Upload SSL Layer MOS checkpoints (part 19)

results_tencent/w2v2_xlsr_1b/layer34/run1/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:daafc5d61a19dfb697bb03c6ff44d9b58fcf682917efaae08ff0d19d446c8025
+size 4060422

results_tencent/w2v2_xlsr_1b/layer34/run1/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 8000
+INFO:root:Num val speech clips: 2000
+INFO:root:Num test speech clips: 1563
+INFO:root:Device=cuda
+INFO:root:Average loss=0.4902742062211037
+INFO:root:
+[Valid][0][UTT][ MSE = 0.2398 | LCC = 0.9484 | SRCC = 0.9459 ]
+INFO:root:Average loss=0.14633447909355163
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2604 | LCC = 0.9567 | SRCC = 0.9579 ]
+INFO:root:Average loss=0.12721437513828276
+INFO:root:
+[Valid][2][UTT][ MSE = 0.1669 | LCC = 0.9626 | SRCC = 0.9621 ]
+INFO:root:Average loss=0.1046346583366394
+INFO:root:
+[Valid][3][UTT][ MSE = 0.1866 | LCC = 0.9657 | SRCC = 0.9651 ]
+INFO:root:Average loss=0.09733275064826012
+INFO:root:
+[Valid][4][UTT][ MSE = 0.1638 | LCC = 0.9664 | SRCC = 0.9660 ]
+INFO:root:Average loss=0.08789659485220909
+INFO:root:
+[Valid][5][UTT][ MSE = 0.1676 | LCC = 0.9684 | SRCC = 0.9675 ]
+INFO:root:Average loss=0.0830150817334652
+INFO:root:
+[Valid][6][UTT][ MSE = 0.1765 | LCC = 0.9677 | SRCC = 0.9679 ]
+INFO:root:Average loss=0.07978170904517173
+INFO:root:
+[Valid][7][UTT][ MSE = 0.1479 | LCC = 0.9699 | SRCC = 0.9690 ]
+INFO:root:Average loss=0.07524063646793365
+INFO:root:
+[Valid][8][UTT][ MSE = 0.1509 | LCC = 0.9680 | SRCC = 0.9665 ]
+INFO:root:Average loss=0.07233297297358512
+INFO:root:
+[Valid][9][UTT][ MSE = 0.1749 | LCC = 0.9703 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.06812826901674271
+INFO:root:
+[Valid][10][UTT][ MSE = 0.1709 | LCC = 0.9708 | SRCC = 0.9702 ]
+INFO:root:Average loss=0.06281923732161522
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1504 | LCC = 0.9711 | SRCC = 0.9704 ]
+INFO:root:Average loss=0.06082053729891777
+INFO:root:
+[Valid][12][UTT][ MSE = 0.1516 | LCC = 0.9705 | SRCC = 0.9699 ]
+INFO:root:Average loss=0.05907168531417847
+INFO:root:
+[Valid][13][UTT][ MSE = 0.1739 | LCC = 0.9711 | SRCC = 0.9702 ]
+INFO:root:Average loss=0.05746838785707951
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1801 | LCC = 0.9712 | SRCC = 0.9697 ]
+INFO:root:Average loss=0.05068235886096954
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1560 | LCC = 0.9709 | SRCC = 0.9695 ]
+INFO:root:Average loss=0.05124501751363277
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1653 | LCC = 0.9714 | SRCC = 0.9705 ]
+INFO:root:Average loss=0.04573335160315037
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1542 | LCC = 0.9713 | SRCC = 0.9702 ]
+INFO:root:Average loss=0.045032204672694205
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1484 | LCC = 0.9710 | SRCC = 0.9701 ]
+INFO:root:Average loss=0.04412492543458939
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1536 | LCC = 0.9713 | SRCC = 0.9705 ]
+INFO:root:Average loss=0.04120438221096993
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1473 | LCC = 0.9714 | SRCC = 0.9704 ]
+INFO:root:Average loss=0.04032501465082169
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1548 | LCC = 0.9712 | SRCC = 0.9700 ]
+INFO:root:Average loss=0.03921673908829689
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1411 | LCC = 0.9723 | SRCC = 0.9711 ]
+INFO:root:Average loss=0.03591878201067448
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1573 | LCC = 0.9710 | SRCC = 0.9699 ]
+INFO:root:Average loss=0.03771319682896137
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1470 | LCC = 0.9718 | SRCC = 0.9706 ]
+INFO:root:Average loss=0.03572441041469574
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1560 | LCC = 0.9713 | SRCC = 0.9704 ]
+INFO:root:Average loss=0.03770526441931724
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1440 | LCC = 0.9716 | SRCC = 0.9704 ]
+INFO:root:Average loss=0.03133926630020142
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1340 | LCC = 0.9704 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.031310425609350205
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1377 | LCC = 0.9717 | SRCC = 0.9707 ]
+INFO:root:Average loss=0.030254511252045632
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1485 | LCC = 0.9717 | SRCC = 0.9703 ]
+INFO:root:Average loss=0.028796360448002815
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1493 | LCC = 0.9713 | SRCC = 0.9695 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1423 | LCC = 0.9713 | SRCC = 0.9707 ]

results_tencent/w2v2_xlsr_1b/layer34/run2/config.gin

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

results_tencent/w2v2_xlsr_1b/layer34/run2/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:abfe23df20637d2229ab176812129e6bd0599c366e2b20c580fa535c2a9ebdea
+size 4058699

results_tencent/w2v2_xlsr_1b/layer34/run2/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:53f2f520bb57a11f1f04be9999c6d7d7444c4ec5e08d7a29296dec1b317ab0ff
+size 4060358

results_tencent/w2v2_xlsr_1b/layer34/run2/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 8000
+INFO:root:Num val speech clips: 2000
+INFO:root:Num test speech clips: 1563
+INFO:root:Device=cuda
+INFO:root:Average loss=0.5116374804973602
+INFO:root:
+[Valid][0][UTT][ MSE = 0.2605 | LCC = 0.9452 | SRCC = 0.9420 ]
+INFO:root:Average loss=0.17314399302005767
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2513 | LCC = 0.9569 | SRCC = 0.9561 ]
+INFO:root:Average loss=0.12890104061365126
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2070 | LCC = 0.9627 | SRCC = 0.9625 ]
+INFO:root:Average loss=0.11005250978469848
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2378 | LCC = 0.9647 | SRCC = 0.9645 ]
+INFO:root:Average loss=0.0985694120824337
+INFO:root:
+[Valid][4][UTT][ MSE = 0.1762 | LCC = 0.9658 | SRCC = 0.9656 ]
+INFO:root:Average loss=0.09373472476005554
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2205 | LCC = 0.9663 | SRCC = 0.9663 ]
+INFO:root:Average loss=0.08342433172464371
+INFO:root:
+[Valid][6][UTT][ MSE = 0.1952 | LCC = 0.9666 | SRCC = 0.9667 ]
+INFO:root:Average loss=0.08070914933085442
+INFO:root:
+[Valid][7][UTT][ MSE = 0.1980 | LCC = 0.9677 | SRCC = 0.9670 ]
+INFO:root:Average loss=0.0756270144879818
+INFO:root:
+[Valid][8][UTT][ MSE = 0.1873 | LCC = 0.9693 | SRCC = 0.9685 ]
+INFO:root:Average loss=0.07366727724671364
+INFO:root:
+[Valid][9][UTT][ MSE = 0.1528 | LCC = 0.9683 | SRCC = 0.9689 ]
+INFO:root:Average loss=0.06635770091414452
+INFO:root:
+[Valid][10][UTT][ MSE = 0.1696 | LCC = 0.9685 | SRCC = 0.9679 ]
+INFO:root:Average loss=0.06216112345457077
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1881 | LCC = 0.9706 | SRCC = 0.9695 ]
+INFO:root:Average loss=0.06074040991067886
+INFO:root:
+[Valid][12][UTT][ MSE = 0.1688 | LCC = 0.9704 | SRCC = 0.9701 ]
+INFO:root:Average loss=0.060610357597470285
+INFO:root:
+[Valid][13][UTT][ MSE = 0.1765 | LCC = 0.9706 | SRCC = 0.9696 ]
+INFO:root:Average loss=0.055302672609686854
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1679 | LCC = 0.9703 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.05374186682701111
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1651 | LCC = 0.9705 | SRCC = 0.9694 ]
+INFO:root:Average loss=0.05052183696627617
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1975 | LCC = 0.9703 | SRCC = 0.9693 ]
+INFO:root:Average loss=0.05052878153324127
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1635 | LCC = 0.9702 | SRCC = 0.9697 ]
+INFO:root:Average loss=0.048052635431289674
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1695 | LCC = 0.9715 | SRCC = 0.9701 ]
+INFO:root:Average loss=0.04415951931476593
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1717 | LCC = 0.9712 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.04442810708284378
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1783 | LCC = 0.9718 | SRCC = 0.9707 ]
+INFO:root:Average loss=0.039961826995015146
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1973 | LCC = 0.9715 | SRCC = 0.9710 ]
+INFO:root:Average loss=0.03955658057332039
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1821 | LCC = 0.9697 | SRCC = 0.9684 ]
+INFO:root:Average loss=0.038611474484205245
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1539 | LCC = 0.9718 | SRCC = 0.9708 ]
+INFO:root:Average loss=0.0374235537648201
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1552 | LCC = 0.9692 | SRCC = 0.9690 ]
+INFO:root:Average loss=0.035631714433431626
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1569 | LCC = 0.9705 | SRCC = 0.9703 ]
+INFO:root:Average loss=0.034970214545726776
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1726 | LCC = 0.9710 | SRCC = 0.9700 ]
+INFO:root:Average loss=0.03262574368715286
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1751 | LCC = 0.9700 | SRCC = 0.9691 ]
+INFO:root:Average loss=0.03202556210756302
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1588 | LCC = 0.9708 | SRCC = 0.9702 ]
+INFO:root:Average loss=0.030695222318172455
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1656 | LCC = 0.9713 | SRCC = 0.9699 ]
+INFO:root:Average loss=0.03096596737205982
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1641 | LCC = 0.9709 | SRCC = 0.9698 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1746 | LCC = 0.9733 | SRCC = 0.9725 ]

