init
Browse files- experiment_cache/cluster/w2v_bert_se.expresso.speaker_id.csv +0 -3
- experiment_cache/cluster/w2v_bert_se.expresso.style.csv +0 -3
- experiment_cache/cluster/w2v_bert_se.voxceleb1-test-split.speaker_id.csv +0 -3
- experiment_cache/embeddings/w2v_bert_se.expresso.json +0 -3
- experiment_cache/embeddings/w2v_bert_se.voxceleb1-test-split.json +0 -3
- experiment_cache/figure/2d.latent_space.w2v_bert_se.expresso.speaker_id.png +0 -3
- experiment_cache/figure/2d.latent_space.w2v_bert_se.expresso.style.png +0 -3
- experiment_cache/figure/2d.latent_space.w2v_bert_se.voxceleb1-test-split.speaker_id.png +0 -3
- experiment_cache/tsne/w2v_bert_se.expresso.speaker_id.npy +0 -3
- experiment_cache/tsne/w2v_bert_se.expresso.style.npy +0 -3
- experiment_cache/tsne/w2v_bert_se.voxceleb1-test-split.speaker_id.npy +0 -3
- experiment_speaker_verification.py +29 -28
- model_speaker_embedding.py +1 -4
- test.py +0 -1
experiment_cache/cluster/w2v_bert_se.expresso.speaker_id.csv
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:5a7b8722c152981c4e514dfaf90eb17dbefbc653dbee44c59bee558aa2a9cb52
|
3 |
-
size 162416
|
|
|
|
|
|
|
|
experiment_cache/cluster/w2v_bert_se.expresso.style.csv
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:201cafca59ca308d8882d3434dd2637fe49a0c5b11b6cb8bccec7396bfc170be
|
3 |
-
size 181626
|
|
|
|
|
|
|
|
experiment_cache/cluster/w2v_bert_se.voxceleb1-test-split.speaker_id.csv
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:d0a9a82594fdef76a933e5a21534a8af37487f7ec1b6b2e4662be6950c52ce76
|
3 |
-
size 123308
|
|
|
|
|
|
|
|
experiment_cache/embeddings/w2v_bert_se.expresso.json
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:6adea5c716277b2525cf78b38234da223ff9af3722635aeae19ac26e78226ef5
|
3 |
-
size 268823097
|
|
|
|
|
|
|
|
experiment_cache/embeddings/w2v_bert_se.voxceleb1-test-split.json
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:03324e439f856b30e136b3b13fe9d841919b792aeaca8a6f7d65787977dca1dc
|
3 |
-
size 112682448
|
|
|
|
|
|
|
|
experiment_cache/figure/2d.latent_space.w2v_bert_se.expresso.speaker_id.png
DELETED
Git LFS Details
|
experiment_cache/figure/2d.latent_space.w2v_bert_se.expresso.style.png
DELETED
Git LFS Details
|
experiment_cache/figure/2d.latent_space.w2v_bert_se.voxceleb1-test-split.speaker_id.png
DELETED
Git LFS Details
|
experiment_cache/tsne/w2v_bert_se.expresso.speaker_id.npy
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:e186ef5932df77876df41323523fdef80e17d9ae2db2e1f6452a66cd5c4d0a61
|
3 |
-
size 93048
|
|
|
|
|
|
|
|
experiment_cache/tsne/w2v_bert_se.expresso.style.npy
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:600c2be510bb6044ab2d38a87a88ac64224d832f622b0bc8647e5c07ff2c03de
|
3 |
-
size 93048
|
|
|
|
|
|
|
|
experiment_cache/tsne/w2v_bert_se.voxceleb1-test-split.speaker_id.npy
