Upload SSL Layer MOS checkpoints (part 20)

results_tencent/w2v2_xlsr_300m/layer23/run1/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9deed34687949bbdffbdbca2eaf80b31ca9490445ba34ab25b807c663d058f47
+size 3405062

results_tencent/w2v2_xlsr_300m/layer23/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=1.0563081212043761
+INFO:root:
+[Valid][0][UTT][ MSE = 0.7386 | LCC = 0.8590 | SRCC = 0.8606 ]
+INFO:root:Average loss=0.43636824917793277
+INFO:root:
+[Valid][1][UTT][ MSE = 0.4267 | LCC = 0.8996 | SRCC = 0.8935 ]
+INFO:root:Average loss=0.3665706579685211
+INFO:root:
+[Valid][2][UTT][ MSE = 0.5825 | LCC = 0.8866 | SRCC = 0.9005 ]
+INFO:root:Average loss=0.3331493958234787
+INFO:root:
+[Valid][3][UTT][ MSE = 0.3987 | LCC = 0.9092 | SRCC = 0.9098 ]
+INFO:root:Average loss=0.2601349173784256
+INFO:root:
+[Valid][4][UTT][ MSE = 0.3710 | LCC = 0.9199 | SRCC = 0.9199 ]
+INFO:root:Average loss=0.2473576443195343
+INFO:root:
+[Valid][5][UTT][ MSE = 0.3600 | LCC = 0.9305 | SRCC = 0.9290 ]
+INFO:root:Average loss=0.2470013654232025
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2677 | LCC = 0.9316 | SRCC = 0.9299 ]
+INFO:root:Average loss=0.2507935168147087
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2492 | LCC = 0.9306 | SRCC = 0.9307 ]
+INFO:root:Average loss=0.19358640390634538
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2317 | LCC = 0.9388 | SRCC = 0.9382 ]
+INFO:root:Average loss=0.18437114834785462
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2824 | LCC = 0.9446 | SRCC = 0.9433 ]
+INFO:root:Average loss=0.2348708925843239
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2946 | LCC = 0.9454 | SRCC = 0.9464 ]
+INFO:root:Average loss=0.17242506766319274
+INFO:root:
+[Valid][11][UTT][ MSE = 0.2381 | LCC = 0.9443 | SRCC = 0.9461 ]
+INFO:root:Average loss=0.1622229313850403
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2086 | LCC = 0.9495 | SRCC = 0.9496 ]
+INFO:root:Average loss=0.1574323083758354
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2053 | LCC = 0.9508 | SRCC = 0.9516 ]
+INFO:root:Average loss=0.15850027018785476
+INFO:root:
+[Valid][14][UTT][ MSE = 0.3115 | LCC = 0.9474 | SRCC = 0.9513 ]
+INFO:root:Average loss=0.15702237820625306
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2225 | LCC = 0.9545 | SRCC = 0.9541 ]
+INFO:root:Average loss=0.1442300369143486
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1936 | LCC = 0.9506 | SRCC = 0.9516 ]
+INFO:root:Average loss=0.13388900846242904
+INFO:root:
+[Valid][17][UTT][ MSE = 0.2077 | LCC = 0.9542 | SRCC = 0.9550 ]
+INFO:root:Average loss=0.13456004190444945
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1951 | LCC = 0.9569 | SRCC = 0.9564 ]
+INFO:root:Average loss=0.12940751343965531
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1985 | LCC = 0.9515 | SRCC = 0.9543 ]
+INFO:root:Average loss=0.12696247392892837
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2164 | LCC = 0.9575 | SRCC = 0.9577 ]
+INFO:root:Average loss=0.12185577318072319
+INFO:root:
+[Valid][21][UTT][ MSE = 0.2530 | LCC = 0.9567 | SRCC = 0.9574 ]
+INFO:root:Average loss=0.11650395280122756
+INFO:root:
+[Valid][22][UTT][ MSE = 0.2103 | LCC = 0.9586 | SRCC = 0.9585 ]
+INFO:root:Average loss=0.12136834174394608
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2055 | LCC = 0.9589 | SRCC = 0.9581 ]
+INFO:root:Average loss=0.1141653277873993
+INFO:root:
+[Valid][24][UTT][ MSE = 0.2030 | LCC = 0.9571 | SRCC = 0.9574 ]
+INFO:root:Average loss=0.11773455655574798
+INFO:root:
+[Valid][25][UTT][ MSE = 0.2126 | LCC = 0.9608 | SRCC = 0.9602 ]
+INFO:root:Average loss=0.12199297505617142
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2463 | LCC = 0.9562 | SRCC = 0.9571 ]
+INFO:root:Average loss=0.10990665045380592
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1957 | LCC = 0.9615 | SRCC = 0.9606 ]
+INFO:root:Average loss=0.10368599903583527
+INFO:root:
+[Valid][28][UTT][ MSE = 0.2125 | LCC = 0.9611 | SRCC = 0.9612 ]
+INFO:root:Average loss=0.10295946589112281
+INFO:root:
+[Valid][29][UTT][ MSE = 0.2154 | LCC = 0.9609 | SRCC = 0.9608 ]
+INFO:root:Average loss=0.10441702657938004
+INFO:root:
+[Valid][30][UTT][ MSE = 0.2320 | LCC = 0.9566 | SRCC = 0.9582 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2017 | LCC = 0.9608 | SRCC = 0.9615 ]

results_tencent/w2v2_xlsr_300m/layer23/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_300m/layer23/run2/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b593f01a812561d51caac313bc5c6bd12401973794f032e49f13e1f2fe69ed9
+size 3403339

results_tencent/w2v2_xlsr_300m/layer23/run2/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0e7f331b68e09a285786ae12faee10f37f40d1ec8e825a5af4f258d78e42f59a
+size 3405062

results_tencent/w2v2_xlsr_300m/layer23/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=1.2454368045330049
+INFO:root:
+[Valid][0][UTT][ MSE = 0.9272 | LCC = 0.7573 | SRCC = 0.7894 ]
+INFO:root:Average loss=0.4070986392498016
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3697 | LCC = 0.8888 | SRCC = 0.8883 ]
+INFO:root:Average loss=0.3334553905725479
+INFO:root:
+[Valid][2][UTT][ MSE = 0.5089 | LCC = 0.8980 | SRCC = 0.9068 ]
+INFO:root:Average loss=0.2786888217926025
+INFO:root:
+[Valid][3][UTT][ MSE = 0.3937 | LCC = 0.9223 | SRCC = 0.9196 ]
+INFO:root:Average loss=0.2641449633836746
+INFO:root:
+[Valid][4][UTT][ MSE = 0.3269 | LCC = 0.9274 | SRCC = 0.9255 ]
+INFO:root:Average loss=0.2518278401494026
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2548 | LCC = 0.9168 | SRCC = 0.9214 ]
+INFO:root:Average loss=0.23646544587612153
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2542 | LCC = 0.9330 | SRCC = 0.9331 ]
+INFO:root:Average loss=0.2051917288303375
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2680 | LCC = 0.9354 | SRCC = 0.9349 ]
+INFO:root:Average loss=0.2318065767288208
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2200 | LCC = 0.9394 | SRCC = 0.9400 ]
+INFO:root:Average loss=0.18956960862874986
+INFO:root:
+[Valid][9][UTT][ MSE = 0.3810 | LCC = 0.9370 | SRCC = 0.9452 ]
+INFO:root:Average loss=0.20585258382558821
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2534 | LCC = 0.9421 | SRCC = 0.9434 ]
+INFO:root:Average loss=0.18163397318124772
+INFO:root:
+[Valid][11][UTT][ MSE = 0.2232 | LCC = 0.9471 | SRCC = 0.9473 ]
+INFO:root:Average loss=0.1720084508061409
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2954 | LCC = 0.9487 | SRCC = 0.9500 ]
+INFO:root:Average loss=0.14915228682756423
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2298 | LCC = 0.9519 | SRCC = 0.9517 ]
+INFO:root:Average loss=0.14277833980321883
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2364 | LCC = 0.9460 | SRCC = 0.9457 ]
+INFO:root:Average loss=0.1492177627682686
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2449 | LCC = 0.9534 | SRCC = 0.9544 ]
+INFO:root:Average loss=0.13858777248859405
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2090 | LCC = 0.9544 | SRCC = 0.9542 ]
+INFO:root:Average loss=0.14254310512542726
+INFO:root:
+[Valid][17][UTT][ MSE = 0.2187 | LCC = 0.9552 | SRCC = 0.9554 ]
+INFO:root:Average loss=0.13297844725847244
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1940 | LCC = 0.9534 | SRCC = 0.9547 ]
+INFO:root:Average loss=0.12449122136831284
+INFO:root:
+[Valid][19][UTT][ MSE = 0.2208 | LCC = 0.9564 | SRCC = 0.9567 ]
+INFO:root:Average loss=0.12975050085783005
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2192 | LCC = 0.9539 | SRCC = 0.9550 ]
+INFO:root:Average loss=0.12197551834583283
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1840 | LCC = 0.9570 | SRCC = 0.9582 ]
+INFO:root:Average loss=0.12703324615955353
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1805 | LCC = 0.9524 | SRCC = 0.9542 ]
+INFO:root:Average loss=0.12633594971895218
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2874 | LCC = 0.9577 | SRCC = 0.9599 ]
+INFO:root:Average loss=0.11747282934188842
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1975 | LCC = 0.9587 | SRCC = 0.9593 ]
+INFO:root:Average loss=0.11518598806858063
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1968 | LCC = 0.9603 | SRCC = 0.9604 ]
+INFO:root:Average loss=0.1197339205443859
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2631 | LCC = 0.9585 | SRCC = 0.9606 ]
+INFO:root:Average loss=0.10964244675636292
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1846 | LCC = 0.9610 | SRCC = 0.9612 ]
+INFO:root:Average loss=0.11100984755158425
+INFO:root:
+[Valid][28][UTT][ MSE = 0.2261 | LCC = 0.9599 | SRCC = 0.9605 ]
+INFO:root:Average loss=0.10690061169862747
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1774 | LCC = 0.9595 | SRCC = 0.9594 ]
+INFO:root:Average loss=0.1042147558927536
+INFO:root:
+[Valid][30][UTT][ MSE = 0.2075 | LCC = 0.9612 | SRCC = 0.9615 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2153 | LCC = 0.9599 | SRCC = 0.9620 ]

