Upload SSL Layer MOS checkpoints (part 9)

results_bvcc/w2v2_large_asr_en/layer10/run2/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:17c954b27d304f577e9fd998036c8afa372b2fb1bb1ecb167af33267fe87ebd6
+size 3403339

results_bvcc/w2v2_large_asr_en/layer10/run2/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b982e79fea04f796dbed29dabf4b3392178d6f9164e38998293ad06c9387d387
+size 3404998

results_bvcc/w2v2_large_asr_en/layer10/run2/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.6746184119047263
+INFO:root:
+[Valid][0][UTT][ MSE = 0.4778 | LCC = 0.7159 | SRCC = 0.7216 ]
+INFO:root:Average loss=0.3387069020133752
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3338 | LCC = 0.8053 | SRCC = 0.8036 ]
+INFO:root:Average loss=0.2847767711067811
+INFO:root:
+[Valid][2][UTT][ MSE = 0.3085 | LCC = 0.8174 | SRCC = 0.8155 ]
+INFO:root:Average loss=0.26873792765232235
+INFO:root:
+[Valid][3][UTT][ MSE = 0.3536 | LCC = 0.8174 | SRCC = 0.8175 ]
+INFO:root:Average loss=0.23792060521932748
+INFO:root:
+[Valid][4][UTT][ MSE = 0.3200 | LCC = 0.8162 | SRCC = 0.8128 ]
+INFO:root:Average loss=0.2266076065790959
+INFO:root:
+[Valid][5][UTT][ MSE = 0.3025 | LCC = 0.8133 | SRCC = 0.8105 ]
+INFO:root:Average loss=0.21253541111946106
+INFO:root:
+[Valid][6][UTT][ MSE = 0.3221 | LCC = 0.8146 | SRCC = 0.8157 ]
+INFO:root:Average loss=0.19001203135420114
+INFO:root:
+[Valid][7][UTT][ MSE = 0.3453 | LCC = 0.8117 | SRCC = 0.8131 ]
+INFO:root:Average loss=0.17686052477130523
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2949 | LCC = 0.8136 | SRCC = 0.8142 ]
+INFO:root:Average loss=0.16311080323962066
+INFO:root:
+[Valid][9][UTT][ MSE = 0.3003 | LCC = 0.8100 | SRCC = 0.8084 ]
+INFO:root:Average loss=0.1606394347663109
+INFO:root:
+[Valid][10][UTT][ MSE = 0.3156 | LCC = 0.8080 | SRCC = 0.8061 ]
+INFO:root:Average loss=0.14482228400615546
+INFO:root:
+[Valid][11][UTT][ MSE = 0.3110 | LCC = 0.8137 | SRCC = 0.8113 ]
+INFO:root:Average loss=0.14077927630681258
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2824 | LCC = 0.8181 | SRCC = 0.8155 ]
+INFO:root:Average loss=0.1233119808901579
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2920 | LCC = 0.8107 | SRCC = 0.8101 ]
+INFO:root:Average loss=0.12160828308417247
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2835 | LCC = 0.8124 | SRCC = 0.8111 ]
+INFO:root:Average loss=0.11040873868534198
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2900 | LCC = 0.8126 | SRCC = 0.8101 ]
+INFO:root:Average loss=0.10938038973089977
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2914 | LCC = 0.8137 | SRCC = 0.8119 ]
+INFO:root:Average loss=0.10012815907024421
+INFO:root:
+[Valid][17][UTT][ MSE = 0.3084 | LCC = 0.8083 | SRCC = 0.8057 ]
+INFO:root:Average loss=0.10388120755744286
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2943 | LCC = 0.8107 | SRCC = 0.8089 ]
+INFO:root:Average loss=0.08978882174079235
+INFO:root:
+[Valid][19][UTT][ MSE = 0.3115 | LCC = 0.8074 | SRCC = 0.8049 ]
+INFO:root:Average loss=0.08635260518162678
+INFO:root:
+[Valid][20][UTT][ MSE = 0.3136 | LCC = 0.8057 | SRCC = 0.8069 ]
+INFO:root:Average loss=0.07857741916982028
+INFO:root:
+[Valid][21][UTT][ MSE = 0.2885 | LCC = 0.8124 | SRCC = 0.8121 ]
+INFO:root:Average loss=0.08343375440782462
+INFO:root:
+[Valid][22][UTT][ MSE = 0.3734 | LCC = 0.8007 | SRCC = 0.8059 ]
+INFO:root:Average loss=0.08682268590499194
+INFO:root:
+[Valid][23][UTT][ MSE = 0.3107 | LCC = 0.8072 | SRCC = 0.8076 ]
+INFO:root:Average loss=0.07026724560329548
+INFO:root:
+[Valid][24][UTT][ MSE = 0.2972 | LCC = 0.8077 | SRCC = 0.8065 ]
+INFO:root:Average loss=0.0699538896099115
+INFO:root:
+[Valid][25][UTT][ MSE = 0.3158 | LCC = 0.8078 | SRCC = 0.8084 ]
+INFO:root:Average loss=0.0652948336628003
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2908 | LCC = 0.8137 | SRCC = 0.8124 ]
+INFO:root:Average loss=0.06554413371934341
+INFO:root:
+[Valid][27][UTT][ MSE = 0.2928 | LCC = 0.8084 | SRCC = 0.8075 ]
+INFO:root:Average loss=0.06051074364819588
+INFO:root:
+[Valid][28][UTT][ MSE = 0.3119 | LCC = 0.8020 | SRCC = 0.8009 ]
+INFO:root:Average loss=0.0663526530067126
+INFO:root:
+[Valid][29][UTT][ MSE = 0.3253 | LCC = 0.8010 | SRCC = 0.8017 ]
+INFO:root:Average loss=0.06058197826720201
+INFO:root:
+[Valid][30][UTT][ MSE = 0.3325 | LCC = 0.8038 | SRCC = 0.8048 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2816 | LCC = 0.8313 | SRCC = 0.8339 ]

results_bvcc/w2v2_large_asr_en/layer10/run3/config.gin

@@ -0,0 +1,31 @@
+# Training config on BVCC dataset.
+
+# ------------------------------
+# Architecture
+# ------------------------------
+ProjectionHead:
+    conv_channels = (32, 32)
+    dense_neurons = (64, 32, 1)
+    use_poolings = (True, True)
+    kernel_size = 5
+    pool_size = 5
+    
+# ------------------------------
+# Dataset
+# ------------------------------
+Bvcc:
+    debug = False
+
+get_dataloader:
+    num_workers = 12
+    shuffle = True
+
+# ------------------------------
+# Training loop
+# ------------------------------
+TrainingLoop:
+    num_epochs = 30
+    learning_rate = 1e-4
+    dataset_cls = @Bvcc
+    weight_decay = 0.0
+    batch_size_train = 64

results_bvcc/w2v2_large_asr_en/layer10/run3/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7a9677dbfb8f2aad402e8378478a102d4599225892a987b760e678b5ed4ccc42
+size 3403339

results_bvcc/w2v2_large_asr_en/layer10/run3/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bef1cdb12deabd38fd53fc24ae27e1f98b2f9318a55b33d24b2f0916d0fb429c
+size 3404998

results_bvcc/w2v2_large_asr_en/layer10/run3/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.5988704699736375
+INFO:root:
+[Valid][0][UTT][ MSE = 0.4256 | LCC = 0.7561 | SRCC = 0.7585 ]
+INFO:root:Average loss=0.3351873986614056
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3306 | LCC = 0.8009 | SRCC = 0.8010 ]
+INFO:root:Average loss=0.2886387335184293
+INFO:root:
+[Valid][2][UTT][ MSE = 0.3175 | LCC = 0.8080 | SRCC = 0.8062 ]
+INFO:root:Average loss=0.2622057657975417
+INFO:root:
+[Valid][3][UTT][ MSE = 0.3165 | LCC = 0.8089 | SRCC = 0.8082 ]
+INFO:root:Average loss=0.2366573931888128
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2980 | LCC = 0.8144 | SRCC = 0.8138 ]
+INFO:root:Average loss=0.22310085451373687
+INFO:root:
+[Valid][5][UTT][ MSE = 0.3147 | LCC = 0.8087 | SRCC = 0.8086 ]
+INFO:root:Average loss=0.23775893927384645
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2993 | LCC = 0.8173 | SRCC = 0.8149 ]
+INFO:root:Average loss=0.19699565645975944
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2875 | LCC = 0.8165 | SRCC = 0.8149 ]
+INFO:root:Average loss=0.17750538675448832
+INFO:root:
+[Valid][8][UTT][ MSE = 0.3055 | LCC = 0.8035 | SRCC = 0.8018 ]
+INFO:root:Average loss=0.18402894739157113
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2866 | LCC = 0.8115 | SRCC = 0.8086 ]
+INFO:root:Average loss=0.15437620715835157
+INFO:root:
+[Valid][10][UTT][ MSE = 0.3052 | LCC = 0.8091 | SRCC = 0.8077 ]
+INFO:root:Average loss=0.15082182897589144
+INFO:root:
+[Valid][11][UTT][ MSE = 0.3023 | LCC = 0.8042 | SRCC = 0.8035 ]
+INFO:root:Average loss=0.14588640639797235
+INFO:root:
+[Valid][12][UTT][ MSE = 0.3007 | LCC = 0.8087 | SRCC = 0.8076 ]
+INFO:root:Average loss=0.13715299237997103
+INFO:root:
+[Valid][13][UTT][ MSE = 0.3086 | LCC = 0.8106 | SRCC = 0.8090 ]
+INFO:root:Average loss=0.12994531336694193
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2964 | LCC = 0.8083 | SRCC = 0.8081 ]
+INFO:root:Average loss=0.11570002855016635
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2924 | LCC = 0.8070 | SRCC = 0.8052 ]
+INFO:root:Average loss=0.10925124208323467
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2987 | LCC = 0.8022 | SRCC = 0.8017 ]
+INFO:root:Average loss=0.10947629188497861
+INFO:root:
+[Valid][17][UTT][ MSE = 0.2971 | LCC = 0.8047 | SRCC = 0.8042 ]
+INFO:root:Average loss=0.11201404579556905
+INFO:root:
+[Valid][18][UTT][ MSE = 0.3016 | LCC = 0.8077 | SRCC = 0.8076 ]
+INFO:root:Average loss=0.09224934384035759
+INFO:root:
+[Valid][19][UTT][ MSE = 0.2914 | LCC = 0.8080 | SRCC = 0.8071 ]
+INFO:root:Average loss=0.08938650796428704
+INFO:root:
+[Valid][20][UTT][ MSE = 0.3218 | LCC = 0.8049 | SRCC = 0.8047 ]
+INFO:root:Average loss=0.09217797339153595
+INFO:root:
+[Valid][21][UTT][ MSE = 0.3163 | LCC = 0.8044 | SRCC = 0.8039 ]
+INFO:root:Average loss=0.08819882896466133
+INFO:root:
+[Valid][22][UTT][ MSE = 0.3013 | LCC = 0.8039 | SRCC = 0.8029 ]
+INFO:root:Average loss=0.07946886528187837
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2983 | LCC = 0.8056 | SRCC = 0.8044 ]
+INFO:root:Average loss=0.07416030201010215
+INFO:root:
+[Valid][24][UTT][ MSE = 0.3147 | LCC = 0.8069 | SRCC = 0.8071 ]
+INFO:root:Average loss=0.07000843511942105
+INFO:root:
+[Valid][25][UTT][ MSE = 0.3102 | LCC = 0.8058 | SRCC = 0.8061 ]
+INFO:root:Average loss=0.07103851313392322
+INFO:root:
+[Valid][26][UTT][ MSE = 0.3075 | LCC = 0.8059 | SRCC = 0.8064 ]
+INFO:root:Average loss=0.07397129558599912
+INFO:root:
+[Valid][27][UTT][ MSE = 0.3003 | LCC = 0.8026 | SRCC = 0.8048 ]
+INFO:root:Average loss=0.0688283399034005
+INFO:root:
+[Valid][28][UTT][ MSE = 0.2973 | LCC = 0.8031 | SRCC = 0.8037 ]
+INFO:root:Average loss=0.06074103789451795
+INFO:root:
+[Valid][29][UTT][ MSE = 0.3053 | LCC = 0.8039 | SRCC = 0.8037 ]
+INFO:root:Average loss=0.05725609396512692
+INFO:root:
+[Valid][30][UTT][ MSE = 0.3043 | LCC = 0.8031 | SRCC = 0.8028 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2946 | LCC = 0.8287 | SRCC = 0.8330 ]