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:5d09ce160684b68806d052b6c699819e41410df0a47fe5f1bb51bb4039b1d9c6
|
3 |
-
size 39120
|
|
|
|
|
|
|
|
experiment_speaker_verification.py
CHANGED
@@ -16,10 +16,11 @@ from datasets import load_dataset
|
|
16 |
|
17 |
from model_meta_voice import MetaVoiceEmbedding
|
18 |
from model_pyannote_embedding import PyannoteEmbedding
|
19 |
-
from model_w2v_bert import W2VBERTEmbedding
|
20 |
from model_clap import CLAPEmbedding, CLAPGeneralEmbedding
|
21 |
-
from
|
22 |
-
|
|
|
|
|
23 |
|
24 |
|
25 |
def get_embedding(model_class, model_name: str, dataset_name: str, data_split: str):
|
@@ -118,68 +119,68 @@ def analyze_embedding(model_name: str, dataset_name: str, n_shot: int = 5, n_cro
|
|
118 |
if __name__ == '__main__':
|
119 |
# get_embedding(MetaVoiceEmbedding, "meta_voice_se", "asahi417/voxceleb1-test-split", "test")
|
120 |
# get_embedding(PyannoteEmbedding, "pyannote_se", "asahi417/voxceleb1-test-split", "test")
|
121 |
-
# get_embedding(W2VBERTEmbedding, "w2v_bert_se", "asahi417/voxceleb1-test-split", "test")
|
122 |
# get_embedding(CLAPEmbedding, "clap_se", "asahi417/voxceleb1-test-split", "test")
|
123 |
# get_embedding(CLAPGeneralEmbedding, "clap_general_se", "asahi417/voxceleb1-test-split", "test")
|
124 |
# get_embedding(HuBERTBaseEmbedding, "hubert_base_se", "asahi417/voxceleb1-test-split", "test")
|
125 |
# get_embedding(HuBERTLargeEmbedding, "hubert_large_se", "asahi417/voxceleb1-test-split", "test")
|
126 |
# get_embedding(HuBERTXLEmbedding, "hubert_xl_se", "asahi417/voxceleb1-test-split", "test")
|
127 |
-
get_embedding(
|
128 |
-
get_embedding(
|
129 |
-
get_embedding(
|
130 |
-
get_embedding(
|
|
|
131 |
|
132 |
# get_embedding(MetaVoiceEmbedding, "meta_voice_se", "ylacombe/expresso", "train")
|
133 |
# get_embedding(PyannoteEmbedding, "pyannote_se", "ylacombe/expresso", "train")
|
134 |
-
# get_embedding(W2VBERTEmbedding, "w2v_bert_se", "ylacombe/expresso", "train")
|
135 |
# get_embedding(CLAPEmbedding, "clap_se", "ylacombe/expresso", "train")
|
136 |
# get_embedding(CLAPGeneralEmbedding, "clap_general_se", "ylacombe/expresso", "train")
|
137 |
# get_embedding(HuBERTBaseEmbedding, "hubert_base_se", "ylacombe/expresso", "train")
|
138 |
# get_embedding(HuBERTLargeEmbedding, "hubert_large_se", "ylacombe/expresso", "train")
|
139 |
# get_embedding(HuBERTXLEmbedding, "hubert_xl_se", "ylacombe/expresso", "train")
|
140 |
-
get_embedding(
|
141 |
-
get_embedding(
|
142 |
-
get_embedding(
|
143 |
-
get_embedding(
|
|
|
144 |
|
145 |
# cluster_embedding("meta_voice_se", "asahi417/voxceleb1-test-split", "speaker_id")
|
146 |
# cluster_embedding("pyannote_se", "asahi417/voxceleb1-test-split", "speaker_id")
|
147 |
-
# cluster_embedding("w2v_bert_se", "asahi417/voxceleb1-test-split", "speaker_id")
|
148 |
# cluster_embedding("clap_se", "asahi417/voxceleb1-test-split", "speaker_id")
|