results_tencent/w2v2_xlsr_1b/layer34/run3/config.gin

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

results_tencent/w2v2_xlsr_1b/layer34/run3/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:420f117e0bef9ec02074948b2e72ab0273af756c09b6da9cae22f932e33cbbea
+size 4058699

results_tencent/w2v2_xlsr_1b/layer34/run3/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:25c7f5658c0535e04cad24513dd3de1ef2bf2dcadeea569d6b8ac5bf8c5a313f
+size 4060358

results_tencent/w2v2_xlsr_1b/layer34/run3/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 8000
+INFO:root:Num val speech clips: 2000
+INFO:root:Num test speech clips: 1563
+INFO:root:Device=cuda
+INFO:root:Average loss=0.42322697162628176
+INFO:root:
+[Valid][0][UTT][ MSE = 0.2565 | LCC = 0.9407 | SRCC = 0.9397 ]
+INFO:root:Average loss=0.1641413117647171
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2434 | LCC = 0.9571 | SRCC = 0.9567 ]
+INFO:root:Average loss=0.12665840870141984
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2400 | LCC = 0.9621 | SRCC = 0.9604 ]
+INFO:root:Average loss=0.11475816321372986
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2087 | LCC = 0.9644 | SRCC = 0.9643 ]
+INFO:root:Average loss=0.10551588690280914
+INFO:root:
+[Valid][4][UTT][ MSE = 0.1569 | LCC = 0.9667 | SRCC = 0.9660 ]
+INFO:root:Average loss=0.09330324819684029
+INFO:root:
+[Valid][5][UTT][ MSE = 0.1858 | LCC = 0.9684 | SRCC = 0.9679 ]
+INFO:root:Average loss=0.08807834824919701
+INFO:root:
+[Valid][6][UTT][ MSE = 0.1634 | LCC = 0.9681 | SRCC = 0.9683 ]
+INFO:root:Average loss=0.07908114370703698
+INFO:root:
+[Valid][7][UTT][ MSE = 0.1693 | LCC = 0.9696 | SRCC = 0.9692 ]
+INFO:root:Average loss=0.07634212982654572
+INFO:root:
+[Valid][8][UTT][ MSE = 0.1892 | LCC = 0.9700 | SRCC = 0.9697 ]
+INFO:root:Average loss=0.07429390037059784
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2025 | LCC = 0.9702 | SRCC = 0.9699 ]
+INFO:root:Average loss=0.0714883383512497
+INFO:root:
+[Valid][10][UTT][ MSE = 0.1644 | LCC = 0.9706 | SRCC = 0.9700 ]
+INFO:root:Average loss=0.06761336946487427
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1512 | LCC = 0.9696 | SRCC = 0.9701 ]
+INFO:root:Average loss=0.06911549091339111
+INFO:root:
+[Valid][12][UTT][ MSE = 0.1901 | LCC = 0.9710 | SRCC = 0.9702 ]
+INFO:root:Average loss=0.05884688803553581
+INFO:root:
+[Valid][13][UTT][ MSE = 0.1729 | LCC = 0.9717 | SRCC = 0.9707 ]
+INFO:root:Average loss=0.05701669503748417
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1604 | LCC = 0.9712 | SRCC = 0.9702 ]
+INFO:root:Average loss=0.05438076056540012
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1503 | LCC = 0.9723 | SRCC = 0.9714 ]
+INFO:root:Average loss=0.052790969014167785
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1933 | LCC = 0.9715 | SRCC = 0.9717 ]
+INFO:root:Average loss=0.05152842557430267
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1698 | LCC = 0.9725 | SRCC = 0.9717 ]
+INFO:root:Average loss=0.048267154276371
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1493 | LCC = 0.9708 | SRCC = 0.9697 ]
+INFO:root:Average loss=0.04684811107814312
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1916 | LCC = 0.9712 | SRCC = 0.9707 ]
+INFO:root:Average loss=0.044876496985554695
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1547 | LCC = 0.9724 | SRCC = 0.9712 ]
+INFO:root:Average loss=0.04433149087429047
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1653 | LCC = 0.9709 | SRCC = 0.9699 ]
+INFO:root:Average loss=0.04125556680560112
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1612 | LCC = 0.9724 | SRCC = 0.9713 ]
+INFO:root:Average loss=0.03947563420236111
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1622 | LCC = 0.9726 | SRCC = 0.9717 ]
+INFO:root:Average loss=0.036119799092411996
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1686 | LCC = 0.9722 | SRCC = 0.9709 ]
+INFO:root:Average loss=0.036942376121878624
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1474 | LCC = 0.9721 | SRCC = 0.9712 ]
+INFO:root:Average loss=0.03647490967810154
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1468 | LCC = 0.9715 | SRCC = 0.9712 ]
+INFO:root:Average loss=0.03421395343542099
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1701 | LCC = 0.9723 | SRCC = 0.9713 ]
+INFO:root:Average loss=0.032851971536874774
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1552 | LCC = 0.9720 | SRCC = 0.9710 ]
+INFO:root:Average loss=0.03334848453104496
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1555 | LCC = 0.9725 | SRCC = 0.9713 ]
+INFO:root:Average loss=0.0340045747384429
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1596 | LCC = 0.9725 | SRCC = 0.9715 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1614 | LCC = 0.9727 | SRCC = 0.9722 ]

results_tencent/w2v2_xlsr_1b/layer34/run4/config.gin

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

results_tencent/w2v2_xlsr_1b/layer34/run4/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:df6558a0a0b1b3e9512c07a3f4084d8d6897db66a5edc40d4d541dd8dd46095b
+size 4058699

results_tencent/w2v2_xlsr_1b/layer34/run4/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c98a20d728561cb7c5b8562fef7f4f43af97c06067c01bf69c4e5d3837cc3a8e
+size 4060358

