Upload SSL Layer MOS checkpoints (part 10)

results_bvcc/w2v2_large_xlsr53/layer20/run2/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5d14589be817c0d74382b4c45df57ba4f96a2d9a394b0defdfa20c7967059784
+size 3403339

results_bvcc/w2v2_large_xlsr53/layer20/run2/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:31e283b49ce56846e5ea60201716f2775e95f09dbc524bb890f76e317df5e738
+size 3404998

results_bvcc/w2v2_large_xlsr53/layer20/run2/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.5037704167457727
+INFO:root:
+[Valid][0][UTT][ MSE = 0.3710 | LCC = 0.8137 | SRCC = 0.8130 ]
+INFO:root:Average loss=0.3218344642947882
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3017 | LCC = 0.8344 | SRCC = 0.8325 ]
+INFO:root:Average loss=0.28683051887231
+INFO:root:
+[Valid][2][UTT][ MSE = 0.3121 | LCC = 0.8354 | SRCC = 0.8332 ]
+INFO:root:Average loss=0.2438610845651382
+INFO:root:
+[Valid][3][UTT][ MSE = 0.3209 | LCC = 0.8407 | SRCC = 0.8401 ]
+INFO:root:Average loss=0.24924463224716675
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2590 | LCC = 0.8458 | SRCC = 0.8458 ]
+INFO:root:Average loss=0.2369476317977294
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2576 | LCC = 0.8437 | SRCC = 0.8408 ]
+INFO:root:Average loss=0.22474074516540918
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2660 | LCC = 0.8524 | SRCC = 0.8500 ]
+INFO:root:Average loss=0.2444543399107762
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2875 | LCC = 0.8515 | SRCC = 0.8516 ]
+INFO:root:Average loss=0.20904724892133322
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2772 | LCC = 0.8513 | SRCC = 0.8500 ]
+INFO:root:Average loss=0.22854098400626427
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2458 | LCC = 0.8576 | SRCC = 0.8539 ]
+INFO:root:Average loss=0.20622509364516306
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2966 | LCC = 0.8542 | SRCC = 0.8541 ]
+INFO:root:Average loss=0.24844215485530022
+INFO:root:
+[Valid][11][UTT][ MSE = 0.4012 | LCC = 0.8432 | SRCC = 0.8474 ]
+INFO:root:Average loss=0.23113030615525368
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2578 | LCC = 0.8575 | SRCC = 0.8540 ]
+INFO:root:Average loss=0.20823164876454917
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2423 | LCC = 0.8570 | SRCC = 0.8573 ]
+INFO:root:Average loss=0.19476968174179396
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2662 | LCC = 0.8567 | SRCC = 0.8562 ]
+INFO:root:Average loss=0.20127781767111558
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2434 | LCC = 0.8608 | SRCC = 0.8588 ]
+INFO:root:Average loss=0.1766820180301483
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2376 | LCC = 0.8623 | SRCC = 0.8589 ]
+INFO:root:Average loss=0.17697289051153722
+INFO:root:
+[Valid][17][UTT][ MSE = 0.2839 | LCC = 0.8550 | SRCC = 0.8540 ]
+INFO:root:Average loss=0.17354973529775938
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2558 | LCC = 0.8608 | SRCC = 0.8563 ]
+INFO:root:Average loss=0.17916993128183561
+INFO:root:
+[Valid][19][UTT][ MSE = 0.2381 | LCC = 0.8615 | SRCC = 0.8606 ]
+INFO:root:Average loss=0.17042449068946716
+INFO:root:
+[Valid][20][UTT][ MSE = 0.3470 | LCC = 0.8488 | SRCC = 0.8525 ]
+INFO:root:Average loss=0.1760022763449412
+INFO:root:
+[Valid][21][UTT][ MSE = 0.2639 | LCC = 0.8549 | SRCC = 0.8551 ]
+INFO:root:Average loss=0.16853642702484742
+INFO:root:
+[Valid][22][UTT][ MSE = 0.2303 | LCC = 0.8604 | SRCC = 0.8592 ]
+INFO:root:Average loss=0.15448664902494505
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2323 | LCC = 0.8632 | SRCC = 0.8597 ]
+INFO:root:Average loss=0.15746389988523263
+INFO:root:
+[Valid][24][UTT][ MSE = 0.3181 | LCC = 0.8536 | SRCC = 0.8569 ]
+INFO:root:Average loss=0.15692691256602606
+INFO:root:
+[Valid][25][UTT][ MSE = 0.2275 | LCC = 0.8571 | SRCC = 0.8586 ]
+INFO:root:Average loss=0.14679662816417524
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2730 | LCC = 0.8571 | SRCC = 0.8559 ]
+INFO:root:Average loss=0.13968717698485422
+INFO:root:
+[Valid][27][UTT][ MSE = 0.2873 | LCC = 0.8537 | SRCC = 0.8529 ]
+INFO:root:Average loss=0.14161112551123667
+INFO:root:
+[Valid][28][UTT][ MSE = 0.2696 | LCC = 0.8556 | SRCC = 0.8553 ]
+INFO:root:Average loss=0.13464587267774802
+INFO:root:
+[Valid][29][UTT][ MSE = 0.2552 | LCC = 0.8610 | SRCC = 0.8577 ]
+INFO:root:Average loss=0.134915314328212
+INFO:root:
+[Valid][30][UTT][ MSE = 0.2346 | LCC = 0.8566 | SRCC = 0.8554 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2229 | LCC = 0.8692 | SRCC = 0.8727 ]

results_bvcc/w2v2_large_xlsr53/layer20/run3/config.gin

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

results_bvcc/w2v2_large_xlsr53/layer20/run3/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7b5850ead9cca3ebc32292cd021f6c60bc35c88612ac958069def7a13b4be03
+size 3403339

results_bvcc/w2v2_large_xlsr53/layer20/run3/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:48420bd691f9a26d2a20781ee333fe6be696d58cbe7c68542d5149ad2b63fc7c
+size 3404998

results_bvcc/w2v2_large_xlsr53/layer20/run3/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.5425823359535291
+INFO:root:
+[Valid][0][UTT][ MSE = 0.3698 | LCC = 0.8129 | SRCC = 0.8186 ]
+INFO:root:Average loss=0.32132799713275373
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3541 | LCC = 0.8196 | SRCC = 0.8255 ]
+INFO:root:Average loss=0.2719362173707057
+INFO:root:
+[Valid][2][UTT][ MSE = 0.3146 | LCC = 0.8396 | SRCC = 0.8391 ]
+INFO:root:Average loss=0.2679524010954759
+INFO:root:
+[Valid][3][UTT][ MSE = 0.3305 | LCC = 0.8328 | SRCC = 0.8417 ]
+INFO:root:Average loss=0.25979573260515165
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2647 | LCC = 0.8423 | SRCC = 0.8431 ]
+INFO:root:Average loss=0.2486117742000482
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2708 | LCC = 0.8522 | SRCC = 0.8505 ]
+INFO:root:Average loss=0.24040728425368285
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2817 | LCC = 0.8404 | SRCC = 0.8411 ]
+INFO:root:Average loss=0.22360487377796417
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2636 | LCC = 0.8546 | SRCC = 0.8516 ]
+INFO:root:Average loss=0.2194032772229268
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2495 | LCC = 0.8537 | SRCC = 0.8512 ]
+INFO:root:Average loss=0.2205862891024504
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2866 | LCC = 0.8585 | SRCC = 0.8552 ]
+INFO:root:Average loss=0.20756436807987017
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2407 | LCC = 0.8547 | SRCC = 0.8531 ]
+INFO:root:Average loss=0.2210563491934385
+INFO:root:
+[Valid][11][UTT][ MSE = 0.2822 | LCC = 0.8585 | SRCC = 0.8546 ]
+INFO:root:Average loss=0.20477527876694998
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2332 | LCC = 0.8603 | SRCC = 0.8590 ]
+INFO:root:Average loss=0.19850699737285957
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2565 | LCC = 0.8544 | SRCC = 0.8524 ]
+INFO:root:Average loss=0.1866154365050487
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2960 | LCC = 0.8608 | SRCC = 0.8589 ]
+INFO:root:Average loss=0.18801097056040397
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2596 | LCC = 0.8573 | SRCC = 0.8543 ]
+INFO:root:Average loss=0.1865861190435214
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2432 | LCC = 0.8593 | SRCC = 0.8560 ]
+INFO:root:Average loss=0.19375459047464225
+INFO:root:
+[Valid][17][UTT][ MSE = 0.2574 | LCC = 0.8593 | SRCC = 0.8575 ]
+INFO:root:Average loss=0.18797000516683626
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2607 | LCC = 0.8636 | SRCC = 0.8605 ]
+INFO:root:Average loss=0.1724727220642261
+INFO:root:
+[Valid][19][UTT][ MSE = 0.2477 | LCC = 0.8620 | SRCC = 0.8591 ]
+INFO:root:Average loss=0.16185139960203415
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2719 | LCC = 0.8507 | SRCC = 0.8535 ]
+INFO:root:Average loss=0.1838340662801877
+INFO:root:
+[Valid][21][UTT][ MSE = 0.2447 | LCC = 0.8636 | SRCC = 0.8593 ]
+INFO:root:Average loss=0.17008591290467825
+INFO:root:
+[Valid][22][UTT][ MSE = 0.3163 | LCC = 0.8507 | SRCC = 0.8530 ]
+INFO:root:Average loss=0.16352493315935135
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2311 | LCC = 0.8599 | SRCC = 0.8568 ]
+INFO:root:Average loss=0.15196033337941536
+INFO:root:
+[Valid][24][UTT][ MSE = 0.2397 | LCC = 0.8593 | SRCC = 0.8537 ]
+INFO:root:Average loss=0.14883078491458526
+INFO:root:
+[Valid][25][UTT][ MSE = 0.2508 | LCC = 0.8544 | SRCC = 0.8538 ]
+INFO:root:Average loss=0.15664823410602716
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2321 | LCC = 0.8595 | SRCC = 0.8555 ]
+INFO:root:Average loss=0.1461077043070243
+INFO:root:
+[Valid][27][UTT][ MSE = 0.2510 | LCC = 0.8575 | SRCC = 0.8537 ]
+INFO:root:Average loss=0.13786816090727463
+INFO:root:
+[Valid][28][UTT][ MSE = 0.2476 | LCC = 0.8555 | SRCC = 0.8539 ]
+INFO:root:Average loss=0.13456128919735932
+INFO:root:
+[Valid][29][UTT][ MSE = 0.2449 | LCC = 0.8544 | SRCC = 0.8482 ]
+INFO:root:Average loss=0.13000222629843614
+INFO:root:
+[Valid][30][UTT][ MSE = 0.2385 | LCC = 0.8561 | SRCC = 0.8531 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2623 | LCC = 0.8731 | SRCC = 0.8742 ]