results_tencent/w2v2_xlsr_300m/layer23/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_300m/layer23/run3/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bb453cf8d22c715acdbfc7a44d78c26f4604b6ef363206f5c7b9125be580437a
+size 3403339

results_tencent/w2v2_xlsr_300m/layer23/run3/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b2e395f1a1a88a57c0b9762c97f53edde8094ad76a76e45587acef062a39651b
+size 3404998

results_tencent/w2v2_xlsr_300m/layer23/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=1.0217993781566619
+INFO:root:
+[Valid][0][UTT][ MSE = 1.0591 | LCC = 0.8105 | SRCC = 0.8568 ]
+INFO:root:Average loss=0.4108892401456833
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3454 | LCC = 0.8973 | SRCC = 0.8937 ]
+INFO:root:Average loss=0.2908277262449265
+INFO:root:
+[Valid][2][UTT][ MSE = 0.4217 | LCC = 0.9067 | SRCC = 0.9095 ]
+INFO:root:Average loss=0.3161389358043671
+INFO:root:
+[Valid][3][UTT][ MSE = 0.4445 | LCC = 0.9103 | SRCC = 0.9153 ]
+INFO:root:Average loss=0.30991377222537997
+INFO:root:
+[Valid][4][UTT][ MSE = 0.3158 | LCC = 0.9249 | SRCC = 0.9241 ]
+INFO:root:Average loss=0.255293332695961
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2536 | LCC = 0.9333 | SRCC = 0.9319 ]
+INFO:root:Average loss=0.21495477044582367
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2286 | LCC = 0.9373 | SRCC = 0.9361 ]
+INFO:root:Average loss=0.19873052734136581
+INFO:root:
+[Valid][7][UTT][ MSE = 0.3152 | LCC = 0.9417 | SRCC = 0.9403 ]
+INFO:root:Average loss=0.18445929199457167
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2543 | LCC = 0.9417 | SRCC = 0.9419 ]
+INFO:root:Average loss=0.18127670788764955
+INFO:root:
+[Valid][9][UTT][ MSE = 0.3192 | LCC = 0.9437 | SRCC = 0.9436 ]
+INFO:root:Average loss=0.1928513303399086
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2450 | LCC = 0.9396 | SRCC = 0.9439 ]
+INFO:root:Average loss=0.16134409826993942
+INFO:root:
+[Valid][11][UTT][ MSE = 0.2407 | LCC = 0.9483 | SRCC = 0.9472 ]
+INFO:root:Average loss=0.17079754358530044
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2452 | LCC = 0.9476 | SRCC = 0.9474 ]
+INFO:root:Average loss=0.15236402332782745
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2353 | LCC = 0.9485 | SRCC = 0.9489 ]
+INFO:root:Average loss=0.14256081676483154
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1954 | LCC = 0.9518 | SRCC = 0.9512 ]
+INFO:root:Average loss=0.1471487997174263
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1737 | LCC = 0.9513 | SRCC = 0.9529 ]
+INFO:root:Average loss=0.1566015036702156
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2357 | LCC = 0.9533 | SRCC = 0.9523 ]
+INFO:root:Average loss=0.13364748448133468
+INFO:root:
+[Valid][17][UTT][ MSE = 0.2799 | LCC = 0.9519 | SRCC = 0.9543 ]
+INFO:root:Average loss=0.15910709178447724
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2889 | LCC = 0.9459 | SRCC = 0.9503 ]
+INFO:root:Average loss=0.14192144161462783
+INFO:root:
+[Valid][19][UTT][ MSE = 0.2777 | LCC = 0.9451 | SRCC = 0.9495 ]
+INFO:root:Average loss=0.13185440748929977
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2025 | LCC = 0.9559 | SRCC = 0.9554 ]
+INFO:root:Average loss=0.12572974821925162
+INFO:root:
+[Valid][21][UTT][ MSE = 0.2025 | LCC = 0.9531 | SRCC = 0.9543 ]
+INFO:root:Average loss=0.12733065366744994
+INFO:root:
+[Valid][22][UTT][ MSE = 0.2168 | LCC = 0.9511 | SRCC = 0.9532 ]
+INFO:root:Average loss=0.11917251497507095
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2036 | LCC = 0.9563 | SRCC = 0.9554 ]
+INFO:root:Average loss=0.11531247371435166
+INFO:root:
+[Valid][24][UTT][ MSE = 0.2819 | LCC = 0.9509 | SRCC = 0.9522 ]
+INFO:root:Average loss=0.12601213324069976
+INFO:root:
+[Valid][25][UTT][ MSE = 0.2440 | LCC = 0.9578 | SRCC = 0.9572 ]
+INFO:root:Average loss=0.12161339432001114
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2255 | LCC = 0.9585 | SRCC = 0.9577 ]
+INFO:root:Average loss=0.1213982959985733
+INFO:root:
+[Valid][27][UTT][ MSE = 0.2173 | LCC = 0.9592 | SRCC = 0.9581 ]
+INFO:root:Average loss=0.11161489087343215
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1747 | LCC = 0.9583 | SRCC = 0.9578 ]
+INFO:root:Average loss=0.11380220684409141
+INFO:root:
+[Valid][29][UTT][ MSE = 0.2964 | LCC = 0.9556 | SRCC = 0.9580 ]
+INFO:root:Average loss=0.10564400154352188
+INFO:root:
+[Valid][30][UTT][ MSE = 0.2176 | LCC = 0.9596 | SRCC = 0.9592 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2175 | LCC = 0.9600 | SRCC = 0.9606 ]

results_tencent/w2v2_xlsr_300m/layer23/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_300m/layer23/run4/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3a66760664507800bd9cd629fc085c12ea15f1ae5d31c4e062684591f1cfb71c
+size 3403339

results_tencent/w2v2_xlsr_300m/layer23/run4/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c3e0be75784aed4989877cfd11a53e63f20cfeb72c41b1bcc4a6e4dfc44b2f8f
+size 3405062