results_bvcc/w2v2_large_asr_en/layer10/run4/config.gin

@@ -0,0 +1,31 @@
+# Training config on BVCC dataset.
+
+# ------------------------------
+# Architecture
+# ------------------------------
+ProjectionHead:
+    conv_channels = (32, 32)
+    dense_neurons = (64, 32, 1)
+    use_poolings = (True, True)
+    kernel_size = 5
+    pool_size = 5
+    
+# ------------------------------
+# Dataset
+# ------------------------------
+Bvcc:
+    debug = False
+
+get_dataloader:
+    num_workers = 12
+    shuffle = True
+
+# ------------------------------
+# Training loop
+# ------------------------------
+TrainingLoop:
+    num_epochs = 30
+    learning_rate = 1e-4
+    dataset_cls = @Bvcc
+    weight_decay = 0.0
+    batch_size_train = 64

results_bvcc/w2v2_large_asr_en/layer10/run4/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:43b280e478e0984619e04a636aa21c3ca58b59766f3a1b907a3e1a28a2bf5628
+size 3403339

results_bvcc/w2v2_large_asr_en/layer10/run4/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e0908d121f460742f853bf06173266c352b2723d6d316379565c5b860f3781a2
+size 3405062

results_bvcc/w2v2_large_asr_en/layer10/run4/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.6203618072546445
+INFO:root:
+[Valid][0][UTT][ MSE = 0.4357 | LCC = 0.7424 | SRCC = 0.7445 ]
+INFO:root:Average loss=0.32978219195054126
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3305 | LCC = 0.8015 | SRCC = 0.8003 ]
+INFO:root:Average loss=0.2830068669640101
+INFO:root:
+[Valid][2][UTT][ MSE = 0.3209 | LCC = 0.8173 | SRCC = 0.8138 ]
+INFO:root:Average loss=0.2473831717402507
+INFO:root:
+[Valid][3][UTT][ MSE = 0.3117 | LCC = 0.8149 | SRCC = 0.8109 ]
+INFO:root:Average loss=0.2436703187532914
+INFO:root:
+[Valid][4][UTT][ MSE = 0.3327 | LCC = 0.8152 | SRCC = 0.8133 ]
+INFO:root:Average loss=0.22828983591917235
+INFO:root:
+[Valid][5][UTT][ MSE = 0.3274 | LCC = 0.8206 | SRCC = 0.8169 ]
+INFO:root:Average loss=0.19720400258516654
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2941 | LCC = 0.8195 | SRCC = 0.8157 ]
+INFO:root:Average loss=0.199496636310449
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2920 | LCC = 0.8163 | SRCC = 0.8138 ]
+INFO:root:Average loss=0.17800079887876144
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2984 | LCC = 0.8118 | SRCC = 0.8102 ]
+INFO:root:Average loss=0.16389273231228194
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2954 | LCC = 0.8107 | SRCC = 0.8117 ]
+INFO:root:Average loss=0.15842864261223719
+INFO:root:
+[Valid][10][UTT][ MSE = 0.3438 | LCC = 0.8145 | SRCC = 0.8152 ]
+INFO:root:Average loss=0.1473681323039226
+INFO:root:
+[Valid][11][UTT][ MSE = 0.3195 | LCC = 0.8119 | SRCC = 0.8122 ]
+INFO:root:Average loss=0.14767956275206345
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2986 | LCC = 0.8129 | SRCC = 0.8106 ]
+INFO:root:Average loss=0.12661838818054932
+INFO:root:
+[Valid][13][UTT][ MSE = 0.3090 | LCC = 0.8052 | SRCC = 0.8046 ]
+INFO:root:Average loss=0.12022751426467529
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2973 | LCC = 0.8087 | SRCC = 0.8068 ]
+INFO:root:Average loss=0.11529687218941174
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2894 | LCC = 0.8115 | SRCC = 0.8105 ]
+INFO:root:Average loss=0.11008151047504865
+INFO:root:
+[Valid][16][UTT][ MSE = 0.3617 | LCC = 0.8013 | SRCC = 0.8013 ]
+INFO:root:Average loss=0.10258800918475175
+INFO:root:
+[Valid][17][UTT][ MSE = 0.3020 | LCC = 0.8053 | SRCC = 0.8036 ]
+INFO:root:Average loss=0.09492622760052864
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2955 | LCC = 0.8096 | SRCC = 0.8073 ]
+INFO:root:Average loss=0.09255884143595512
+INFO:root:
+[Valid][19][UTT][ MSE = 0.3150 | LCC = 0.8041 | SRCC = 0.8029 ]
+INFO:root:Average loss=0.0843467849951524
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2996 | LCC = 0.8023 | SRCC = 0.8005 ]
+INFO:root:Average loss=0.0875364528157008
+INFO:root:
+[Valid][21][UTT][ MSE = 0.3082 | LCC = 0.8034 | SRCC = 0.8041 ]
+INFO:root:Average loss=0.08333408980606458
+INFO:root:
+[Valid][22][UTT][ MSE = 0.3099 | LCC = 0.8018 | SRCC = 0.8019 ]
+INFO:root:Average loss=0.0776115395128727
+INFO:root:
+[Valid][23][UTT][ MSE = 0.3091 | LCC = 0.7993 | SRCC = 0.7975 ]
+INFO:root:Average loss=0.07417628866357681
+INFO:root:
+[Valid][24][UTT][ MSE = 0.3151 | LCC = 0.7950 | SRCC = 0.7930 ]
+INFO:root:Average loss=0.07346136371294658
+INFO:root:
+[Valid][25][UTT][ MSE = 0.3040 | LCC = 0.8005 | SRCC = 0.7995 ]
+INFO:root:Average loss=0.06646742208454853
+INFO:root:
+[Valid][26][UTT][ MSE = 0.3142 | LCC = 0.8004 | SRCC = 0.7992 ]
+INFO:root:Average loss=0.06499218501341648
+INFO:root:
+[Valid][27][UTT][ MSE = 0.3218 | LCC = 0.8002 | SRCC = 0.8004 ]
+INFO:root:Average loss=0.06706518684633267
+INFO:root:
+[Valid][28][UTT][ MSE = 0.3453 | LCC = 0.7962 | SRCC = 0.7969 ]
+INFO:root:Average loss=0.06496112478467134
+INFO:root:
+[Valid][29][UTT][ MSE = 0.3409 | LCC = 0.8005 | SRCC = 0.8014 ]
+INFO:root:Average loss=0.05908839113246172
+INFO:root:
+[Valid][30][UTT][ MSE = 0.3119 | LCC = 0.7975 | SRCC = 0.7962 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.3395 | LCC = 0.8158 | SRCC = 0.8257 ]

results_bvcc/w2v2_large_asr_en/layer12/run0/config.gin

@@ -0,0 +1,31 @@
+# Training config on BVCC dataset.
+
+# ------------------------------
+# Architecture
+# ------------------------------
+ProjectionHead:
+    conv_channels = (32, 32)
+    dense_neurons = (64, 32, 1)
+    use_poolings = (True, True)
+    kernel_size = 5
+    pool_size = 5
+    
+# ------------------------------
+# Dataset
+# ------------------------------
+Bvcc:
+    debug = False
+
+get_dataloader:
+    num_workers = 12
+    shuffle = True
+
+# ------------------------------
+# Training loop
+# ------------------------------
+TrainingLoop:
+    num_epochs = 30
+    learning_rate = 1e-4
+    dataset_cls = @Bvcc
+    weight_decay = 0.0
+    batch_size_train = 64

results_bvcc/w2v2_large_asr_en/layer12/run0/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f9627174152b51416edf38ab0aaeb16341f92d0a3eae0b4a5fea045918080aa2
+size 3403339

results_bvcc/w2v2_large_asr_en/layer12/run0/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b3b06a62fff74787b637a935ffe19ad02e0bfb8a2f69a6353632c73c4a258b3
+size 3404998