149 |
# cluster_embedding("clap_general_se", "asahi417/voxceleb1-test-split", "speaker_id")
|
150 |
# cluster_embedding("hubert_base_se", "asahi417/voxceleb1-test-split", "speaker_id")
|
151 |
# cluster_embedding("hubert_large_se", "asahi417/voxceleb1-test-split", "speaker_id")
|
152 |
# cluster_embedding("hubert_xl_se", "asahi417/voxceleb1-test-split", "speaker_id")
|
153 |
-
cluster_embedding("
|
154 |
-
cluster_embedding("
|
155 |
-
cluster_embedding("
|
156 |
-
cluster_embedding("
|
|
|
157 |
|
158 |
# cluster_embedding("meta_voice_se", "ylacombe/expresso", "speaker_id")
|
159 |
# cluster_embedding("pyannote_se", "ylacombe/expresso", "speaker_id")
|
160 |
-
# cluster_embedding("w2v_bert_se", "ylacombe/expresso", "speaker_id")
|
161 |
# cluster_embedding("clap_se", "ylacombe/expresso", "speaker_id")
|
162 |
# cluster_embedding("clap_general_se", "ylacombe/expresso", "speaker_id")
|
163 |
# cluster_embedding("hubert_base_se", "ylacombe/expresso", "speaker_id")
|
164 |
# cluster_embedding("hubert_large_se", "ylacombe/expresso", "speaker_id")
|
165 |
# cluster_embedding("hubert_xl_se", "ylacombe/expresso", "speaker_id")
|
166 |
-
cluster_embedding("
|
167 |
-
cluster_embedding("
|
168 |
-
cluster_embedding("
|
169 |
-
cluster_embedding("
|
|
|
170 |
|
171 |
# cluster_embedding("meta_voice_se", "ylacombe/expresso", "style")
|
172 |
# cluster_embedding("pyannote_se", "ylacombe/expresso", "style")
|
173 |
-
# cluster_embedding("w2v_bert_se", "ylacombe/expresso", "style")
|
174 |
# cluster_embedding("clap_se", "ylacombe/expresso", "style")
|
175 |
# cluster_embedding("clap_general_se", "ylacombe/expresso", "style")
|
176 |
# cluster_embedding("hubert_base_se", "ylacombe/expresso", "style")
|
177 |
# cluster_embedding("hubert_large_se", "ylacombe/expresso", "style")
|
178 |
# cluster_embedding("hubert_xl_se", "ylacombe/expresso", "style")
|
179 |
-
cluster_embedding("
|
180 |
-
cluster_embedding("
|
181 |
-
cluster_embedding("
|
182 |
-
cluster_embedding("
|
|
|
183 |
|
184 |
|
185 |
|
|
|
16 |
|
17 |
from model_meta_voice import MetaVoiceEmbedding
|
18 |
from model_pyannote_embedding import PyannoteEmbedding
|
|
|
19 |
from model_clap import CLAPEmbedding, CLAPGeneralEmbedding
|
20 |
+
from model_speaker_embedding import (
|
21 |
+
W2VBERTEmbedding, Wav2VecEmbedding, XLSR300MEmbedding, XLSR1BEmbedding, XLSR2BEmbedding,
|
22 |
+
HuBERTBaseEmbedding, HuBERTLargeEmbedding, HuBERTXLEmbedding
|
23 |
+
)
|
24 |
|
25 |
|
26 |
def get_embedding(model_class, model_name: str, dataset_name: str, data_split: str):
|
|
|
119 |
if __name__ == '__main__':
|
120 |
# get_embedding(MetaVoiceEmbedding, "meta_voice_se", "asahi417/voxceleb1-test-split", "test")
|
121 |
# get_embedding(PyannoteEmbedding, "pyannote_se", "asahi417/voxceleb1-test-split", "test")
|
|
|
122 |