results_tencent/w2v2_xlsr_300m/layer23/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=0.8145056846141815
+INFO:root:
+[Valid][0][UTT][ MSE = 0.5079 | LCC = 0.8813 | SRCC = 0.8778 ]
+INFO:root:Average loss=0.3912081800699234
+INFO:root:
+[Valid][1][UTT][ MSE = 0.4546 | LCC = 0.8927 | SRCC = 0.8958 ]
+INFO:root:Average loss=0.3569305625557899
+INFO:root:
+[Valid][2][UTT][ MSE = 0.3900 | LCC = 0.9121 | SRCC = 0.9098 ]
+INFO:root:Average loss=0.37009758031368256
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2886 | LCC = 0.9206 | SRCC = 0.9167 ]
+INFO:root:Average loss=0.2958523344993591
+INFO:root:
+[Valid][4][UTT][ MSE = 0.4682 | LCC = 0.9135 | SRCC = 0.9183 ]
+INFO:root:Average loss=0.22553248322010042
+INFO:root:
+[Valid][5][UTT][ MSE = 0.5679 | LCC = 0.9123 | SRCC = 0.9226 ]
+INFO:root:Average loss=0.2567094199061394
+INFO:root:
+[Valid][6][UTT][ MSE = 0.3188 | LCC = 0.9405 | SRCC = 0.9380 ]
+INFO:root:Average loss=0.22448823195695877
+INFO:root:
+[Valid][7][UTT][ MSE = 0.3060 | LCC = 0.9410 | SRCC = 0.9395 ]
+INFO:root:Average loss=0.1885664335489273
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2902 | LCC = 0.9383 | SRCC = 0.9398 ]
+INFO:root:Average loss=0.19456501942873
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2486 | LCC = 0.9459 | SRCC = 0.9457 ]
+INFO:root:Average loss=0.17806837284564972
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2564 | LCC = 0.9453 | SRCC = 0.9455 ]
+INFO:root:Average loss=0.15499526584148407
+INFO:root:
+[Valid][11][UTT][ MSE = 0.2147 | LCC = 0.9500 | SRCC = 0.9502 ]
+INFO:root:Average loss=0.16285975557565688
+INFO:root:
+[Valid][12][UTT][ MSE = 0.3589 | LCC = 0.9488 | SRCC = 0.9494 ]
+INFO:root:Average loss=0.16125369787216187
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2220 | LCC = 0.9510 | SRCC = 0.9511 ]
+INFO:root:Average loss=0.15557553827762605
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2578 | LCC = 0.9507 | SRCC = 0.9515 ]
+INFO:root:Average loss=0.1599581662416458
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2192 | LCC = 0.9520 | SRCC = 0.9525 ]
+INFO:root:Average loss=0.13801892495155335
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2273 | LCC = 0.9542 | SRCC = 0.9537 ]
+INFO:root:Average loss=0.13788513362407684
+INFO:root:
+[Valid][17][UTT][ MSE = 0.2258 | LCC = 0.9538 | SRCC = 0.9534 ]
+INFO:root:Average loss=0.1377027438879013
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2296 | LCC = 0.9512 | SRCC = 0.9506 ]
+INFO:root:Average loss=0.12967479968070983
+INFO:root:
+[Valid][19][UTT][ MSE = 0.2255 | LCC = 0.9553 | SRCC = 0.9554 ]
+INFO:root:Average loss=0.1240396209359169
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2883 | LCC = 0.9521 | SRCC = 0.9537 ]
+INFO:root:Average loss=0.12480794876813889
+INFO:root:
+[Valid][21][UTT][ MSE = 0.2768 | LCC = 0.9582 | SRCC = 0.9584 ]
+INFO:root:Average loss=0.12560943567752839
+INFO:root:
+[Valid][22][UTT][ MSE = 0.2291 | LCC = 0.9550 | SRCC = 0.9553 ]
+INFO:root:Average loss=0.1200638672709465
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2564 | LCC = 0.9553 | SRCC = 0.9568 ]
+INFO:root:Average loss=0.13097032684087753
+INFO:root:
+[Valid][24][UTT][ MSE = 0.2212 | LCC = 0.9524 | SRCC = 0.9533 ]
+INFO:root:Average loss=0.12304448008537293
+INFO:root:
+[Valid][25][UTT][ MSE = 0.2513 | LCC = 0.9527 | SRCC = 0.9550 ]
+INFO:root:Average loss=0.11969953566789628
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2351 | LCC = 0.9596 | SRCC = 0.9595 ]
+INFO:root:Average loss=0.11048496359586715
+INFO:root:
+[Valid][27][UTT][ MSE = 0.2014 | LCC = 0.9593 | SRCC = 0.9589 ]
+INFO:root:Average loss=0.11392447927594185
+INFO:root:
+[Valid][28][UTT][ MSE = 0.2106 | LCC = 0.9587 | SRCC = 0.9597 ]
+INFO:root:Average loss=0.10840985369682311
+INFO:root:
+[Valid][29][UTT][ MSE = 0.2144 | LCC = 0.9603 | SRCC = 0.9598 ]
+INFO:root:Average loss=0.1037284859418869
+INFO:root:
+[Valid][30][UTT][ MSE = 0.2233 | LCC = 0.9595 | SRCC = 0.9591 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2146 | LCC = 0.9598 | SRCC = 0.9605 ]

results_tencent/w2v2_xlsr_300m/layer3/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_300m/layer3/run0/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7d4d14ae671b197d8d8d4109c100984582b288e047ee8c08c6b977cdcc74ace9
+size 3403339

results_tencent/w2v2_xlsr_300m/layer3/run0/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bd84e031d1e5a2a42d589559e6fad5afc96763ed128f23626e2a985e9aa5e1db
+size 3405062

results_tencent/w2v2_xlsr_300m/layer3/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=0.4928909802436829
+INFO:root:
+[Valid][0][UTT][ MSE = 0.2999 | LCC = 0.9365 | SRCC = 0.9372 ]
+INFO:root:Average loss=0.16937899965047837
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2725 | LCC = 0.9498 | SRCC = 0.9495 ]
+INFO:root:Average loss=0.13991226106882096
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2283 | LCC = 0.9555 | SRCC = 0.9554 ]
+INFO:root:Average loss=0.1192230771780014
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2289 | LCC = 0.9601 | SRCC = 0.9603 ]
+INFO:root:Average loss=0.1098528333902359
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2098 | LCC = 0.9632 | SRCC = 0.9627 ]
+INFO:root:Average loss=0.09928976985812187
+INFO:root:
+[Valid][5][UTT][ MSE = 0.1914 | LCC = 0.9636 | SRCC = 0.9632 ]
+INFO:root:Average loss=0.08686692520976066
+INFO:root:
+[Valid][6][UTT][ MSE = 0.1928 | LCC = 0.9659 | SRCC = 0.9655 ]
+INFO:root:Average loss=0.08223827317357063
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2097 | LCC = 0.9663 | SRCC = 0.9661 ]
+INFO:root:Average loss=0.07617785930633544
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2168 | LCC = 0.9658 | SRCC = 0.9655 ]
+INFO:root:Average loss=0.07284880968928337
+INFO:root:
+[Valid][9][UTT][ MSE = 0.1634 | LCC = 0.9659 | SRCC = 0.9659 ]
+INFO:root:Average loss=0.0717342868745327
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2133 | LCC = 0.9675 | SRCC = 0.9672 ]
+INFO:root:Average loss=0.0682714629471302
+INFO:root:
+[Valid][11][UTT][ MSE = 0.2210 | LCC = 0.9684 | SRCC = 0.9691 ]
+INFO:root:Average loss=0.06349933075904846
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2104 | LCC = 0.9688 | SRCC = 0.9687 ]
+INFO:root:Average loss=0.06211875455081463
+INFO:root:
+[Valid][13][UTT][ MSE = 0.1813 | LCC = 0.9678 | SRCC = 0.9668 ]
+INFO:root:Average loss=0.057986576482653616
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1774 | LCC = 0.9704 | SRCC = 0.9691 ]
+INFO:root:Average loss=0.05545867249369621
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1887 | LCC = 0.9707 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.05239409136772156
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2099 | LCC = 0.9685 | SRCC = 0.9690 ]
+INFO:root:Average loss=0.04977992175519466
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1832 | LCC = 0.9700 | SRCC = 0.9689 ]
+INFO:root:Average loss=0.046591076463460925
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1702 | LCC = 0.9707 | SRCC = 0.9696 ]
+INFO:root:Average loss=0.04589277376234532
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1565 | LCC = 0.9706 | SRCC = 0.9693 ]
+INFO:root:Average loss=0.04347516365349293
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1732 | LCC = 0.9714 | SRCC = 0.9702 ]
+INFO:root:Average loss=0.043134723335504534
+INFO:root:
+[Valid][21][UTT][ MSE = 0.2012 | LCC = 0.9705 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.043562618657946584
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1732 | LCC = 0.9713 | SRCC = 0.9704 ]
+INFO:root:Average loss=0.03857028645277023
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1833 | LCC = 0.9714 | SRCC = 0.9702 ]
+INFO:root:Average loss=0.03828483219444752
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1784 | LCC = 0.9701 | SRCC = 0.9692 ]
+INFO:root:Average loss=0.03631746079027653
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1682 | LCC = 0.9707 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.03477946110069752
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1900 | LCC = 0.9716 | SRCC = 0.9703 ]
+INFO:root:Average loss=0.0341524433195591
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1786 | LCC = 0.9721 | SRCC = 0.9710 ]
+INFO:root:Average loss=0.034333244621753696
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1568 | LCC = 0.9712 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.03195479291677475
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1698 | LCC = 0.9702 | SRCC = 0.9688 ]
+INFO:root:Average loss=0.031488272041082385
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1602 | LCC = 0.9714 | SRCC = 0.9706 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1861 | LCC = 0.9690 | SRCC = 0.9692 ]