results_bvcc/w2v2_large_xlsr53/layer20/run4/config.gin

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

results_bvcc/w2v2_large_xlsr53/layer20/run4/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c50488e9e8b9f83e3743a1e0aa60b31d7e3c38e45c421b06da3c734e2743a1f1
+size 3403339

results_bvcc/w2v2_large_xlsr53/layer20/run4/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7f557d716907f5344d87c1ad00ae44e9c99276ded924802742444e4ad13ddeb2
+size 3404998

results_bvcc/w2v2_large_xlsr53/layer20/run4/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.5252037472449816
+INFO:root:
+[Valid][0][UTT][ MSE = 0.3260 | LCC = 0.8098 | SRCC = 0.8132 ]
+INFO:root:Average loss=0.3006973887483279
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2792 | LCC = 0.8349 | SRCC = 0.8337 ]
+INFO:root:Average loss=0.28456209275202876
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2737 | LCC = 0.8432 | SRCC = 0.8399 ]
+INFO:root:Average loss=0.2644761571517357
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2677 | LCC = 0.8417 | SRCC = 0.8395 ]
+INFO:root:Average loss=0.2482989699794696
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2976 | LCC = 0.8481 | SRCC = 0.8462 ]
+INFO:root:Average loss=0.23510840726204407
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2639 | LCC = 0.8520 | SRCC = 0.8486 ]
+INFO:root:Average loss=0.26927776319476276
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2799 | LCC = 0.8532 | SRCC = 0.8506 ]
+INFO:root:Average loss=0.22184919241147164
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2623 | LCC = 0.8512 | SRCC = 0.8503 ]
+INFO:root:Average loss=0.22584270418454439
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2872 | LCC = 0.8481 | SRCC = 0.8487 ]
+INFO:root:Average loss=0.213427450412359
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2583 | LCC = 0.8548 | SRCC = 0.8506 ]
+INFO:root:Average loss=0.20632134740933394
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2672 | LCC = 0.8518 | SRCC = 0.8513 ]
+INFO:root:Average loss=0.20528878816045248
+INFO:root:
+[Valid][11][UTT][ MSE = 0.2751 | LCC = 0.8532 | SRCC = 0.8495 ]
+INFO:root:Average loss=0.20711010006757882
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2598 | LCC = 0.8545 | SRCC = 0.8502 ]
+INFO:root:Average loss=0.20164973871448102
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2799 | LCC = 0.8524 | SRCC = 0.8528 ]
+INFO:root:Average loss=0.1846661360409015
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2438 | LCC = 0.8565 | SRCC = 0.8528 ]
+INFO:root:Average loss=0.18137922004247323
+INFO:root:
+[Valid][15][UTT][ MSE = 0.3013 | LCC = 0.8507 | SRCC = 0.8487 ]
+INFO:root:Average loss=0.19190689692130455
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2794 | LCC = 0.8563 | SRCC = 0.8547 ]
+INFO:root:Average loss=0.18160534296662378
+INFO:root:
+[Valid][17][UTT][ MSE = 0.2580 | LCC = 0.8580 | SRCC = 0.8546 ]
+INFO:root:Average loss=0.16761686232609627
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2677 | LCC = 0.8524 | SRCC = 0.8545 ]
+INFO:root:Average loss=0.17268649517343596
+INFO:root:
+[Valid][19][UTT][ MSE = 0.2744 | LCC = 0.8553 | SRCC = 0.8557 ]
+INFO:root:Average loss=0.16771187575963828
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2467 | LCC = 0.8589 | SRCC = 0.8576 ]
+INFO:root:Average loss=0.16012380635127044
+INFO:root:
+[Valid][21][UTT][ MSE = 0.2472 | LCC = 0.8602 | SRCC = 0.8587 ]
+INFO:root:Average loss=0.1710203097989926
+INFO:root:
+[Valid][22][UTT][ MSE = 0.2308 | LCC = 0.8568 | SRCC = 0.8537 ]
+INFO:root:Average loss=0.15017850888081086
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2880 | LCC = 0.8546 | SRCC = 0.8551 ]
+INFO:root:Average loss=0.14663071156694338
+INFO:root:
+[Valid][24][UTT][ MSE = 0.2700 | LCC = 0.8556 | SRCC = 0.8555 ]
+INFO:root:Average loss=0.14466455894020888
+INFO:root:
+[Valid][25][UTT][ MSE = 0.2646 | LCC = 0.8584 | SRCC = 0.8580 ]
+INFO:root:Average loss=0.1419999205913299
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2210 | LCC = 0.8589 | SRCC = 0.8565 ]
+INFO:root:Average loss=0.14327592144791895
+INFO:root:
+[Valid][27][UTT][ MSE = 0.2841 | LCC = 0.8573 | SRCC = 0.8544 ]
+INFO:root:Average loss=0.13189556769644603
+INFO:root:
+[Valid][28][UTT][ MSE = 0.2464 | LCC = 0.8553 | SRCC = 0.8540 ]
+INFO:root:Average loss=0.12708164933018196
+INFO:root:
+[Valid][29][UTT][ MSE = 0.2558 | LCC = 0.8581 | SRCC = 0.8559 ]
+INFO:root:Average loss=0.1344688210922938
+INFO:root:
+[Valid][30][UTT][ MSE = 0.2375 | LCC = 0.8540 | SRCC = 0.8516 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2276 | LCC = 0.8744 | SRCC = 0.8738 ]

results_bvcc/w2v2_large_xlsr53/layer22/run0/config.gin

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

results_bvcc/w2v2_large_xlsr53/layer22/run0/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1a56f815795e931cc454630646df20dca752d5208210c5deb5516fcfe734a0c
+size 3403339

results_bvcc/w2v2_large_xlsr53/layer22/run0/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a8e5f07ead89a4d410d80619c52e4a0485ab0e3c54493edec80831b864f3770d
+size 3405062