results_bvcc/w2v2_large_asr_en/layer12/run0/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.5967773099740347
+INFO:root:
+[Valid][0][UTT][ MSE = 0.5083 | LCC = 0.7323 | SRCC = 0.7392 ]
+INFO:root:Average loss=0.34689008731108445
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3440 | LCC = 0.7944 | SRCC = 0.7930 ]
+INFO:root:Average loss=0.3308933778451039
+INFO:root:
+[Valid][2][UTT][ MSE = 0.3955 | LCC = 0.7990 | SRCC = 0.8004 ]
+INFO:root:Average loss=0.2722178880984967
+INFO:root:
+[Valid][3][UTT][ MSE = 0.3819 | LCC = 0.8036 | SRCC = 0.8026 ]
+INFO:root:Average loss=0.2590073903019612
+INFO:root:
+[Valid][4][UTT][ MSE = 0.3180 | LCC = 0.8129 | SRCC = 0.8098 ]
+INFO:root:Average loss=0.23053675374159446
+INFO:root:
+[Valid][5][UTT][ MSE = 0.3261 | LCC = 0.8118 | SRCC = 0.8087 ]
+INFO:root:Average loss=0.21345278983696914
+INFO:root:
+[Valid][6][UTT][ MSE = 0.3180 | LCC = 0.8190 | SRCC = 0.8155 ]
+INFO:root:Average loss=0.20548403683381203
+INFO:root:
+[Valid][7][UTT][ MSE = 0.3376 | LCC = 0.8126 | SRCC = 0.8108 ]
+INFO:root:Average loss=0.20480541846691033
+INFO:root:
+[Valid][8][UTT][ MSE = 0.3139 | LCC = 0.8094 | SRCC = 0.8083 ]
+INFO:root:Average loss=0.17525660494963327
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2926 | LCC = 0.8121 | SRCC = 0.8099 ]
+INFO:root:Average loss=0.17095294269995812
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2990 | LCC = 0.8079 | SRCC = 0.8063 ]
+INFO:root:Average loss=0.15322914127355966
+INFO:root:
+[Valid][11][UTT][ MSE = 0.3106 | LCC = 0.8120 | SRCC = 0.8091 ]
+INFO:root:Average loss=0.14001862017007974
+INFO:root:
+[Valid][12][UTT][ MSE = 0.3499 | LCC = 0.8112 | SRCC = 0.8095 ]
+INFO:root:Average loss=0.15172623604153976
+INFO:root:
+[Valid][13][UTT][ MSE = 0.3013 | LCC = 0.8063 | SRCC = 0.8045 ]
+INFO:root:Average loss=0.14074587439879394
+INFO:root:
+[Valid][14][UTT][ MSE = 0.3060 | LCC = 0.8110 | SRCC = 0.8091 ]
+INFO:root:Average loss=0.12415926702893697
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2993 | LCC = 0.8074 | SRCC = 0.8067 ]
+INFO:root:Average loss=0.1146744348299809
+INFO:root:
+[Valid][16][UTT][ MSE = 0.3094 | LCC = 0.8033 | SRCC = 0.8015 ]
+INFO:root:Average loss=0.11406851951510479
+INFO:root:
+[Valid][17][UTT][ MSE = 0.3587 | LCC = 0.7996 | SRCC = 0.7986 ]
+INFO:root:Average loss=0.10659033136490063
+INFO:root:
+[Valid][18][UTT][ MSE = 0.3100 | LCC = 0.8036 | SRCC = 0.8026 ]
+INFO:root:Average loss=0.09453476110521035
+INFO:root:
+[Valid][19][UTT][ MSE = 0.3329 | LCC = 0.8067 | SRCC = 0.8045 ]
+INFO:root:Average loss=0.08874362950714734
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2957 | LCC = 0.8076 | SRCC = 0.8051 ]
+INFO:root:Average loss=0.09242981037077232
+INFO:root:
+[Valid][21][UTT][ MSE = 0.3054 | LCC = 0.8065 | SRCC = 0.8043 ]
+INFO:root:Average loss=0.07949087367607997
+INFO:root:
+[Valid][22][UTT][ MSE = 0.3520 | LCC = 0.8035 | SRCC = 0.8004 ]
+INFO:root:Average loss=0.07928717561448231
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2984 | LCC = 0.8094 | SRCC = 0.8073 ]
+INFO:root:Average loss=0.08720068447291851
+INFO:root:
+[Valid][24][UTT][ MSE = 0.3126 | LCC = 0.8090 | SRCC = 0.8061 ]
+INFO:root:Average loss=0.07371282224089672
+INFO:root:
+[Valid][25][UTT][ MSE = 0.3219 | LCC = 0.8051 | SRCC = 0.8041 ]
+INFO:root:Average loss=0.07177434422266789
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2913 | LCC = 0.8100 | SRCC = 0.8077 ]
+INFO:root:Average loss=0.06913609372881743
+INFO:root:
+[Valid][27][UTT][ MSE = 0.3401 | LCC = 0.8019 | SRCC = 0.7991 ]
+INFO:root:Average loss=0.0761683588513197
+INFO:root:
+[Valid][28][UTT][ MSE = 0.3065 | LCC = 0.8055 | SRCC = 0.8031 ]
+INFO:root:Average loss=0.0654251968058256
+INFO:root:
+[Valid][29][UTT][ MSE = 0.3148 | LCC = 0.8042 | SRCC = 0.8018 ]
+INFO:root:Average loss=0.06390051720425105
+INFO:root:
+[Valid][30][UTT][ MSE = 0.3053 | LCC = 0.8099 | SRCC = 0.8086 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2847 | LCC = 0.8197 | SRCC = 0.8195 ]

results_bvcc/w2v2_large_asr_en/layer12/run1/config.gin

@@ -0,0 +1,31 @@
+# Training config on BVCC dataset.
+
+# ------------------------------
+# Architecture
+# ------------------------------
+ProjectionHead:
+    conv_channels = (32, 32)
+    dense_neurons = (64, 32, 1)
+    use_poolings = (True, True)
+    kernel_size = 5
+    pool_size = 5
+    
+# ------------------------------
+# Dataset
+# ------------------------------
+Bvcc:
+    debug = False
+
+get_dataloader:
+    num_workers = 12
+    shuffle = True
+
+# ------------------------------
+# Training loop
+# ------------------------------
+TrainingLoop:
+    num_epochs = 30
+    learning_rate = 1e-4
+    dataset_cls = @Bvcc
+    weight_decay = 0.0
+    batch_size_train = 64

results_bvcc/w2v2_large_asr_en/layer12/run1/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bf84ab06f58f2a54803a6a19b3781544137bdae1ecba533d2c3d8b94ea169253
+size 3403339

results_bvcc/w2v2_large_asr_en/layer12/run1/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3196ed7c16941f06cf9acdbc5fae9e8942902b9695a99be04a78db14bdab4d47
+size 3404998

results_bvcc/w2v2_large_asr_en/layer12/run1/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.6111231052722687
+INFO:root:
+[Valid][0][UTT][ MSE = 0.4574 | LCC = 0.7163 | SRCC = 0.7186 ]
+INFO:root:Average loss=0.3580839347380858
+INFO:root:
+[Valid][1][UTT][ MSE = 0.4009 | LCC = 0.7808 | SRCC = 0.7832 ]
+INFO:root:Average loss=0.30512720556595385
+INFO:root:
+[Valid][2][UTT][ MSE = 0.3525 | LCC = 0.8025 | SRCC = 0.8021 ]
+INFO:root:Average loss=0.2711172887148001
+INFO:root:
+[Valid][3][UTT][ MSE = 0.3829 | LCC = 0.8011 | SRCC = 0.8019 ]
+INFO:root:Average loss=0.2620871812105179
+INFO:root:
+[Valid][4][UTT][ MSE = 0.4072 | LCC = 0.8052 | SRCC = 0.8094 ]
+INFO:root:Average loss=0.2431554207817102
+INFO:root:
+[Valid][5][UTT][ MSE = 0.3058 | LCC = 0.8136 | SRCC = 0.8097 ]
+INFO:root:Average loss=0.2137223243331298
+INFO:root:
+[Valid][6][UTT][ MSE = 0.3034 | LCC = 0.8146 | SRCC = 0.8108 ]
+INFO:root:Average loss=0.2017128905042624
+INFO:root:
+[Valid][7][UTT][ MSE = 0.3873 | LCC = 0.8064 | SRCC = 0.8097 ]
+INFO:root:Average loss=0.18478004701244524
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2929 | LCC = 0.8128 | SRCC = 0.8109 ]
+INFO:root:Average loss=0.17394910303828043
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2902 | LCC = 0.8135 | SRCC = 0.8111 ]
+INFO:root:Average loss=0.1696313469646833
+INFO:root:
+[Valid][10][UTT][ MSE = 0.3878 | LCC = 0.8061 | SRCC = 0.8076 ]
+INFO:root:Average loss=0.17139079574591073
+INFO:root:
+[Valid][11][UTT][ MSE = 0.3245 | LCC = 0.8067 | SRCC = 0.8059 ]
+INFO:root:Average loss=0.13853010191367224
+INFO:root:
+[Valid][12][UTT][ MSE = 0.3798 | LCC = 0.7954 | SRCC = 0.7975 ]
+INFO:root:Average loss=0.13903944729230341
+INFO:root:
+[Valid][13][UTT][ MSE = 0.3053 | LCC = 0.8020 | SRCC = 0.7972 ]
+INFO:root:Average loss=0.12900117956675017
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2957 | LCC = 0.8049 | SRCC = 0.8025 ]
+INFO:root:Average loss=0.12103551416060863
+INFO:root:
+[Valid][15][UTT][ MSE = 0.3212 | LCC = 0.8053 | SRCC = 0.8034 ]
+INFO:root:Average loss=0.11534078113543682
+INFO:root:
+[Valid][16][UTT][ MSE = 0.3003 | LCC = 0.8066 | SRCC = 0.8030 ]
+INFO:root:Average loss=0.10717813364970379
+INFO:root:
+[Valid][17][UTT][ MSE = 0.3113 | LCC = 0.8017 | SRCC = 0.7972 ]
+INFO:root:Average loss=0.10493040657960452
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2933 | LCC = 0.8079 | SRCC = 0.8048 ]
+INFO:root:Average loss=0.0925721837541996
+INFO:root:
+[Valid][19][UTT][ MSE = 0.3083 | LCC = 0.8038 | SRCC = 0.7998 ]
+INFO:root:Average loss=0.08859112156698337
+INFO:root:
+[Valid][20][UTT][ MSE = 0.3068 | LCC = 0.7973 | SRCC = 0.7938 ]
+INFO:root:Average loss=0.09077216226321
+INFO:root:
+[Valid][21][UTT][ MSE = 0.3093 | LCC = 0.8010 | SRCC = 0.7955 ]
+INFO:root:Average loss=0.08821797834183925
+INFO:root:
+[Valid][22][UTT][ MSE = 0.3173 | LCC = 0.8019 | SRCC = 0.8002 ]
+INFO:root:Average loss=0.07885123287829068
+INFO:root:
+[Valid][23][UTT][ MSE = 0.3267 | LCC = 0.7952 | SRCC = 0.7919 ]
+INFO:root:Average loss=0.07573024904689728
+INFO:root:
+[Valid][24][UTT][ MSE = 0.3180 | LCC = 0.7988 | SRCC = 0.7968 ]
+INFO:root:Average loss=0.06840286609262992
+INFO:root:
+[Valid][25][UTT][ MSE = 0.3259 | LCC = 0.7977 | SRCC = 0.7973 ]
+INFO:root:Average loss=0.06431310022106537
+INFO:root:
+[Valid][26][UTT][ MSE = 0.3008 | LCC = 0.8013 | SRCC = 0.7996 ]
+INFO:root:Average loss=0.06694942450103088
+INFO:root:
+[Valid][27][UTT][ MSE = 0.3120 | LCC = 0.7936 | SRCC = 0.7910 ]
+INFO:root:Average loss=0.061848485173705295
+INFO:root:
+[Valid][28][UTT][ MSE = 0.2987 | LCC = 0.8031 | SRCC = 0.8002 ]
+INFO:root:Average loss=0.05977314920761646
+INFO:root:
+[Valid][29][UTT][ MSE = 0.3274 | LCC = 0.7991 | SRCC = 0.7974 ]
+INFO:root:Average loss=0.05841148921694511
+INFO:root:
+[Valid][30][UTT][ MSE = 0.3039 | LCC = 0.8013 | SRCC = 0.7976 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2870 | LCC = 0.8221 | SRCC = 0.8224 ]