results_tencent/w2v2_xlsr_300m/layer3/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_300m/layer3/run1/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:06f970ccbfeeeda29451662d84976e2f04a8b09cb3b2810cb82e2e6e76166755
+size 3403339

results_tencent/w2v2_xlsr_300m/layer3/run1/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:26498e14c298d2fbb45f2bc25c7a245339cba729de222da271f483f6695bf26f
+size 3405062

results_tencent/w2v2_xlsr_300m/layer3/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=0.39172349286079405
+INFO:root:
+[Valid][0][UTT][ MSE = 0.2452 | LCC = 0.9335 | SRCC = 0.9357 ]
+INFO:root:Average loss=0.1631543598175049
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2577 | LCC = 0.9484 | SRCC = 0.9463 ]
+INFO:root:Average loss=0.1373210073709488
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2732 | LCC = 0.9541 | SRCC = 0.9538 ]
+INFO:root:Average loss=0.11968734902143478
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2440 | LCC = 0.9566 | SRCC = 0.9559 ]
+INFO:root:Average loss=0.10614920368790627
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2279 | LCC = 0.9601 | SRCC = 0.9598 ]
+INFO:root:Average loss=0.09772597977519035
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2263 | LCC = 0.9620 | SRCC = 0.9615 ]
+INFO:root:Average loss=0.08985240116715432
+INFO:root:
+[Valid][6][UTT][ MSE = 0.1695 | LCC = 0.9616 | SRCC = 0.9620 ]
+INFO:root:Average loss=0.08811028820276261
+INFO:root:
+[Valid][7][UTT][ MSE = 0.1973 | LCC = 0.9642 | SRCC = 0.9635 ]
+INFO:root:Average loss=0.07966661924123764
+INFO:root:
+[Valid][8][UTT][ MSE = 0.1957 | LCC = 0.9650 | SRCC = 0.9642 ]
+INFO:root:Average loss=0.07141376048326492
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2288 | LCC = 0.9645 | SRCC = 0.9640 ]
+INFO:root:Average loss=0.0724522342979908
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2128 | LCC = 0.9648 | SRCC = 0.9642 ]
+INFO:root:Average loss=0.06668364444375038
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1947 | LCC = 0.9634 | SRCC = 0.9622 ]
+INFO:root:Average loss=0.06506802225112915
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2270 | LCC = 0.9649 | SRCC = 0.9657 ]
+INFO:root:Average loss=0.06406673233211041
+INFO:root:
+[Valid][13][UTT][ MSE = 0.1766 | LCC = 0.9649 | SRCC = 0.9640 ]
+INFO:root:Average loss=0.05752970385551453
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2125 | LCC = 0.9666 | SRCC = 0.9662 ]
+INFO:root:Average loss=0.05704319806396961
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1692 | LCC = 0.9679 | SRCC = 0.9671 ]
+INFO:root:Average loss=0.054945798367261885
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1877 | LCC = 0.9676 | SRCC = 0.9667 ]
+INFO:root:Average loss=0.05074971768260002
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1854 | LCC = 0.9671 | SRCC = 0.9669 ]
+INFO:root:Average loss=0.05246031494438648
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1725 | LCC = 0.9685 | SRCC = 0.9674 ]
+INFO:root:Average loss=0.04877910551428795
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1944 | LCC = 0.9677 | SRCC = 0.9671 ]
+INFO:root:Average loss=0.04537329433858395
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1706 | LCC = 0.9662 | SRCC = 0.9658 ]
+INFO:root:Average loss=0.04382076473534107
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1881 | LCC = 0.9682 | SRCC = 0.9674 ]
+INFO:root:Average loss=0.04116334687173367
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1758 | LCC = 0.9688 | SRCC = 0.9676 ]
+INFO:root:Average loss=0.04096766270697117
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2053 | LCC = 0.9677 | SRCC = 0.9676 ]
+INFO:root:Average loss=0.040873097628355025
+INFO:root:
+[Valid][24][UTT][ MSE = 0.2228 | LCC = 0.9660 | SRCC = 0.9669 ]
+INFO:root:Average loss=0.03826784770190716
+INFO:root:
+[Valid][25][UTT][ MSE = 0.2157 | LCC = 0.9681 | SRCC = 0.9679 ]
+INFO:root:Average loss=0.03710025091469288
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1836 | LCC = 0.9699 | SRCC = 0.9686 ]
+INFO:root:Average loss=0.033851901218295094
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1744 | LCC = 0.9698 | SRCC = 0.9679 ]
+INFO:root:Average loss=0.03558759294450283
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1923 | LCC = 0.9677 | SRCC = 0.9667 ]
+INFO:root:Average loss=0.03433390028774738
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1760 | LCC = 0.9694 | SRCC = 0.9682 ]
+INFO:root:Average loss=0.03310309068858624
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1745 | LCC = 0.9694 | SRCC = 0.9676 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1879 | LCC = 0.9685 | SRCC = 0.9683 ]

results_tencent/w2v2_xlsr_300m/layer3/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_300m/layer3/run2/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a8bc80f0dcac434ef2ad23e8b15029f942a75a7659999b36bcf19b35129d420c
+size 3403339

results_tencent/w2v2_xlsr_300m/layer3/run2/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4a5ef4e1cf484f54a66add24d67127da52aeb8609558fbb7c01a7bc25c2a8f17
+size 3405062