results_tencent/w2v2_xlsr_1b/layer34/run4/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 8000
+INFO:root:Num val speech clips: 2000
+INFO:root:Num test speech clips: 1563
+INFO:root:Device=cuda
+INFO:root:Average loss=0.36400548738241195
+INFO:root:
+[Valid][0][UTT][ MSE = 0.2443 | LCC = 0.9433 | SRCC = 0.9420 ]
+INFO:root:Average loss=0.15119187492132186
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2264 | LCC = 0.9552 | SRCC = 0.9540 ]
+INFO:root:Average loss=0.12893920147418975
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2173 | LCC = 0.9599 | SRCC = 0.9593 ]
+INFO:root:Average loss=0.11330500611662865
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2049 | LCC = 0.9610 | SRCC = 0.9615 ]
+INFO:root:Average loss=0.09980326175689697
+INFO:root:
+[Valid][4][UTT][ MSE = 0.1914 | LCC = 0.9647 | SRCC = 0.9642 ]
+INFO:root:Average loss=0.08943218106031418
+INFO:root:
+[Valid][5][UTT][ MSE = 0.1918 | LCC = 0.9655 | SRCC = 0.9650 ]
+INFO:root:Average loss=0.0829474957883358
+INFO:root:
+[Valid][6][UTT][ MSE = 0.1767 | LCC = 0.9657 | SRCC = 0.9655 ]
+INFO:root:Average loss=0.08013744083046913
+INFO:root:
+[Valid][7][UTT][ MSE = 0.1753 | LCC = 0.9635 | SRCC = 0.9643 ]
+INFO:root:Average loss=0.07496358072757721
+INFO:root:
+[Valid][8][UTT][ MSE = 0.1987 | LCC = 0.9680 | SRCC = 0.9674 ]
+INFO:root:Average loss=0.0705914244055748
+INFO:root:
+[Valid][9][UTT][ MSE = 0.1751 | LCC = 0.9681 | SRCC = 0.9674 ]
+INFO:root:Average loss=0.06566153863072395
+INFO:root:
+[Valid][10][UTT][ MSE = 0.1477 | LCC = 0.9667 | SRCC = 0.9675 ]
+INFO:root:Average loss=0.06612320005893707
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1635 | LCC = 0.9683 | SRCC = 0.9677 ]
+INFO:root:Average loss=0.06122040030360222
+INFO:root:
+[Valid][12][UTT][ MSE = 0.1631 | LCC = 0.9681 | SRCC = 0.9678 ]
+INFO:root:Average loss=0.05771169379353523
+INFO:root:
+[Valid][13][UTT][ MSE = 0.1870 | LCC = 0.9687 | SRCC = 0.9677 ]
+INFO:root:Average loss=0.05371959680318832
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1564 | LCC = 0.9684 | SRCC = 0.9674 ]
+INFO:root:Average loss=0.05002111430466175
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2096 | LCC = 0.9673 | SRCC = 0.9676 ]
+INFO:root:Average loss=0.050281098887324335
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2241 | LCC = 0.9676 | SRCC = 0.9676 ]
+INFO:root:Average loss=0.048399248272180555
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1610 | LCC = 0.9685 | SRCC = 0.9677 ]
+INFO:root:Average loss=0.04781105010211468
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1644 | LCC = 0.9691 | SRCC = 0.9683 ]
+INFO:root:Average loss=0.045207404017448424
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1852 | LCC = 0.9687 | SRCC = 0.9672 ]
+INFO:root:Average loss=0.04095040114223957
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1715 | LCC = 0.9685 | SRCC = 0.9679 ]
+INFO:root:Average loss=0.03902539610862732
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1677 | LCC = 0.9688 | SRCC = 0.9678 ]
+INFO:root:Average loss=0.038662344694137575
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1551 | LCC = 0.9690 | SRCC = 0.9681 ]
+INFO:root:Average loss=0.03937659129500389
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1581 | LCC = 0.9686 | SRCC = 0.9677 ]
+INFO:root:Average loss=0.0358126219958067
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1585 | LCC = 0.9679 | SRCC = 0.9673 ]
+INFO:root:Average loss=0.03809351152181625
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1512 | LCC = 0.9686 | SRCC = 0.9682 ]
+INFO:root:Average loss=0.034593225583434105
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1873 | LCC = 0.9682 | SRCC = 0.9672 ]
+INFO:root:Average loss=0.03302263800799847
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1403 | LCC = 0.9678 | SRCC = 0.9671 ]
+INFO:root:Average loss=0.03317894513905048
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1478 | LCC = 0.9679 | SRCC = 0.9678 ]
+INFO:root:Average loss=0.031097458362579346
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1637 | LCC = 0.9684 | SRCC = 0.9677 ]
+INFO:root:Average loss=0.029349580004811287
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1649 | LCC = 0.9693 | SRCC = 0.9685 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1617 | LCC = 0.9708 | SRCC = 0.9705 ]

results_tencent/w2v2_xlsr_1b/layer36/run0/config.gin

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

results_tencent/w2v2_xlsr_1b/layer36/run0/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c3de9439ad96ddcb9aa14b1035b927b3b9a10843b32745befdea5e3c47d99832
+size 4058699

results_tencent/w2v2_xlsr_1b/layer36/run0/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b33b6a2f48b87897d70f89deb210c93b16b419199afe32ae7815c2e3dab3f6e7
+size 4060358

results_tencent/w2v2_xlsr_1b/layer36/run0/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 8000
+INFO:root:Num val speech clips: 2000
+INFO:root:Num test speech clips: 1563
+INFO:root:Device=cuda
+INFO:root:Average loss=0.37084116476774215
+INFO:root:
+[Valid][0][UTT][ MSE = 0.2625 | LCC = 0.9456 | SRCC = 0.9431 ]
+INFO:root:Average loss=0.1466196327805519
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2394 | LCC = 0.9583 | SRCC = 0.9575 ]
+INFO:root:Average loss=0.11966153651475907
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2065 | LCC = 0.9630 | SRCC = 0.9634 ]
+INFO:root:Average loss=0.09997559505701065
+INFO:root:
+[Valid][3][UTT][ MSE = 0.1902 | LCC = 0.9657 | SRCC = 0.9652 ]
+INFO:root:Average loss=0.09148479464650154
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2435 | LCC = 0.9668 | SRCC = 0.9661 ]
+INFO:root:Average loss=0.08862799698114396
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2219 | LCC = 0.9672 | SRCC = 0.9669 ]
+INFO:root:Average loss=0.07906556087732315
+INFO:root:
+[Valid][6][UTT][ MSE = 0.1955 | LCC = 0.9687 | SRCC = 0.9682 ]
+INFO:root:Average loss=0.0771837994158268
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2188 | LCC = 0.9675 | SRCC = 0.9672 ]
+INFO:root:Average loss=0.07137568792700767
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2150 | LCC = 0.9686 | SRCC = 0.9682 ]
+INFO:root:Average loss=0.06630247557163238
+INFO:root:
+[Valid][9][UTT][ MSE = 0.1683 | LCC = 0.9695 | SRCC = 0.9690 ]
+INFO:root:Average loss=0.06341636164486408
+INFO:root:
+[Valid][10][UTT][ MSE = 0.1664 | LCC = 0.9703 | SRCC = 0.9696 ]
+INFO:root:Average loss=0.05839603883028031
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1710 | LCC = 0.9699 | SRCC = 0.9692 ]
+INFO:root:Average loss=0.05651375888288021
+INFO:root:
+[Valid][12][UTT][ MSE = 0.1611 | LCC = 0.9697 | SRCC = 0.9692 ]
+INFO:root:Average loss=0.05258131051063537
+INFO:root:
+[Valid][13][UTT][ MSE = 0.1918 | LCC = 0.9702 | SRCC = 0.9694 ]
+INFO:root:Average loss=0.04937686942517758
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1657 | LCC = 0.9704 | SRCC = 0.9695 ]
+INFO:root:Average loss=0.05143510685861111
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1500 | LCC = 0.9700 | SRCC = 0.9691 ]
+INFO:root:Average loss=0.04847771035134792
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1584 | LCC = 0.9710 | SRCC = 0.9700 ]
+INFO:root:Average loss=0.04459230253100395
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1822 | LCC = 0.9692 | SRCC = 0.9691 ]
+INFO:root:Average loss=0.04214941976964474
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2095 | LCC = 0.9691 | SRCC = 0.9685 ]
+INFO:root:Average loss=0.04159797143936157
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1819 | LCC = 0.9707 | SRCC = 0.9699 ]
+INFO:root:Average loss=0.03915795786678791
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1957 | LCC = 0.9713 | SRCC = 0.9706 ]
+INFO:root:Average loss=0.03757145771384239
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1768 | LCC = 0.9703 | SRCC = 0.9699 ]
+INFO:root:Average loss=0.0359332632124424
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1579 | LCC = 0.9701 | SRCC = 0.9699 ]
+INFO:root:Average loss=0.0351613128632307
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1718 | LCC = 0.9705 | SRCC = 0.9697 ]
+INFO:root:Average loss=0.034543736264109615
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1587 | LCC = 0.9710 | SRCC = 0.9702 ]
+INFO:root:Average loss=0.03201238319277763
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1876 | LCC = 0.9711 | SRCC = 0.9705 ]
+INFO:root:Average loss=0.032399823412299154
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1858 | LCC = 0.9718 | SRCC = 0.9708 ]
+INFO:root:Average loss=0.028996627867221833
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1793 | LCC = 0.9706 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.028090556278824806
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1949 | LCC = 0.9704 | SRCC = 0.9696 ]
+INFO:root:Average loss=0.029231697738170623
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1894 | LCC = 0.9710 | SRCC = 0.9702 ]
+INFO:root:Average loss=0.02756267496943474
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1757 | LCC = 0.9710 | SRCC = 0.9701 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1820 | LCC = 0.9722 | SRCC = 0.9716 ]