results_bvcc/w2v2_large_xlsr53/layer22/run0/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.814239657842196
+INFO:root:
+[Valid][0][UTT][ MSE = 0.8270 | LCC = -0.0026 | SRCC = 0.0041 ]
+INFO:root:Average loss=0.8085363232172452
+INFO:root:
+[Valid][1][UTT][ MSE = 0.8259 | LCC = 0.0493 | SRCC = 0.0578 ]
+INFO:root:Average loss=0.8069201738406451
+INFO:root:
+[Valid][2][UTT][ MSE = 0.8247 | LCC = 0.1962 | SRCC = 0.1724 ]
+INFO:root:Average loss=0.804259663208937
+INFO:root:
+[Valid][3][UTT][ MSE = 0.8243 | LCC = 0.1360 | SRCC = 0.0895 ]
+INFO:root:Average loss=0.7963876479711288
+INFO:root:
+[Valid][4][UTT][ MSE = 0.8091 | LCC = 0.2472 | SRCC = 0.2498 ]
+INFO:root:Average loss=0.7853928804397583
+INFO:root:
+[Valid][5][UTT][ MSE = 0.7979 | LCC = 0.2328 | SRCC = 0.2393 ]
+INFO:root:Average loss=0.7715532069022839
+INFO:root:
+[Valid][6][UTT][ MSE = 0.7915 | LCC = 0.2385 | SRCC = 0.2449 ]
+INFO:root:Average loss=0.7595658860145471
+INFO:root:
+[Valid][7][UTT][ MSE = 0.7796 | LCC = 0.2547 | SRCC = 0.2535 ]
+INFO:root:Average loss=0.7552185222888604
+INFO:root:
+[Valid][8][UTT][ MSE = 0.7782 | LCC = 0.2615 | SRCC = 0.2638 ]
+INFO:root:Average loss=0.7525569139382778
+INFO:root:
+[Valid][9][UTT][ MSE = 0.7804 | LCC = 0.2499 | SRCC = 0.2511 ]
+INFO:root:Average loss=0.7490798051540668
+INFO:root:
+[Valid][10][UTT][ MSE = 0.7807 | LCC = 0.2506 | SRCC = 0.2477 ]
+INFO:root:Average loss=0.7523076916352297
+INFO:root:
+[Valid][11][UTT][ MSE = 0.7763 | LCC = 0.2525 | SRCC = 0.2499 ]
+INFO:root:Average loss=0.7473140374208108
+INFO:root:
+[Valid][12][UTT][ MSE = 0.7775 | LCC = 0.2565 | SRCC = 0.2576 ]
+INFO:root:Average loss=0.7457434053604419
+INFO:root:
+[Valid][13][UTT][ MSE = 0.7724 | LCC = 0.2590 | SRCC = 0.2566 ]
+INFO:root:Average loss=0.7413836671755865
+INFO:root:
+[Valid][14][UTT][ MSE = 0.7757 | LCC = 0.2561 | SRCC = 0.2580 ]
+INFO:root:Average loss=0.746223298402933
+INFO:root:
+[Valid][15][UTT][ MSE = 0.7791 | LCC = 0.2531 | SRCC = 0.2559 ]
+INFO:root:Average loss=0.7420797638404064
+INFO:root:
+[Valid][16][UTT][ MSE = 0.7749 | LCC = 0.2586 | SRCC = 0.2605 ]
+INFO:root:Average loss=0.7429421719832298
+INFO:root:
+[Valid][17][UTT][ MSE = 0.7730 | LCC = 0.2617 | SRCC = 0.2634 ]
+INFO:root:Average loss=0.7376499966933177
+INFO:root:
+[Valid][18][UTT][ MSE = 0.7722 | LCC = 0.2626 | SRCC = 0.2612 ]
+INFO:root:Average loss=0.7371719991549467
+INFO:root:
+[Valid][19][UTT][ MSE = 0.7723 | LCC = 0.2641 | SRCC = 0.2609 ]
+INFO:root:Average loss=0.7375246989421356
+INFO:root:
+[Valid][20][UTT][ MSE = 0.7737 | LCC = 0.2596 | SRCC = 0.2569 ]
+INFO:root:Average loss=0.7357758390597808
+INFO:root:
+[Valid][21][UTT][ MSE = 0.7709 | LCC = 0.2627 | SRCC = 0.2593 ]
+INFO:root:Average loss=0.7394503248043549
+INFO:root:
+[Valid][22][UTT][ MSE = 0.7721 | LCC = 0.2609 | SRCC = 0.2602 ]
+INFO:root:Average loss=0.7379284046399288
+INFO:root:
+[Valid][23][UTT][ MSE = 0.7725 | LCC = 0.2575 | SRCC = 0.2581 ]
+INFO:root:Average loss=0.7340310563643774
+INFO:root:
+[Valid][24][UTT][ MSE = 0.7736 | LCC = 0.2577 | SRCC = 0.2566 ]
+INFO:root:Average loss=0.732508102288613
+INFO:root:
+[Valid][25][UTT][ MSE = 0.7724 | LCC = 0.2564 | SRCC = 0.2555 ]
+INFO:root:Average loss=0.7312567891218723
+INFO:root:
+[Valid][26][UTT][ MSE = 0.7739 | LCC = 0.2568 | SRCC = 0.2540 ]
+INFO:root:Average loss=0.7326350448987423
+INFO:root:
+[Valid][27][UTT][ MSE = 0.7737 | LCC = 0.2552 | SRCC = 0.2567 ]
+INFO:root:Average loss=0.7306141421580926
+INFO:root:
+[Valid][28][UTT][ MSE = 0.7751 | LCC = 0.2552 | SRCC = 0.2572 ]
+INFO:root:Average loss=0.7304448515940936
+INFO:root:
+[Valid][29][UTT][ MSE = 0.7763 | LCC = 0.2522 | SRCC = 0.2621 ]
+INFO:root:Average loss=0.7258755343082623
+INFO:root:
+[Valid][30][UTT][ MSE = 0.7713 | LCC = 0.2600 | SRCC = 0.2612 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.8067 | LCC = 0.2564 | SRCC = 0.2638 ]

results_bvcc/w2v2_large_xlsr53/layer22/run1/config.gin

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

results_bvcc/w2v2_large_xlsr53/layer22/run1/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2a4207c575fda017ff98028b9abcd06aa3d9d89765d6e50c7f88d11e2c0f6f64
+size 3403339

results_bvcc/w2v2_large_xlsr53/layer22/run1/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:13807fed47bdbbf469ba629fdbafecfe255caee85de995c0a35f4000aaa2fed2
+size 3405062

results_bvcc/w2v2_large_xlsr53/layer22/run1/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.8181307224126962
+INFO:root:
+[Valid][0][UTT][ MSE = 0.8383 | LCC = 0.1263 | SRCC = 0.1136 ]
+INFO:root:Average loss=0.8058509589769901
+INFO:root:
+[Valid][1][UTT][ MSE = 0.8313 | LCC = 0.1621 | SRCC = 0.1426 ]
+INFO:root:Average loss=0.8050016615635309
+INFO:root:
+[Valid][2][UTT][ MSE = 0.8302 | LCC = 0.1852 | SRCC = 0.1710 ]
+INFO:root:Average loss=0.8022013230201526
+INFO:root:
+[Valid][3][UTT][ MSE = 0.8238 | LCC = 0.2060 | SRCC = 0.2061 ]
+INFO:root:Average loss=0.7932518109297141
+INFO:root:
+[Valid][4][UTT][ MSE = 0.8095 | LCC = 0.2203 | SRCC = 0.2173 ]
+INFO:root:Average loss=0.7824825889024979
+INFO:root:
+[Valid][5][UTT][ MSE = 0.8238 | LCC = 0.2229 | SRCC = 0.2354 ]
+INFO:root:Average loss=0.7676528906210874
+INFO:root:
+[Valid][6][UTT][ MSE = 0.7966 | LCC = 0.2300 | SRCC = 0.2333 ]
+INFO:root:Average loss=0.7637531146025046
+INFO:root:
+[Valid][7][UTT][ MSE = 0.7890 | LCC = 0.2374 | SRCC = 0.2388 ]
+INFO:root:Average loss=0.7614126335351895
+INFO:root:
+[Valid][8][UTT][ MSE = 0.7893 | LCC = 0.2364 | SRCC = 0.2401 ]
+INFO:root:Average loss=0.7546045344609481
+INFO:root:
+[Valid][9][UTT][ MSE = 0.7927 | LCC = 0.2397 | SRCC = 0.2433 ]
+INFO:root:Average loss=0.7521014488660372
+INFO:root:
+[Valid][10][UTT][ MSE = 0.7971 | LCC = 0.2430 | SRCC = 0.2452 ]
+INFO:root:Average loss=0.7542945650907663
+INFO:root:
+[Valid][11][UTT][ MSE = 0.8035 | LCC = 0.2403 | SRCC = 0.2467 ]
+INFO:root:Average loss=0.7496093186812524
+INFO:root:
+[Valid][12][UTT][ MSE = 0.8264 | LCC = 0.2363 | SRCC = 0.2470 ]
+INFO:root:Average loss=0.7481262401128427
+INFO:root:
+[Valid][13][UTT][ MSE = 0.8146 | LCC = 0.2366 | SRCC = 0.2440 ]
+INFO:root:Average loss=0.7483717394180787
+INFO:root:
+[Valid][14][UTT][ MSE = 0.7988 | LCC = 0.2392 | SRCC = 0.2478 ]
+INFO:root:Average loss=0.7470266200028933
+INFO:root:
+[Valid][15][UTT][ MSE = 0.7773 | LCC = 0.2472 | SRCC = 0.2479 ]
+INFO:root:Average loss=0.7472251134041028
+INFO:root:
+[Valid][16][UTT][ MSE = 0.8218 | LCC = 0.2336 | SRCC = 0.2444 ]
+INFO:root:Average loss=0.7448329543456053
+INFO:root:
+[Valid][17][UTT][ MSE = 0.8404 | LCC = 0.2386 | SRCC = 0.2471 ]
+INFO:root:Average loss=0.7468257385950822
+INFO:root:
+[Valid][18][UTT][ MSE = 0.8102 | LCC = 0.2413 | SRCC = 0.2495 ]
+INFO:root:Average loss=0.7398762595959198
+INFO:root:
+[Valid][19][UTT][ MSE = 0.7956 | LCC = 0.2506 | SRCC = 0.2513 ]
+INFO:root:Average loss=0.739950743623269
+INFO:root:
+[Valid][20][UTT][ MSE = 0.7798 | LCC = 0.2558 | SRCC = 0.2576 ]
+INFO:root:Average loss=0.74148706977184
+INFO:root:
+[Valid][21][UTT][ MSE = 0.7825 | LCC = 0.2547 | SRCC = 0.2602 ]
+INFO:root:Average loss=0.7412474002593603
+INFO:root:
+[Valid][22][UTT][ MSE = 0.8015 | LCC = 0.2487 | SRCC = 0.2586 ]
+INFO:root:Average loss=0.7376955036933606
+INFO:root:
+[Valid][23][UTT][ MSE = 0.7965 | LCC = 0.2547 | SRCC = 0.2586 ]
+INFO:root:Average loss=0.7389899431130825
+INFO:root:
+[Valid][24][UTT][ MSE = 0.8086 | LCC = 0.2474 | SRCC = 0.2556 ]
+INFO:root:Average loss=0.7341073384651771
+INFO:root:
+[Valid][25][UTT][ MSE = 0.7827 | LCC = 0.2564 | SRCC = 0.2639 ]
+INFO:root:Average loss=0.7389515500802261
+INFO:root:
+[Valid][26][UTT][ MSE = 0.7884 | LCC = 0.2528 | SRCC = 0.2571 ]
+INFO:root:Average loss=0.7339698744890018
+INFO:root:
+[Valid][27][UTT][ MSE = 0.7711 | LCC = 0.2622 | SRCC = 0.2609 ]
+INFO:root:Average loss=0.737748962182265
+INFO:root:
+[Valid][28][UTT][ MSE = 0.7770 | LCC = 0.2527 | SRCC = 0.2605 ]
+INFO:root:Average loss=0.7384612575555459
+INFO:root:
+[Valid][29][UTT][ MSE = 0.7779 | LCC = 0.2590 | SRCC = 0.2643 ]
+INFO:root:Average loss=0.7393181075652441
+INFO:root:
+[Valid][30][UTT][ MSE = 0.8223 | LCC = 0.2459 | SRCC = 0.2583 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.7996 | LCC = 0.2422 | SRCC = 0.2504 ]