results_tencent/w2v2_xlsr_300m/layer3/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=0.32500236386060716
+INFO:root:
+[Valid][0][UTT][ MSE = 0.2885 | LCC = 0.9391 | SRCC = 0.9385 ]
+INFO:root:Average loss=0.1558499710559845
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2587 | LCC = 0.9454 | SRCC = 0.9486 ]
+INFO:root:Average loss=0.14717653560638427
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2308 | LCC = 0.9543 | SRCC = 0.9554 ]
+INFO:root:Average loss=0.11835578638315201
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2476 | LCC = 0.9586 | SRCC = 0.9592 ]
+INFO:root:Average loss=0.10746299874782562
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2186 | LCC = 0.9583 | SRCC = 0.9602 ]
+INFO:root:Average loss=0.09926072120666504
+INFO:root:
+[Valid][5][UTT][ MSE = 0.1998 | LCC = 0.9634 | SRCC = 0.9630 ]
+INFO:root:Average loss=0.08872406920790672
+INFO:root:
+[Valid][6][UTT][ MSE = 0.1973 | LCC = 0.9637 | SRCC = 0.9632 ]
+INFO:root:Average loss=0.08390092438459397
+INFO:root:
+[Valid][7][UTT][ MSE = 0.1869 | LCC = 0.9633 | SRCC = 0.9646 ]
+INFO:root:Average loss=0.07669991993904114
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2116 | LCC = 0.9656 | SRCC = 0.9659 ]
+INFO:root:Average loss=0.07126889047026634
+INFO:root:
+[Valid][9][UTT][ MSE = 0.1746 | LCC = 0.9659 | SRCC = 0.9649 ]
+INFO:root:Average loss=0.06808758071064949
+INFO:root:
+[Valid][10][UTT][ MSE = 0.1890 | LCC = 0.9682 | SRCC = 0.9678 ]
+INFO:root:Average loss=0.07008291900157929
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1670 | LCC = 0.9680 | SRCC = 0.9672 ]
+INFO:root:Average loss=0.06043029496073723
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2078 | LCC = 0.9666 | SRCC = 0.9666 ]
+INFO:root:Average loss=0.059582085132598876
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2207 | LCC = 0.9656 | SRCC = 0.9672 ]
+INFO:root:Average loss=0.06040261626243591
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1618 | LCC = 0.9674 | SRCC = 0.9673 ]
+INFO:root:Average loss=0.05360208167135715
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1690 | LCC = 0.9683 | SRCC = 0.9672 ]
+INFO:root:Average loss=0.052096041172742844
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1796 | LCC = 0.9685 | SRCC = 0.9681 ]
+INFO:root:Average loss=0.048393024265766145
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1609 | LCC = 0.9681 | SRCC = 0.9675 ]
+INFO:root:Average loss=0.04572081921994686
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1854 | LCC = 0.9679 | SRCC = 0.9672 ]
+INFO:root:Average loss=0.04716121120750904
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1496 | LCC = 0.9687 | SRCC = 0.9676 ]
+INFO:root:Average loss=0.041934609800577165
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1696 | LCC = 0.9680 | SRCC = 0.9676 ]
+INFO:root:Average loss=0.040934306874871255
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1456 | LCC = 0.9690 | SRCC = 0.9683 ]
+INFO:root:Average loss=0.03817667020857334
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1534 | LCC = 0.9689 | SRCC = 0.9686 ]
+INFO:root:Average loss=0.03709316331148148
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1706 | LCC = 0.9693 | SRCC = 0.9683 ]
+INFO:root:Average loss=0.03549412325024605
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1677 | LCC = 0.9696 | SRCC = 0.9687 ]
+INFO:root:Average loss=0.03586291381716728
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1550 | LCC = 0.9690 | SRCC = 0.9686 ]
+INFO:root:Average loss=0.03558068524301052
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1646 | LCC = 0.9688 | SRCC = 0.9679 ]
+INFO:root:Average loss=0.03305184856057167
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1894 | LCC = 0.9691 | SRCC = 0.9689 ]
+INFO:root:Average loss=0.03446996255218983
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1821 | LCC = 0.9677 | SRCC = 0.9676 ]
+INFO:root:Average loss=0.032194572389125826
+INFO:root:
+[Valid][29][UTT][ MSE = 0.2019 | LCC = 0.9685 | SRCC = 0.9686 ]
+INFO:root:Average loss=0.030766275979578494
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1831 | LCC = 0.9691 | SRCC = 0.9688 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1712 | LCC = 0.9675 | SRCC = 0.9672 ]

results_tencent/w2v2_xlsr_300m/layer3/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_300m/layer3/run3/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4cafcac12c56a0906e74f8f43cdf7d6286d09c61c69db21f4a884bce1ba3dd44
+size 3403339

results_tencent/w2v2_xlsr_300m/layer3/run3/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2c2e119336f1e19c811374bd91b6bba64fd605663728098526b86877a34c93e5
+size 3405062

results_tencent/w2v2_xlsr_300m/layer3/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=0.4237779357433319
+INFO:root:
+[Valid][0][UTT][ MSE = 0.2507 | LCC = 0.9407 | SRCC = 0.9412 ]
+INFO:root:Average loss=0.15904512071609497
+INFO:root:
+[Valid][1][UTT][ MSE = 0.1991 | LCC = 0.9490 | SRCC = 0.9508 ]
+INFO:root:Average loss=0.12917388647794723
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2041 | LCC = 0.9577 | SRCC = 0.9569 ]
+INFO:root:Average loss=0.11946277451515198
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2158 | LCC = 0.9588 | SRCC = 0.9586 ]
+INFO:root:Average loss=0.10236752167344093
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2171 | LCC = 0.9620 | SRCC = 0.9612 ]
+INFO:root:Average loss=0.09482676592469215
+INFO:root:
+[Valid][5][UTT][ MSE = 0.1740 | LCC = 0.9630 | SRCC = 0.9627 ]
+INFO:root:Average loss=0.09287394666671753
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2303 | LCC = 0.9630 | SRCC = 0.9636 ]
+INFO:root:Average loss=0.080492874532938
+INFO:root:
+[Valid][7][UTT][ MSE = 0.1874 | LCC = 0.9646 | SRCC = 0.9643 ]
+INFO:root:Average loss=0.07866597092151642
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2078 | LCC = 0.9654 | SRCC = 0.9661 ]
+INFO:root:Average loss=0.0753565459549427
+INFO:root:
+[Valid][9][UTT][ MSE = 0.1660 | LCC = 0.9678 | SRCC = 0.9673 ]
+INFO:root:Average loss=0.06905877244472504
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2352 | LCC = 0.9674 | SRCC = 0.9670 ]
+INFO:root:Average loss=0.06652959322929382
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1574 | LCC = 0.9678 | SRCC = 0.9670 ]
+INFO:root:Average loss=0.06260098233819007
+INFO:root:
+[Valid][12][UTT][ MSE = 0.1575 | LCC = 0.9675 | SRCC = 0.9663 ]
+INFO:root:Average loss=0.0616500555574894
+INFO:root:
+[Valid][13][UTT][ MSE = 0.1422 | LCC = 0.9684 | SRCC = 0.9682 ]
+INFO:root:Average loss=0.05589770913124084
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1801 | LCC = 0.9684 | SRCC = 0.9676 ]
+INFO:root:Average loss=0.05502425518631935
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1770 | LCC = 0.9687 | SRCC = 0.9681 ]
+INFO:root:Average loss=0.053068888947367665
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1768 | LCC = 0.9675 | SRCC = 0.9669 ]
+INFO:root:Average loss=0.047760760962963104
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1555 | LCC = 0.9669 | SRCC = 0.9675 ]
+INFO:root:Average loss=0.04732796381413937
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1410 | LCC = 0.9684 | SRCC = 0.9677 ]
+INFO:root:Average loss=0.046957701101899146
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1656 | LCC = 0.9698 | SRCC = 0.9687 ]
+INFO:root:Average loss=0.04378133215010166
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1886 | LCC = 0.9693 | SRCC = 0.9685 ]
+INFO:root:Average loss=0.0433526478856802
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1493 | LCC = 0.9690 | SRCC = 0.9679 ]
+INFO:root:Average loss=0.040885006234049795
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1614 | LCC = 0.9694 | SRCC = 0.9679 ]
+INFO:root:Average loss=0.03903745400905609
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1486 | LCC = 0.9690 | SRCC = 0.9685 ]
+INFO:root:Average loss=0.03657390807569027
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1741 | LCC = 0.9701 | SRCC = 0.9688 ]
+INFO:root:Average loss=0.039520827695727345
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1467 | LCC = 0.9693 | SRCC = 0.9679 ]
+INFO:root:Average loss=0.034794993579387666
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1615 | LCC = 0.9705 | SRCC = 0.9692 ]
+INFO:root:Average loss=0.033548924967646596
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1373 | LCC = 0.9692 | SRCC = 0.9691 ]
+INFO:root:Average loss=0.03349556158483028
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1646 | LCC = 0.9684 | SRCC = 0.9678 ]
+INFO:root:Average loss=0.033362504348158836
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1690 | LCC = 0.9712 | SRCC = 0.9702 ]
+INFO:root:Average loss=0.03324629755318165
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1324 | LCC = 0.9701 | SRCC = 0.9700 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1724 | LCC = 0.9692 | SRCC = 0.9691 ]

results_tencent/w2v2_xlsr_300m/layer3/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_300m/layer3/run4/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cc90c8e13e29b36108ea187c896e9630c05a14eb2d56e0b8d3f21926dc39901a
+size 3403339