results_tencent/w2v2_xlsr_1b/layer36/run1/config.gin

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

results_tencent/w2v2_xlsr_1b/layer36/run1/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aee5d8fe62198099bf1497d3c85640e789ab5b3bb0f3415b6090c86dd873b73b
+size 4058699

results_tencent/w2v2_xlsr_1b/layer36/run1/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:336196f3a4b86b38c3dfcf82e27a83a10128a1321cb50b8379c4b9830c94ea53
+size 4060358

results_tencent/w2v2_xlsr_1b/layer36/run1/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 8000
+INFO:root:Num val speech clips: 2000
+INFO:root:Num test speech clips: 1563
+INFO:root:Device=cuda
+INFO:root:Average loss=0.39617545187473296
+INFO:root:
+[Valid][0][UTT][ MSE = 0.2443 | LCC = 0.9484 | SRCC = 0.9477 ]
+INFO:root:Average loss=0.14206642138957978
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2131 | LCC = 0.9578 | SRCC = 0.9575 ]
+INFO:root:Average loss=0.11275172126293183
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2335 | LCC = 0.9636 | SRCC = 0.9629 ]
+INFO:root:Average loss=0.09932926201820373
+INFO:root:
+[Valid][3][UTT][ MSE = 0.1807 | LCC = 0.9650 | SRCC = 0.9646 ]
+INFO:root:Average loss=0.09484005585312844
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2079 | LCC = 0.9673 | SRCC = 0.9666 ]
+INFO:root:Average loss=0.08880963152647019
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2065 | LCC = 0.9680 | SRCC = 0.9672 ]
+INFO:root:Average loss=0.08077442383766174
+INFO:root:
+[Valid][6][UTT][ MSE = 0.1892 | LCC = 0.9688 | SRCC = 0.9683 ]
+INFO:root:Average loss=0.07647999602556228
+INFO:root:
+[Valid][7][UTT][ MSE = 0.1796 | LCC = 0.9693 | SRCC = 0.9685 ]
+INFO:root:Average loss=0.06898756206035614
+INFO:root:
+[Valid][8][UTT][ MSE = 0.1517 | LCC = 0.9705 | SRCC = 0.9699 ]
+INFO:root:Average loss=0.0667625637948513
+INFO:root:
+[Valid][9][UTT][ MSE = 0.1771 | LCC = 0.9703 | SRCC = 0.9697 ]
+INFO:root:Average loss=0.06127203559875488
+INFO:root:
+[Valid][10][UTT][ MSE = 0.1989 | LCC = 0.9700 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.06175964760780334
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1400 | LCC = 0.9700 | SRCC = 0.9695 ]
+INFO:root:Average loss=0.058717340111732484
+INFO:root:
+[Valid][12][UTT][ MSE = 0.1345 | LCC = 0.9696 | SRCC = 0.9703 ]
+INFO:root:Average loss=0.05347913739085197
+INFO:root:
+[Valid][13][UTT][ MSE = 0.1880 | LCC = 0.9705 | SRCC = 0.9694 ]
+INFO:root:Average loss=0.051421783030033114
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1511 | LCC = 0.9712 | SRCC = 0.9705 ]
+INFO:root:Average loss=0.04982315538823605
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1833 | LCC = 0.9716 | SRCC = 0.9706 ]
+INFO:root:Average loss=0.04831311704218388
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1784 | LCC = 0.9709 | SRCC = 0.9706 ]
+INFO:root:Average loss=0.046446365997195246
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1412 | LCC = 0.9709 | SRCC = 0.9709 ]
+INFO:root:Average loss=0.04322810117900371
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1654 | LCC = 0.9710 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.0410259318202734
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1631 | LCC = 0.9712 | SRCC = 0.9704 ]
+INFO:root:Average loss=0.03919383545219898
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1579 | LCC = 0.9706 | SRCC = 0.9696 ]
+INFO:root:Average loss=0.03788477848470211
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1510 | LCC = 0.9714 | SRCC = 0.9703 ]
+INFO:root:Average loss=0.03533661440014839
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1507 | LCC = 0.9709 | SRCC = 0.9700 ]
+INFO:root:Average loss=0.03758320347964764
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1496 | LCC = 0.9712 | SRCC = 0.9703 ]
+INFO:root:Average loss=0.0353399672806263
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1368 | LCC = 0.9708 | SRCC = 0.9701 ]
+INFO:root:Average loss=0.035094783678650855
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1531 | LCC = 0.9719 | SRCC = 0.9710 ]
+INFO:root:Average loss=0.031619231402873996
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1663 | LCC = 0.9724 | SRCC = 0.9713 ]
+INFO:root:Average loss=0.029193493619561195
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1574 | LCC = 0.9715 | SRCC = 0.9707 ]
+INFO:root:Average loss=0.03056273479759693
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1506 | LCC = 0.9712 | SRCC = 0.9702 ]
+INFO:root:Average loss=0.028620364226400852
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1528 | LCC = 0.9706 | SRCC = 0.9701 ]
+INFO:root:Average loss=0.029940906956791878
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1478 | LCC = 0.9709 | SRCC = 0.9701 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1640 | LCC = 0.9727 | SRCC = 0.9721 ]

results_tencent/w2v2_xlsr_1b/layer36/run2/config.gin

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

results_tencent/w2v2_xlsr_1b/layer36/run2/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:674d0ab958539b5080b868effa331dd344e43dba7c26a2e04395e1e9e1966012
+size 4058699

results_tencent/w2v2_xlsr_1b/layer36/run2/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aebe68c5e5e9efcc95728859252b408c01eaad77e6482d708669c331199afb96
+size 4060358

