init
Browse files- .gitattributes +4 -0
- .hyperparams.yaml.swp +0 -0
- embedding_model.ckpt +3 -0
- hyperparams.yaml +42 -0
.gitattributes
CHANGED
@@ -32,3 +32,7 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
32 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
33 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
34 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
32 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
33 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
34 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
35 |
+
classifier.ckpt filter=lfs diff=lfs merge=lfs -text
|
36 |
+
embedding_model.ckpt filter=lfs diff=lfs merge=lfs -text
|
37 |
+
label_encoder.ckpt filter=lfs diff=lfs merge=lfs -text
|
38 |
+
mean_var_norm_emb.ckpt filter=lfs diff=lfs merge=lfs -text
|
.hyperparams.yaml.swp
ADDED
Binary file (4.1 kB). View file
|
|
embedding_model.ckpt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:82e2efb502c70ac145176ae6698077f55eae03e42d6768c2cf4c2532797dd3f5
|
3 |
+
size 83313275
|
hyperparams.yaml
ADDED
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# ############################################################################
|
2 |
+
# Model: ECAPA big for Speaker verification
|
3 |
+
# ############################################################################
|
4 |
+
|
5 |
+
# Feature parameters
|
6 |
+
n_mels: 80
|
7 |
+
|
8 |
+
# Pretrain folder (HuggingFace)
|
9 |
+
pretrained_path: Ubenwa/ecapa-voxceleb-ft2-cryceleb
|
10 |
+
|
11 |
+
|
12 |
+
# Model params
|
13 |
+
compute_features: !new:speechbrain.lobes.features.Fbank
|
14 |
+
n_mels: !ref <n_mels>
|
15 |
+
|
16 |
+
mean_var_norm: !new:speechbrain.processing.features.InputNormalization
|
17 |
+
norm_type: sentence
|
18 |
+
std_norm: False
|
19 |
+
|
20 |
+
embedding_model: !new:speechbrain.lobes.models.ECAPA_TDNN.ECAPA_TDNN
|
21 |
+
input_size: !ref <n_mels>
|
22 |
+
channels: [1024, 1024, 1024, 1024, 3072]
|
23 |
+
kernel_sizes: [5, 3, 3, 3, 1]
|
24 |
+
dilations: [1, 2, 3, 4, 1]
|
25 |
+
attention_channels: 128
|
26 |
+
lin_neurons: 192
|
27 |
+
|
28 |
+
modules:
|
29 |
+
compute_features: !ref <compute_features>
|
30 |
+
mean_var_norm: !ref <mean_var_norm>
|
31 |
+
embedding_model: !ref <embedding_model>
|
32 |
+
|
33 |
+
label_encoder: !new:speechbrain.dataio.encoder.CategoricalEncoder
|
34 |
+
|
35 |
+
|
36 |
+
pretrainer: !new:speechbrain.utils.parameter_transfer.Pretrainer
|
37 |
+
loadables:
|
38 |
+
embedding_model: !ref <embedding_model>
|
39 |
+
|
40 |
+
paths:
|
41 |
+
embedding_model: !ref <pretrained_path>/embedding_model.ckpt
|
42 |
+
|