden, argument 33 self.rnn = rnn.RNN(num_hidden, num_layers, dropout=dropout, 36 self.rnn = rnn.RNN(num_hidden, num_layers, 'tanh', dropout=dropout, 39 self.rnn = rnn.LSTM(num_hidden, num_layers, dropout=dropout, 42 self.rnn = rnn.GRU(num_hidden, num_layers, dropout=dropout, 49 self.decoder = nn.Dense(vocab_size, in_units=num_hidden, 52 self.decoder = nn.Dense(vocab_size, in_units=num_hidden) 54 self.num_hidden = num_hidden 60 decoded = self.decoder(output.reshape((-1, self.num_hidden)))
|
/dports/misc/py-mxnet/incubator-mxnet-1.9.0/example/gluon/word_language_model/ |
H A D | model.py | 25 def __init__(self, mode, vocab_size, num_embed, num_hidden, argument 33 self.rnn = rnn.RNN(num_hidden, num_layers, dropout=dropout, 36 self.rnn = rnn.RNN(num_hidden, num_layers, 'tanh', dropout=dropout, 39 self.rnn = rnn.LSTM(num_hidden, num_layers, dropout=dropout, 42 self.rnn = rnn.GRU(num_hidden, num_layers, dropout=dropout, 49 self.decoder = nn.Dense(vocab_size, in_units=num_hidden, 52 self.decoder = nn.Dense(vocab_size, in_units=num_hidden) 54 self.num_hidden = num_hidden 60 decoded = self.decoder(output.reshape((-1, self.num_hidden)))
|
/dports/misc/mxnet/incubator-mxnet-1.9.0/example/ctc/ |
H A D | lstm.py | 34 def _lstm(num_hidden, indata, prev_state, param, seqidx, layeridx): argument 39 num_hidden=num_hidden * 4, 44 num_hidden=num_hidden * 4, 58 def _lstm_unroll_base(num_lstm_layer, seq_len, num_hidden): argument 81 num_hidden=num_hidden, 92 pred_fc = mx.sym.FullyConnected(data=hidden_concat, num_hidden=11, name="pred_fc") 129 def lstm_unroll(num_lstm_layer, seq_len, num_hidden, num_label, loss_type=None): argument 148 pred = _lstm_unroll_base(num_lstm_layer, seq_len, num_hidden) 158 def init_states(batch_size, num_lstm_layer, num_hidden): argument 172 init_c = [('l%d_init_c' % l, (batch_size, num_hidden)) for l in range(num_lstm_layer)] [all …]
|
/dports/misc/py-mxnet/incubator-mxnet-1.9.0/example/ctc/ |
H A D | lstm.py | 34 def _lstm(num_hidden, indata, prev_state, param, seqidx, layeridx): argument 39 num_hidden=num_hidden * 4, 44 num_hidden=num_hidden * 4, 58 def _lstm_unroll_base(num_lstm_layer, seq_len, num_hidden): argument 81 num_hidden=num_hidden, 92 pred_fc = mx.sym.FullyConnected(data=hidden_concat, num_hidden=11, name="pred_fc") 129 def lstm_unroll(num_lstm_layer, seq_len, num_hidden, num_label, loss_type=None): argument 148 pred = _lstm_unroll_base(num_lstm_layer, seq_len, num_hidden) 158 def init_states(batch_size, num_lstm_layer, num_hidden): argument 172 init_c = [('l%d_init_c' % l, (batch_size, num_hidden)) for l in range(num_lstm_layer)] [all …]
|
/dports/misc/tvm/incubator-tvm-0.6.1/tests/python/unittest/ |
H A D | test_schedule_lstm.py | 22 num_hidden = 1152 26 Wi2h = tvm.placeholder((4, num_hidden, num_input), name="Wi2h") 27 Wh2h = tvm.placeholder((4, num_hidden, num_hidden), name="Wh2h") 29 s_state_h = tvm.placeholder((num_step, batch_size, num_hidden)) 30 s_state_c = tvm.placeholder((num_step, batch_size, num_hidden)) 31 s_init_c = tvm.compute((1, batch_size, num_hidden), 33 s_init_h = tvm.compute((1, batch_size, num_hidden), 38 (num_step, 4, batch_size, num_hidden), 41 k = tvm.reduce_axis((0, num_hidden), name="ki2h") 43 (num_step, 4, batch_size, num_hidden), [all …]