results_tencent/w2v2_xlsr_300m/layer3/run4/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cf1e9fb85fbb2be8d46594a6cbbe6859145b7660baf40bd32a662ff7d9c850e7
+size 3405062

results_tencent/w2v2_xlsr_300m/layer3/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=0.48466843366622925
+INFO:root:
+[Valid][0][UTT][ MSE = 0.2569 | LCC = 0.9384 | SRCC = 0.9373 ]
+INFO:root:Average loss=0.1775839799642563
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2046 | LCC = 0.9454 | SRCC = 0.9473 ]
+INFO:root:Average loss=0.15533198940753937
+INFO:root:
+[Valid][2][UTT][ MSE = 0.1944 | LCC = 0.9498 | SRCC = 0.9514 ]
+INFO:root:Average loss=0.1221833969950676
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2344 | LCC = 0.9492 | SRCC = 0.9519 ]
+INFO:root:Average loss=0.10968787693977355
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2175 | LCC = 0.9615 | SRCC = 0.9612 ]
+INFO:root:Average loss=0.101326884329319
+INFO:root:
+[Valid][5][UTT][ MSE = 0.1744 | LCC = 0.9627 | SRCC = 0.9621 ]
+INFO:root:Average loss=0.09600243681669235
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2125 | LCC = 0.9635 | SRCC = 0.9636 ]
+INFO:root:Average loss=0.08750571379065514
+INFO:root:
+[Valid][7][UTT][ MSE = 0.1742 | LCC = 0.9622 | SRCC = 0.9645 ]
+INFO:root:Average loss=0.07722979310154915
+INFO:root:
+[Valid][8][UTT][ MSE = 0.1612 | LCC = 0.9616 | SRCC = 0.9618 ]
+INFO:root:Average loss=0.07694434255361557
+INFO:root:
+[Valid][9][UTT][ MSE = 0.1627 | LCC = 0.9646 | SRCC = 0.9658 ]
+INFO:root:Average loss=0.07097282034158707
+INFO:root:
+[Valid][10][UTT][ MSE = 0.1605 | LCC = 0.9661 | SRCC = 0.9662 ]
+INFO:root:Average loss=0.0660415803194046
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1641 | LCC = 0.9675 | SRCC = 0.9671 ]
+INFO:root:Average loss=0.0639457170367241
+INFO:root:
+[Valid][12][UTT][ MSE = 0.1561 | LCC = 0.9655 | SRCC = 0.9669 ]
+INFO:root:Average loss=0.06160608339309692
+INFO:root:
+[Valid][13][UTT][ MSE = 0.1573 | LCC = 0.9687 | SRCC = 0.9678 ]
+INFO:root:Average loss=0.05722728738188743
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1853 | LCC = 0.9683 | SRCC = 0.9671 ]
+INFO:root:Average loss=0.056382120430469515
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1571 | LCC = 0.9669 | SRCC = 0.9681 ]
+INFO:root:Average loss=0.052327201813459394
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1402 | LCC = 0.9684 | SRCC = 0.9677 ]
+INFO:root:Average loss=0.050134144946932795
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1659 | LCC = 0.9675 | SRCC = 0.9683 ]
+INFO:root:Average loss=0.04756868174672127
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1527 | LCC = 0.9689 | SRCC = 0.9685 ]
+INFO:root:Average loss=0.04681389385461807
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1563 | LCC = 0.9690 | SRCC = 0.9682 ]
+INFO:root:Average loss=0.044286971300840375
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1652 | LCC = 0.9688 | SRCC = 0.9680 ]
+INFO:root:Average loss=0.04239480862021446
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1540 | LCC = 0.9690 | SRCC = 0.9683 ]
+INFO:root:Average loss=0.04106458252668381
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1500 | LCC = 0.9682 | SRCC = 0.9682 ]
+INFO:root:Average loss=0.040112100258469584
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1790 | LCC = 0.9659 | SRCC = 0.9662 ]
+INFO:root:Average loss=0.03831818917393685
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1645 | LCC = 0.9691 | SRCC = 0.9681 ]
+INFO:root:Average loss=0.03647170528769493
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1448 | LCC = 0.9681 | SRCC = 0.9677 ]
+INFO:root:Average loss=0.03561024002730846
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1561 | LCC = 0.9694 | SRCC = 0.9684 ]
+INFO:root:Average loss=0.035114300280809405
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1830 | LCC = 0.9688 | SRCC = 0.9684 ]
+INFO:root:Average loss=0.03342070460319519
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1677 | LCC = 0.9686 | SRCC = 0.9676 ]
+INFO:root:Average loss=0.032618674844503405
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1756 | LCC = 0.9692 | SRCC = 0.9681 ]
+INFO:root:Average loss=0.03046013058722019
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1590 | LCC = 0.9696 | SRCC = 0.9688 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1594 | LCC = 0.9686 | SRCC = 0.9681 ]

results_tencent/w2v2_xlsr_300m/layer4/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_300m/layer4/run0/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b131ec31001310df34289a8edb00d57a85b28f1284d59dead3ba3805a72e01c
+size 3403339

results_tencent/w2v2_xlsr_300m/layer4/run0/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:78f13040df8e49be355170ba7e86fe6473da9a7c2f4e585862f13a3cc6387d18
+size 3405062

results_tencent/w2v2_xlsr_300m/layer4/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=0.40181518894433976
+INFO:root:
+[Valid][0][UTT][ MSE = 0.2499 | LCC = 0.9466 | SRCC = 0.9464 ]
+INFO:root:Average loss=0.15385954695940018
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2499 | LCC = 0.9537 | SRCC = 0.9529 ]
+INFO:root:Average loss=0.12373063308000565
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2109 | LCC = 0.9604 | SRCC = 0.9592 ]
+INFO:root:Average loss=0.10653454419970512
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2147 | LCC = 0.9632 | SRCC = 0.9622 ]
+INFO:root:Average loss=0.09398544239997864
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2260 | LCC = 0.9643 | SRCC = 0.9635 ]
+INFO:root:Average loss=0.0881934364438057
+INFO:root:
+[Valid][5][UTT][ MSE = 0.1943 | LCC = 0.9658 | SRCC = 0.9660 ]
+INFO:root:Average loss=0.07958787786960601
+INFO:root:
+[Valid][6][UTT][ MSE = 0.1783 | LCC = 0.9675 | SRCC = 0.9673 ]
+INFO:root:Average loss=0.0724699057340622
+INFO:root:
+[Valid][7][UTT][ MSE = 0.1683 | LCC = 0.9653 | SRCC = 0.9650 ]
+INFO:root:Average loss=0.06957830360531807
+INFO:root:
+[Valid][8][UTT][ MSE = 0.1751 | LCC = 0.9687 | SRCC = 0.9681 ]
+INFO:root:Average loss=0.06896216365695
+INFO:root:
+[Valid][9][UTT][ MSE = 0.1896 | LCC = 0.9698 | SRCC = 0.9689 ]
+INFO:root:Average loss=0.06770285993814468
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2146 | LCC = 0.9696 | SRCC = 0.9693 ]
+INFO:root:Average loss=0.06310671600699425
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1457 | LCC = 0.9690 | SRCC = 0.9686 ]
+INFO:root:Average loss=0.056994130730628964
+INFO:root:
+[Valid][12][UTT][ MSE = 0.1926 | LCC = 0.9696 | SRCC = 0.9701 ]
+INFO:root:Average loss=0.054625910580158234
+INFO:root:
+[Valid][13][UTT][ MSE = 0.1578 | LCC = 0.9700 | SRCC = 0.9691 ]
+INFO:root:Average loss=0.051796904265880586
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2039 | LCC = 0.9675 | SRCC = 0.9674 ]
+INFO:root:Average loss=0.04977180987596512
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1849 | LCC = 0.9679 | SRCC = 0.9680 ]
+INFO:root:Average loss=0.044429125517606735
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1734 | LCC = 0.9703 | SRCC = 0.9696 ]
+INFO:root:Average loss=0.043082465261220934
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1581 | LCC = 0.9691 | SRCC = 0.9685 ]
+INFO:root:Average loss=0.0434044725894928
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1267 | LCC = 0.9706 | SRCC = 0.9701 ]
+INFO:root:Average loss=0.03978926730155945
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1833 | LCC = 0.9699 | SRCC = 0.9694 ]
+INFO:root:Average loss=0.03919456271827221
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1736 | LCC = 0.9704 | SRCC = 0.9701 ]
+INFO:root:Average loss=0.0359546305090189
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1468 | LCC = 0.9701 | SRCC = 0.9694 ]
+INFO:root:Average loss=0.03708182345330715
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1795 | LCC = 0.9711 | SRCC = 0.9705 ]
+INFO:root:Average loss=0.035054993987083435
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1552 | LCC = 0.9707 | SRCC = 0.9700 ]
+INFO:root:Average loss=0.03389630497992039
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1574 | LCC = 0.9701 | SRCC = 0.9700 ]
+INFO:root:Average loss=0.03245997132360935
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1844 | LCC = 0.9701 | SRCC = 0.9694 ]
+INFO:root:Average loss=0.030413863480091095
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1572 | LCC = 0.9710 | SRCC = 0.9705 ]
+INFO:root:Average loss=0.031395073607563974
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1631 | LCC = 0.9700 | SRCC = 0.9689 ]
+INFO:root:Average loss=0.02919469544291496
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1755 | LCC = 0.9701 | SRCC = 0.9692 ]
+INFO:root:Average loss=0.02769296584278345
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1592 | LCC = 0.9714 | SRCC = 0.9706 ]
+INFO:root:Average loss=0.026416000045835972
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1723 | LCC = 0.9710 | SRCC = 0.9704 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1614 | LCC = 0.9701 | SRCC = 0.9701 ]

