Merge pull request #66 from hyperion-ml/lachesis-1

jesus-villalba · web-flow · commit 307129f793a3 · 2021-07-04T12:27:04.000-04:00
fixed feat config file in sre20cts global configs
diff --git a/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_effnetb4_v2_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.01.amp.v1.sh b/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_effnetb4_v2_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.01.amp.v1.sh
@@ -1,7 +1,7 @@
 # ResNet34 x-vector with mixed precision training
 
 # acoustic features
-feat_config=conf/fbank64_mvn_8k.yaml
+feat_config=conf/fbank64_stmn_8k.yaml
 feat_type=fbank64_stmn
 
 
diff --git a/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_res2net50w26s4_eina_hln_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.01.amp.v1.sh b/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_res2net50w26s4_eina_hln_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.01.amp.v1.sh
@@ -1,7 +1,7 @@
 # Res2Net50 x-vector with mixed precision training
 
 # acoustic features
-feat_config=conf/fbank64_mvn_8k.yaml
+feat_config=conf/fbank64_stmn_8k.yaml
 feat_type=fbank64_stmn
 
 
diff --git a/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_resnet34_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.01.amp.v1.sh b/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_resnet34_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.01.amp.v1.sh
@@ -1,7 +1,7 @@
 # ResNet34 x-vector with mixed precision training
 
 # acoustic features
-feat_config=conf/fbank64_mvn_8k.yaml
+feat_config=conf/fbank64_stmn_8k.yaml
 feat_type=fbank64_stmn
 
 
diff --git a/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_resnet34_arcs30m0.3_trn_alllangs_nocv_nocnceleb_highnoise_adam_lr0.01.amp.v1.sh b/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_resnet34_arcs30m0.3_trn_alllangs_nocv_nocnceleb_highnoise_adam_lr0.01.amp.v1.sh
@@ -1,7 +1,7 @@
 # ResNet34 x-vector with mixed precision training
 
 # acoustic features
-feat_config=conf/fbank64_mvn_8k.yaml
+feat_config=conf/fbank64_stmn_8k.yaml
 feat_type=fbank64_stmn
 
 
diff --git a/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_resnet34_eina_hln_arcs30m0.3_trn_alllangs_nocv_adam_lr0.01.amp.v1.sh b/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_resnet34_eina_hln_arcs30m0.3_trn_alllangs_nocv_adam_lr0.01.amp.v1.sh
@@ -1,7 +1,7 @@
 # ResNet34 x-vector with mixed precision training
 
 # acoustic features
-feat_config=conf/fbank64_mvn_8k.yaml
+feat_config=conf/fbank64_stmn_8k.yaml
 feat_type=fbank64_stmn
 
 
diff --git a/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_resnet34_eina_hln_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.01.amp.v1.sh b/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_resnet34_eina_hln_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.01.amp.v1.sh
@@ -1,7 +1,7 @@
 # ResNet34 x-vector with mixed precision training
 
 # acoustic features
-feat_config=conf/fbank64_mvn_8k.yaml
+feat_config=conf/fbank64_stmn_8k.yaml
 feat_type=fbank64_stmn
 
 
diff --git a/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_resnet34_eina_hln_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.01.amp.v3.sh b/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_resnet34_eina_hln_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.01.amp.v3.sh
@@ -1,7 +1,7 @@
 # ResNet34 x-vector with mixed precision training
 
 # acoustic features
-feat_config=conf/fbank64_mvn_8k.yaml
+feat_config=conf/fbank64_stmn_8k.yaml
 feat_type=fbank64_stmn
 
 
diff --git a/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_resnet34_eina_hln_bmhah64d8192_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.01.amp.v2.sh b/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_resnet34_eina_hln_bmhah64d8192_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.01.amp.v2.sh
@@ -1,7 +1,7 @@
 # ResNet34 x-vector with mixed precision training
 
 # acoustic features
-feat_config=conf/fbank64_mvn_8k.yaml
+feat_config=conf/fbank64_stmn_8k.yaml
 feat_type=fbank64_stmn
 
 
diff --git a/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_resnet34_eina_hln_chattstatsi128_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.01.amp.v1.sh b/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_resnet34_eina_hln_chattstatsi128_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.01.amp.v1.sh
@@ -1,7 +1,7 @@
 # ResNet34 x-vector with mixed precision training
 