results_bvcc/w2v2_large_xlsr53/layer22/run2/config.gin

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

results_bvcc/w2v2_large_xlsr53/layer22/run2/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:600e13b24b4df5a9bad0859914464388871330115787498ebd9cf9ef13c194bc
+size 3403339

results_bvcc/w2v2_large_xlsr53/layer22/run2/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bc6b697fc358ef048a7d3407572dcc3c9f3ee6f595b36f5cc664485cc6f7138c
+size 3404998

results_bvcc/w2v2_large_xlsr53/layer22/run2/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.8170791971377838
+INFO:root:
+[Valid][0][UTT][ MSE = 0.8325 | LCC = 0.1026 | SRCC = 0.1114 ]
+INFO:root:Average loss=0.8104762167502673
+INFO:root:
+[Valid][1][UTT][ MSE = 0.8263 | LCC = -0.0209 | SRCC = -0.0027 ]
+INFO:root:Average loss=0.8071949627154913
+INFO:root:
+[Valid][2][UTT][ MSE = 0.8266 | LCC = 0.0121 | SRCC = 0.0360 ]
+INFO:root:Average loss=0.8096001698420598
+INFO:root:
+[Valid][3][UTT][ MSE = 0.8256 | LCC = 0.0345 | SRCC = 0.0465 ]
+INFO:root:Average loss=0.8068318619177892
+INFO:root:
+[Valid][4][UTT][ MSE = 0.8247 | LCC = 0.0963 | SRCC = 0.1003 ]
+INFO:root:Average loss=0.8068208495775858
+INFO:root:
+[Valid][5][UTT][ MSE = 0.8289 | LCC = 0.1201 | SRCC = 0.1209 ]
+INFO:root:Average loss=0.8082066774368286
+INFO:root:
+[Valid][6][UTT][ MSE = 0.8299 | LCC = 0.1215 | SRCC = 0.1197 ]
+INFO:root:Average loss=0.7999909252692492
+INFO:root:
+[Valid][7][UTT][ MSE = 0.8185 | LCC = 0.2234 | SRCC = 0.2272 ]
+INFO:root:Average loss=0.7974462914161193
+INFO:root:
+[Valid][8][UTT][ MSE = 0.8092 | LCC = 0.2301 | SRCC = 0.2401 ]
+INFO:root:Average loss=0.7786302451904004
+INFO:root:
+[Valid][9][UTT][ MSE = 0.8018 | LCC = 0.2356 | SRCC = 0.2532 ]
+INFO:root:Average loss=0.7637408131208175
+INFO:root:
+[Valid][10][UTT][ MSE = 0.8020 | LCC = 0.2301 | SRCC = 0.2432 ]
+INFO:root:Average loss=0.7688980881984417
+INFO:root:
+[Valid][11][UTT][ MSE = 0.7847 | LCC = 0.2382 | SRCC = 0.2453 ]
+INFO:root:Average loss=0.7564181311008258
+INFO:root:
+[Valid][12][UTT][ MSE = 0.7874 | LCC = 0.2371 | SRCC = 0.2423 ]
+INFO:root:Average loss=0.7545683070635184
+INFO:root:
+[Valid][13][UTT][ MSE = 0.8091 | LCC = 0.2361 | SRCC = 0.2466 ]
+INFO:root:Average loss=0.7512389161647894
+INFO:root:
+[Valid][14][UTT][ MSE = 0.7793 | LCC = 0.2450 | SRCC = 0.2449 ]
+INFO:root:Average loss=0.7586557078055847
+INFO:root:
+[Valid][15][UTT][ MSE = 0.7864 | LCC = 0.2386 | SRCC = 0.2431 ]
+INFO:root:Average loss=0.7566420504680047
+INFO:root:
+[Valid][16][UTT][ MSE = 0.7858 | LCC = 0.2423 | SRCC = 0.2426 ]
+INFO:root:Average loss=0.7498466303715339
+INFO:root:
+[Valid][17][UTT][ MSE = 0.7781 | LCC = 0.2444 | SRCC = 0.2435 ]
+INFO:root:Average loss=0.7462894381620945
+INFO:root:
+[Valid][18][UTT][ MSE = 0.8138 | LCC = 0.2368 | SRCC = 0.2465 ]
+INFO:root:Average loss=0.7527802273249015
+INFO:root:
+[Valid][19][UTT][ MSE = 0.7938 | LCC = 0.2362 | SRCC = 0.2415 ]
+INFO:root:Average loss=0.7473375663543359
+INFO:root:
+[Valid][20][UTT][ MSE = 0.7803 | LCC = 0.2437 | SRCC = 0.2459 ]
+INFO:root:Average loss=0.7387752280785487
+INFO:root:
+[Valid][21][UTT][ MSE = 0.7824 | LCC = 0.2478 | SRCC = 0.2484 ]
+INFO:root:Average loss=0.7506021536313571
+INFO:root:
+[Valid][22][UTT][ MSE = 0.7798 | LCC = 0.2467 | SRCC = 0.2469 ]
+INFO:root:Average loss=0.742162606655023
+INFO:root:
+[Valid][23][UTT][ MSE = 0.7770 | LCC = 0.2476 | SRCC = 0.2459 ]
+INFO:root:Average loss=0.7410742491483688
+INFO:root:
+[Valid][24][UTT][ MSE = 0.7751 | LCC = 0.2503 | SRCC = 0.2458 ]
+INFO:root:Average loss=0.7445911574058044
+INFO:root:
+[Valid][25][UTT][ MSE = 0.7795 | LCC = 0.2432 | SRCC = 0.2463 ]
+INFO:root:Average loss=0.7379241952529321
+INFO:root:
+[Valid][26][UTT][ MSE = 0.7901 | LCC = 0.2368 | SRCC = 0.2378 ]
+INFO:root:Average loss=0.7360416750113169
+INFO:root:
+[Valid][27][UTT][ MSE = 0.7746 | LCC = 0.2503 | SRCC = 0.2480 ]
+INFO:root:Average loss=0.7346521952213385
+INFO:root:
+[Valid][28][UTT][ MSE = 0.7962 | LCC = 0.2415 | SRCC = 0.2478 ]
+INFO:root:Average loss=0.7360071792052343
+INFO:root:
+[Valid][29][UTT][ MSE = 0.7771 | LCC = 0.2491 | SRCC = 0.2513 ]
+INFO:root:Average loss=0.7348292752718314
+INFO:root:
+[Valid][30][UTT][ MSE = 0.7890 | LCC = 0.2410 | SRCC = 0.2432 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.7907 | LCC = 0.2665 | SRCC = 0.2687 ]