results_tencent/w2v2_xlsr_1b/layer36/run2/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 8000
+INFO:root:Num val speech clips: 2000
+INFO:root:Num test speech clips: 1563
+INFO:root:Device=cuda
+INFO:root:Average loss=0.39079192459583284
+INFO:root:
+[Valid][0][UTT][ MSE = 0.2299 | LCC = 0.9470 | SRCC = 0.9461 ]
+INFO:root:Average loss=0.14307945865392685
+INFO:root:
+[Valid][1][UTT][ MSE = 0.1950 | LCC = 0.9590 | SRCC = 0.9584 ]
+INFO:root:Average loss=0.11326974377036095
+INFO:root:
+[Valid][2][UTT][ MSE = 0.1983 | LCC = 0.9648 | SRCC = 0.9645 ]
+INFO:root:Average loss=0.10428494682908059
+INFO:root:
+[Valid][3][UTT][ MSE = 0.1877 | LCC = 0.9669 | SRCC = 0.9663 ]
+INFO:root:Average loss=0.0958240289092064
+INFO:root:
+[Valid][4][UTT][ MSE = 0.1907 | LCC = 0.9672 | SRCC = 0.9668 ]
+INFO:root:Average loss=0.0870006841123104
+INFO:root:
+[Valid][5][UTT][ MSE = 0.1801 | LCC = 0.9684 | SRCC = 0.9676 ]
+INFO:root:Average loss=0.07758983084559441
+INFO:root:
+[Valid][6][UTT][ MSE = 0.1796 | LCC = 0.9686 | SRCC = 0.9683 ]
+INFO:root:Average loss=0.07353946301341056
+INFO:root:
+[Valid][7][UTT][ MSE = 0.1427 | LCC = 0.9690 | SRCC = 0.9691 ]
+INFO:root:Average loss=0.07332936948537827
+INFO:root:
+[Valid][8][UTT][ MSE = 0.1850 | LCC = 0.9704 | SRCC = 0.9699 ]
+INFO:root:Average loss=0.06465735644102097
+INFO:root:
+[Valid][9][UTT][ MSE = 0.1705 | LCC = 0.9703 | SRCC = 0.9697 ]
+INFO:root:Average loss=0.06295517334342003
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2091 | LCC = 0.9699 | SRCC = 0.9701 ]
+INFO:root:Average loss=0.06022040620446205
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1607 | LCC = 0.9707 | SRCC = 0.9700 ]
+INFO:root:Average loss=0.05581982740759849
+INFO:root:
+[Valid][12][UTT][ MSE = 0.1366 | LCC = 0.9705 | SRCC = 0.9702 ]
+INFO:root:Average loss=0.054982634544372556
+INFO:root:
+[Valid][13][UTT][ MSE = 0.1969 | LCC = 0.9704 | SRCC = 0.9707 ]
+INFO:root:Average loss=0.051424546033143996
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1768 | LCC = 0.9701 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.04961366702616215
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1573 | LCC = 0.9711 | SRCC = 0.9703 ]
+INFO:root:Average loss=0.04846008837223053
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1439 | LCC = 0.9701 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.04522567541897297
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1720 | LCC = 0.9714 | SRCC = 0.9705 ]
+INFO:root:Average loss=0.04323581247031689
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1728 | LCC = 0.9712 | SRCC = 0.9703 ]
+INFO:root:Average loss=0.0419912174642086
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1694 | LCC = 0.9702 | SRCC = 0.9700 ]
+INFO:root:Average loss=0.04146806475520134
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1639 | LCC = 0.9717 | SRCC = 0.9712 ]
+INFO:root:Average loss=0.03771034741401672
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1709 | LCC = 0.9705 | SRCC = 0.9699 ]
+INFO:root:Average loss=0.03704496978223324
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1590 | LCC = 0.9712 | SRCC = 0.9702 ]
+INFO:root:Average loss=0.03763911803066731
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1647 | LCC = 0.9711 | SRCC = 0.9701 ]
+INFO:root:Average loss=0.03491347037255764
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1566 | LCC = 0.9718 | SRCC = 0.9709 ]
+INFO:root:Average loss=0.031737336680293084
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1625 | LCC = 0.9718 | SRCC = 0.9708 ]
+INFO:root:Average loss=0.031511312022805216
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1631 | LCC = 0.9711 | SRCC = 0.9702 ]
+INFO:root:Average loss=0.031939387530088424
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1582 | LCC = 0.9720 | SRCC = 0.9710 ]
+INFO:root:Average loss=0.029642372965812683
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1690 | LCC = 0.9703 | SRCC = 0.9697 ]
+INFO:root:Average loss=0.029086005866527556
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1586 | LCC = 0.9709 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.02773505213856697
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1517 | LCC = 0.9713 | SRCC = 0.9701 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1552 | LCC = 0.9730 | SRCC = 0.9730 ]

results_tencent/w2v2_xlsr_1b/layer36/run3/config.gin

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

results_tencent/w2v2_xlsr_1b/layer36/run3/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2a1daa260328be8ea79aa3878e76e9315ed54ea79c913a35085a7d1693fae976
+size 4058699

results_tencent/w2v2_xlsr_1b/layer36/run3/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:198775b37aaf963a4c02686b1ab9cdbe38bcc96c40b482744bdd1855111404dc
+size 4060358

results_tencent/w2v2_xlsr_1b/layer36/run3/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 8000
+INFO:root:Num val speech clips: 2000
+INFO:root:Num test speech clips: 1563
+INFO:root:Device=cuda
+INFO:root:Average loss=0.4184150598049164
+INFO:root:
+[Valid][0][UTT][ MSE = 0.2139 | LCC = 0.9482 | SRCC = 0.9464 ]
+INFO:root:Average loss=0.14041236561536788
+INFO:root:
+[Valid][1][UTT][ MSE = 0.1773 | LCC = 0.9571 | SRCC = 0.9560 ]
+INFO:root:Average loss=0.11671010273694991
+INFO:root:
+[Valid][2][UTT][ MSE = 0.1658 | LCC = 0.9622 | SRCC = 0.9618 ]
+INFO:root:Average loss=0.10370067405700684
+INFO:root:
+[Valid][3][UTT][ MSE = 0.1990 | LCC = 0.9648 | SRCC = 0.9637 ]
+INFO:root:Average loss=0.09313969212770462
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2338 | LCC = 0.9672 | SRCC = 0.9661 ]
+INFO:root:Average loss=0.0937761844098568
+INFO:root:
+[Valid][5][UTT][ MSE = 0.1695 | LCC = 0.9682 | SRCC = 0.9672 ]
+INFO:root:Average loss=0.07943975445628167
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2063 | LCC = 0.9687 | SRCC = 0.9680 ]
+INFO:root:Average loss=0.07610993379354478
+INFO:root:
+[Valid][7][UTT][ MSE = 0.1912 | LCC = 0.9695 | SRCC = 0.9686 ]
+INFO:root:Average loss=0.07089036694169044
+INFO:root:
+[Valid][8][UTT][ MSE = 0.1854 | LCC = 0.9697 | SRCC = 0.9689 ]
+INFO:root:Average loss=0.06716691708564758
+INFO:root:
+[Valid][9][UTT][ MSE = 0.1834 | LCC = 0.9697 | SRCC = 0.9690 ]
+INFO:root:Average loss=0.06476069831848144
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2299 | LCC = 0.9705 | SRCC = 0.9693 ]
+INFO:root:Average loss=0.060296593576669694
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1845 | LCC = 0.9701 | SRCC = 0.9694 ]
+INFO:root:Average loss=0.056426486283540724
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2203 | LCC = 0.9705 | SRCC = 0.9697 ]
+INFO:root:Average loss=0.05369547292590141
+INFO:root:
+[Valid][13][UTT][ MSE = 0.1798 | LCC = 0.9708 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.05339738672971726
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1827 | LCC = 0.9716 | SRCC = 0.9705 ]
+INFO:root:Average loss=0.05019618168473244
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1734 | LCC = 0.9702 | SRCC = 0.9689 ]
+INFO:root:Average loss=0.047284602239727976
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1486 | LCC = 0.9685 | SRCC = 0.9694 ]
+INFO:root:Average loss=0.04673340328037739
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1482 | LCC = 0.9703 | SRCC = 0.9696 ]
+INFO:root:Average loss=0.043822299867868425
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1762 | LCC = 0.9699 | SRCC = 0.9686 ]
+INFO:root:Average loss=0.04166319189965725
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1655 | LCC = 0.9700 | SRCC = 0.9696 ]
+INFO:root:Average loss=0.0402757787257433
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1405 | LCC = 0.9697 | SRCC = 0.9692 ]
+INFO:root:Average loss=0.04029739107191563
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1739 | LCC = 0.9710 | SRCC = 0.9700 ]
+INFO:root:Average loss=0.03568760584294796
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1773 | LCC = 0.9706 | SRCC = 0.9697 ]
+INFO:root:Average loss=0.035622006043791773
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1587 | LCC = 0.9711 | SRCC = 0.9699 ]
+INFO:root:Average loss=0.03358796517550945
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1808 | LCC = 0.9689 | SRCC = 0.9685 ]
+INFO:root:Average loss=0.034545952156186104
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1915 | LCC = 0.9714 | SRCC = 0.9703 ]
+INFO:root:Average loss=0.034197140738368034
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1638 | LCC = 0.9717 | SRCC = 0.9703 ]
+INFO:root:Average loss=0.029682789251208305
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1876 | LCC = 0.9709 | SRCC = 0.9700 ]
+INFO:root:Average loss=0.030314480409026148
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1664 | LCC = 0.9712 | SRCC = 0.9700 ]
+INFO:root:Average loss=0.029289673045277596
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1840 | LCC = 0.9707 | SRCC = 0.9695 ]
+INFO:root:Average loss=0.028404358118772507
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1673 | LCC = 0.9722 | SRCC = 0.9707 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1632 | LCC = 0.9729 | SRCC = 0.9724 ]

results_tencent/w2v2_xlsr_1b/layer36/run4/config.gin

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

results_tencent/w2v2_xlsr_1b/layer36/run4/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d597d38f707dae6769e94de7e0e27f302bee8ca5363749ca5d635305208af238
+size 4058699

results_tencent/w2v2_xlsr_1b/layer36/run4/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:10027e9d8351b24020792ccfe100ddc34e9fb239262aa83b05002e1790cd383a
+size 4060358

