Upload 22 files

.gitattributes

@@ -33,3 +33,9 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+models/byol_tgt_layer_6/checkpoint filter=lfs diff=lfs merge=lfs -text
+models/byol/checkpoint filter=lfs diff=lfs merge=lfs -text
+models/byol/checkpoint_sid filter=lfs diff=lfs merge=lfs -text
+models/dino/checkpoint filter=lfs diff=lfs merge=lfs -text
+models/hubert/checkpoint_sid filter=lfs diff=lfs merge=lfs -text
+models/vg-hubert_3/checkpoint_sid filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,11 @@
+---
+license: mit
+datasets:
+- openslr/librispeech_asr
+language:
+- en
+---
+
+# Self-Supervised Syllable Discovery Based on Speaker-Disentangled HuBERT
+
+This is the official model repository of [a GitHub project](https://github.com/ryota-komatsu/speaker_disentangled_hubert).

models/byol/checkpoint

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5a615c08753d835e1b45960c2c8d5723bd1e9186a9749bd3e637aa6f5d2f0ced
+size 1444845288

models/byol/checkpoint_sid

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:46e3da2370c3fd40f30ad0cd7977e197aeacc4f9d483aacbc3895ba242b9a47d
+size 383784814

models/byol/quantizer1.joblib

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7c329cacd4f3f84387e6a16b623935082e330f3e45e0068fc79029935af4ae1d
+size 50398058

models/byol/quantizer2.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c77213ed62a78b046e0267baad2f575b68fa6a979dfced83ab83537965f56a33
+size 131200

models/byol_tgt_layer_6/checkpoint

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:113dbcb3b7bf5605465c9c5523b7123fef13561f9ec73ac1421d3b01e7d7b069
+size 1457428200

models/byol_tgt_layer_6/quantizer1.joblib

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d2536bade704d94ce9d761330c6201fc78483e1c58f89ed8d0d547d458e279c8
+size 50398058

models/byol_tgt_layer_6/quantizer2.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f6ebc267396f21ba9a60c2f8abce3044988edd4c3ecfca7ae654f4a119967634
+size 131200

models/dino/checkpoint

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0d074285e61d15e8eb49ffa81be7061eb04568896d1dfb11c91c907ee603d49e
+size 1516041480

models/dino/quantizer1.joblib

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:eb4199b374ea13b20a145d5d4123a728f6933c0fa1accabdedaa8f492628726a
+size 50398058

models/dino/quantizer2.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d0e7347fd57e478a9e228e546f150cd3f2591999ac408d74087218f632354cc3
+size 131200

models/hubert/checkpoint_sid

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2562f5d5c0c8aee8559f3bf1f64c55b8564242ebde01801293897d58b62fcd8b
+size 383784814

models/hubert/quantizer1.joblib

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:730d7408a8f48d91185eadee509a39ad93bfc9b799b524c4890a4371e8b0e5f8
+size 50398058

models/hubert/quantizer2.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:62db8041c381d3e26caeb429d53edfa59f7c385a7d097b0532a7b19b445aa810
+size 131200

models/vg-hubert_3/checkpoint_sid

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f8be075bab4362d94f8cc14d4603532a126174885fa6aca14ba59da17cdb460b
+size 383781359

models/vg-hubert_3/quantizer1.joblib

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e30a79b1273298b09b3d08671bb635270b15532d7601eba61bd695142e0ce853
+size 50398058

models/vg-hubert_3/quantizer2.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6cfbb8dda51794e2630d3d26d3fc91a577c84335933401b0c0c66107c608cf58
+size 131200

results/byol.json

@@ -0,0 +1,20 @@
+{
+    "segmentation": {
+        "match_gt": 110641,
+        "match_pred": 110641,
+        "gt_len": 163763,
+        "pred_len": 150931,
+        "prec": 0.7330568272919413,
+        "recall": 0.6756165922705373,
+        "f1": 0.7031656148512523,
+        "os": -0.07835713805926858,
+        "r1": 0.3337131047696781,
+        "r2": -0.1739668436338245,
+        "r_val": 0.7461600257982487
+    },
+    "clustering": {
+        "syllable_purity": 0.5935364727608494,
+        "cluster_purity": 0.4447403934938561,
+        "mutual_info": 5.07649445946761
+    }
+}

results/byol_tgt_layer_6.json

@@ -0,0 +1,20 @@
+{
+    "segmentation": {
+        "match_gt": 37215,
+        "match_pred": 37215,
+        "gt_len": 163763,
+        "pred_len": 81606,
+        "prec": 0.456032644658481,
+        "recall": 0.22724913441986286,
+        "f1": 0.303339052610558,
+        "os": -0.5016823091907208,
+        "r1": 0.9213191844359832,
+        "r2": -0.19167441438940436,
+        "r_val": 0.4435032005873063
+    },
+    "clustering": {
+        "syllable_purity": 0.10408265804357068,
+        "cluster_purity": 0.1427706419654475,
+        "mutual_info": 2.166068535065666
+    }
+}

results/dino.json

@@ -0,0 +1,20 @@
+{
+    "segmentation": {
+        "match_gt": 106629,
+        "match_pred": 106629,
+        "gt_len": 163763,
+        "pred_len": 165706,
+        "prec": 0.6434830362207765,
+        "recall": 0.6511177738561214,
+        "f1": 0.6472778926090162,
+        "os": 0.011864706924030521,
+        "r1": 0.34908391396554156,
+        "r2": -0.2550866026645681,
+        "r_val": 0.6979147416849452
+    },
+    "clustering": {
+        "syllable_purity": 0.590599474096453,
+        "cluster_purity": 0.42873727201521766,
+        "mutual_info": 5.060749923481224
+    }
+}

results/hubert.json

@@ -0,0 +1,20 @@
+{
+    "segmentation": {
+        "match_gt": 51481,
+        "match_pred": 51481,
+        "gt_len": 163763,
+        "pred_len": 100135,
+        "prec": 0.514115943476307,
+        "recall": 0.3143628292105054,
+        "f1": 0.39015831874436335,
+        "os": -0.38853709323839936,
+        "r1": 0.7880732217188751,
+        "r2": -0.2100814795274286,
+        "r_val": 0.5009226493768482
+    },
+    "clustering": {
+        "syllable_purity": 0.33106618025927986,
+        "cluster_purity": 0.28381664914775656,
+        "mutual_info": 3.5403592325037194
+    }
+}

results/vg-hubert_3.json

@@ -0,0 +1,20 @@
+{
+    "segmentation": {
+        "match_gt": 105266,
+        "match_pred": 105266,
+        "gt_len": 163763,
+        "pred_len": 161292,
+        "prec": 0.6526424125189098,
+        "recall": 0.6427947704915029,
+        "f1": 0.6476811616495669,
+        "os": -0.015088878440184916,
+        "r1": 0.357523775769949,
+        "r2": -0.2419127917952011,
+        "r_val": 0.700281716217425
+    },
+    "clustering": {
+        "syllable_purity": 0.534248207514771,
+        "cluster_purity": 0.4357403682391943,
+        "mutual_info": 4.6595905283191055
+    }
+}