results_bvcc/w2v2_large_xlsr53/layer22/run3/config.gin

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

results_bvcc/w2v2_large_xlsr53/layer22/run3/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9d90223c11566e5b9366d1f191619706c6868e92086587ede254eea29795319c
+size 3403339

results_bvcc/w2v2_large_xlsr53/layer22/run3/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5055a81939fd491e91cffaa54a4a2efad6ceca33ada7ff52dff767a619970f25
+size 3405062

results_bvcc/w2v2_large_xlsr53/layer22/run3/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.8139951924482981
+INFO:root:
+[Valid][0][UTT][ MSE = 0.8538 | LCC = 0.0628 | SRCC = 0.0433 ]
+INFO:root:Average loss=0.8084817964297074
+INFO:root:
+[Valid][1][UTT][ MSE = 0.8346 | LCC = 0.1495 | SRCC = 0.1408 ]
+INFO:root:Average loss=0.8109443149505517
+INFO:root:
+[Valid][2][UTT][ MSE = 0.8399 | LCC = 0.1735 | SRCC = 0.1573 ]
+INFO:root:Average loss=0.8042151973797724
+INFO:root:
+[Valid][3][UTT][ MSE = 0.8378 | LCC = 0.2081 | SRCC = 0.2045 ]
+INFO:root:Average loss=0.8034162750610938
+INFO:root:
+[Valid][4][UTT][ MSE = 0.8440 | LCC = 0.2189 | SRCC = 0.2294 ]
+INFO:root:Average loss=0.807106227446825
+INFO:root:
+[Valid][5][UTT][ MSE = 0.8303 | LCC = 0.2063 | SRCC = 0.2054 ]
+INFO:root:Average loss=0.7967051993577908
+INFO:root:
+[Valid][6][UTT][ MSE = 0.8262 | LCC = 0.2212 | SRCC = 0.2231 ]
+INFO:root:Average loss=0.7873269961430476
+INFO:root:
+[Valid][7][UTT][ MSE = 0.8369 | LCC = 0.2284 | SRCC = 0.2433 ]
+INFO:root:Average loss=0.7755554616451263
+INFO:root:
+[Valid][8][UTT][ MSE = 0.8117 | LCC = 0.2389 | SRCC = 0.2573 ]
+INFO:root:Average loss=0.7619453622744634
+INFO:root:
+[Valid][9][UTT][ MSE = 0.7796 | LCC = 0.2543 | SRCC = 0.2549 ]
+INFO:root:Average loss=0.7669727420195555
+INFO:root:
+[Valid][10][UTT][ MSE = 0.7817 | LCC = 0.2582 | SRCC = 0.2596 ]
+INFO:root:Average loss=0.7573283819051889
+INFO:root:
+[Valid][11][UTT][ MSE = 0.8115 | LCC = 0.2542 | SRCC = 0.2628 ]
+INFO:root:Average loss=0.753503374564342
+INFO:root:
+[Valid][12][UTT][ MSE = 0.7973 | LCC = 0.2575 | SRCC = 0.2668 ]
+INFO:root:Average loss=0.7453485207679944
+INFO:root:
+[Valid][13][UTT][ MSE = 0.8421 | LCC = 0.2580 | SRCC = 0.2705 ]
+INFO:root:Average loss=0.7502139508724213
+INFO:root:
+[Valid][14][UTT][ MSE = 0.7675 | LCC = 0.2719 | SRCC = 0.2685 ]
+INFO:root:Average loss=0.7436514829213803
+INFO:root:
+[Valid][15][UTT][ MSE = 0.7981 | LCC = 0.2540 | SRCC = 0.2611 ]
+INFO:root:Average loss=0.7450853158266116
+INFO:root:
+[Valid][16][UTT][ MSE = 0.7642 | LCC = 0.2791 | SRCC = 0.2766 ]
+INFO:root:Average loss=0.7439048061004052
+INFO:root:
+[Valid][17][UTT][ MSE = 0.7791 | LCC = 0.2748 | SRCC = 0.2763 ]
+INFO:root:Average loss=0.7414584794105628
+INFO:root:
+[Valid][18][UTT][ MSE = 0.7793 | LCC = 0.2727 | SRCC = 0.2757 ]
+INFO:root:Average loss=0.7410066070464941
+INFO:root:
+[Valid][19][UTT][ MSE = 0.7671 | LCC = 0.2795 | SRCC = 0.2782 ]
+INFO:root:Average loss=0.7383616111981564
+INFO:root:
+[Valid][20][UTT][ MSE = 0.7835 | LCC = 0.2746 | SRCC = 0.2757 ]
+INFO:root:Average loss=0.7370491738502796
+INFO:root:
+[Valid][21][UTT][ MSE = 0.8018 | LCC = 0.2653 | SRCC = 0.2673 ]
+INFO:root:Average loss=0.7324779484516535
+INFO:root:
+[Valid][22][UTT][ MSE = 0.7692 | LCC = 0.2761 | SRCC = 0.2786 ]
+INFO:root:Average loss=0.740232072579555
+INFO:root:
+[Valid][23][UTT][ MSE = 0.7750 | LCC = 0.2804 | SRCC = 0.2792 ]
+INFO:root:Average loss=0.7392700532307992
+INFO:root:
+[Valid][24][UTT][ MSE = 0.7744 | LCC = 0.2807 | SRCC = 0.2760 ]
+INFO:root:Average loss=0.7341036876806846
+INFO:root:
+[Valid][25][UTT][ MSE = 0.7650 | LCC = 0.2814 | SRCC = 0.2765 ]
+INFO:root:Average loss=0.7312420346798041
+INFO:root:
+[Valid][26][UTT][ MSE = 0.7766 | LCC = 0.2829 | SRCC = 0.2840 ]
+INFO:root:Average loss=0.7305280428666335
+INFO:root:
+[Valid][27][UTT][ MSE = 0.7855 | LCC = 0.2772 | SRCC = 0.2782 ]
+INFO:root:Average loss=0.7277241028272189
+INFO:root:
+[Valid][28][UTT][ MSE = 0.7805 | LCC = 0.2800 | SRCC = 0.2774 ]
+INFO:root:Average loss=0.7310245992281498
+INFO:root:
+[Valid][29][UTT][ MSE = 0.7570 | LCC = 0.2933 | SRCC = 0.2889 ]
+INFO:root:Average loss=0.7278549052201785
+INFO:root:
+[Valid][30][UTT][ MSE = 0.7954 | LCC = 0.2802 | SRCC = 0.2838 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.7960 | LCC = 0.2681 | SRCC = 0.2799 ]

results_bvcc/w2v2_large_xlsr53/layer22/run4/config.gin

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

results_bvcc/w2v2_large_xlsr53/layer22/run4/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e9bb001e0a7e711f1bb9275c034935d862ccdc20710f9458f8eefba1a22f9319
+size 3403339

results_bvcc/w2v2_large_xlsr53/layer22/run4/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:439a309d13eef5a0952916c0c3681ace23ef98e55fffd31d55e0902364494747
+size 3405062