|
/dports/misc/py-tvm/incubator-tvm-0.6.1/tests/python/unittest/ |
H A D | test_schedule_lstm.py | 22 num_hidden = 1152 26 Wi2h = tvm.placeholder((4, num_hidden, num_input), name="Wi2h") 27 Wh2h = tvm.placeholder((4, num_hidden, num_hidden), name="Wh2h") 29 s_state_h = tvm.placeholder((num_step, batch_size, num_hidden)) 30 s_state_c = tvm.placeholder((num_step, batch_size, num_hidden)) 31 s_init_c = tvm.compute((1, batch_size, num_hidden), 33 s_init_h = tvm.compute((1, batch_size, num_hidden), 38 (num_step, 4, batch_size, num_hidden), 41 k = tvm.reduce_axis((0, num_hidden), name="ki2h") 43 (num_step, 4, batch_size, num_hidden), [all …]
|
/dports/misc/mxnet/incubator-mxnet-1.9.0/3rdparty/tvm/tests/python/unittest/ |
H A D | test_te_schedule_lstm.py | 24 num_hidden = 1152 28 Wi2h = te.placeholder((4, num_hidden, num_input), name="Wi2h") 29 Wh2h = te.placeholder((4, num_hidden, num_hidden), name="Wh2h") 31 s_state_h = te.placeholder((num_step, batch_size, num_hidden)) 32 s_state_c = te.placeholder((num_step, batch_size, num_hidden)) 33 s_init_c = te.compute((1, batch_size, num_hidden), lambda *i: 0.0, name="init_c") 34 s_init_h = te.compute((1, batch_size, num_hidden), lambda *i: 0.0, name="init_h") 38 (num_step, 4, batch_size, num_hidden), 42 k = te.reduce_axis((0, num_hidden), name="ki2h") 44 (num_step, 4, batch_size, num_hidden), [all …]
|
/dports/misc/mxnet/incubator-mxnet-1.9.0/example/rnn/bucketing/ |
H A D | cudnn_rnn_bucketing.py | 106 cell.add(mx.rnn.FusedRNNCell(args.num_hidden, num_layers=1, 123 shape=(-1, args.num_hidden*(1+args.bidirectional))) 183 … cell = mx.rnn.LSTMCell(num_hidden=args.num_hidden, prefix='%s_%dl0_'%(args.rnntype,i)) 187 … mx.rnn.LSTMCell(num_hidden=args.num_hidden, prefix='%s_%dr0_'%(args.rnntype,i)), 190 … cell = mx.rnn.GRUCell(num_hidden=args.num_hidden, prefix='%s_%dl0_'%(args.rnntype,i)) 194 … mx.rnn.GRUCell(num_hidden=args.num_hidden, prefix='%s_%dr0_'%(args.rnntype,i)), 197 …cell = mx.rnn.RNNCell(num_hidden=args.num_hidden, activation='tanh', prefix='%s_%dl0_'%(args.rnnty… 201 … mx.rnn.RNNCell(num_hidden=args.num_hidden, activation='tanh', prefix='%s_%dr0_'%(args.rnntype,i)), 204 …cell = mx.rnn.RNNCell(num_hidden=args.num_hidden, activation='relu', prefix='%s_%dl0_'%(args.rnnty… 208 … mx.rnn.RNNCell(num_hidden=args.num_hidden, activation='relu', prefix='%s_%dr0_'%(args.rnntype,i)), [all …]
|
/dports/misc/py-mxnet/incubator-mxnet-1.9.0/example/rnn/bucketing/ |
