Upload SSL Layer MOS checkpoints (part 18)

results_tencent/w2v2_large_asr_en/layer0/run0/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f44dee60f861cfe5dde45b2487f4a9158d7d0225ce4f8e6a280877f6f168acf3
+size 3403339

results_tencent/w2v2_large_asr_en/layer0/run0/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:58a78ba6675856841641c2e13a55ee3fb0acf2a01be27ebdf6e0bc75c85e9155
+size 3404998

results_tencent/w2v2_large_asr_en/layer0/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=0.6649895780086518
+INFO:root:
+[Valid][0][UTT][ MSE = 0.5021 | LCC = 0.8662 | SRCC = 0.8684 ]
+INFO:root:Average loss=0.36423822593688965
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3618 | LCC = 0.8960 | SRCC = 0.8962 ]
+INFO:root:Average loss=0.34296851766109465
+INFO:root:
+[Valid][2][UTT][ MSE = 0.3343 | LCC = 0.9080 | SRCC = 0.9090 ]
+INFO:root:Average loss=0.2726272573471069
+INFO:root:
+[Valid][3][UTT][ MSE = 0.3904 | LCC = 0.9064 | SRCC = 0.9085 ]
+INFO:root:Average loss=0.2669645261764526
+INFO:root:
+[Valid][4][UTT][ MSE = 0.3021 | LCC = 0.9187 | SRCC = 0.9204 ]
+INFO:root:Average loss=0.2269453490972519
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2834 | LCC = 0.9185 | SRCC = 0.9199 ]
+INFO:root:Average loss=0.21508347779512405
+INFO:root:
+[Valid][6][UTT][ MSE = 0.3365 | LCC = 0.9255 | SRCC = 0.9266 ]
+INFO:root:Average loss=0.207648442029953
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2773 | LCC = 0.9282 | SRCC = 0.9301 ]
+INFO:root:Average loss=0.18569231790304183
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2999 | LCC = 0.9301 | SRCC = 0.9312 ]
+INFO:root:Average loss=0.17414556181430815
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2337 | LCC = 0.9339 | SRCC = 0.9356 ]
+INFO:root:Average loss=0.1659562236070633
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2474 | LCC = 0.9327 | SRCC = 0.9336 ]
+INFO:root:Average loss=0.158993832051754
+INFO:root:
+[Valid][11][UTT][ MSE = 0.2262 | LCC = 0.9356 | SRCC = 0.9357 ]
+INFO:root:Average loss=0.15157693582773207
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2717 | LCC = 0.9352 | SRCC = 0.9362 ]
+INFO:root:Average loss=0.1491793937087059
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2489 | LCC = 0.9392 | SRCC = 0.9401 ]
+INFO:root:Average loss=0.1365955706834793
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2534 | LCC = 0.9349 | SRCC = 0.9373 ]
+INFO:root:Average loss=0.13709089648723602
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2557 | LCC = 0.9383 | SRCC = 0.9394 ]
+INFO:root:Average loss=0.12561885887384414
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2210 | LCC = 0.9391 | SRCC = 0.9391 ]
+INFO:root:Average loss=0.12346147406101227
+INFO:root:
+[Valid][17][UTT][ MSE = 0.2727 | LCC = 0.9422 | SRCC = 0.9433 ]
+INFO:root:Average loss=0.1230960151553154
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2270 | LCC = 0.9409 | SRCC = 0.9413 ]
+INFO:root:Average loss=0.11931769216060638
+INFO:root:
+[Valid][19][UTT][ MSE = 0.2474 | LCC = 0.9398 | SRCC = 0.9401 ]
+INFO:root:Average loss=0.11049323427677155
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2099 | LCC = 0.9368 | SRCC = 0.9388 ]
+INFO:root:Average loss=0.10819276887178421
+INFO:root:
+[Valid][21][UTT][ MSE = 0.2211 | LCC = 0.9432 | SRCC = 0.9435 ]
+INFO:root:Average loss=0.1075104551911354
+INFO:root:
+[Valid][22][UTT][ MSE = 0.2665 | LCC = 0.9410 | SRCC = 0.9428 ]
+INFO:root:Average loss=0.10115887171030044
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2205 | LCC = 0.9448 | SRCC = 0.9444 ]
+INFO:root:Average loss=0.09877768251299858
+INFO:root:
+[Valid][24][UTT][ MSE = 0.2133 | LCC = 0.9466 | SRCC = 0.9471 ]
+INFO:root:Average loss=0.0910660448372364
+INFO:root:
+[Valid][25][UTT][ MSE = 0.2269 | LCC = 0.9449 | SRCC = 0.9449 ]
+INFO:root:Average loss=0.08806130051612854
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2242 | LCC = 0.9458 | SRCC = 0.9458 ]
+INFO:root:Average loss=0.08847611147165299
+INFO:root:
+[Valid][27][UTT][ MSE = 0.2219 | LCC = 0.9423 | SRCC = 0.9422 ]
+INFO:root:Average loss=0.08456234943866729
+INFO:root:
+[Valid][28][UTT][ MSE = 0.2153 | LCC = 0.9445 | SRCC = 0.9442 ]
+INFO:root:Average loss=0.08081513851881027
+INFO:root:
+[Valid][29][UTT][ MSE = 0.2277 | LCC = 0.9467 | SRCC = 0.9470 ]
+INFO:root:Average loss=0.08008565375208855
+INFO:root:
+[Valid][30][UTT][ MSE = 0.2023 | LCC = 0.9461 | SRCC = 0.9467 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2323 | LCC = 0.9470 | SRCC = 0.9481 ]