results_bvcc/w2v2_large_xlsr53/layer22/run4/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.8078734744817783
+INFO:root:
+[Valid][0][UTT][ MSE = 0.8438 | LCC = 0.1551 | SRCC = 0.1338 ]
+INFO:root:Average loss=0.8061453310342935
+INFO:root:
+[Valid][1][UTT][ MSE = 0.8527 | LCC = 0.1858 | SRCC = 0.1696 ]
+INFO:root:Average loss=0.7980850124970461
+INFO:root:
+[Valid][2][UTT][ MSE = 0.8782 | LCC = 0.2039 | SRCC = 0.2082 ]
+INFO:root:Average loss=0.7904722224443387
+INFO:root:
+[Valid][3][UTT][ MSE = 0.8657 | LCC = 0.2205 | SRCC = 0.2218 ]
+INFO:root:Average loss=0.7704049562796568
+INFO:root:
+[Valid][4][UTT][ MSE = 0.8042 | LCC = 0.2416 | SRCC = 0.2469 ]
+INFO:root:Average loss=0.7644153917447115
+INFO:root:
+[Valid][5][UTT][ MSE = 0.8250 | LCC = 0.2385 | SRCC = 0.2437 ]
+INFO:root:Average loss=0.7613810132711362
+INFO:root:
+[Valid][6][UTT][ MSE = 0.8584 | LCC = 0.2353 | SRCC = 0.2405 ]
+INFO:root:Average loss=0.7635564146897732
+INFO:root:
+[Valid][7][UTT][ MSE = 0.8602 | LCC = 0.2334 | SRCC = 0.2399 ]
+INFO:root:Average loss=0.7606048110203866
+INFO:root:
+[Valid][8][UTT][ MSE = 0.7999 | LCC = 0.2399 | SRCC = 0.2370 ]
+INFO:root:Average loss=0.7518155513665615
+INFO:root:
+[Valid][9][UTT][ MSE = 0.7966 | LCC = 0.2453 | SRCC = 0.2419 ]
+INFO:root:Average loss=0.7531586511012835
+INFO:root:
+[Valid][10][UTT][ MSE = 0.8018 | LCC = 0.2439 | SRCC = 0.2402 ]
+INFO:root:Average loss=0.7529040429836664
+INFO:root:
+[Valid][11][UTT][ MSE = 0.8175 | LCC = 0.2530 | SRCC = 0.2528 ]
+INFO:root:Average loss=0.7498894028174572
+INFO:root:
+[Valid][12][UTT][ MSE = 0.8310 | LCC = 0.2520 | SRCC = 0.2505 ]
+INFO:root:Average loss=0.7508819340131222
+INFO:root:
+[Valid][13][UTT][ MSE = 0.8223 | LCC = 0.2518 | SRCC = 0.2502 ]
+INFO:root:Average loss=0.7533572163337316
+INFO:root:
+[Valid][14][UTT][ MSE = 0.8383 | LCC = 0.2479 | SRCC = 0.2476 ]
+INFO:root:Average loss=0.7443092568562582
+INFO:root:
+[Valid][15][UTT][ MSE = 0.8035 | LCC = 0.2544 | SRCC = 0.2475 ]
+INFO:root:Average loss=0.7448368473694875
+INFO:root:
+[Valid][16][UTT][ MSE = 0.8218 | LCC = 0.2568 | SRCC = 0.2515 ]
+INFO:root:Average loss=0.748010686574838
+INFO:root:
+[Valid][17][UTT][ MSE = 0.8373 | LCC = 0.2558 | SRCC = 0.2524 ]
+INFO:root:Average loss=0.7410632696671363
+INFO:root:
+[Valid][18][UTT][ MSE = 0.8248 | LCC = 0.2561 | SRCC = 0.2532 ]
+INFO:root:Average loss=0.7445670266946157
+INFO:root:
+[Valid][19][UTT][ MSE = 0.8269 | LCC = 0.2542 | SRCC = 0.2514 ]
+INFO:root:Average loss=0.74035588823832
+INFO:root:
+[Valid][20][UTT][ MSE = 0.8644 | LCC = 0.2530 | SRCC = 0.2545 ]
+INFO:root:Average loss=0.7474632763709778
+INFO:root:
+[Valid][21][UTT][ MSE = 0.8147 | LCC = 0.2602 | SRCC = 0.2546 ]
+INFO:root:Average loss=0.7406386037667593
+INFO:root:
+[Valid][22][UTT][ MSE = 0.8054 | LCC = 0.2618 | SRCC = 0.2536 ]
+INFO:root:Average loss=0.7379722109971902
+INFO:root:
+[Valid][23][UTT][ MSE = 0.8150 | LCC = 0.2635 | SRCC = 0.2614 ]
+INFO:root:Average loss=0.7341487545233506
+INFO:root:
+[Valid][24][UTT][ MSE = 0.8073 | LCC = 0.2630 | SRCC = 0.2580 ]
+INFO:root:Average loss=0.7340960743335577
+INFO:root:
+[Valid][25][UTT][ MSE = 0.7963 | LCC = 0.2616 | SRCC = 0.2544 ]
+INFO:root:Average loss=0.7333975468690579
+INFO:root:
+[Valid][26][UTT][ MSE = 0.8081 | LCC = 0.2641 | SRCC = 0.2592 ]
+INFO:root:Average loss=0.7372842583900843
+INFO:root:
+[Valid][27][UTT][ MSE = 0.8026 | LCC = 0.2604 | SRCC = 0.2548 ]
+INFO:root:Average loss=0.7322733646784073
+INFO:root:
+[Valid][28][UTT][ MSE = 0.8087 | LCC = 0.2649 | SRCC = 0.2605 ]
+INFO:root:Average loss=0.7382883513585116
+INFO:root:
+[Valid][29][UTT][ MSE = 0.7805 | LCC = 0.2660 | SRCC = 0.2607 ]
+INFO:root:Average loss=0.7296412785847982
+INFO:root:
+[Valid][30][UTT][ MSE = 0.7854 | LCC = 0.2664 | SRCC = 0.2601 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.7988 | LCC = 0.2486 | SRCC = 0.2584 ]

results_bvcc/w2v2_large_xlsr53/layer4/run0/config.gin

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

results_bvcc/w2v2_large_xlsr53/layer4/run0/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0d17ddd1fc840a4c8efb2d94b5c79a8182ea133d0da68d1c78d08949e45d54f9
+size 3403339

results_bvcc/w2v2_large_xlsr53/layer4/run0/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c77faa9c79a76bc831692b5f9b595ac4ccb9406a66d60eb59c0a3cdc68cdec7d
+size 3405062

results_bvcc/w2v2_large_xlsr53/layer4/run0/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.547156467460669
+INFO:root:
+[Valid][0][UTT][ MSE = 0.3832 | LCC = 0.7870 | SRCC = 0.7937 ]
+INFO:root:Average loss=0.29149841994811326
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3145 | LCC = 0.8253 | SRCC = 0.8260 ]
+INFO:root:Average loss=0.24849581298155662
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2810 | LCC = 0.8434 | SRCC = 0.8447 ]
+INFO:root:Average loss=0.22738109395290032
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2621 | LCC = 0.8573 | SRCC = 0.8536 ]
+INFO:root:Average loss=0.21490304286663348
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2838 | LCC = 0.8643 | SRCC = 0.8606 ]
+INFO:root:Average loss=0.20760062088569006
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2279 | LCC = 0.8673 | SRCC = 0.8642 ]
+INFO:root:Average loss=0.20366449520373955
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2377 | LCC = 0.8735 | SRCC = 0.8702 ]
+INFO:root:Average loss=0.17576329687084907
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2163 | LCC = 0.8735 | SRCC = 0.8715 ]
+INFO:root:Average loss=0.17690044412246117
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2238 | LCC = 0.8779 | SRCC = 0.8757 ]
+INFO:root:Average loss=0.1625153881808122
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2058 | LCC = 0.8757 | SRCC = 0.8747 ]
+INFO:root:Average loss=0.16646776797297674
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2525 | LCC = 0.8692 | SRCC = 0.8712 ]
+INFO:root:Average loss=0.16911816062071386
+INFO:root:
+[Valid][11][UTT][ MSE = 0.2083 | LCC = 0.8794 | SRCC = 0.8774 ]
+INFO:root:Average loss=0.14906550905643365
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2191 | LCC = 0.8795 | SRCC = 0.8771 ]
+INFO:root:Average loss=0.14371993793891027
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2117 | LCC = 0.8791 | SRCC = 0.8769 ]
+INFO:root:Average loss=0.1384089978841635
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2030 | LCC = 0.8787 | SRCC = 0.8776 ]
+INFO:root:Average loss=0.12990750276889557
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2243 | LCC = 0.8789 | SRCC = 0.8775 ]
+INFO:root:Average loss=0.13416247537884957
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1994 | LCC = 0.8776 | SRCC = 0.8764 ]
+INFO:root:Average loss=0.1280193323126206
+INFO:root:
+[Valid][17][UTT][ MSE = 0.2107 | LCC = 0.8757 | SRCC = 0.8748 ]
+INFO:root:Average loss=0.11507375280444439
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2014 | LCC = 0.8787 | SRCC = 0.8778 ]
+INFO:root:Average loss=0.11986773652143967
+INFO:root:
+[Valid][19][UTT][ MSE = 0.2211 | LCC = 0.8760 | SRCC = 0.8761 ]
+INFO:root:Average loss=0.10640933001652741
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2026 | LCC = 0.8763 | SRCC = 0.8763 ]
+INFO:root:Average loss=0.10712833387347367
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1972 | LCC = 0.8812 | SRCC = 0.8802 ]
+INFO:root:Average loss=0.10519853802636647
+INFO:root:
+[Valid][22][UTT][ MSE = 0.2054 | LCC = 0.8747 | SRCC = 0.8757 ]
+INFO:root:Average loss=0.09490100838817082
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2048 | LCC = 0.8798 | SRCC = 0.8781 ]
+INFO:root:Average loss=0.09653763964963265
+INFO:root:
+[Valid][24][UTT][ MSE = 0.2104 | LCC = 0.8747 | SRCC = 0.8743 ]
+INFO:root:Average loss=0.09158640760832872
+INFO:root:
+[Valid][25][UTT][ MSE = 0.2154 | LCC = 0.8728 | SRCC = 0.8724 ]
+INFO:root:Average loss=0.08465714289400822
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2098 | LCC = 0.8769 | SRCC = 0.8759 ]
+INFO:root:Average loss=0.0891813850746705
+INFO:root:
+[Valid][27][UTT][ MSE = 0.2106 | LCC = 0.8787 | SRCC = 0.8780 ]
+INFO:root:Average loss=0.09102073861047244
+INFO:root:
+[Valid][28][UTT][ MSE = 0.2081 | LCC = 0.8761 | SRCC = 0.8754 ]
+INFO:root:Average loss=0.08620971317092578
+INFO:root:
+[Valid][29][UTT][ MSE = 0.2109 | LCC = 0.8755 | SRCC = 0.8746 ]
+INFO:root:Average loss=0.07545238857467969
+INFO:root:
+[Valid][30][UTT][ MSE = 0.2072 | LCC = 0.8736 | SRCC = 0.8727 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2096 | LCC = 0.8749 | SRCC = 0.8742 ]

