Aku Rouhe commited on
Commit
f6bcbf4
1 Parent(s): d3f34b7

Hyperparams

Browse files
Files changed (2) hide show
  1. .gitattributes +1 -0
  2. hyperparams.yaml +15 -6
.gitattributes CHANGED
@@ -25,3 +25,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
25
  *.zip filter=lfs diff=lfs merge=lfs -text
26
  *.zstandard filter=lfs diff=lfs merge=lfs -text
27
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
25
  *.zip filter=lfs diff=lfs merge=lfs -text
26
  *.zstandard filter=lfs diff=lfs merge=lfs -text
27
  *tfevents* filter=lfs diff=lfs merge=lfs -text
28
+ *.ckpt filter=lfs diff=lfs merge=lfs -text
hyperparams.yaml CHANGED
@@ -102,6 +102,11 @@ seq_lin: !new:speechbrain.nnet.linear.Linear
102
  input_size: !ref <dec_neurons>
103
  n_neurons: !ref <output_neurons>
104
 
 
 
 
 
 
105
  # Final softmax (for log posteriors computation).
106
  log_softmax: !new:speechbrain.nnet.activations.Softmax
107
  apply_log: True
@@ -115,10 +120,18 @@ model: !new:torch.nn.ModuleList
115
  - !ref <ctc_lin>
116
  - !ref <seq_lin>
117
 
 
 
 
 
 
 
 
118
  test_search: !new:speechbrain.decoders.S2SRNNBeamSearcher
119
  embedding: !ref <embedding>
120
  decoder: !ref <decoder>
121
  linear: !ref <seq_lin>
 
122
  bos_index: !ref <bos_index>
123
  eos_index: !ref <eos_index>
124
  blank_index: !ref <blank_index>
@@ -135,14 +148,10 @@ test_search: !new:speechbrain.decoders.S2SRNNBeamSearcher
135
  # Objects in "modules" dict will have their parameters moved to the correct
136
  # device, as well as having train()/eval() called on them by the Brain class
137
  modules:
138
- encoder: !ref <encoder>
139
- embedding: !ref <embedding>
140
- dec: !ref <decoder>
141
  decoder: !ref <test_search>
142
- seq_lin: !ref <seq_lin>
143
- normalize: !ref <normalize>
144
 
145
- pretrainer !new:speechbrain.utils.parameter_transfer
146
  loadables:
147
  model: !ref <model>
148
  normalizer: !ref <normalize>
 
102
  input_size: !ref <dec_neurons>
103
  n_neurons: !ref <output_neurons>
104
 
105
+ # Linear transformation on the top of the encoder.
106
+ ctc_lin: !new:speechbrain.nnet.linear.Linear
107
+ input_size: !ref <dnn_neurons>
108
+ n_neurons: !ref <output_neurons>
109
+
110
  # Final softmax (for log posteriors computation).
111
  log_softmax: !new:speechbrain.nnet.activations.Softmax
112
  apply_log: True
 
120
  - !ref <ctc_lin>
121
  - !ref <seq_lin>
122
 
123
+ full_encode_step: !new:speechbrain.nnet.containers.LengthsCapableSequential
124
+ input_shape: [null, null, !ref <n_mels>]
125
+ compute_features: !ref <compute_features>
126
+ normalize: !ref <normalize>
127
+ model: !ref <encoder>
128
+
129
+
130
  test_search: !new:speechbrain.decoders.S2SRNNBeamSearcher
131
  embedding: !ref <embedding>
132
  decoder: !ref <decoder>
133
  linear: !ref <seq_lin>
134
+ ctc_linear: !ref <ctc_lin>
135
  bos_index: !ref <bos_index>
136
  eos_index: !ref <eos_index>
137
  blank_index: !ref <blank_index>
 
148
  # Objects in "modules" dict will have their parameters moved to the correct
149
  # device, as well as having train()/eval() called on them by the Brain class
150
  modules:
151
+ encoder: !ref <full_encode_step>
 
 
152
  decoder: !ref <test_search>
 
 
153
 
154
+ pretrainer: !new:speechbrain.utils.parameter_transfer.Pretrainer
155
  loadables:
156
  model: !ref <model>
157
  normalizer: !ref <normalize>