# get_embedding(CLAPEmbedding, "clap_se", "asahi417/voxceleb1-test-split", "test")
|
123 |
# get_embedding(CLAPGeneralEmbedding, "clap_general_se", "asahi417/voxceleb1-test-split", "test")
|
124 |
# get_embedding(HuBERTBaseEmbedding, "hubert_base_se", "asahi417/voxceleb1-test-split", "test")
|
125 |
# get_embedding(HuBERTLargeEmbedding, "hubert_large_se", "asahi417/voxceleb1-test-split", "test")
|
126 |
# get_embedding(HuBERTXLEmbedding, "hubert_xl_se", "asahi417/voxceleb1-test-split", "test")
|
127 |
+
get_embedding(W2VBERTEmbedding, "w2v_bert_se", "asahi417/voxceleb1-test-split", "test")
|
128 |
+
# get_embedding(Wav2VecEmbedding, "wav2vec_se", "asahi417/voxceleb1-test-split", "test")
|
129 |
+
# get_embedding(XLSR300MEmbedding, "xlsr_300m_se", "asahi417/voxceleb1-test-split", "test")
|
130 |
+
# get_embedding(XLSR1BEmbedding, "xlsr_1b_se", "asahi417/voxceleb1-test-split", "test")
|
131 |
+
# get_embedding(XLSR2BEmbedding, "xlsr_2b_se", "asahi417/voxceleb1-test-split", "test")
|
132 |
|
133 |
# get_embedding(MetaVoiceEmbedding, "meta_voice_se", "ylacombe/expresso", "train")
|
134 |
# get_embedding(PyannoteEmbedding, "pyannote_se", "ylacombe/expresso", "train")
|
|
|
135 |
# get_embedding(CLAPEmbedding, "clap_se", "ylacombe/expresso", "train")
|
136 |
# get_embedding(CLAPGeneralEmbedding, "clap_general_se", "ylacombe/expresso", "train")
|
137 |
# get_embedding(HuBERTBaseEmbedding, "hubert_base_se", "ylacombe/expresso", "train")
|
138 |
# get_embedding(HuBERTLargeEmbedding, "hubert_large_se", "ylacombe/expresso", "train")
|
139 |
# get_embedding(HuBERTXLEmbedding, "hubert_xl_se", "ylacombe/expresso", "train")
|
140 |
+
get_embedding(W2VBERTEmbedding, "w2v_bert_se", "ylacombe/expresso", "train")
|
141 |
+
# get_embedding(Wav2VecEmbedding, "wav2vec_se", "ylacombe/expresso", "train")
|
142 |
+
# get_embedding(XLSR300MEmbedding, "xlsr_300m_se", "ylacombe/expresso", "train")
|
143 |
+
# get_embedding(XLSR1BEmbedding, "xlsr_1b_se", "ylacombe/expresso", "train")
|
144 |
+
# get_embedding(XLSR2BEmbedding, "xlsr_2b_se", "ylacombe/expresso", "train")
|
145 |
|
146 |
# cluster_embedding("meta_voice_se", "asahi417/voxceleb1-test-split", "speaker_id")
|
147 |
# cluster_embedding("pyannote_se", "asahi417/voxceleb1-test-split", "speaker_id")
|
|
|
148 |
# cluster_embedding("clap_se", "asahi417/voxceleb1-test-split", "speaker_id")
|
149 |
# cluster_embedding("clap_general_se", "asahi417/voxceleb1-test-split", "speaker_id")
|
150 |
# cluster_embedding("hubert_base_se", "asahi417/voxceleb1-test-split", "speaker_id")
|
151 |
# cluster_embedding("hubert_large_se", "asahi417/voxceleb1-test-split", "speaker_id")
|
152 |
# cluster_embedding("hubert_xl_se", "asahi417/voxceleb1-test-split", "speaker_id")
|
153 |
+
cluster_embedding("w2v_bert_se", "asahi417/voxceleb1-test-split", "speaker_id")
|
154 |
+
# cluster_embedding("wav2vec_se", "asahi417/voxceleb1-test-split", "speaker_id")
|