results_bvcc/w2v2_large_xlsr53/layer4/run1/config.gin

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

results_bvcc/w2v2_large_xlsr53/layer4/run1/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cf2cd0da48458a8dad5a0439c598dc91722a10bd72eb1293d122b1b3df500b43
+size 3403339

results_bvcc/w2v2_large_xlsr53/layer4/run1/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ce6fd6a252785a6f331f9946f5b53c13f73e0ee5bc262d421079252183450b33
+size 3405062

results_bvcc/w2v2_large_xlsr53/layer4/run1/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.5605818294943907
+INFO:root:
+[Valid][0][UTT][ MSE = 0.3382 | LCC = 0.7913 | SRCC = 0.7967 ]
+INFO:root:Average loss=0.30025573762563557
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3533 | LCC = 0.8229 | SRCC = 0.8292 ]
+INFO:root:Average loss=0.2813934796513655
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2893 | LCC = 0.8451 | SRCC = 0.8444 ]
+INFO:root:Average loss=0.2493511744034596
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2820 | LCC = 0.8520 | SRCC = 0.8513 ]
+INFO:root:Average loss=0.22732358005566475
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2426 | LCC = 0.8572 | SRCC = 0.8562 ]
+INFO:root:Average loss=0.2069613333695974
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2915 | LCC = 0.8623 | SRCC = 0.8646 ]
+INFO:root:Average loss=0.19090187750183618
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2433 | LCC = 0.8679 | SRCC = 0.8667 ]
+INFO:root:Average loss=0.1918311948195482
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2164 | LCC = 0.8719 | SRCC = 0.8712 ]
+INFO:root:Average loss=0.17321445105167535
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2449 | LCC = 0.8719 | SRCC = 0.8702 ]
+INFO:root:Average loss=0.16294093773915216
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2188 | LCC = 0.8742 | SRCC = 0.8726 ]
+INFO:root:Average loss=0.16203909204938474
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2208 | LCC = 0.8726 | SRCC = 0.8716 ]
+INFO:root:Average loss=0.15740516571662364
+INFO:root:
+[Valid][11][UTT][ MSE = 0.2292 | LCC = 0.8721 | SRCC = 0.8704 ]
+INFO:root:Average loss=0.15466257461752647
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2270 | LCC = 0.8726 | SRCC = 0.8720 ]
+INFO:root:Average loss=0.14520136543955559
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2052 | LCC = 0.8769 | SRCC = 0.8781 ]
+INFO:root:Average loss=0.1345938279842719
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2260 | LCC = 0.8744 | SRCC = 0.8731 ]
+INFO:root:Average loss=0.13297921266311255
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2284 | LCC = 0.8768 | SRCC = 0.8740 ]
+INFO:root:Average loss=0.1225662044225595
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2110 | LCC = 0.8779 | SRCC = 0.8772 ]
+INFO:root:Average loss=0.11552520105854058
+INFO:root:
+[Valid][17][UTT][ MSE = 0.2273 | LCC = 0.8745 | SRCC = 0.8736 ]
+INFO:root:Average loss=0.12532876174037272
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2443 | LCC = 0.8745 | SRCC = 0.8753 ]
+INFO:root:Average loss=0.11093265993090776
+INFO:root:
+[Valid][19][UTT][ MSE = 0.2307 | LCC = 0.8756 | SRCC = 0.8744 ]
+INFO:root:Average loss=0.10863688277701537
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2385 | LCC = 0.8771 | SRCC = 0.8765 ]
+INFO:root:Average loss=0.10018666957815488
+INFO:root:
+[Valid][21][UTT][ MSE = 0.2663 | LCC = 0.8731 | SRCC = 0.8738 ]
+INFO:root:Average loss=0.10150594340685086
+INFO:root:
+[Valid][22][UTT][ MSE = 0.2035 | LCC = 0.8754 | SRCC = 0.8743 ]
+INFO:root:Average loss=0.09620927025874455
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2666 | LCC = 0.8717 | SRCC = 0.8743 ]
+INFO:root:Average loss=0.08765348242834592
+INFO:root:
+[Valid][24][UTT][ MSE = 0.2430 | LCC = 0.8761 | SRCC = 0.8768 ]
+INFO:root:Average loss=0.09197935820198976
+INFO:root:
+[Valid][25][UTT][ MSE = 0.2052 | LCC = 0.8761 | SRCC = 0.8753 ]
+INFO:root:Average loss=0.08101947677250092
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2142 | LCC = 0.8765 | SRCC = 0.8780 ]
+INFO:root:Average loss=0.07929057675676468
+INFO:root:
+[Valid][27][UTT][ MSE = 0.2166 | LCC = 0.8781 | SRCC = 0.8765 ]
+INFO:root:Average loss=0.08396360980203518
+INFO:root:
+[Valid][28][UTT][ MSE = 0.2919 | LCC = 0.8682 | SRCC = 0.8720 ]
+INFO:root:Average loss=0.074722120204033
+INFO:root:
+[Valid][29][UTT][ MSE = 0.2033 | LCC = 0.8764 | SRCC = 0.8761 ]
+INFO:root:Average loss=0.0711737318107715
+INFO:root:
+[Valid][30][UTT][ MSE = 0.2367 | LCC = 0.8755 | SRCC = 0.8754 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2166 | LCC = 0.8761 | SRCC = 0.8773 ]

results_bvcc/w2v2_large_xlsr53/layer4/run2/config.gin

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

results_bvcc/w2v2_large_xlsr53/layer4/run2/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4988351cc534c94f99f7559f933edb0081f47cafa055d2a3e17c09d9ed472b3a
+size 3403339

results_bvcc/w2v2_large_xlsr53/layer4/run2/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ababe9b3184cb9bc8887520d5ba3bf14d8176e8d09fbc642ac6cc11c9daddcd0
+size 3405062