 # acoustic features
-feat_config=conf/fbank64_mvn_8k.yaml
+feat_config=conf/fbank64_stmn_8k.yaml
 feat_type=fbank64_stmn
 
 
diff --git a/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_resnet34_eina_hln_chbattstatsi128_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.01.amp.v1.sh b/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_resnet34_eina_hln_chbattstatsi128_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.01.amp.v1.sh
@@ -1,7 +1,7 @@
 # ResNet34 x-vector with mixed precision training
 
 # acoustic features
-feat_config=conf/fbank64_mvn_8k.yaml
+feat_config=conf/fbank64_stmn_8k.yaml
 feat_type=fbank64_stmn
 
 
diff --git a/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_transformer_lac6b12d2048h8ff2048_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.005.amp.v1.sh b/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_transformer_lac6b12d2048h8ff2048_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.005.amp.v1.sh
@@ -5,7 +5,7 @@
 # input is downsampled x4 by conv network
 
 # acoustic features
-feat_config=conf/fbank64_mvn_8k.yaml
+feat_config=conf/fbank64_stmn_8k.yaml
 feat_type=fbank64_stmn
 
 
diff --git a/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_transformer_lac6b12d720h8ff2880_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.005.amp.v1.sh b/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_transformer_lac6b12d720h8ff2880_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.005.amp.v1.sh
@@ -5,7 +5,7 @@
 # input is downsampled x4 by conv network
 
 # acoustic features
-feat_config=conf/fbank64_mvn_8k.yaml
+feat_config=conf/fbank64_stmn_8k.yaml
 feat_type=fbank64_stmn
 
 
diff --git a/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_transformer_lac6b6d512h8ff2048_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.005.amp.v1.sh b/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_transformer_lac6b6d512h8ff2048_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.005.amp.v1.sh
@@ -5,7 +5,7 @@
 # input is downsampled x4 by conv network
 
 # acoustic features
-feat_config=conf/fbank64_mvn_8k.yaml
+feat_config=conf/fbank64_stmn_8k.yaml
 feat_type=fbank64_stmn
 
 
diff --git a/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_transformer_lac6b8d608h8ff2432_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.005.amp.v1.sh b/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_transformer_lac6b8d608h8ff2432_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.005.amp.v1.sh
@@ -5,7 +5,7 @@
 # input is downsampled x4 by conv network
 
 # acoustic features
-feat_config=conf/fbank64_mvn_8k.yaml
+feat_config=conf/fbank64_stmn_8k.yaml
 feat_type=fbank64_stmn
 
 
diff --git a/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_tseres2net50w26s4_r256_chattstatsi128_eina_hln_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.01.amp.v1.sh b/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_tseres2net50w26s4_r256_chattstatsi128_eina_hln_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.01.amp.v1.sh
@@ -1,7 +1,7 @@
 # Res2Net50 x-vector with mixed precision training
 
 # acoustic features
-feat_config=conf/fbank64_mvn_8k.yaml
+feat_config=conf/fbank64_stmn_8k.yaml
 feat_type=fbank64_stmn
 
 
diff --git a/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_tseres2net50w26s4_r256_chattstatsi128_eina_hln_arcs30m0.3_trn_alllangs_nocveng_adam_lr0.01.amp.v1.sh b/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_tseres2net50w26s4_r256_chattstatsi128_eina_hln_arcs30m0.3_trn_alllangs_nocveng_adam_lr0.01.amp.v1.sh
@@ -1,7 +1,7 @@
 # Res2Net50 x-vector with mixed precision training
 
 # acoustic features
-feat_config=conf/fbank64_mvn_8k.yaml
+feat_config=conf/fbank64_stmn_8k.yaml
 feat_type=fbank64_stmn
 
 
diff --git a/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_tseresnet34_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.01.amp.v1.sh b/egs/sre20-cts/v1/global_conf/config_fbank64_stmn_tseresnet34_arcs30m0.3_trn_alllangs_nocv_nocnceleb_adam_lr0.01.amp.v1.sh
@@ -1,7 +1,7 @@
 # ResNet34 x-vector with mixed precision training
 
 # acoustic features
-feat_config=conf/fbank64_mvn_8k.yaml
+feat_config=conf/fbank64_stmn_8k.yaml
 feat_type=fbank64_stmn
 
 