results_bvcc/w2v2_large_asr_en/layer12/run2/config.gin

@@ -0,0 +1,31 @@
+# Training config on BVCC dataset.
+
+# ------------------------------
+# Architecture
+# ------------------------------
+ProjectionHead:
+    conv_channels = (32, 32)
+    dense_neurons = (64, 32, 1)
+    use_poolings = (True, True)
+    kernel_size = 5
+    pool_size = 5
+    
+# ------------------------------
+# Dataset
+# ------------------------------
+Bvcc:
+    debug = False
+
+get_dataloader:
+    num_workers = 12
+    shuffle = True
+
+# ------------------------------
+# Training loop
+# ------------------------------
+TrainingLoop:
+    num_epochs = 30
+    learning_rate = 1e-4
+    dataset_cls = @Bvcc
+    weight_decay = 0.0
+    batch_size_train = 64

results_bvcc/w2v2_large_asr_en/layer12/run2/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:04f9c1c017313ab3328b7a6db815c9180fbf9bd7746903a73d48991660515a71
+size 3403339

results_bvcc/w2v2_large_asr_en/layer12/run2/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e50e5dd0821a67cae239020a0d55a9bd863f2b5c1b7db8d8f732ccf9dd190188
+size 3404998

results_bvcc/w2v2_large_asr_en/layer12/run2/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.6102118430993496
+INFO:root:
+[Valid][0][UTT][ MSE = 0.4486 | LCC = 0.7210 | SRCC = 0.7226 ]
+INFO:root:Average loss=0.3598540024115489
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3737 | LCC = 0.7838 | SRCC = 0.7838 ]
+INFO:root:Average loss=0.29358902401649034
+INFO:root:
+[Valid][2][UTT][ MSE = 0.3682 | LCC = 0.7965 | SRCC = 0.7936 ]
+INFO:root:Average loss=0.2608952222344203
+INFO:root:
+[Valid][3][UTT][ MSE = 0.3139 | LCC = 0.8086 | SRCC = 0.8059 ]
+INFO:root:Average loss=0.24057510724434486
+INFO:root:
+[Valid][4][UTT][ MSE = 0.3437 | LCC = 0.8072 | SRCC = 0.8060 ]
+INFO:root:Average loss=0.22202213547932795
+INFO:root:
+[Valid][5][UTT][ MSE = 0.3303 | LCC = 0.8044 | SRCC = 0.8016 ]
+INFO:root:Average loss=0.22276107336466128
+INFO:root:
+[Valid][6][UTT][ MSE = 0.3215 | LCC = 0.8071 | SRCC = 0.8033 ]
+INFO:root:Average loss=0.19395450770090789
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2991 | LCC = 0.8101 | SRCC = 0.8059 ]
+INFO:root:Average loss=0.18139779625030664
+INFO:root:
+[Valid][8][UTT][ MSE = 0.3013 | LCC = 0.8105 | SRCC = 0.8059 ]
+INFO:root:Average loss=0.16541293788796815
+INFO:root:
+[Valid][9][UTT][ MSE = 0.3383 | LCC = 0.8089 | SRCC = 0.8058 ]
+INFO:root:Average loss=0.15609066302959734
+INFO:root:
+[Valid][10][UTT][ MSE = 0.3098 | LCC = 0.8048 | SRCC = 0.8012 ]
+INFO:root:Average loss=0.14942565073187536
+INFO:root:
+[Valid][11][UTT][ MSE = 0.3018 | LCC = 0.8046 | SRCC = 0.8013 ]
+INFO:root:Average loss=0.1444778493963755
+INFO:root:
+[Valid][12][UTT][ MSE = 0.3218 | LCC = 0.8104 | SRCC = 0.8074 ]
+INFO:root:Average loss=0.12929204392891663
+INFO:root:
+[Valid][13][UTT][ MSE = 0.3533 | LCC = 0.8064 | SRCC = 0.8051 ]
+INFO:root:Average loss=0.12945508059018698
+INFO:root:
+[Valid][14][UTT][ MSE = 0.3040 | LCC = 0.8066 | SRCC = 0.8034 ]
+INFO:root:Average loss=0.1215602186245796
+INFO:root:
+[Valid][15][UTT][ MSE = 0.3055 | LCC = 0.8009 | SRCC = 0.7961 ]
+INFO:root:Average loss=0.10900326655843319
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2961 | LCC = 0.8056 | SRCC = 0.8030 ]
+INFO:root:Average loss=0.1064690011434066
+INFO:root:
+[Valid][17][UTT][ MSE = 0.3581 | LCC = 0.8007 | SRCC = 0.8006 ]
+INFO:root:Average loss=0.10853447392582893
+INFO:root:
+[Valid][18][UTT][ MSE = 0.3901 | LCC = 0.8002 | SRCC = 0.8011 ]
+INFO:root:Average loss=0.08927819065940686
+INFO:root:
+[Valid][19][UTT][ MSE = 0.3061 | LCC = 0.8089 | SRCC = 0.8049 ]
+INFO:root:Average loss=0.08615353049185032
+INFO:root:
+[Valid][20][UTT][ MSE = 0.3125 | LCC = 0.8016 | SRCC = 0.7982 ]
+INFO:root:Average loss=0.0913308357867675
+INFO:root:
+[Valid][21][UTT][ MSE = 0.2913 | LCC = 0.8074 | SRCC = 0.8046 ]
+INFO:root:Average loss=0.07851233495733677
+INFO:root:
+[Valid][22][UTT][ MSE = 0.3270 | LCC = 0.7990 | SRCC = 0.7969 ]
+INFO:root:Average loss=0.07828337899767436
+INFO:root:
+[Valid][23][UTT][ MSE = 0.3264 | LCC = 0.8047 | SRCC = 0.8044 ]
+INFO:root:Average loss=0.07540841166598675
+INFO:root:
+[Valid][24][UTT][ MSE = 0.2951 | LCC = 0.8084 | SRCC = 0.8064 ]
+INFO:root:Average loss=0.0709887038056667
+INFO:root:
+[Valid][25][UTT][ MSE = 0.3058 | LCC = 0.8052 | SRCC = 0.8026 ]
+INFO:root:Average loss=0.07145582717389633
+INFO:root:
+[Valid][26][UTT][ MSE = 0.3043 | LCC = 0.8051 | SRCC = 0.8032 ]
+INFO:root:Average loss=0.06620600533026916
+INFO:root:
+[Valid][27][UTT][ MSE = 0.3009 | LCC = 0.8102 | SRCC = 0.8087 ]
+INFO:root:Average loss=0.06880018468468617
+INFO:root:
+[Valid][28][UTT][ MSE = 0.3437 | LCC = 0.8032 | SRCC = 0.8023 ]
+INFO:root:Average loss=0.06038604270762358
+INFO:root:
+[Valid][29][UTT][ MSE = 0.3027 | LCC = 0.8078 | SRCC = 0.8054 ]
+INFO:root:Average loss=0.05491556771672689
+INFO:root:
+[Valid][30][UTT][ MSE = 0.3178 | LCC = 0.8009 | SRCC = 0.8005 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2999 | LCC = 0.8180 | SRCC = 0.8209 ]

results_bvcc/w2v2_large_asr_en/layer12/run3/config.gin

@@ -0,0 +1,31 @@
+# Training config on BVCC dataset.
+
+# ------------------------------
+# Architecture
+# ------------------------------
+ProjectionHead:
+    conv_channels = (32, 32)
+    dense_neurons = (64, 32, 1)
+    use_poolings = (True, True)
+    kernel_size = 5
+    pool_size = 5
+    
+# ------------------------------
+# Dataset
+# ------------------------------
+Bvcc:
+    debug = False
+
+get_dataloader:
+    num_workers = 12
+    shuffle = True
+
+# ------------------------------
+# Training loop
+# ------------------------------
+TrainingLoop:
+    num_epochs = 30
+    learning_rate = 1e-4
+    dataset_cls = @Bvcc
+    weight_decay = 0.0
+    batch_size_train = 64

results_bvcc/w2v2_large_asr_en/layer12/run3/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0ead20be6dba04f1b5410d68166d64d5cff49fbe9a78bbfda959b188c7f6222b
+size 3403339

results_bvcc/w2v2_large_asr_en/layer12/run3/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b2d9689b85d9f33cec017b4875d26aa8b002a174bb54cf0aa34e88f48d904b9
+size 3405062