results_bvcc/w2v2_large_xlsr53/layer4/run2/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.4532563615685854
+INFO:root:
+[Valid][0][UTT][ MSE = 0.3356 | LCC = 0.7989 | SRCC = 0.8013 ]
+INFO:root:Average loss=0.27491417641823107
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2828 | LCC = 0.8347 | SRCC = 0.8370 ]
+INFO:root:Average loss=0.2513206410102355
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2710 | LCC = 0.8484 | SRCC = 0.8482 ]
+INFO:root:Average loss=0.2428954717440483
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2799 | LCC = 0.8471 | SRCC = 0.8524 ]
+INFO:root:Average loss=0.21514941904789361
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2534 | LCC = 0.8650 | SRCC = 0.8626 ]
+INFO:root:Average loss=0.2048763797069207
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2392 | LCC = 0.8661 | SRCC = 0.8668 ]
+INFO:root:Average loss=0.19460664336115885
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2166 | LCC = 0.8683 | SRCC = 0.8700 ]
+INFO:root:Average loss=0.18061789364004746
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2214 | LCC = 0.8693 | SRCC = 0.8684 ]
+INFO:root:Average loss=0.18510188811864609
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2186 | LCC = 0.8743 | SRCC = 0.8720 ]
+INFO:root:Average loss=0.1711409091949463
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2395 | LCC = 0.8728 | SRCC = 0.8728 ]
+INFO:root:Average loss=0.1823367731502423
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2216 | LCC = 0.8695 | SRCC = 0.8729 ]
+INFO:root:Average loss=0.15918540228635836
+INFO:root:
+[Valid][11][UTT][ MSE = 0.2110 | LCC = 0.8734 | SRCC = 0.8718 ]
+INFO:root:Average loss=0.15191630302713469
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2003 | LCC = 0.8755 | SRCC = 0.8753 ]
+INFO:root:Average loss=0.15265449738273254
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2462 | LCC = 0.8677 | SRCC = 0.8726 ]
+INFO:root:Average loss=0.1382217956468081
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2130 | LCC = 0.8727 | SRCC = 0.8747 ]
+INFO:root:Average loss=0.14303912518498224
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2350 | LCC = 0.8722 | SRCC = 0.8728 ]
+INFO:root:Average loss=0.13195646449159354
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2163 | LCC = 0.8779 | SRCC = 0.8764 ]
+INFO:root:Average loss=0.12995244027712408
+INFO:root:
+[Valid][17][UTT][ MSE = 0.2148 | LCC = 0.8735 | SRCC = 0.8752 ]
+INFO:root:Average loss=0.12340931288706951
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2682 | LCC = 0.8657 | SRCC = 0.8737 ]
+INFO:root:Average loss=0.11715158571799596
+INFO:root:
+[Valid][19][UTT][ MSE = 0.2107 | LCC = 0.8726 | SRCC = 0.8745 ]
+INFO:root:Average loss=0.1106030118580048
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2178 | LCC = 0.8728 | SRCC = 0.8728 ]
+INFO:root:Average loss=0.11555583459826615
+INFO:root:
+[Valid][21][UTT][ MSE = 0.2155 | LCC = 0.8780 | SRCC = 0.8771 ]
+INFO:root:Average loss=0.10848726771580867
+INFO:root:
+[Valid][22][UTT][ MSE = 0.2078 | LCC = 0.8758 | SRCC = 0.8753 ]
+INFO:root:Average loss=0.09563501962484458
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2425 | LCC = 0.8681 | SRCC = 0.8676 ]
+INFO:root:Average loss=0.09557310367623965
+INFO:root:
+[Valid][24][UTT][ MSE = 0.2283 | LCC = 0.8726 | SRCC = 0.8758 ]
+INFO:root:Average loss=0.09602287803322841
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1953 | LCC = 0.8783 | SRCC = 0.8786 ]
+INFO:root:Average loss=0.10340599371836735
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2294 | LCC = 0.8741 | SRCC = 0.8754 ]
+INFO:root:Average loss=0.08820458687841892
+INFO:root:
+[Valid][27][UTT][ MSE = 0.2308 | LCC = 0.8733 | SRCC = 0.8737 ]
+INFO:root:Average loss=0.08390403567598416
+INFO:root:
+[Valid][28][UTT][ MSE = 0.2481 | LCC = 0.8685 | SRCC = 0.8749 ]
+INFO:root:Average loss=0.0964965601093494
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1953 | LCC = 0.8769 | SRCC = 0.8774 ]
+INFO:root:Average loss=0.08484634895546314
+INFO:root:
+[Valid][30][UTT][ MSE = 0.2053 | LCC = 0.8762 | SRCC = 0.8762 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2247 | LCC = 0.8717 | SRCC = 0.8723 ]

results_bvcc/w2v2_large_xlsr53/layer4/run3/config.gin

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

results_bvcc/w2v2_large_xlsr53/layer4/run3/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:727657e5e90f3d64442de165ff4ede62d7b7457f843b30dba402b9401a31d0b3
+size 3403339

results_bvcc/w2v2_large_xlsr53/layer4/run3/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b1f800e6f5de54cefe7b88d3f824da19de7ff6faf367c840ec7f221ca3160784
+size 3405062

results_bvcc/w2v2_large_xlsr53/layer4/run3/train.log

@@ -0,0 +1,99 @@
+INFO:root:Num train speech clips: 4974
+INFO:root:Num val speech clips: 1066
+INFO:root:Num test speech clips: 1066
+INFO:root:Device=cuda
+INFO:root:Average loss=0.4696754650809826
+INFO:root:
+[Valid][0][UTT][ MSE = 0.3065 | LCC = 0.8023 | SRCC = 0.8093 ]
+INFO:root:Average loss=0.2954771388799716
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2991 | LCC = 0.8411 | SRCC = 0.8407 ]
+INFO:root:Average loss=0.28332846611738205
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2869 | LCC = 0.8463 | SRCC = 0.8464 ]
+INFO:root:Average loss=0.23383696663838166
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2663 | LCC = 0.8586 | SRCC = 0.8569 ]
+INFO:root:Average loss=0.21505269962243545
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2557 | LCC = 0.8593 | SRCC = 0.8594 ]
+INFO:root:Average loss=0.21830975111478415
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2455 | LCC = 0.8675 | SRCC = 0.8651 ]
+INFO:root:Average loss=0.1987616410240149
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2317 | LCC = 0.8693 | SRCC = 0.8681 ]
+INFO:root:Average loss=0.19079558522655413
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2670 | LCC = 0.8720 | SRCC = 0.8704 ]
+INFO:root:Average loss=0.18426570716576698
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2334 | LCC = 0.8724 | SRCC = 0.8717 ]
+INFO:root:Average loss=0.17354323495274934
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2210 | LCC = 0.8753 | SRCC = 0.8748 ]
+INFO:root:Average loss=0.18379609582897943
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2884 | LCC = 0.8663 | SRCC = 0.8725 ]
+INFO:root:Average loss=0.1759594639715476
+INFO:root:
+[Valid][11][UTT][ MSE = 0.2341 | LCC = 0.8718 | SRCC = 0.8731 ]
+INFO:root:Average loss=0.16178548221404737
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2050 | LCC = 0.8786 | SRCC = 0.8777 ]
+INFO:root:Average loss=0.14725517749022216
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2202 | LCC = 0.8749 | SRCC = 0.8736 ]
+INFO:root:Average loss=0.1469860810499925
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2186 | LCC = 0.8779 | SRCC = 0.8773 ]
+INFO:root:Average loss=0.1449679702711411
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1947 | LCC = 0.8786 | SRCC = 0.8785 ]
+INFO:root:Average loss=0.13312828435729712
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2659 | LCC = 0.8658 | SRCC = 0.8687 ]
+INFO:root:Average loss=0.1286121530410571
+INFO:root:
+[Valid][17][UTT][ MSE = 0.2298 | LCC = 0.8746 | SRCC = 0.8741 ]
+INFO:root:Average loss=0.12523656710982323
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2136 | LCC = 0.8763 | SRCC = 0.8749 ]
+INFO:root:Average loss=0.12012366854991668
+INFO:root:
+[Valid][19][UTT][ MSE = 0.2106 | LCC = 0.8773 | SRCC = 0.8772 ]
+INFO:root:Average loss=0.11537392714466804
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2178 | LCC = 0.8739 | SRCC = 0.8743 ]
+INFO:root:Average loss=0.10499710245774342
+INFO:root:
+[Valid][21][UTT][ MSE = 0.2007 | LCC = 0.8805 | SRCC = 0.8794 ]
+INFO:root:Average loss=0.10581345703357305
+INFO:root:
+[Valid][22][UTT][ MSE = 0.2155 | LCC = 0.8743 | SRCC = 0.8756 ]
+INFO:root:Average loss=0.10889474092385708
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2006 | LCC = 0.8767 | SRCC = 0.8754 ]
+INFO:root:Average loss=0.1006790762528395
+INFO:root:
+[Valid][24][UTT][ MSE = 0.2121 | LCC = 0.8718 | SRCC = 0.8693 ]
+INFO:root:Average loss=0.09053596796897742
+INFO:root:
+[Valid][25][UTT][ MSE = 0.2317 | LCC = 0.8743 | SRCC = 0.8740 ]
+INFO:root:Average loss=0.09128491265269426
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2291 | LCC = 0.8756 | SRCC = 0.8734 ]
+INFO:root:Average loss=0.08415590680371492
+INFO:root:
+[Valid][27][UTT][ MSE = 0.2174 | LCC = 0.8759 | SRCC = 0.8760 ]
+INFO:root:Average loss=0.08523786067962646
+INFO:root:
+[Valid][28][UTT][ MSE = 0.2122 | LCC = 0.8769 | SRCC = 0.8763 ]
+INFO:root:Average loss=0.08090076237343825
+INFO:root:
+[Valid][29][UTT][ MSE = 0.2105 | LCC = 0.8771 | SRCC = 0.8764 ]
+INFO:root:Average loss=0.0761953240308242
+INFO:root:
+[Valid][30][UTT][ MSE = 0.2198 | LCC = 0.8731 | SRCC = 0.8756 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2211 | LCC = 0.8703 | SRCC = 0.8700 ]

results_bvcc/w2v2_large_xlsr53/layer4/run4/config.gin

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

results_bvcc/w2v2_large_xlsr53/layer4/run4/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f32c3616db4d0fbebd275bee428cbbddb517f648dd559fd48498d502ae6040be
+size 3403339

results_bvcc/w2v2_large_xlsr53/layer4/run4/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:904fa759cb4ec83f126ca39d143affdeea20223e7a8c03010188a94ce91d7de4
+size 3405062