results_tencent/w2v2_xlsr_300m/layer4/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_300m/layer4/run1/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:918f90d9cc4ede4fafb0658cd8360bca794ee85c2d36ca61193bdad4496030fd
+size 3403339

results_tencent/w2v2_xlsr_300m/layer4/run1/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bf0f34a9b9d10698f48e5cbde6a7f30fdeccc79700ecd089fcd2ec5862f0f330
+size 3405062

results_tencent/w2v2_xlsr_300m/layer4/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=0.3539437585473061
+INFO:root:
+[Valid][0][UTT][ MSE = 0.2316 | LCC = 0.9463 | SRCC = 0.9447 ]
+INFO:root:Average loss=0.1384958703517914
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2892 | LCC = 0.9547 | SRCC = 0.9543 ]
+INFO:root:Average loss=0.12476984417438507
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2154 | LCC = 0.9593 | SRCC = 0.9591 ]
+INFO:root:Average loss=0.10756545868515968
+INFO:root:
+[Valid][3][UTT][ MSE = 0.1943 | LCC = 0.9620 | SRCC = 0.9620 ]
+INFO:root:Average loss=0.09650271216034889
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2152 | LCC = 0.9637 | SRCC = 0.9640 ]
+INFO:root:Average loss=0.08482027754187584
+INFO:root:
+[Valid][5][UTT][ MSE = 0.1902 | LCC = 0.9665 | SRCC = 0.9656 ]
+INFO:root:Average loss=0.0782379894554615
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2176 | LCC = 0.9645 | SRCC = 0.9638 ]
+INFO:root:Average loss=0.07717566844820976
+INFO:root:
+[Valid][7][UTT][ MSE = 0.1921 | LCC = 0.9676 | SRCC = 0.9668 ]
+INFO:root:Average loss=0.07340138745307923
+INFO:root:
+[Valid][8][UTT][ MSE = 0.1647 | LCC = 0.9670 | SRCC = 0.9672 ]
+INFO:root:Average loss=0.06561676475405694
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2143 | LCC = 0.9690 | SRCC = 0.9686 ]
+INFO:root:Average loss=0.06151456007361412
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2067 | LCC = 0.9689 | SRCC = 0.9683 ]
+INFO:root:Average loss=0.06133492562174797
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1873 | LCC = 0.9685 | SRCC = 0.9681 ]
+INFO:root:Average loss=0.057051753878593445
+INFO:root:
+[Valid][12][UTT][ MSE = 0.1582 | LCC = 0.9691 | SRCC = 0.9681 ]
+INFO:root:Average loss=0.055188780531287195
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2032 | LCC = 0.9690 | SRCC = 0.9684 ]
+INFO:root:Average loss=0.05235090309381485
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1706 | LCC = 0.9701 | SRCC = 0.9689 ]
+INFO:root:Average loss=0.04848231323063373
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1827 | LCC = 0.9699 | SRCC = 0.9689 ]
+INFO:root:Average loss=0.048031270012259486
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1583 | LCC = 0.9691 | SRCC = 0.9689 ]
+INFO:root:Average loss=0.045001750349998475
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1901 | LCC = 0.9692 | SRCC = 0.9680 ]
+INFO:root:Average loss=0.04325820714235306
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1839 | LCC = 0.9689 | SRCC = 0.9685 ]
+INFO:root:Average loss=0.04364639917016029
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1854 | LCC = 0.9699 | SRCC = 0.9693 ]
+INFO:root:Average loss=0.037001041427254674
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2028 | LCC = 0.9692 | SRCC = 0.9684 ]
+INFO:root:Average loss=0.03794203832745552
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1609 | LCC = 0.9695 | SRCC = 0.9689 ]
+INFO:root:Average loss=0.035563049912452696
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1807 | LCC = 0.9696 | SRCC = 0.9685 ]
+INFO:root:Average loss=0.034318716898560524
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1774 | LCC = 0.9681 | SRCC = 0.9675 ]
+INFO:root:Average loss=0.033422301664948466
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1669 | LCC = 0.9702 | SRCC = 0.9695 ]
+INFO:root:Average loss=0.033809678971767426
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1618 | LCC = 0.9696 | SRCC = 0.9687 ]
+INFO:root:Average loss=0.03175129644572735
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1680 | LCC = 0.9699 | SRCC = 0.9695 ]
+INFO:root:Average loss=0.032571035385131834
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1877 | LCC = 0.9698 | SRCC = 0.9689 ]
+INFO:root:Average loss=0.03213731575012207
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1871 | LCC = 0.9694 | SRCC = 0.9686 ]
+INFO:root:Average loss=0.028882778719067574
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1790 | LCC = 0.9699 | SRCC = 0.9688 ]
+INFO:root:Average loss=0.02779819478094578
+INFO:root:
+[Valid][30][UTT][ MSE = 0.2094 | LCC = 0.9689 | SRCC = 0.9680 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1690 | LCC = 0.9696 | SRCC = 0.9695 ]

results_tencent/w2v2_xlsr_300m/layer4/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_300m/layer4/run2/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:98f495e1ff221ff61745f039741a8891784fa8ee6b1b76ce4231bd693429eb0f
+size 3403339

results_tencent/w2v2_xlsr_300m/layer4/run2/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:806938213893bb7a5d2fc6b8204779591bc184b270209e9de8bae042cd2df46e
+size 3405062

