tests/nnapi/specs/skip/V1_1/lstm2_state_relaxed.mod.py

   1 #
   2 # Copyright (C) 2018 The Android Open Source Project
   3 #
   4 # Licensed under the Apache License, Version 2.0 (the "License");
   5 # you may not use this file except in compliance with the License.
   6 # You may obtain a copy of the License at
   7 #
   8 #      http://www.apache.org/licenses/LICENSE-2.0
   9 #
  10 # Unless required by applicable law or agreed to in writing, software
  11 # distributed under the License is distributed on an "AS IS" BASIS,
  12 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  13 # See the License for the specific language governing permissions and
  14 # limitations under the License.
  15 #
  16
  17 # LSTM Test, With Cifg, With Peephole, No Projection, No Clipping.
  18
  19 model = Model()
  20
  21 n_batch = 1
  22 n_input = 2
  23 # n_cell and n_output have the same size when there is no projection.
  24 n_cell = 4
  25 n_output = 4
  26
  27 input = Input("input", "TENSOR_FLOAT32", "{%d, %d}" % (n_batch, n_input))
  28
  29 input_to_input_weights = Input("input_to_input_weights", "TENSOR_FLOAT32", "{%d, %d}" % (n_cell, n_input))
  30 input_to_forget_weights = Input("input_to_forget_weights", "TENSOR_FLOAT32", "{%d, %d}" % (n_cell, n_input))
  31 input_to_cell_weights = Input("input_to_cell_weights", "TENSOR_FLOAT32", "{%d, %d}" % (n_cell, n_input))
  32 input_to_output_weights = Input("input_to_output_weights", "TENSOR_FLOAT32", "{%d, %d}" % (n_cell, n_input))
  33
  34 recurrent_to_input_weights = Input("recurrent_to_intput_weights", "TENSOR_FLOAT32", "{%d, %d}" % (n_cell, n_output))
  35 recurrent_to_forget_weights = Input("recurrent_to_forget_weights", "TENSOR_FLOAT32", "{%d, %d}" % (n_cell, n_output))
  36 recurrent_to_cell_weights = Input("recurrent_to_cell_weights", "TENSOR_FLOAT32", "{%d, %d}" % (n_cell, n_output))
  37 recurrent_to_output_weights = Input("recurrent_to_output_weights", "TENSOR_FLOAT32", "{%d, %d}" % (n_cell, n_output))
  38
  39 cell_to_input_weights = Input("cell_to_input_weights", "TENSOR_FLOAT32", "{0}")
  40 cell_to_forget_weights = Input("cell_to_forget_weights", "TENSOR_FLOAT32", "{%d}" % (n_cell))
  41 cell_to_output_weights = Input("cell_to_output_weights", "TENSOR_FLOAT32", "{%d}" % (n_cell))
  42
  43 input_gate_bias = Input("input_gate_bias", "TENSOR_FLOAT32", "{%d}"%(n_cell))
  44 forget_gate_bias = Input("forget_gate_bias", "TENSOR_FLOAT32", "{%d}"%(n_cell))
  45 cell_gate_bias = Input("cell_gate_bias", "TENSOR_FLOAT32", "{%d}"%(n_cell))
  46 output_gate_bias = Input("output_gate_bias", "TENSOR_FLOAT32", "{%d}"%(n_cell))
  47
  48 projection_weights = Input("projection_weights", "TENSOR_FLOAT32", "{0,0}")
  49 projection_bias = Input("projection_bias", "TENSOR_FLOAT32", "{0}")
  50
  51 output_state_in = Input("output_state_in", "TENSOR_FLOAT32", "{%d, %d}" % (n_batch, n_output))
  52 cell_state_in = Input("cell_state_in", "TENSOR_FLOAT32", "{%d, %d}" % (n_batch, n_cell))
  53
  54 activation_param = Int32Scalar("activation_param", 4)  # Tanh
  55 cell_clip_param = Float32Scalar("cell_clip_param", 0.)
  56 proj_clip_param = Float32Scalar("proj_clip_param", 0.)