H A D | cudnn_rnn_bucketing.py | 106 cell.add(mx.rnn.FusedRNNCell(args.num_hidden, num_layers=1, 123 shape=(-1, args.num_hidden*(1+args.bidirectional))) 183 … cell = mx.rnn.LSTMCell(num_hidden=args.num_hidden, prefix='%s_%dl0_'%(args.rnntype,i)) 187 … mx.rnn.LSTMCell(num_hidden=args.num_hidden, prefix='%s_%dr0_'%(args.rnntype,i)), 190 … cell = mx.rnn.GRUCell(num_hidden=args.num_hidden, prefix='%s_%dl0_'%(args.rnntype,i)) 194 … mx.rnn.GRUCell(num_hidden=args.num_hidden, prefix='%s_%dr0_'%(args.rnntype,i)), 197 …cell = mx.rnn.RNNCell(num_hidden=args.num_hidden, activation='tanh', prefix='%s_%dl0_'%(args.rnnty… 201 … mx.rnn.RNNCell(num_hidden=args.num_hidden, activation='tanh', prefix='%s_%dr0_'%(args.rnntype,i)), 204 …cell = mx.rnn.RNNCell(num_hidden=args.num_hidden, activation='relu', prefix='%s_%dl0_'%(args.rnnty… 208 … mx.rnn.RNNCell(num_hidden=args.num_hidden, activation='relu', prefix='%s_%dr0_'%(args.rnntype,i)), [all …]
|
/dports/misc/mxnet/incubator-mxnet-1.9.0/example/rnn/old/ |
H A D | gru.py | 34 def gru(num_hidden, indata, prev_state, param, seqidx, layeridx, dropout=0.): argument 46 num_hidden=num_hidden * 2, 51 num_hidden=num_hidden * 2, 62 num_hidden=num_hidden, 68 num_hidden=num_hidden, 76 num_hidden, num_embed, num_label, dropout=0.): argument 112 next_state = gru(num_hidden, indata=hidden, 124 pred = mx.sym.FullyConnected(data=hidden_concat, num_hidden=num_label,
|
/dports/misc/py-mxnet/incubator-mxnet-1.9.0/example/rnn/old/ |
H A D | gru.py | 34 def gru(num_hidden, indata, prev_state, param, seqidx, layeridx, dropout=0.): argument 46 num_hidden=num_hidden * 2, 51 num_hidden=num_hidden * 2, 62 num_hidden=num_hidden, 68 num_hidden=num_hidden, 76 num_hidden, num_embed, num_label, dropout=0.): argument 112 next_state = gru(num_hidden, indata=hidden, 124 pred = mx.sym.FullyConnected(data=hidden_concat, num_hidden=num_label,
|
/dports/misc/tvm/incubator-tvm-0.6.1/topi/recipe/rnn/ |
H A D | lstm.py | 62 num_hidden = 1152 // 2 68 Wh2h = tvm.placeholder((4, num_hidden, num_hidden), name="Wh2h") 70 s_state_h = tvm.placeholder((num_step, batch_size, num_hidden)) 71 s_state_c = tvm.placeholder((num_step, batch_size, num_hidden)) 72 s_init_c = tvm.compute((1, batch_size, num_hidden), 74 s_init_h = tvm.compute((1, batch_size, num_hidden), 77 k = tvm.reduce_axis((0, num_hidden), name="ki2h") 79 (num_step, batch_size, 4, num_hidden), 85 gshape = (num_step, batch_size, num_hidden) 171 (num_step, batch_size, num_hidden)).astype("float32") [all …]
|
/dports/misc/mxnet/incubator-mxnet-1.9.0/3rdparty/tvm/apps/topi_recipe/rnn/ |
H A D | lstm.py | 63 num_hidden = 1152 // 2 67 Xi2h = te.placeholder((num_step, batch_size, 4, num_hidden), name="Xi2h") 69 Wh2h = te.placeholder((4, num_hidden, num_hidden), name="Wh2h") 71 s_state_h = te.placeholder((num_step, batch_size, num_hidden)) 72 s_state_c = te.placeholder((num_step, batch_size, num_hidden)) 76 k = te.reduce_axis((0, num_hidden), name="ki2h") 78 (num_step, batch_size, 4, num_hidden), 84 gshape = (num_step, batch_size, num_hidden) 176 scan_h_np = np.zeros((num_step, batch_size, num_hidden)).astype("float32") 177 scan_c_np = np.zeros((num_step, batch_size, num_hidden)).astype("float32") [all …]
|
/dports/misc/py-tvm/incubator-tvm-0.6.1/topi/recipe/rnn/ |