results_tencent/w2v2_xlsr_300m/layer4/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=0.32357763147354124
+INFO:root:
+[Valid][0][UTT][ MSE = 0.2447 | LCC = 0.9479 | SRCC = 0.9466 ]
+INFO:root:Average loss=0.13941112178564072
+INFO:root:
+[Valid][1][UTT][ MSE = 0.1826 | LCC = 0.9575 | SRCC = 0.9580 ]
+INFO:root:Average loss=0.11757403671741486
+INFO:root:
+[Valid][2][UTT][ MSE = 0.1995 | LCC = 0.9617 | SRCC = 0.9618 ]
+INFO:root:Average loss=0.10048690873384476
+INFO:root:
+[Valid][3][UTT][ MSE = 0.1944 | LCC = 0.9629 | SRCC = 0.9618 ]
+INFO:root:Average loss=0.09427812054753304
+INFO:root:
+[Valid][4][UTT][ MSE = 0.1960 | LCC = 0.9649 | SRCC = 0.9644 ]
+INFO:root:Average loss=0.08512272840738297
+INFO:root:
+[Valid][5][UTT][ MSE = 0.1736 | LCC = 0.9666 | SRCC = 0.9663 ]
+INFO:root:Average loss=0.07750110748410224
+INFO:root:
+[Valid][6][UTT][ MSE = 0.1773 | LCC = 0.9687 | SRCC = 0.9681 ]
+INFO:root:Average loss=0.0740789364874363
+INFO:root:
+[Valid][7][UTT][ MSE = 0.1501 | LCC = 0.9660 | SRCC = 0.9673 ]
+INFO:root:Average loss=0.07480200892686843
+INFO:root:
+[Valid][8][UTT][ MSE = 0.1988 | LCC = 0.9689 | SRCC = 0.9682 ]
+INFO:root:Average loss=0.06501302102208137
+INFO:root:
+[Valid][9][UTT][ MSE = 0.1840 | LCC = 0.9699 | SRCC = 0.9694 ]
+INFO:root:Average loss=0.06238117054104805
+INFO:root:
+[Valid][10][UTT][ MSE = 0.1899 | LCC = 0.9701 | SRCC = 0.9689 ]
+INFO:root:Average loss=0.05812867943942547
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1971 | LCC = 0.9698 | SRCC = 0.9689 ]
+INFO:root:Average loss=0.05790045550465584
+INFO:root:
+[Valid][12][UTT][ MSE = 0.1662 | LCC = 0.9693 | SRCC = 0.9683 ]
+INFO:root:Average loss=0.05575731548666954
+INFO:root:
+[Valid][13][UTT][ MSE = 0.1630 | LCC = 0.9706 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.05430857688188553
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1771 | LCC = 0.9698 | SRCC = 0.9695 ]
+INFO:root:Average loss=0.04787358658015728
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2146 | LCC = 0.9681 | SRCC = 0.9680 ]
+INFO:root:Average loss=0.048314673706889154
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1617 | LCC = 0.9711 | SRCC = 0.9702 ]
+INFO:root:Average loss=0.04674104622006416
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1683 | LCC = 0.9716 | SRCC = 0.9705 ]
+INFO:root:Average loss=0.04354212300479412
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1583 | LCC = 0.9708 | SRCC = 0.9702 ]
+INFO:root:Average loss=0.03950455245375633
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1576 | LCC = 0.9702 | SRCC = 0.9697 ]
+INFO:root:Average loss=0.03978278426826
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1449 | LCC = 0.9706 | SRCC = 0.9701 ]
+INFO:root:Average loss=0.03667403574287891
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1688 | LCC = 0.9707 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.03677007168531418
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1630 | LCC = 0.9719 | SRCC = 0.9709 ]
+INFO:root:Average loss=0.037206030666828156
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1850 | LCC = 0.9712 | SRCC = 0.9704 ]
+INFO:root:Average loss=0.035217587411403656
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1671 | LCC = 0.9717 | SRCC = 0.9705 ]
+INFO:root:Average loss=0.03294680121541023
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1450 | LCC = 0.9711 | SRCC = 0.9702 ]
+INFO:root:Average loss=0.031606700390577315
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1594 | LCC = 0.9714 | SRCC = 0.9705 ]
+INFO:root:Average loss=0.030644863337278366
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1525 | LCC = 0.9701 | SRCC = 0.9688 ]
+INFO:root:Average loss=0.030930347740650177
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1553 | LCC = 0.9716 | SRCC = 0.9707 ]
+INFO:root:Average loss=0.030007797315716744
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1727 | LCC = 0.9712 | SRCC = 0.9702 ]
+INFO:root:Average loss=0.02821310679614544
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1762 | LCC = 0.9710 | SRCC = 0.9700 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1662 | LCC = 0.9700 | SRCC = 0.9697 ]

results_tencent/w2v2_xlsr_300m/layer4/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_300m/layer4/run3/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f990918d7cdf395d8d730c19d765a27e43c3a2fc4181de8f68d4313921c721a2
+size 3403339

results_tencent/w2v2_xlsr_300m/layer4/run3/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a54c9aa6d917e1900690447e108bdcd499123c77003342eab4d9045a9ec4d9fa
+size 3405062

results_tencent/w2v2_xlsr_300m/layer4/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=0.3493855581283569
+INFO:root:
+[Valid][0][UTT][ MSE = 0.2403 | LCC = 0.9463 | SRCC = 0.9453 ]
+INFO:root:Average loss=0.15164289849996568
+INFO:root:
+[Valid][1][UTT][ MSE = 0.1953 | LCC = 0.9565 | SRCC = 0.9561 ]
+INFO:root:Average loss=0.11794343787431717
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2122 | LCC = 0.9624 | SRCC = 0.9619 ]
+INFO:root:Average loss=0.10819954192638397
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2554 | LCC = 0.9632 | SRCC = 0.9631 ]
+INFO:root:Average loss=0.09795257690548896
+INFO:root:
+[Valid][4][UTT][ MSE = 0.1999 | LCC = 0.9662 | SRCC = 0.9657 ]
+INFO:root:Average loss=0.08717562344670296
+INFO:root:
+[Valid][5][UTT][ MSE = 0.1745 | LCC = 0.9669 | SRCC = 0.9668 ]
+INFO:root:Average loss=0.07944501149654389
+INFO:root:
+[Valid][6][UTT][ MSE = 0.1817 | LCC = 0.9689 | SRCC = 0.9681 ]
+INFO:root:Average loss=0.07397128212451935
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2258 | LCC = 0.9671 | SRCC = 0.9679 ]
+INFO:root:Average loss=0.0712431848347187
+INFO:root:
+[Valid][8][UTT][ MSE = 0.1866 | LCC = 0.9688 | SRCC = 0.9682 ]
+INFO:root:Average loss=0.07080922010540962
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2336 | LCC = 0.9699 | SRCC = 0.9692 ]
+INFO:root:Average loss=0.06404505583643913
+INFO:root:
+[Valid][10][UTT][ MSE = 0.1647 | LCC = 0.9697 | SRCC = 0.9687 ]
+INFO:root:Average loss=0.05922862908244133
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1842 | LCC = 0.9693 | SRCC = 0.9689 ]
+INFO:root:Average loss=0.06013742625713348
+INFO:root:
+[Valid][12][UTT][ MSE = 0.1932 | LCC = 0.9693 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.05613620123267174
+INFO:root:
+[Valid][13][UTT][ MSE = 0.1834 | LCC = 0.9711 | SRCC = 0.9700 ]
+INFO:root:Average loss=0.05181649950146675
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1994 | LCC = 0.9698 | SRCC = 0.9690 ]
+INFO:root:Average loss=0.05046484811604023
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1915 | LCC = 0.9699 | SRCC = 0.9692 ]
+INFO:root:Average loss=0.048300379991531375
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1946 | LCC = 0.9709 | SRCC = 0.9701 ]
+INFO:root:Average loss=0.04513438878953457
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1508 | LCC = 0.9704 | SRCC = 0.9697 ]
+INFO:root:Average loss=0.04444257880747318
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2046 | LCC = 0.9697 | SRCC = 0.9688 ]
+INFO:root:Average loss=0.04042323687672615
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1672 | LCC = 0.9708 | SRCC = 0.9698 ]
+INFO:root:Average loss=0.03882869379222393
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1756 | LCC = 0.9701 | SRCC = 0.9692 ]
+INFO:root:Average loss=0.037710220754146574
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1897 | LCC = 0.9702 | SRCC = 0.9692 ]
+INFO:root:Average loss=0.03739005798101425
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1870 | LCC = 0.9713 | SRCC = 0.9704 ]
+INFO:root:Average loss=0.03796876099705696
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1810 | LCC = 0.9695 | SRCC = 0.9682 ]
+INFO:root:Average loss=0.033981205120682714
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1909 | LCC = 0.9694 | SRCC = 0.9685 ]
+INFO:root:Average loss=0.03287234637141228
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1695 | LCC = 0.9689 | SRCC = 0.9682 ]
+INFO:root:Average loss=0.031535907059907914
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1660 | LCC = 0.9708 | SRCC = 0.9703 ]
+INFO:root:Average loss=0.03071501253545284
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1736 | LCC = 0.9712 | SRCC = 0.9701 ]
+INFO:root:Average loss=0.027936208471655847
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1413 | LCC = 0.9698 | SRCC = 0.9689 ]
+INFO:root:Average loss=0.028565396949648856
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1831 | LCC = 0.9693 | SRCC = 0.9683 ]
+INFO:root:Average loss=0.03051776148378849
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1636 | LCC = 0.9695 | SRCC = 0.9689 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1884 | LCC = 0.9699 | SRCC = 0.9695 ]