155 |
+
# cluster_embedding("xlsr_300m_se", "asahi417/voxceleb1-test-split", "speaker_id")
|
156 |
+
# cluster_embedding("xlsr_1b_se", "asahi417/voxceleb1-test-split", "speaker_id")
|
157 |
+
# cluster_embedding("xlsr_2b_se", "asahi417/voxceleb1-test-split", "speaker_id")
|
158 |
|
159 |
# cluster_embedding("meta_voice_se", "ylacombe/expresso", "speaker_id")
|
160 |
# cluster_embedding("pyannote_se", "ylacombe/expresso", "speaker_id")
|
|
|
161 |
# cluster_embedding("clap_se", "ylacombe/expresso", "speaker_id")
|
162 |
# cluster_embedding("clap_general_se", "ylacombe/expresso", "speaker_id")
|
163 |
# cluster_embedding("hubert_base_se", "ylacombe/expresso", "speaker_id")
|
164 |
# cluster_embedding("hubert_large_se", "ylacombe/expresso", "speaker_id")
|
165 |
# cluster_embedding("hubert_xl_se", "ylacombe/expresso", "speaker_id")
|
166 |
+
cluster_embedding("w2v_bert_se", "ylacombe/expresso", "speaker_id")
|
167 |
+
# cluster_embedding("wav2vec_se", "ylacombe/expresso", "speaker_id")
|
168 |
+
# cluster_embedding("xlsr_300m_se", "ylacombe/expresso", "speaker_id")
|
169 |
+
# cluster_embedding("xlsr_1b_se", "ylacombe/expresso", "speaker_id")
|
170 |
+
# cluster_embedding("xlsr_2b_se", "ylacombe/expresso", "speaker_id")
|
171 |
|
172 |
# cluster_embedding("meta_voice_se", "ylacombe/expresso", "style")
|
173 |
# cluster_embedding("pyannote_se", "ylacombe/expresso", "style")
|
|
|
174 |
# cluster_embedding("clap_se", "ylacombe/expresso", "style")
|
175 |
# cluster_embedding("clap_general_se", "ylacombe/expresso", "style")
|
176 |
# cluster_embedding("hubert_base_se", "ylacombe/expresso", "style")
|
177 |
# cluster_embedding("hubert_large_se", "ylacombe/expresso", "style")
|
178 |
# cluster_embedding("hubert_xl_se", "ylacombe/expresso", "style")
|
179 |
+
cluster_embedding("w2v_bert_se", "ylacombe/expresso", "style")
|
180 |
+
# cluster_embedding("wav2vec_se", "ylacombe/expresso", "style")
|
181 |
+
# cluster_embedding("xlsr_300m_se", "ylacombe/expresso", "style")
|
182 |
+
# cluster_embedding("xlsr_1b_se", "ylacombe/expresso", "style")
|
183 |
+
# cluster_embedding("xlsr_2b_se", "ylacombe/expresso", "style")
|
184 |
|
185 |
|
186 |
|
model_speaker_embedding.py
CHANGED
@@ -1,7 +1,4 @@
|
|
1 |
-
"""Meta's w2vBERT based speaker embedding.
|
2 |
-
- feature dimension: 1024
|
3 |
-
- source: https://huggingface.co/facebook/w2v-bert-2.0
|
4 |
-
"""
|
5 |
from typing import Optional
|
6 |
|
7 |
import torch
|
|
|
1 |
+
"""Meta's w2vBERT based speaker embedding."""
|
|
|
|
|
|
|
2 |
from typing import Optional
|
3 |
|
4 |
import torch
|
test.py
CHANGED
@@ -33,6 +33,5 @@ def test():
|
|
33 |
print(v.shape)
|
34 |
|
35 |
|
36 |
-
|
37 |
if __name__ == '__main__':
|
38 |
test()
|
|
|
33 |
print(v.shape)
|
34 |
|
35 |
|
|
|
36 |
if __name__ == '__main__':
|
37 |
test()
|