results_bvcc/w2v2_large_asr_en/layer12/run3/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.7038338230206416
+INFO:root:
+[Valid][0][UTT][ MSE = 0.4931 | LCC = 0.6788 | SRCC = 0.6801 ]
+INFO:root:Average loss=0.37384560455878574
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3693 | LCC = 0.7838 | SRCC = 0.7838 ]
+INFO:root:Average loss=0.2999933856802109
+INFO:root:
+[Valid][2][UTT][ MSE = 0.3384 | LCC = 0.8002 | SRCC = 0.7987 ]
+INFO:root:Average loss=0.26500188291837007
+INFO:root:
+[Valid][3][UTT][ MSE = 0.3288 | LCC = 0.8070 | SRCC = 0.8044 ]
+INFO:root:Average loss=0.23847780701441643
+INFO:root:
+[Valid][4][UTT][ MSE = 0.3540 | LCC = 0.8071 | SRCC = 0.8052 ]
+INFO:root:Average loss=0.22249199602848443
+INFO:root:
+[Valid][5][UTT][ MSE = 0.3009 | LCC = 0.8112 | SRCC = 0.8076 ]
+INFO:root:Average loss=0.20372035392583945
+INFO:root:
+[Valid][6][UTT][ MSE = 0.3045 | LCC = 0.8111 | SRCC = 0.8069 ]
+INFO:root:Average loss=0.19250235897608292
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2994 | LCC = 0.8121 | SRCC = 0.8074 ]
+INFO:root:Average loss=0.17996701588615394
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2776 | LCC = 0.8202 | SRCC = 0.8168 ]
+INFO:root:Average loss=0.1668266461063654
+INFO:root:
+[Valid][9][UTT][ MSE = 0.3254 | LCC = 0.8164 | SRCC = 0.8138 ]
+INFO:root:Average loss=0.15905942366673395
+INFO:root:
+[Valid][10][UTT][ MSE = 0.3296 | LCC = 0.8108 | SRCC = 0.8096 ]
+INFO:root:Average loss=0.15147841702669096
+INFO:root:
+[Valid][11][UTT][ MSE = 0.3526 | LCC = 0.8057 | SRCC = 0.8044 ]
+INFO:root:Average loss=0.1517454889149238
+INFO:root:
+[Valid][12][UTT][ MSE = 0.3294 | LCC = 0.8116 | SRCC = 0.8085 ]
+INFO:root:Average loss=0.14102680484453836
+INFO:root:
+[Valid][13][UTT][ MSE = 0.3441 | LCC = 0.8147 | SRCC = 0.8140 ]
+INFO:root:Average loss=0.12603258274686643
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2879 | LCC = 0.8125 | SRCC = 0.8083 ]
+INFO:root:Average loss=0.11384060959785412
+INFO:root:
+[Valid][15][UTT][ MSE = 0.3237 | LCC = 0.8094 | SRCC = 0.8075 ]
+INFO:root:Average loss=0.10927917483525398
+INFO:root:
+[Valid][16][UTT][ MSE = 0.3029 | LCC = 0.8081 | SRCC = 0.8050 ]
+INFO:root:Average loss=0.10366906598210335
+INFO:root:
+[Valid][17][UTT][ MSE = 0.3050 | LCC = 0.8042 | SRCC = 0.8025 ]
+INFO:root:Average loss=0.09829619240302306
+INFO:root:
+[Valid][18][UTT][ MSE = 0.3236 | LCC = 0.8103 | SRCC = 0.8062 ]
+INFO:root:Average loss=0.09296508334003963
+INFO:root:
+[Valid][19][UTT][ MSE = 0.3247 | LCC = 0.8074 | SRCC = 0.8044 ]
+INFO:root:Average loss=0.08567066232745464
+INFO:root:
+[Valid][20][UTT][ MSE = 0.3026 | LCC = 0.8062 | SRCC = 0.8014 ]
+INFO:root:Average loss=0.08407249258687863
+INFO:root:
+[Valid][21][UTT][ MSE = 0.3115 | LCC = 0.8053 | SRCC = 0.8025 ]
+INFO:root:Average loss=0.07837912526268226
+INFO:root:
+[Valid][22][UTT][ MSE = 0.2995 | LCC = 0.8018 | SRCC = 0.7982 ]
+INFO:root:Average loss=0.08392994831769894
+INFO:root:
+[Valid][23][UTT][ MSE = 0.3279 | LCC = 0.8054 | SRCC = 0.8005 ]
+INFO:root:Average loss=0.07431051794152993
+INFO:root:
+[Valid][24][UTT][ MSE = 0.2946 | LCC = 0.8070 | SRCC = 0.8038 ]
+INFO:root:Average loss=0.06799868661432694
+INFO:root:
+[Valid][25][UTT][ MSE = 0.3061 | LCC = 0.7993 | SRCC = 0.7971 ]
+INFO:root:Average loss=0.0673973303861343
+INFO:root:
+[Valid][26][UTT][ MSE = 0.3030 | LCC = 0.8015 | SRCC = 0.7973 ]
+INFO:root:Average loss=0.06311425772041847
+INFO:root:
+[Valid][27][UTT][ MSE = 0.3116 | LCC = 0.8084 | SRCC = 0.8064 ]
+INFO:root:Average loss=0.06494387826667382
+INFO:root:
+[Valid][28][UTT][ MSE = 0.3072 | LCC = 0.8091 | SRCC = 0.8060 ]
+INFO:root:Average loss=0.059478260385684475
+INFO:root:
+[Valid][29][UTT][ MSE = 0.3006 | LCC = 0.8058 | SRCC = 0.8045 ]
+INFO:root:Average loss=0.059121602143232636
+INFO:root:
+[Valid][30][UTT][ MSE = 0.2950 | LCC = 0.8070 | SRCC = 0.8054 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.3062 | LCC = 0.8157 | SRCC = 0.8164 ]

results_bvcc/w2v2_large_asr_en/layer12/run4/config.gin

@@ -0,0 +1,31 @@
+# Training config on BVCC dataset.
+
+# ------------------------------
+# Architecture
+# ------------------------------
+ProjectionHead:
+    conv_channels = (32, 32)
+    dense_neurons = (64, 32, 1)
+    use_poolings = (True, True)
+    kernel_size = 5
+    pool_size = 5
+    
+# ------------------------------
+# Dataset
+# ------------------------------
+Bvcc:
+    debug = False
+
+get_dataloader:
+    num_workers = 12
+    shuffle = True
+
+# ------------------------------
+# Training loop
+# ------------------------------
+TrainingLoop:
+    num_epochs = 30
+    learning_rate = 1e-4
+    dataset_cls = @Bvcc
+    weight_decay = 0.0
+    batch_size_train = 64

results_bvcc/w2v2_large_asr_en/layer12/run4/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da8afe3d99c6ced5a81feaa639a55c8e1a9b589fbf4b0a8699e50494160966a1
+size 3403339

results_bvcc/w2v2_large_asr_en/layer12/run4/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8d0807b5465cb26fcd978057b9d0de1c7cdb5d993e499ad0e61943909c16ea76
+size 3405062

results_bvcc/w2v2_large_asr_en/layer12/run4/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.629870287500895
+INFO:root:
+[Valid][0][UTT][ MSE = 0.4628 | LCC = 0.7109 | SRCC = 0.7140 ]
+INFO:root:Average loss=0.3486146888671777
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3833 | LCC = 0.7779 | SRCC = 0.7778 ]
+INFO:root:Average loss=0.30247746923795116
+INFO:root:
+[Valid][2][UTT][ MSE = 0.3435 | LCC = 0.7981 | SRCC = 0.7969 ]
+INFO:root:Average loss=0.2903971492480009
+INFO:root:
+[Valid][3][UTT][ MSE = 0.3342 | LCC = 0.8005 | SRCC = 0.7989 ]
+INFO:root:Average loss=0.2524496593918556
+INFO:root:
+[Valid][4][UTT][ MSE = 0.3906 | LCC = 0.8058 | SRCC = 0.8060 ]
+INFO:root:Average loss=0.22911507846453252
+INFO:root:
+[Valid][5][UTT][ MSE = 0.3181 | LCC = 0.8064 | SRCC = 0.8041 ]
+INFO:root:Average loss=0.21391650709586266
+INFO:root:
+[Valid][6][UTT][ MSE = 0.3005 | LCC = 0.8122 | SRCC = 0.8103 ]
+INFO:root:Average loss=0.1907470042888935
+INFO:root:
+[Valid][7][UTT][ MSE = 0.3211 | LCC = 0.8106 | SRCC = 0.8086 ]
+INFO:root:Average loss=0.182094951852774
+INFO:root:
+[Valid][8][UTT][ MSE = 0.3109 | LCC = 0.8112 | SRCC = 0.8094 ]
+INFO:root:Average loss=0.17697418882296637
+INFO:root:
+[Valid][9][UTT][ MSE = 0.3565 | LCC = 0.8053 | SRCC = 0.8055 ]
+INFO:root:Average loss=0.15253328579740646
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2970 | LCC = 0.8096 | SRCC = 0.8092 ]
+INFO:root:Average loss=0.15185313031841546
+INFO:root:
+[Valid][11][UTT][ MSE = 0.3611 | LCC = 0.8006 | SRCC = 0.7988 ]
+INFO:root:Average loss=0.13487681068288973
+INFO:root:
+[Valid][12][UTT][ MSE = 0.3007 | LCC = 0.8081 | SRCC = 0.8063 ]
+INFO:root:Average loss=0.12277116760229453
+INFO:root:
+[Valid][13][UTT][ MSE = 0.3128 | LCC = 0.8080 | SRCC = 0.8051 ]
+INFO:root:Average loss=0.12296248007661258
+INFO:root:
+[Valid][14][UTT][ MSE = 0.3159 | LCC = 0.8034 | SRCC = 0.7999 ]
+INFO:root:Average loss=0.10632928155171566
+INFO:root:
+[Valid][15][UTT][ MSE = 0.3184 | LCC = 0.8101 | SRCC = 0.8085 ]
+INFO:root:Average loss=0.11399256285184468
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2917 | LCC = 0.8103 | SRCC = 0.8100 ]
+INFO:root:Average loss=0.10046655785005826
+INFO:root:
+[Valid][17][UTT][ MSE = 0.3252 | LCC = 0.8092 | SRCC = 0.8083 ]
+INFO:root:Average loss=0.0938771465458931
+INFO:root:
+[Valid][18][UTT][ MSE = 0.3440 | LCC = 0.8056 | SRCC = 0.8035 ]
+INFO:root:Average loss=0.09173524800019386
+INFO:root:
+[Valid][19][UTT][ MSE = 0.3307 | LCC = 0.8088 | SRCC = 0.8072 ]
+INFO:root:Average loss=0.08001724840738834
+INFO:root:
+[Valid][20][UTT][ MSE = 0.3070 | LCC = 0.8075 | SRCC = 0.8050 ]
+INFO:root:Average loss=0.08038300457291114
+INFO:root:
+[Valid][21][UTT][ MSE = 0.3170 | LCC = 0.8071 | SRCC = 0.8057 ]
+INFO:root:Average loss=0.0751321111829617
+INFO:root:
+[Valid][22][UTT][ MSE = 0.3115 | LCC = 0.8099 | SRCC = 0.8073 ]
+INFO:root:Average loss=0.07465068246118534
+INFO:root:
+[Valid][23][UTT][ MSE = 0.3101 | LCC = 0.8038 | SRCC = 0.8005 ]
+INFO:root:Average loss=0.06820170103739469
+INFO:root:
+[Valid][24][UTT][ MSE = 0.3353 | LCC = 0.8043 | SRCC = 0.8019 ]
+INFO:root:Average loss=0.06471494991236772
+INFO:root:
+[Valid][25][UTT][ MSE = 0.2934 | LCC = 0.8094 | SRCC = 0.8074 ]
+INFO:root:Average loss=0.0650003196623845
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2972 | LCC = 0.8107 | SRCC = 0.8071 ]
+INFO:root:Average loss=0.06452292547776149
+INFO:root:
+[Valid][27][UTT][ MSE = 0.3088 | LCC = 0.8027 | SRCC = 0.8000 ]
+INFO:root:Average loss=0.06001786782573431
+INFO:root:
+[Valid][28][UTT][ MSE = 0.3289 | LCC = 0.8055 | SRCC = 0.8035 ]
+INFO:root:Average loss=0.05812227386885729
+INFO:root:
+[Valid][29][UTT][ MSE = 0.3155 | LCC = 0.8026 | SRCC = 0.8003 ]
+INFO:root:Average loss=0.055573841031545244
+INFO:root:
+[Valid][30][UTT][ MSE = 0.3125 | LCC = 0.8098 | SRCC = 0.8081 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.3076 | LCC = 0.8107 | SRCC = 0.8140 ]