results_tencent/w2v2_large_asr_en/layer0/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_large_asr_en/layer0/run1/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:621abede2a5b23eb12d984e3d8d5ed7f6c63e4b51f8732f2562cc3b51c33b469
+size 3403339

results_tencent/w2v2_large_asr_en/layer0/run1/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e72a903b536b993be1ef9291259bb1d2c32406d78f44746ab26cf5d6e5b6f116
+size 3405062

results_tencent/w2v2_large_asr_en/layer0/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=0.5740894581079483
+INFO:root:
+[Valid][0][UTT][ MSE = 0.3895 | LCC = 0.8815 | SRCC = 0.8794 ]
+INFO:root:Average loss=0.33942470848560335
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3404 | LCC = 0.9064 | SRCC = 0.9070 ]
+INFO:root:Average loss=0.3118090524673462
+INFO:root:
+[Valid][2][UTT][ MSE = 0.3112 | LCC = 0.9094 | SRCC = 0.9135 ]
+INFO:root:Average loss=0.2815272874832153
+INFO:root:
+[Valid][3][UTT][ MSE = 0.3010 | LCC = 0.9181 | SRCC = 0.9195 ]
+INFO:root:Average loss=0.24935106205940247
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2845 | LCC = 0.9242 | SRCC = 0.9261 ]
+INFO:root:Average loss=0.21995245516300202
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2458 | LCC = 0.9232 | SRCC = 0.9250 ]
+INFO:root:Average loss=0.2059981990456581
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2909 | LCC = 0.9264 | SRCC = 0.9283 ]
+INFO:root:Average loss=0.188407022356987
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2773 | LCC = 0.9291 | SRCC = 0.9319 ]
+INFO:root:Average loss=0.18127622067928315
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2528 | LCC = 0.9309 | SRCC = 0.9322 ]
+INFO:root:Average loss=0.1760287303328514
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2519 | LCC = 0.9310 | SRCC = 0.9342 ]
+INFO:root:Average loss=0.17421934288740157
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2777 | LCC = 0.9329 | SRCC = 0.9346 ]
+INFO:root:Average loss=0.16470972645282744
+INFO:root:
+[Valid][11][UTT][ MSE = 0.2320 | LCC = 0.9374 | SRCC = 0.9389 ]
+INFO:root:Average loss=0.15133613926172257
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2863 | LCC = 0.9390 | SRCC = 0.9406 ]
+INFO:root:Average loss=0.14636920660734176
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2580 | LCC = 0.9366 | SRCC = 0.9379 ]
+INFO:root:Average loss=0.13522404074668884
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2360 | LCC = 0.9392 | SRCC = 0.9405 ]
+INFO:root:Average loss=0.1288481270670891
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2471 | LCC = 0.9382 | SRCC = 0.9410 ]
+INFO:root:Average loss=0.12814946675300598
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2587 | LCC = 0.9423 | SRCC = 0.9434 ]
+INFO:root:Average loss=0.12289551103115082
+INFO:root:
+[Valid][17][UTT][ MSE = 0.2572 | LCC = 0.9385 | SRCC = 0.9405 ]
+INFO:root:Average loss=0.11703359311819077
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2786 | LCC = 0.9360 | SRCC = 0.9370 ]