H A D | lstm.py | 62 num_hidden = 1152 // 2 68 Wh2h = tvm.placeholder((4, num_hidden, num_hidden), name="Wh2h") 70 s_state_h = tvm.placeholder((num_step, batch_size, num_hidden)) 71 s_state_c = tvm.placeholder((num_step, batch_size, num_hidden)) 72 s_init_c = tvm.compute((1, batch_size, num_hidden), 74 s_init_h = tvm.compute((1, batch_size, num_hidden), 77 k = tvm.reduce_axis((0, num_hidden), name="ki2h") 79 (num_step, batch_size, 4, num_hidden), 85 gshape = (num_step, batch_size, num_hidden) 171 (num_step, batch_size, num_hidden)).astype("float32") [all …]
|
/dports/misc/mxnet/incubator-mxnet-1.9.0/perl-package/AI-MXNet/t/ |
H A D | test_symbol.t | 35 my $net2 = mx->symbol->FullyConnected(name=>'fc3', num_hidden=>10); 37 $net2 = mx->symbol->FullyConnected(data=>$net2, name=>'fc4', num_hidden=>20); 133 my $num_hidden = 128; 139 my $x2h = mx->symbol->FullyConnected(data=>$data, name=>'x2h', num_hidden=>$num_hidden); 140 my $h2h = mx->symbol->FullyConnected(data=>$prev, name=>'h2h', num_hidden=>$num_hidden); 152 is_deeply($arg_shapes{x2h_weight}, [$num_hidden, $num_dim]); 160 is_deeply($arg_shapes{x2h_weight}, [$num_hidden, $num_dim]); 161 is_deeply($arg_shapes{h2h_weight}, [$num_hidden, $num_hidden]); 335 "num_hidden": "128" 380 "num_hidden": "64" [all …]
|
/dports/misc/py-mxnet/incubator-mxnet-1.9.0/perl-package/AI-MXNet/t/ |
H A D | test_symbol.t | 35 my $net2 = mx->symbol->FullyConnected(name=>'fc3', num_hidden=>10); 37 $net2 = mx->symbol->FullyConnected(data=>$net2, name=>'fc4', num_hidden=>20); 133 my $num_hidden = 128; 139 my $x2h = mx->symbol->FullyConnected(data=>$data, name=>'x2h', num_hidden=>$num_hidden); 140 my $h2h = mx->symbol->FullyConnected(data=>$prev, name=>'h2h', num_hidden=>$num_hidden); 152 is_deeply($arg_shapes{x2h_weight}, [$num_hidden, $num_dim]); 160 is_deeply($arg_shapes{x2h_weight}, [$num_hidden, $num_dim]); 161 is_deeply($arg_shapes{h2h_weight}, [$num_hidden, $num_hidden]); 335 "num_hidden": "128" 380 "num_hidden": "64" [all …]
|
/dports/misc/tvm/incubator-tvm-0.6.1/tests/python/frontend/mxnet/model_zoo/ |
H A D | mlp.py | 27 fc1 = mx.symbol.FullyConnected(data = data, name='fc1', num_hidden=128, flatten=False) 29 fc2 = mx.symbol.FullyConnected(data = act1, name = 'fc2', num_hidden = 64, flatten=False) 31 … fc3 = mx.symbol.FullyConnected(data = act2, name='fc3', num_hidden=num_classes, flatten=False) 34 fc1 = mx.symbol.FullyConnected(data = data, name='fc1', num_hidden=128) 36 fc2 = mx.symbol.FullyConnected(data = act1, name = 'fc2', num_hidden = 64) 38 fc3 = mx.symbol.FullyConnected(data = act2, name='fc3', num_hidden=num_classes)
|
/dports/misc/mxnet/incubator-mxnet-1.9.0/3rdparty/tvm/tests/python/frontend/mxnet/model_zoo/ |
H A D | mlp.py | 28 fc1 = mx.symbol.FullyConnected(data=data, name="fc1", num_hidden=128, flatten=False) 30 fc2 = mx.symbol.FullyConnected(data=act1, name="fc2", num_hidden=64, flatten=False) 32 fc3 = mx.symbol.FullyConnected(data=act2, name="fc3", num_hidden=num_classes, flatten=False) 35 fc1 = mx.symbol.FullyConnected(data=data, name="fc1", num_hidden=128) 37 fc2 = mx.symbol.FullyConnected(data=act1, name="fc2", num_hidden=64) 39 fc3 = mx.symbol.FullyConnected(data=act2, name="fc3", num_hidden=num_classes)
|