results_bvcc/w2v2_large_asr_en/layer14/run0/config.gin

@@ -0,0 +1,31 @@
+# Training config on BVCC dataset.
+
+# ------------------------------
+# Architecture
+# ------------------------------
+ProjectionHead:
+    conv_channels = (32, 32)
+    dense_neurons = (64, 32, 1)
+    use_poolings = (True, True)
+    kernel_size = 5
+    pool_size = 5
+    
+# ------------------------------
+# Dataset
+# ------------------------------
+Bvcc:
+    debug = False
+
+get_dataloader:
+    num_workers = 12
+    shuffle = True
+
+# ------------------------------
+# Training loop
+# ------------------------------
+TrainingLoop:
+    num_epochs = 30
+    learning_rate = 1e-4
+    dataset_cls = @Bvcc
+    weight_decay = 0.0
+    batch_size_train = 64

results_bvcc/w2v2_large_asr_en/layer14/run0/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c7bdb91f00f6807e0aa3c8fb169b2e8a9043ac5a04b0d19b446bd071909cd2f4
+size 3403339

results_bvcc/w2v2_large_asr_en/layer14/run0/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3568c4c623091099662ba2f8dbdaf778a4352b39059e1c3faf36bcfa6578647b
+size 3404998

results_bvcc/w2v2_large_asr_en/layer14/run0/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.639124198219715
+INFO:root:
+[Valid][0][UTT][ MSE = 0.5824 | LCC = 0.6603 | SRCC = 0.6654 ]
+INFO:root:Average loss=0.40829688004958326
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3725 | LCC = 0.7650 | SRCC = 0.7667 ]
+INFO:root:Average loss=0.3168532964892877
+INFO:root:
+[Valid][2][UTT][ MSE = 0.3455 | LCC = 0.7900 | SRCC = 0.7882 ]
+INFO:root:Average loss=0.2736568292364096
+INFO:root:
+[Valid][3][UTT][ MSE = 0.3320 | LCC = 0.8046 | SRCC = 0.8030 ]
+INFO:root:Average loss=0.2780578547181227
+INFO:root:
+[Valid][4][UTT][ MSE = 0.4968 | LCC = 0.7974 | SRCC = 0.8015 ]
+INFO:root:Average loss=0.24588502771579301
+INFO:root:
+[Valid][5][UTT][ MSE = 0.3295 | LCC = 0.8075 | SRCC = 0.8025 ]
+INFO:root:Average loss=0.2572734610010416
+INFO:root:
+[Valid][6][UTT][ MSE = 0.3514 | LCC = 0.8064 | SRCC = 0.8056 ]
+INFO:root:Average loss=0.21617332186836463
+INFO:root:
+[Valid][7][UTT][ MSE = 0.3010 | LCC = 0.8103 | SRCC = 0.8057 ]
+INFO:root:Average loss=0.19947211979291377
+INFO:root:
+[Valid][8][UTT][ MSE = 0.3068 | LCC = 0.8131 | SRCC = 0.8103 ]
+INFO:root:Average loss=0.18418184152016273
+INFO:root:
+[Valid][9][UTT][ MSE = 0.3275 | LCC = 0.8129 | SRCC = 0.8107 ]
+INFO:root:Average loss=0.18136565587841547
+INFO:root:
+[Valid][10][UTT][ MSE = 0.3026 | LCC = 0.8115 | SRCC = 0.8093 ]
+INFO:root:Average loss=0.17634256021716657
+INFO:root:
+[Valid][11][UTT][ MSE = 0.3040 | LCC = 0.8105 | SRCC = 0.8073 ]
+INFO:root:Average loss=0.16672048908777726
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2918 | LCC = 0.8074 | SRCC = 0.8056 ]
+INFO:root:Average loss=0.1531867396373015
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2858 | LCC = 0.8110 | SRCC = 0.8089 ]
+INFO:root:Average loss=0.13818359909913477
+INFO:root:
+[Valid][14][UTT][ MSE = 0.3116 | LCC = 0.8099 | SRCC = 0.8066 ]
+INFO:root:Average loss=0.13977856437365213
+INFO:root:
+[Valid][15][UTT][ MSE = 0.3141 | LCC = 0.8075 | SRCC = 0.8063 ]
+INFO:root:Average loss=0.12866498329318488
+INFO:root:
+[Valid][16][UTT][ MSE = 0.3146 | LCC = 0.8058 | SRCC = 0.8037 ]
+INFO:root:Average loss=0.12641079996067744
+INFO:root:
+[Valid][17][UTT][ MSE = 0.3072 | LCC = 0.8084 | SRCC = 0.8061 ]
+INFO:root:Average loss=0.11902084497687145
+INFO:root:
+[Valid][18][UTT][ MSE = 0.3461 | LCC = 0.8102 | SRCC = 0.8089 ]
+INFO:root:Average loss=0.10734949982128082
+INFO:root:
+[Valid][19][UTT][ MSE = 0.2973 | LCC = 0.8039 | SRCC = 0.8031 ]
+INFO:root:Average loss=0.10676883839261837
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2959 | LCC = 0.8040 | SRCC = 0.8023 ]
+INFO:root:Average loss=0.09338671265122218
+INFO:root:
+[Valid][21][UTT][ MSE = 0.3056 | LCC = 0.8074 | SRCC = 0.8045 ]
+INFO:root:Average loss=0.11823369424121502
+INFO:root:
+[Valid][22][UTT][ MSE = 0.2920 | LCC = 0.8080 | SRCC = 0.8059 ]
+INFO:root:Average loss=0.09307295757417496
+INFO:root:
+[Valid][23][UTT][ MSE = 0.3335 | LCC = 0.8079 | SRCC = 0.8067 ]
+INFO:root:Average loss=0.08407396402878639
+INFO:root:
+[Valid][24][UTT][ MSE = 0.3107 | LCC = 0.8055 | SRCC = 0.8040 ]
+INFO:root:Average loss=0.08697844788623162
+INFO:root:
+[Valid][25][UTT][ MSE = 0.3272 | LCC = 0.8020 | SRCC = 0.8012 ]
+INFO:root:Average loss=0.0841066506333076
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2910 | LCC = 0.8075 | SRCC = 0.8057 ]
+INFO:root:Average loss=0.0758334105500044
+INFO:root:
+[Valid][27][UTT][ MSE = 0.2999 | LCC = 0.8063 | SRCC = 0.8045 ]
+INFO:root:Average loss=0.07849707659811546
+INFO:root:
+[Valid][28][UTT][ MSE = 0.2987 | LCC = 0.8069 | SRCC = 0.8055 ]
+INFO:root:Average loss=0.07672988012050971
+INFO:root:
+[Valid][29][UTT][ MSE = 0.3049 | LCC = 0.8049 | SRCC = 0.8026 ]
+INFO:root:Average loss=0.06985761306415765
+INFO:root:
+[Valid][30][UTT][ MSE = 0.2994 | LCC = 0.8019 | SRCC = 0.8003 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.3380 | LCC = 0.8146 | SRCC = 0.8211 ]

results_bvcc/w2v2_large_asr_en/layer14/run1/config.gin

@@ -0,0 +1,31 @@
+# Training config on BVCC dataset.
+
+# ------------------------------
+# Architecture
+# ------------------------------
+ProjectionHead:
+    conv_channels = (32, 32)
+    dense_neurons = (64, 32, 1)
+    use_poolings = (True, True)
+    kernel_size = 5
+    pool_size = 5
+    
+# ------------------------------
+# Dataset
+# ------------------------------
+Bvcc:
+    debug = False
+
+get_dataloader:
+    num_workers = 12
+    shuffle = True
+
+# ------------------------------
+# Training loop
+# ------------------------------
+TrainingLoop:
+    num_epochs = 30
+    learning_rate = 1e-4
+    dataset_cls = @Bvcc
+    weight_decay = 0.0
+    batch_size_train = 64

results_bvcc/w2v2_large_asr_en/layer14/run1/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0334e35d89573d14eff4c4c7292361bddda31360cdb14bf0e31f9f9060bb5063
+size 3403339

results_bvcc/w2v2_large_asr_en/layer14/run1/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4e1a8ba90d97d5ef442aa7226f11e56b51a1136d1243f8b231f6d2a049dfda46
+size 3405062

