eden-emu/opus
15
0
Fork 0
mirror of https://github.com/xiph/opus.git synced 2025-05-19 18:08:29 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

mdense max size

Browse source
This commit is contained in:
Jean-Marc Valin 2018-11-24 15:51:08 -05:00
parent 94ac0841df
commit 37fbcaee0b
2 changed files with 6 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

5
dnn/dump_lpcnet.py
View file

@ -14,6 +14,7 @@ import re
max_rnn_neurons = 1
max_conv_inputs = 1
max_mdense_tmp = 1
def printVector(f, vector, name):
v = np.reshape(vector, (-1));
@ -80,6 +81,7 @@ def dump_dense_layer(self, f, hf):
Dense.dump_layer = dump_dense_layer
def dump_mdense_layer(self, f, hf):
global max_mdense_tmp
name = self.name
print("printing layer " + name + " of type " + self.__class__.__name__)
weights = self.get_weights()
@ -87,6 +89,7 @@ def dump_mdense_layer(self, f, hf):
printVector(f, weights[1], name + '_bias')
printVector(f, weights[1], name + '_factor')
activation = self.activation.__name__.upper()
max_mdense_tmp = max(max_mdense_tmp, weights[0].shape[0]*weights[0].shape[2])
f.write('const MDenseLayer {} = {{\n {}_bias,\n {}_weights,\n {}_factor,\n {}, {}, {}, ACTIVATION_{}\n}};\n\n'
.format(name, name, name, name, weights[0].shape[0], weights[0].shape[1], weights[0].shape[2], activation))
hf.write('#define {}_OUT_SIZE {}\n'.format(name.upper(), weights[0].shape[0]))
@ -148,6 +151,8 @@ for i, layer in enumerate(model.layers):
hf.write('#define MAX_RNN_NEURONS {}\n\n'.format(max_rnn_neurons))
hf.write('#define MAX_CONV_INPUTS {}\n\n'.format(max_conv_inputs))
hf.write('#define MAX_MDENSE_TMP {}\n\n'.format(max_mdense_tmp))
hf.write('struct RNNState {\n')
for i, name in enumerate(layer_list):

3
dnn/nnet.c
View file

@ -132,11 +132,10 @@ void compute_mdense(const MDenseLayer *layer, float *output, const float *input)
int i, c;
int N, M, C;
int stride;
float tmp[MAX_MDENSE_TMP];
M = layer->nb_inputs;
N = layer->nb_neurons;
C = layer->nb_channels;
/* FIXME: Make this C90. */
float tmp[N*C];
stride = N*C;
for (i=0;i<N*C;i++)
tmp[i] = layer->bias[i];