results_tencent/w2v2_xlsr_1b/layer36/run4/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 8000
+INFO:root:Num val speech clips: 2000
+INFO:root:Num test speech clips: 1563
+INFO:root:Device=cuda
+INFO:root:Average loss=0.3466171278953552
+INFO:root:
+[Valid][0][UTT][ MSE = 0.2414 | LCC = 0.9474 | SRCC = 0.9473 ]
+INFO:root:Average loss=0.14407474410533905
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2357 | LCC = 0.9567 | SRCC = 0.9586 ]
+INFO:root:Average loss=0.12045472604036331
+INFO:root:
+[Valid][2][UTT][ MSE = 0.1806 | LCC = 0.9602 | SRCC = 0.9620 ]
+INFO:root:Average loss=0.10376573845744133
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2069 | LCC = 0.9658 | SRCC = 0.9656 ]
+INFO:root:Average loss=0.09214080440998078
+INFO:root:
+[Valid][4][UTT][ MSE = 0.1736 | LCC = 0.9660 | SRCC = 0.9656 ]
+INFO:root:Average loss=0.08704169976711273
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2007 | LCC = 0.9672 | SRCC = 0.9676 ]
+INFO:root:Average loss=0.08388708880543709
+INFO:root:
+[Valid][6][UTT][ MSE = 0.1643 | LCC = 0.9688 | SRCC = 0.9686 ]
+INFO:root:Average loss=0.07602933475375176
+INFO:root:
+[Valid][7][UTT][ MSE = 0.1791 | LCC = 0.9682 | SRCC = 0.9677 ]
+INFO:root:Average loss=0.06806914582848549
+INFO:root:
+[Valid][8][UTT][ MSE = 0.1524 | LCC = 0.9688 | SRCC = 0.9690 ]
+INFO:root:Average loss=0.0666362652182579
+INFO:root:
+[Valid][9][UTT][ MSE = 0.1515 | LCC = 0.9693 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.06287260290980339
+INFO:root:
+[Valid][10][UTT][ MSE = 0.1546 | LCC = 0.9701 | SRCC = 0.9695 ]
+INFO:root:Average loss=0.059920053124427795
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1474 | LCC = 0.9698 | SRCC = 0.9701 ]
+INFO:root:Average loss=0.05628760279715061
+INFO:root:
+[Valid][12][UTT][ MSE = 0.1654 | LCC = 0.9707 | SRCC = 0.9703 ]
+INFO:root:Average loss=0.05284775601327419
+INFO:root:
+[Valid][13][UTT][ MSE = 0.1566 | LCC = 0.9706 | SRCC = 0.9691 ]
+INFO:root:Average loss=0.05544387456774712
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1744 | LCC = 0.9709 | SRCC = 0.9700 ]
+INFO:root:Average loss=0.04981437346339226
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1670 | LCC = 0.9708 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.04602447083592415
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1512 | LCC = 0.9711 | SRCC = 0.9700 ]
+INFO:root:Average loss=0.04670155562460423
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1457 | LCC = 0.9710 | SRCC = 0.9705 ]
+INFO:root:Average loss=0.042255980134010315
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1756 | LCC = 0.9702 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.042576082974672316
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1703 | LCC = 0.9709 | SRCC = 0.9700 ]
+INFO:root:Average loss=0.040510559529066084
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1662 | LCC = 0.9713 | SRCC = 0.9704 ]
+INFO:root:Average loss=0.04114079168438912
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1631 | LCC = 0.9718 | SRCC = 0.9708 ]
+INFO:root:Average loss=0.03683482539653778
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1593 | LCC = 0.9716 | SRCC = 0.9704 ]
+INFO:root:Average loss=0.03525456528365612
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1514 | LCC = 0.9714 | SRCC = 0.9707 ]
+INFO:root:Average loss=0.033437778785824775
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1437 | LCC = 0.9712 | SRCC = 0.9703 ]
+INFO:root:Average loss=0.03135680602490902
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1862 | LCC = 0.9706 | SRCC = 0.9702 ]
+INFO:root:Average loss=0.031429137006402014
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1700 | LCC = 0.9715 | SRCC = 0.9704 ]
+INFO:root:Average loss=0.031810793444514276
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1778 | LCC = 0.9701 | SRCC = 0.9700 ]
+INFO:root:Average loss=0.0291527586132288
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1456 | LCC = 0.9713 | SRCC = 0.9701 ]
+INFO:root:Average loss=0.02701096299290657
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1556 | LCC = 0.9709 | SRCC = 0.9697 ]
+INFO:root:Average loss=0.027985470943152905
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1645 | LCC = 0.9710 | SRCC = 0.9702 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1614 | LCC = 0.9726 | SRCC = 0.9724 ]

results_tencent/w2v2_xlsr_1b/layer38/run0/config.gin

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

results_tencent/w2v2_xlsr_1b/layer38/run0/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:59928d25e2a3be9bf81fd1236b3641005b5dd820dccb03732cfbc045f300f84d
+size 4058699

results_tencent/w2v2_xlsr_1b/layer38/run0/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ced1892b6897f7d7e1b781a22ada4790568512a16c2293465bca3e6bd9201d66
+size 4060358

results_tencent/w2v2_xlsr_1b/layer38/run0/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 8000
+INFO:root:Num val speech clips: 2000
+INFO:root:Num test speech clips: 1563
+INFO:root:Device=cuda
+INFO:root:Average loss=0.3558842204809189
+INFO:root:
+[Valid][0][UTT][ MSE = 0.2582 | LCC = 0.9480 | SRCC = 0.9488 ]
+INFO:root:Average loss=0.14679123330116273
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2523 | LCC = 0.9563 | SRCC = 0.9584 ]
+INFO:root:Average loss=0.11954030108451844
+INFO:root:
+[Valid][2][UTT][ MSE = 0.1919 | LCC = 0.9626 | SRCC = 0.9623 ]
+INFO:root:Average loss=0.10363795626163483
+INFO:root:
+[Valid][3][UTT][ MSE = 0.1817 | LCC = 0.9649 | SRCC = 0.9646 ]
+INFO:root:Average loss=0.09424833685159684
+INFO:root:
+[Valid][4][UTT][ MSE = 0.1755 | LCC = 0.9645 | SRCC = 0.9636 ]
+INFO:root:Average loss=0.08609261974692345
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2040 | LCC = 0.9672 | SRCC = 0.9670 ]
+INFO:root:Average loss=0.07944248157739639
+INFO:root:
+[Valid][6][UTT][ MSE = 0.1761 | LCC = 0.9680 | SRCC = 0.9665 ]
+INFO:root:Average loss=0.07284740918874741
+INFO:root:
+[Valid][7][UTT][ MSE = 0.1894 | LCC = 0.9683 | SRCC = 0.9677 ]
+INFO:root:Average loss=0.07072249266505241
+INFO:root:
+[Valid][8][UTT][ MSE = 0.1508 | LCC = 0.9686 | SRCC = 0.9686 ]
+INFO:root:Average loss=0.06409601128101348
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2092 | LCC = 0.9695 | SRCC = 0.9692 ]
+INFO:root:Average loss=0.06285435909032822
+INFO:root:
+[Valid][10][UTT][ MSE = 0.1575 | LCC = 0.9693 | SRCC = 0.9693 ]
+INFO:root:Average loss=0.061157888621091845
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1841 | LCC = 0.9699 | SRCC = 0.9692 ]
+INFO:root:Average loss=0.05656137952208519
+INFO:root:
+[Valid][12][UTT][ MSE = 0.1888 | LCC = 0.9706 | SRCC = 0.9696 ]
+INFO:root:Average loss=0.058424759060144424
+INFO:root:
+[Valid][13][UTT][ MSE = 0.1548 | LCC = 0.9707 | SRCC = 0.9702 ]
+INFO:root:Average loss=0.05241496953368187
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1711 | LCC = 0.9710 | SRCC = 0.9703 ]
+INFO:root:Average loss=0.049481783553957936
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1939 | LCC = 0.9698 | SRCC = 0.9697 ]
+INFO:root:Average loss=0.048216026574373246
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1747 | LCC = 0.9706 | SRCC = 0.9695 ]
+INFO:root:Average loss=0.04518214210867882
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1487 | LCC = 0.9709 | SRCC = 0.9701 ]
+INFO:root:Average loss=0.04323844361305237
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1860 | LCC = 0.9702 | SRCC = 0.9696 ]
+INFO:root:Average loss=0.04365216052532196
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1851 | LCC = 0.9707 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.0375362349152565
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1776 | LCC = 0.9707 | SRCC = 0.9700 ]
+INFO:root:Average loss=0.03631195691227913
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1786 | LCC = 0.9704 | SRCC = 0.9694 ]
+INFO:root:Average loss=0.038143286049365996
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1639 | LCC = 0.9709 | SRCC = 0.9697 ]
+INFO:root:Average loss=0.03450111666321754
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1653 | LCC = 0.9704 | SRCC = 0.9692 ]
+INFO:root:Average loss=0.037062941417098046
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1852 | LCC = 0.9705 | SRCC = 0.9689 ]
+INFO:root:Average loss=0.033990324556827545
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1829 | LCC = 0.9709 | SRCC = 0.9704 ]
+INFO:root:Average loss=0.029204134047031403
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1677 | LCC = 0.9708 | SRCC = 0.9700 ]
+INFO:root:Average loss=0.029512074708938598
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1497 | LCC = 0.9712 | SRCC = 0.9700 ]
+INFO:root:Average loss=0.032467821285128595
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1717 | LCC = 0.9704 | SRCC = 0.9695 ]
+INFO:root:Average loss=0.02685604488104582
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1542 | LCC = 0.9717 | SRCC = 0.9706 ]
+INFO:root:Average loss=0.028194790333509447
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1723 | LCC = 0.9711 | SRCC = 0.9701 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1474 | LCC = 0.9729 | SRCC = 0.9729 ]