results_bvcc/w2v2_large_asr_en/layer14/run1/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.5812411392346407
+INFO:root:
+[Valid][0][UTT][ MSE = 0.4863 | LCC = 0.7401 | SRCC = 0.7450 ]
+INFO:root:Average loss=0.3544054925441742
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3510 | LCC = 0.7938 | SRCC = 0.7942 ]
+INFO:root:Average loss=0.31247991896592653
+INFO:root:
+[Valid][2][UTT][ MSE = 0.3257 | LCC = 0.8021 | SRCC = 0.7997 ]
+INFO:root:Average loss=0.2948195548393788
+INFO:root:
+[Valid][3][UTT][ MSE = 0.3317 | LCC = 0.8104 | SRCC = 0.8059 ]
+INFO:root:Average loss=0.2630130499601364
+INFO:root:
+[Valid][4][UTT][ MSE = 0.3274 | LCC = 0.8069 | SRCC = 0.8040 ]
+INFO:root:Average loss=0.24918348724261308
+INFO:root:
+[Valid][5][UTT][ MSE = 0.3135 | LCC = 0.8088 | SRCC = 0.8055 ]
+INFO:root:Average loss=0.2262976987239642
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2992 | LCC = 0.8145 | SRCC = 0.8107 ]
+INFO:root:Average loss=0.20898243402823424
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2921 | LCC = 0.8181 | SRCC = 0.8129 ]
+INFO:root:Average loss=0.21231674899657568
+INFO:root:
+[Valid][8][UTT][ MSE = 0.3155 | LCC = 0.8145 | SRCC = 0.8144 ]
+INFO:root:Average loss=0.190538444962257
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2829 | LCC = 0.8193 | SRCC = 0.8154 ]
+INFO:root:Average loss=0.18087279672423998
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2871 | LCC = 0.8229 | SRCC = 0.8205 ]
+INFO:root:Average loss=0.16717880161908957
+INFO:root:
+[Valid][11][UTT][ MSE = 0.2927 | LCC = 0.8147 | SRCC = 0.8125 ]
+INFO:root:Average loss=0.15654730997406519
+INFO:root:
+[Valid][12][UTT][ MSE = 0.3133 | LCC = 0.8090 | SRCC = 0.8085 ]
+INFO:root:Average loss=0.15454688276617956
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2922 | LCC = 0.8134 | SRCC = 0.8109 ]
+INFO:root:Average loss=0.14769291667602
+INFO:root:
+[Valid][14][UTT][ MSE = 0.3003 | LCC = 0.8128 | SRCC = 0.8107 ]
+INFO:root:Average loss=0.1449632299825167
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2969 | LCC = 0.8185 | SRCC = 0.8157 ]
+INFO:root:Average loss=0.1303374068095134
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2905 | LCC = 0.8096 | SRCC = 0.8073 ]
+INFO:root:Average loss=0.1257298872447931
+INFO:root:
+[Valid][17][UTT][ MSE = 0.3071 | LCC = 0.8121 | SRCC = 0.8089 ]
+INFO:root:Average loss=0.11752169378674947
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2882 | LCC = 0.8136 | SRCC = 0.8121 ]
+INFO:root:Average loss=0.10992229252289502
+INFO:root:
+[Valid][19][UTT][ MSE = 0.3138 | LCC = 0.8099 | SRCC = 0.8086 ]
+INFO:root:Average loss=0.10096019081389293
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2823 | LCC = 0.8153 | SRCC = 0.8135 ]
+INFO:root:Average loss=0.09729361983063893
+INFO:root:
+[Valid][21][UTT][ MSE = 0.3079 | LCC = 0.8128 | SRCC = 0.8123 ]
+INFO:root:Average loss=0.093237780368863
+INFO:root:
+[Valid][22][UTT][ MSE = 0.3062 | LCC = 0.8121 | SRCC = 0.8110 ]
+INFO:root:Average loss=0.08845462774236997
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2951 | LCC = 0.8120 | SRCC = 0.8110 ]
+INFO:root:Average loss=0.09973976398125672
+INFO:root:
+[Valid][24][UTT][ MSE = 0.2915 | LCC = 0.8118 | SRCC = 0.8111 ]
+INFO:root:Average loss=0.08304178046110348
+INFO:root:
+[Valid][25][UTT][ MSE = 0.3001 | LCC = 0.8034 | SRCC = 0.8015 ]
+INFO:root:Average loss=0.07921191433874461
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2891 | LCC = 0.8084 | SRCC = 0.8081 ]
+INFO:root:Average loss=0.08367712767078327
+INFO:root:
+[Valid][27][UTT][ MSE = 0.3105 | LCC = 0.8090 | SRCC = 0.8105 ]
+INFO:root:Average loss=0.0783615472416083
+INFO:root:
+[Valid][28][UTT][ MSE = 0.3004 | LCC = 0.8082 | SRCC = 0.8082 ]
+INFO:root:Average loss=0.08223956723052722
+INFO:root:
+[Valid][29][UTT][ MSE = 0.2974 | LCC = 0.8095 | SRCC = 0.8100 ]
+INFO:root:Average loss=0.07453711927892306
+INFO:root:
+[Valid][30][UTT][ MSE = 0.3065 | LCC = 0.8109 | SRCC = 0.8105 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2835 | LCC = 0.8207 | SRCC = 0.8208 ]

results_bvcc/w2v2_large_asr_en/layer14/run2/config.gin

@@ -0,0 +1,31 @@
+# Training config on BVCC dataset.
+
+# ------------------------------
+# Architecture
+# ------------------------------
+ProjectionHead:
+    conv_channels = (32, 32)
+    dense_neurons = (64, 32, 1)
+    use_poolings = (True, True)
+    kernel_size = 5
+    pool_size = 5
+    
+# ------------------------------
+# Dataset
+# ------------------------------
+Bvcc:
+    debug = False
+
+get_dataloader:
+    num_workers = 12
+    shuffle = True
+
+# ------------------------------
+# Training loop
+# ------------------------------
+TrainingLoop:
+    num_epochs = 30
+    learning_rate = 1e-4
+    dataset_cls = @Bvcc
+    weight_decay = 0.0
+    batch_size_train = 64

results_bvcc/w2v2_large_asr_en/layer14/run2/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d3dc384c09284d751fa7edb1e270768905ff6dfd7c70bcae7e9f414d6b353c69
+size 3403339

results_bvcc/w2v2_large_asr_en/layer14/run2/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7cd66cf3e7d05c3b94cfe9b9c72c6beaf7479f1479d85db356f58db841f9dad7
+size 3405062

results_bvcc/w2v2_large_asr_en/layer14/run2/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.6524099031319985
+INFO:root:
+[Valid][0][UTT][ MSE = 0.4754 | LCC = 0.6931 | SRCC = 0.6955 ]
+INFO:root:Average loss=0.35758627225191164
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3836 | LCC = 0.7808 | SRCC = 0.7829 ]
+INFO:root:Average loss=0.3075551348618972
+INFO:root:
+[Valid][2][UTT][ MSE = 0.3402 | LCC = 0.7947 | SRCC = 0.7927 ]
+INFO:root:Average loss=0.2832821517800673
+INFO:root:
+[Valid][3][UTT][ MSE = 0.3283 | LCC = 0.8025 | SRCC = 0.8005 ]
+INFO:root:Average loss=0.2530775230664473
+INFO:root:
+[Valid][4][UTT][ MSE = 0.3363 | LCC = 0.8030 | SRCC = 0.8020 ]
+INFO:root:Average loss=0.24408865108703956
+INFO:root:
+[Valid][5][UTT][ MSE = 0.3090 | LCC = 0.8028 | SRCC = 0.8009 ]
+INFO:root:Average loss=0.2182787757080335
+INFO:root:
+[Valid][6][UTT][ MSE = 0.3090 | LCC = 0.8111 | SRCC = 0.8079 ]
+INFO:root:Average loss=0.2112214899598024
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2955 | LCC = 0.8129 | SRCC = 0.8092 ]
+INFO:root:Average loss=0.19962961322222
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2955 | LCC = 0.8131 | SRCC = 0.8103 ]
+INFO:root:Average loss=0.1844892527621526
+INFO:root:
+[Valid][9][UTT][ MSE = 0.3150 | LCC = 0.8071 | SRCC = 0.8049 ]
+INFO:root:Average loss=0.18336004553697047
+INFO:root:
+[Valid][10][UTT][ MSE = 0.3011 | LCC = 0.8135 | SRCC = 0.8102 ]
+INFO:root:Average loss=0.18344622048047873
+INFO:root:
+[Valid][11][UTT][ MSE = 0.3021 | LCC = 0.8163 | SRCC = 0.8123 ]
+INFO:root:Average loss=0.163018013135745
+INFO:root:
+[Valid][12][UTT][ MSE = 0.3010 | LCC = 0.8080 | SRCC = 0.8050 ]
+INFO:root:Average loss=0.14758349515688726
+INFO:root:
+[Valid][13][UTT][ MSE = 0.3488 | LCC = 0.7970 | SRCC = 0.7955 ]
+INFO:root:Average loss=0.13971616642979476
+INFO:root:
+[Valid][14][UTT][ MSE = 0.3724 | LCC = 0.8026 | SRCC = 0.8016 ]
+INFO:root:Average loss=0.14740928291128233
+INFO:root:
+[Valid][15][UTT][ MSE = 0.3598 | LCC = 0.8041 | SRCC = 0.8036 ]
+INFO:root:Average loss=0.1275135239538474
+INFO:root:
+[Valid][16][UTT][ MSE = 0.3341 | LCC = 0.8032 | SRCC = 0.8028 ]
+INFO:root:Average loss=0.12222580993786836
+INFO:root:
+[Valid][17][UTT][ MSE = 0.3528 | LCC = 0.7907 | SRCC = 0.7888 ]
+INFO:root:Average loss=0.115394481863731
+INFO:root:
+[Valid][18][UTT][ MSE = 0.3066 | LCC = 0.7961 | SRCC = 0.7935 ]
+INFO:root:Average loss=0.11262325063729897
+INFO:root:
+[Valid][19][UTT][ MSE = 0.3114 | LCC = 0.7992 | SRCC = 0.7955 ]
+INFO:root:Average loss=0.10322136613420951
+INFO:root:
+[Valid][20][UTT][ MSE = 0.3046 | LCC = 0.7978 | SRCC = 0.7954 ]
+INFO:root:Average loss=0.10801306600945118
+INFO:root:
+[Valid][21][UTT][ MSE = 0.3106 | LCC = 0.7995 | SRCC = 0.7975 ]
+INFO:root:Average loss=0.09470349884568116
+INFO:root:
+[Valid][22][UTT][ MSE = 0.3348 | LCC = 0.8007 | SRCC = 0.8006 ]
+INFO:root:Average loss=0.09117928639245339
+INFO:root:
+[Valid][23][UTT][ MSE = 0.3315 | LCC = 0.7981 | SRCC = 0.7962 ]
+INFO:root:Average loss=0.0903872536161007
+INFO:root:
+[Valid][24][UTT][ MSE = 0.3039 | LCC = 0.8068 | SRCC = 0.8055 ]
+INFO:root:Average loss=0.08537686267533363
+INFO:root:
+[Valid][25][UTT][ MSE = 0.3091 | LCC = 0.8008 | SRCC = 0.7980 ]
+INFO:root:Average loss=0.08537523032954106
+INFO:root:
+[Valid][26][UTT][ MSE = 0.3165 | LCC = 0.8010 | SRCC = 0.7998 ]
+INFO:root:Average loss=0.08757533223774189
+INFO:root:
+[Valid][27][UTT][ MSE = 0.2990 | LCC = 0.8046 | SRCC = 0.8021 ]
+INFO:root:Average loss=0.07680801588755387
+INFO:root:
+[Valid][28][UTT][ MSE = 0.3423 | LCC = 0.8014 | SRCC = 0.8019 ]
+INFO:root:Average loss=0.07683007342693134
+INFO:root:
+[Valid][29][UTT][ MSE = 0.3016 | LCC = 0.8037 | SRCC = 0.8019 ]
+INFO:root:Average loss=0.07444462165809594
+INFO:root:
+[Valid][30][UTT][ MSE = 0.3368 | LCC = 0.8025 | SRCC = 0.8021 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.3370 | LCC = 0.7984 | SRCC = 0.8052 ]