  57
  58 scratch_buffer = IgnoredOutput("scratch_buffer", "TENSOR_FLOAT32", "{%d, %d}" % (n_batch, n_cell * 3))
  59 output_state_out = Output("output_state_out", "TENSOR_FLOAT32", "{%d, %d}" % (n_batch, n_output))
  60 cell_state_out = Output("cell_state_out", "TENSOR_FLOAT32", "{%d, %d}" % (n_batch, n_cell))
  61 output = Output("output", "TENSOR_FLOAT32", "{%d, %d}" % (n_batch, n_output))
  62
  63 model = model.Operation("LSTM",
  64                         input,
  65
  66                         input_to_input_weights,
  67                         input_to_forget_weights,
  68                         input_to_cell_weights,
  69                         input_to_output_weights,
  70
  71                         recurrent_to_input_weights,
  72                         recurrent_to_forget_weights,
  73                         recurrent_to_cell_weights,
  74                         recurrent_to_output_weights,
  75
  76                         cell_to_input_weights,
  77                         cell_to_forget_weights,
  78                         cell_to_output_weights,
  79
  80                         input_gate_bias,
  81                         forget_gate_bias,
  82                         cell_gate_bias,
  83                         output_gate_bias,
  84
  85                         projection_weights,
  86                         projection_bias,
  87
  88                         output_state_in,
  89                         cell_state_in,
  90
  91                         activation_param,
  92                         cell_clip_param,
  93                         proj_clip_param
  94 ).To([scratch_buffer, output_state_out, cell_state_out, output])
  95 model = model.RelaxedExecution(True)
  96
  97 input0 = {input_to_input_weights:[],
  98           input_to_cell_weights: [-0.49770179, -0.27711356, -0.09624726, 0.05100781, 0.04717243, 0.48944736, -0.38535351, -0.17212132],
  99           input_to_forget_weights: [-0.55291498, -0.42866567, 0.13056988, -0.3633365, -0.22755712, 0.28253698, 0.24407166, 0.33826375],
 100           input_to_output_weights: [0.10725588, -0.02335852, -0.55932593, -0.09426838, -0.44257352, 0.54939759, 0.01533556, 0.42751634],
 101
 102           input_gate_bias:  [],
 103           forget_gate_bias: [1.,1.,1.,1.],
 104           cell_gate_bias:   [0.,0.,0.,0.],
 105           output_gate_bias: [0.,0.,0.,0.],
 106
 107           recurrent_to_input_weights: [],
 108           recurrent_to_cell_weights: [
 109               0.54066205, -0.32668582, -0.43562764, -0.56094903, 0.42957711,
 110               0.01841056, -0.32764608, -0.33027974, -0.10826075, 0.20675004,
 111               0.19069612, -0.03026325, -0.54532051, 0.33003211, 0.44901288,
 112               0.21193194],
 113
 114           recurrent_to_forget_weights: [
 115               -0.13832897, -0.0515101, -0.2359007, -0.16661474, -0.14340827,
 116             0.36986142, 0.23414481, 0.55899, 0.10798943, -0.41174671, 0.17751795,
 117             -0.34484994, -0.35874045, -0.11352962, 0.27268326, 0.54058349],
 118
 119           recurrent_to_output_weights: [
 120               0.41613156, 0.42610586, -0.16495961, -0.5663873, 0.30579174, -0.05115908,
 121               -0.33941799, 0.23364776, 0.11178309, 0.09481031, -0.26424935, 0.46261835,
 122               0.50248802, 0.26114327, -0.43736315, 0.33149987],
 123
 124           cell_to_input_weights: [],
 125           cell_to_forget_weights: [0.47485286, -0.51955009, -0.24458408, 0.31544167],
 126           cell_to_output_weights: [-0.17135078, 0.82760304, 0.85573703, -0.77109635],
 127
 128           projection_weights: [],
 129           projection_bias: [],
 130 }
 131
 132 output0 = {
 133     scratch_buffer: [ 0 for x in range(n_batch * n_cell * 3) ],
 134     cell_state_out: [ -0.978419, -0.139203, 0.338163, -0.0983904 ],
 135     output_state_out: [ -0.423122, -0.0121822, 0.24201, -0.0812458 ],
 136 }
 137
 138 input0[input] = [3., 4.]
 139 input0[output_state_in] = [-0.364445, -0.00352185, 0.128866, -0.0516365]
 140 input0[cell_state_in] = [-0.760444, -0.0180416, 0.182264, -0.0649371]
 141 output0[output] = [-0.42312205, -0.01218222, 0.24201041, -0.08124574]
 142 Example((input0, output0))