results_tencent/w2v2_xlsr_1b/layer38/run1/config.gin

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

results_tencent/w2v2_xlsr_1b/layer38/run1/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d2dec282890a96175b1e01b2b2bd475b71821d654a8bec018850fb24b131c066
+size 4058699

results_tencent/w2v2_xlsr_1b/layer38/run1/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c4da0081cdcc580481425017d18abae339e09ecc6bdf13b17abb679199cecd8
+size 4060358

results_tencent/w2v2_xlsr_1b/layer38/run1/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 8000
+INFO:root:Num val speech clips: 2000
+INFO:root:Num test speech clips: 1563
+INFO:root:Device=cuda
+INFO:root:Average loss=0.3856969577074051
+INFO:root:
+[Valid][0][UTT][ MSE = 0.2479 | LCC = 0.9503 | SRCC = 0.9500 ]
+INFO:root:Average loss=0.13848739260435106
+INFO:root:
+[Valid][1][UTT][ MSE = 0.1794 | LCC = 0.9602 | SRCC = 0.9592 ]
+INFO:root:Average loss=0.10929290664196015
+INFO:root:
+[Valid][2][UTT][ MSE = 0.1833 | LCC = 0.9641 | SRCC = 0.9630 ]
+INFO:root:Average loss=0.10242943039536476
+INFO:root:
+[Valid][3][UTT][ MSE = 0.1645 | LCC = 0.9660 | SRCC = 0.9654 ]
+INFO:root:Average loss=0.09316253691911697
+INFO:root:
+[Valid][4][UTT][ MSE = 0.1752 | LCC = 0.9672 | SRCC = 0.9666 ]
+INFO:root:Average loss=0.0799498461484909
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2254 | LCC = 0.9667 | SRCC = 0.9666 ]
+INFO:root:Average loss=0.07767478197813034
+INFO:root:
+[Valid][6][UTT][ MSE = 0.1577 | LCC = 0.9688 | SRCC = 0.9680 ]
+INFO:root:Average loss=0.07336836779117584
+INFO:root:
+[Valid][7][UTT][ MSE = 0.1594 | LCC = 0.9697 | SRCC = 0.9687 ]
+INFO:root:Average loss=0.06617289617657661
+INFO:root:
+[Valid][8][UTT][ MSE = 0.1471 | LCC = 0.9701 | SRCC = 0.9691 ]
+INFO:root:Average loss=0.06345405882596969
+INFO:root:
+[Valid][9][UTT][ MSE = 0.1835 | LCC = 0.9695 | SRCC = 0.9691 ]
+INFO:root:Average loss=0.06454156318306922
+INFO:root:
+[Valid][10][UTT][ MSE = 0.1850 | LCC = 0.9709 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.05887997670471668
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1436 | LCC = 0.9707 | SRCC = 0.9697 ]
+INFO:root:Average loss=0.055705289766192434
+INFO:root:
+[Valid][12][UTT][ MSE = 0.1428 | LCC = 0.9711 | SRCC = 0.9701 ]
+INFO:root:Average loss=0.054259054869413376
+INFO:root:
+[Valid][13][UTT][ MSE = 0.1486 | LCC = 0.9687 | SRCC = 0.9679 ]
+INFO:root:Average loss=0.05171012181043625
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1434 | LCC = 0.9708 | SRCC = 0.9697 ]
+INFO:root:Average loss=0.04836002166569233
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1379 | LCC = 0.9702 | SRCC = 0.9692 ]
+INFO:root:Average loss=0.043975448697805404
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1540 | LCC = 0.9695 | SRCC = 0.9689 ]
+INFO:root:Average loss=0.04366652525961399
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1393 | LCC = 0.9676 | SRCC = 0.9677 ]
+INFO:root:Average loss=0.04300688318908215
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1425 | LCC = 0.9714 | SRCC = 0.9701 ]
+INFO:root:Average loss=0.04092701049149036
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1449 | LCC = 0.9709 | SRCC = 0.9696 ]
+INFO:root:Average loss=0.0406169667840004
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1398 | LCC = 0.9712 | SRCC = 0.9700 ]
+INFO:root:Average loss=0.03672604832053185
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1507 | LCC = 0.9715 | SRCC = 0.9702 ]
+INFO:root:Average loss=0.03712812514603138
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1393 | LCC = 0.9707 | SRCC = 0.9694 ]
+INFO:root:Average loss=0.034318795174360275
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1472 | LCC = 0.9689 | SRCC = 0.9684 ]
+INFO:root:Average loss=0.03358851247280836
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1424 | LCC = 0.9711 | SRCC = 0.9701 ]
+INFO:root:Average loss=0.034150296851992605
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1479 | LCC = 0.9713 | SRCC = 0.9700 ]
+INFO:root:Average loss=0.031058604896068574
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1514 | LCC = 0.9716 | SRCC = 0.9701 ]
+INFO:root:Average loss=0.02965648752450943
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1677 | LCC = 0.9714 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.029592418894171715
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1440 | LCC = 0.9702 | SRCC = 0.9693 ]
+INFO:root:Average loss=0.03005637326091528
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1457 | LCC = 0.9709 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.027972604781389236
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1509 | LCC = 0.9710 | SRCC = 0.9698 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1447 | LCC = 0.9735 | SRCC = 0.9734 ]

results_tencent/w2v2_xlsr_1b/layer38/run2/config.gin

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

results_tencent/w2v2_xlsr_1b/layer38/run2/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:58451ff715e8da175bb5def734bb6524daa0a2b33207923442199818a9fb8a45
+size 4058699

results_tencent/w2v2_xlsr_1b/layer38/run2/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:02995df1b59741c46dda3cd9bf92282b6e621aa291ef3bd66798961efee5de1d
+size 4060358