results_bvcc/w2v2_large_asr_en/layer14/run3/config.gin

@@ -0,0 +1,31 @@
+# Training config on BVCC dataset.
+
+# ------------------------------
+# Architecture
+# ------------------------------
+ProjectionHead:
+    conv_channels = (32, 32)
+    dense_neurons = (64, 32, 1)
+    use_poolings = (True, True)
+    kernel_size = 5
+    pool_size = 5
+    
+# ------------------------------
+# Dataset
+# ------------------------------
+Bvcc:
+    debug = False
+
+get_dataloader:
+    num_workers = 12
+    shuffle = True
+
+# ------------------------------
+# Training loop
+# ------------------------------
+TrainingLoop:
+    num_epochs = 30
+    learning_rate = 1e-4
+    dataset_cls = @Bvcc
+    weight_decay = 0.0
+    batch_size_train = 64

results_bvcc/w2v2_large_asr_en/layer14/run3/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:069389fc67653cde3cc867f7aaba683212efec873d300ef5aee50aa5b581a980
+size 3403339

results_bvcc/w2v2_large_asr_en/layer14/run3/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4ef4b777e9e71d9ac8929c8805232b3010a34334d3386463b4f3820a6efc2dfd
+size 3404998

results_bvcc/w2v2_large_asr_en/layer14/run3/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.6228337616492541
+INFO:root:
+[Valid][0][UTT][ MSE = 0.4639 | LCC = 0.7118 | SRCC = 0.7112 ]
+INFO:root:Average loss=0.36973212048029286
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3787 | LCC = 0.7793 | SRCC = 0.7781 ]
+INFO:root:Average loss=0.31610557245902526
+INFO:root:
+[Valid][2][UTT][ MSE = 0.3625 | LCC = 0.7884 | SRCC = 0.7869 ]
+INFO:root:Average loss=0.2852938834291238
+INFO:root:
+[Valid][3][UTT][ MSE = 0.3119 | LCC = 0.8009 | SRCC = 0.7975 ]
+INFO:root:Average loss=0.2740824589362511
+INFO:root:
+[Valid][4][UTT][ MSE = 0.3700 | LCC = 0.7982 | SRCC = 0.7940 ]
+INFO:root:Average loss=0.24433110864498678
+INFO:root:
+[Valid][5][UTT][ MSE = 0.3307 | LCC = 0.8002 | SRCC = 0.7949 ]
+INFO:root:Average loss=0.23941120295188364
+INFO:root:
+[Valid][6][UTT][ MSE = 0.3721 | LCC = 0.8017 | SRCC = 0.7983 ]
+INFO:root:Average loss=0.2147610577253195
+INFO:root:
+[Valid][7][UTT][ MSE = 0.3154 | LCC = 0.8049 | SRCC = 0.8007 ]
+INFO:root:Average loss=0.1963093800422473
+INFO:root:
+[Valid][8][UTT][ MSE = 0.3047 | LCC = 0.8082 | SRCC = 0.8045 ]
+INFO:root:Average loss=0.19134491806228957
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2998 | LCC = 0.8076 | SRCC = 0.8039 ]
+INFO:root:Average loss=0.17910867795730248
+INFO:root:
+[Valid][10][UTT][ MSE = 0.3235 | LCC = 0.8018 | SRCC = 0.8003 ]
+INFO:root:Average loss=0.1719729864062407
+INFO:root:
+[Valid][11][UTT][ MSE = 0.3029 | LCC = 0.8019 | SRCC = 0.7983 ]
+INFO:root:Average loss=0.15981111025963074
+INFO:root:
+[Valid][12][UTT][ MSE = 0.3317 | LCC = 0.8037 | SRCC = 0.8021 ]
+INFO:root:Average loss=0.1607825325276607
+INFO:root:
+[Valid][13][UTT][ MSE = 0.3118 | LCC = 0.8021 | SRCC = 0.8010 ]
+INFO:root:Average loss=0.15734562745843178
+INFO:root:
+[Valid][14][UTT][ MSE = 0.3096 | LCC = 0.7990 | SRCC = 0.7966 ]
+INFO:root:Average loss=0.1332972825337679
+INFO:root:
+[Valid][15][UTT][ MSE = 0.3301 | LCC = 0.7923 | SRCC = 0.7919 ]
+INFO:root:Average loss=0.1389224115663614
+INFO:root:
+[Valid][16][UTT][ MSE = 0.3122 | LCC = 0.8018 | SRCC = 0.7999 ]
+INFO:root:Average loss=0.13425574289300504
+INFO:root:
+[Valid][17][UTT][ MSE = 0.3329 | LCC = 0.8014 | SRCC = 0.7979 ]
+INFO:root:Average loss=0.11482151206105183
+INFO:root:
+[Valid][18][UTT][ MSE = 0.3155 | LCC = 0.7997 | SRCC = 0.7972 ]
+INFO:root:Average loss=0.11373232744443111
+INFO:root:
+[Valid][19][UTT][ MSE = 0.3101 | LCC = 0.8007 | SRCC = 0.7981 ]
+INFO:root:Average loss=0.10444456740067555
+INFO:root:
+[Valid][20][UTT][ MSE = 0.3060 | LCC = 0.7958 | SRCC = 0.7934 ]
+INFO:root:Average loss=0.09774937857993138
+INFO:root:
+[Valid][21][UTT][ MSE = 0.3169 | LCC = 0.7925 | SRCC = 0.7909 ]
+INFO:root:Average loss=0.09432413472005954
+INFO:root:
+[Valid][22][UTT][ MSE = 0.3318 | LCC = 0.7895 | SRCC = 0.7878 ]
+INFO:root:Average loss=0.09473877371503757
+INFO:root:
+[Valid][23][UTT][ MSE = 0.3360 | LCC = 0.7912 | SRCC = 0.7904 ]
+INFO:root:Average loss=0.09524659310969022
+INFO:root:
+[Valid][24][UTT][ MSE = 0.3147 | LCC = 0.7952 | SRCC = 0.7928 ]
+INFO:root:Average loss=0.08925897604188858
+INFO:root:
+[Valid][25][UTT][ MSE = 0.3078 | LCC = 0.7954 | SRCC = 0.7935 ]
+INFO:root:Average loss=0.08056991270337349
+INFO:root:
+[Valid][26][UTT][ MSE = 0.3018 | LCC = 0.7978 | SRCC = 0.7968 ]
+INFO:root:Average loss=0.07634544687775466
+INFO:root:
+[Valid][27][UTT][ MSE = 0.3187 | LCC = 0.7910 | SRCC = 0.7881 ]
+INFO:root:Average loss=0.07849306880663602
+INFO:root:
+[Valid][28][UTT][ MSE = 0.3045 | LCC = 0.7995 | SRCC = 0.7976 ]
+INFO:root:Average loss=0.07884880671134362
+INFO:root:
+[Valid][29][UTT][ MSE = 0.3241 | LCC = 0.7931 | SRCC = 0.7911 ]
+INFO:root:Average loss=0.0762025519536856
+INFO:root:
+[Valid][30][UTT][ MSE = 0.3172 | LCC = 0.7931 | SRCC = 0.7910 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.3233 | LCC = 0.8182 | SRCC = 0.8232 ]

results_bvcc/w2v2_large_asr_en/layer14/run4/config.gin

@@ -0,0 +1,31 @@
+# Training config on BVCC dataset.
+
+# ------------------------------
+# Architecture
+# ------------------------------
+ProjectionHead:
+    conv_channels = (32, 32)
+    dense_neurons = (64, 32, 1)
+    use_poolings = (True, True)
+    kernel_size = 5
+    pool_size = 5
+    
+# ------------------------------
+# Dataset
+# ------------------------------
+Bvcc:
+    debug = False
+
+get_dataloader:
+    num_workers = 12
+    shuffle = True
+
+# ------------------------------
+# Training loop
+# ------------------------------
+TrainingLoop:
+    num_epochs = 30
+    learning_rate = 1e-4
+    dataset_cls = @Bvcc
+    weight_decay = 0.0
+    batch_size_train = 64

results_bvcc/w2v2_large_asr_en/layer14/run4/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:85672245a0b51022641a0d909fb1c8b224bc0e681ed39290f994f2275b15daf4
+size 3403339

results_bvcc/w2v2_large_asr_en/layer14/run4/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a5afadd72191475a9b936a4ff7e854fbfa08c81774a315128368698ed55c638b
+size 3405062