results_tencent/w2v2_xlsr_1b/layer38/run2/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 8000
+INFO:root:Num val speech clips: 2000
+INFO:root:Num test speech clips: 1563
+INFO:root:Device=cuda
+INFO:root:Average loss=0.3794033288955688
+INFO:root:
+[Valid][0][UTT][ MSE = 0.3109 | LCC = 0.9459 | SRCC = 0.9455 ]
+INFO:root:Average loss=0.1416382457613945
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2201 | LCC = 0.9597 | SRCC = 0.9579 ]
+INFO:root:Average loss=0.11196582877635956
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2155 | LCC = 0.9634 | SRCC = 0.9624 ]
+INFO:root:Average loss=0.09794212204217911
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2410 | LCC = 0.9653 | SRCC = 0.9641 ]
+INFO:root:Average loss=0.09182572892308236
+INFO:root:
+[Valid][4][UTT][ MSE = 0.1583 | LCC = 0.9654 | SRCC = 0.9647 ]
+INFO:root:Average loss=0.0865036818087101
+INFO:root:
+[Valid][5][UTT][ MSE = 0.1957 | LCC = 0.9677 | SRCC = 0.9661 ]
+INFO:root:Average loss=0.07724099358916282
+INFO:root:
+[Valid][6][UTT][ MSE = 0.1629 | LCC = 0.9681 | SRCC = 0.9669 ]
+INFO:root:Average loss=0.07598049944639206
+INFO:root:
+[Valid][7][UTT][ MSE = 0.1692 | LCC = 0.9670 | SRCC = 0.9665 ]
+INFO:root:Average loss=0.0742725749015808
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2247 | LCC = 0.9690 | SRCC = 0.9682 ]
+INFO:root:Average loss=0.06504936426877976
+INFO:root:
+[Valid][9][UTT][ MSE = 0.1646 | LCC = 0.9698 | SRCC = 0.9687 ]
+INFO:root:Average loss=0.06173207664489746
+INFO:root:
+[Valid][10][UTT][ MSE = 0.1995 | LCC = 0.9701 | SRCC = 0.9691 ]
+INFO:root:Average loss=0.05952079963684082
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1696 | LCC = 0.9701 | SRCC = 0.9691 ]
+INFO:root:Average loss=0.05654339304566383
+INFO:root:
+[Valid][12][UTT][ MSE = 0.1872 | LCC = 0.9689 | SRCC = 0.9683 ]
+INFO:root:Average loss=0.05348136024177074
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2044 | LCC = 0.9694 | SRCC = 0.9690 ]
+INFO:root:Average loss=0.05224865095317364
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1704 | LCC = 0.9705 | SRCC = 0.9694 ]
+INFO:root:Average loss=0.04908099475502968
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2130 | LCC = 0.9697 | SRCC = 0.9691 ]
+INFO:root:Average loss=0.04488840971887112
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1875 | LCC = 0.9696 | SRCC = 0.9690 ]
+INFO:root:Average loss=0.04518809826672077
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1737 | LCC = 0.9697 | SRCC = 0.9681 ]
+INFO:root:Average loss=0.04163896660506725
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1696 | LCC = 0.9705 | SRCC = 0.9690 ]
+INFO:root:Average loss=0.03920104588568211
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1682 | LCC = 0.9708 | SRCC = 0.9697 ]
+INFO:root:Average loss=0.040962110802531246
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1575 | LCC = 0.9701 | SRCC = 0.9690 ]
+INFO:root:Average loss=0.03647107282280922
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1839 | LCC = 0.9709 | SRCC = 0.9696 ]
+INFO:root:Average loss=0.0365851811170578
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1741 | LCC = 0.9705 | SRCC = 0.9689 ]
+INFO:root:Average loss=0.036303778678178786
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1897 | LCC = 0.9703 | SRCC = 0.9688 ]
+INFO:root:Average loss=0.03547481192648411
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1538 | LCC = 0.9692 | SRCC = 0.9681 ]
+INFO:root:Average loss=0.0331257444024086
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1548 | LCC = 0.9703 | SRCC = 0.9689 ]
+INFO:root:Average loss=0.031102280110120773
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1731 | LCC = 0.9708 | SRCC = 0.9693 ]
+INFO:root:Average loss=0.03095718766748905
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1862 | LCC = 0.9702 | SRCC = 0.9690 ]
+INFO:root:Average loss=0.03009469159692526
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1750 | LCC = 0.9702 | SRCC = 0.9685 ]
+INFO:root:Average loss=0.02784664736688137
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1705 | LCC = 0.9708 | SRCC = 0.9692 ]
+INFO:root:Average loss=0.02976624257862568
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1592 | LCC = 0.9712 | SRCC = 0.9698 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1504 | LCC = 0.9738 | SRCC = 0.9734 ]

results_tencent/w2v2_xlsr_1b/layer38/run3/config.gin

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

results_tencent/w2v2_xlsr_1b/layer38/run3/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da611060cbeb0c9bf5b0c4f53131c2d9a0ba4609631ec8f1fe82016d693719e3
+size 4058699

results_tencent/w2v2_xlsr_1b/layer38/run3/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:16339bda6cf6b0049d8fa4ee56fc867857126e106323f197385d62ff5f317bc3
+size 4060358

results_tencent/w2v2_xlsr_1b/layer38/run3/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 8000
+INFO:root:Num val speech clips: 2000
+INFO:root:Num test speech clips: 1563
+INFO:root:Device=cuda
+INFO:root:Average loss=0.39518193674087526
+INFO:root:
+[Valid][0][UTT][ MSE = 0.2611 | LCC = 0.9497 | SRCC = 0.9502 ]
+INFO:root:Average loss=0.1414701886177063
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2556 | LCC = 0.9602 | SRCC = 0.9603 ]
+INFO:root:Average loss=0.1182013139128685
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2213 | LCC = 0.9631 | SRCC = 0.9625 ]
+INFO:root:Average loss=0.10103740113973618
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2178 | LCC = 0.9652 | SRCC = 0.9645 ]
+INFO:root:Average loss=0.08940617421269417
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2279 | LCC = 0.9663 | SRCC = 0.9657 ]
+INFO:root:Average loss=0.0874656894505024
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2144 | LCC = 0.9681 | SRCC = 0.9676 ]
+INFO:root:Average loss=0.07892866724729539
+INFO:root:
+[Valid][6][UTT][ MSE = 0.1938 | LCC = 0.9685 | SRCC = 0.9680 ]
+INFO:root:Average loss=0.07517336672544479
+INFO:root:
+[Valid][7][UTT][ MSE = 0.1862 | LCC = 0.9678 | SRCC = 0.9676 ]
+INFO:root:Average loss=0.07207538169622421
+INFO:root:
+[Valid][8][UTT][ MSE = 0.1944 | LCC = 0.9685 | SRCC = 0.9679 ]
+INFO:root:Average loss=0.06571751376986504
+INFO:root:
+[Valid][9][UTT][ MSE = 0.1810 | LCC = 0.9694 | SRCC = 0.9682 ]
+INFO:root:Average loss=0.06114544069766998
+INFO:root:
+[Valid][10][UTT][ MSE = 0.1924 | LCC = 0.9697 | SRCC = 0.9688 ]
+INFO:root:Average loss=0.05673066011071205
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1548 | LCC = 0.9692 | SRCC = 0.9690 ]
+INFO:root:Average loss=0.05721793439984322
+INFO:root:
+[Valid][12][UTT][ MSE = 0.1835 | LCC = 0.9699 | SRCC = 0.9693 ]
+INFO:root:Average loss=0.054944389790296554
+INFO:root:
+[Valid][13][UTT][ MSE = 0.1819 | LCC = 0.9698 | SRCC = 0.9692 ]
+INFO:root:Average loss=0.049349200055003165
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1681 | LCC = 0.9702 | SRCC = 0.9693 ]
+INFO:root:Average loss=0.04871702240407467
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1893 | LCC = 0.9701 | SRCC = 0.9696 ]
+INFO:root:Average loss=0.04712915739417076
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1754 | LCC = 0.9701 | SRCC = 0.9691 ]
+INFO:root:Average loss=0.04491954132914543
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1957 | LCC = 0.9698 | SRCC = 0.9696 ]
+INFO:root:Average loss=0.04172478725016117
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1759 | LCC = 0.9706 | SRCC = 0.9696 ]
+INFO:root:Average loss=0.0407626311480999
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1584 | LCC = 0.9700 | SRCC = 0.9692 ]
+INFO:root:Average loss=0.038834272131323816
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1638 | LCC = 0.9706 | SRCC = 0.9694 ]
+INFO:root:Average loss=0.037612869411706926
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1761 | LCC = 0.9707 | SRCC = 0.9697 ]
+INFO:root:Average loss=0.035039506554603576
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1754 | LCC = 0.9705 | SRCC = 0.9696 ]
+INFO:root:Average loss=0.03711528511345386
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1859 | LCC = 0.9707 | SRCC = 0.9695 ]
+INFO:root:Average loss=0.03994645991921425
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1663 | LCC = 0.9704 | SRCC = 0.9690 ]
+INFO:root:Average loss=0.03346462886035442
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1829 | LCC = 0.9706 | SRCC = 0.9699 ]
+INFO:root:Average loss=0.031186686784029007
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1576 | LCC = 0.9704 | SRCC = 0.9691 ]
+INFO:root:Average loss=0.03445906959474087
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1645 | LCC = 0.9709 | SRCC = 0.9696 ]
+INFO:root:Average loss=0.030764516606926917
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1582 | LCC = 0.9708 | SRCC = 0.9693 ]
+INFO:root:Average loss=0.028681699320673944
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1540 | LCC = 0.9695 | SRCC = 0.9689 ]
+INFO:root:Average loss=0.02761603620648384
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1727 | LCC = 0.9702 | SRCC = 0.9688 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1602 | LCC = 0.9725 | SRCC = 0.9716 ]