+INFO:root:Average loss=0.11796789079904556
+INFO:root:
+[Valid][19][UTT][ MSE = 0.2052 | LCC = 0.9372 | SRCC = 0.9407 ]
+INFO:root:Average loss=0.11379064375162125
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2206 | LCC = 0.9382 | SRCC = 0.9385 ]
+INFO:root:Average loss=0.1046984760761261
+INFO:root:
+[Valid][21][UTT][ MSE = 0.2565 | LCC = 0.9438 | SRCC = 0.9440 ]
+INFO:root:Average loss=0.10058193561434746
+INFO:root:
+[Valid][22][UTT][ MSE = 0.2175 | LCC = 0.9445 | SRCC = 0.9452 ]
+INFO:root:Average loss=0.09745386728644372
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2305 | LCC = 0.9428 | SRCC = 0.9440 ]
+INFO:root:Average loss=0.09204903349280358
+INFO:root:
+[Valid][24][UTT][ MSE = 0.2314 | LCC = 0.9466 | SRCC = 0.9471 ]
+INFO:root:Average loss=0.09020958796143531
+INFO:root:
+[Valid][25][UTT][ MSE = 0.2185 | LCC = 0.9457 | SRCC = 0.9468 ]
+INFO:root:Average loss=0.08709004393219948
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2193 | LCC = 0.9459 | SRCC = 0.9458 ]
+INFO:root:Average loss=0.08658751103281975
+INFO:root:
+[Valid][27][UTT][ MSE = 0.2141 | LCC = 0.9461 | SRCC = 0.9466 ]
+INFO:root:Average loss=0.08099481245875359
+INFO:root:
+[Valid][28][UTT][ MSE = 0.2104 | LCC = 0.9446 | SRCC = 0.9444 ]
+INFO:root:Average loss=0.07930216109752655
+INFO:root:
+[Valid][29][UTT][ MSE = 0.2238 | LCC = 0.9449 | SRCC = 0.9455 ]
+INFO:root:Average loss=0.07841939279437066
+INFO:root:
+[Valid][30][UTT][ MSE = 0.2343 | LCC = 0.9450 | SRCC = 0.9454 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2431 | LCC = 0.9452 | SRCC = 0.9466 ]

results_tencent/w2v2_large_asr_en/layer0/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_large_asr_en/layer0/run2/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5b955b930d0ac4b0f88820c999998baac16d6ddb8928503c3702f5b0a1274903
+size 3403339

results_tencent/w2v2_large_asr_en/layer0/run2/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4ca61e766ae81e8f00c63c2bb5293bd7f2725ec3cf8356b5292527dc04b888e8
+size 3405062

results_tencent/w2v2_large_asr_en/layer0/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=0.709221761226654
+INFO:root:
+[Valid][0][UTT][ MSE = 0.4366 | LCC = 0.8583 | SRCC = 0.8577 ]
+INFO:root:Average loss=0.3729499524831772
+INFO:root:
+[Valid][1][UTT][ MSE = 0.4500 | LCC = 0.8869 | SRCC = 0.8869 ]
+INFO:root:Average loss=0.325014719247818
+INFO:root:
+[Valid][2][UTT][ MSE = 0.3388 | LCC = 0.9091 | SRCC = 0.9101 ]
+INFO:root:Average loss=0.2722231197357178
+INFO:root:
+[Valid][3][UTT][ MSE = 0.3134 | LCC = 0.9185 | SRCC = 0.9205 ]
+INFO:root:Average loss=0.2466454672217369
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2878 | LCC = 0.9220 | SRCC = 0.9232 ]
+INFO:root:Average loss=0.2227160667181015
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2816 | LCC = 0.9283 | SRCC = 0.9294 ]
+INFO:root:Average loss=0.19683788174390793
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2595 | LCC = 0.9282 | SRCC = 0.9294 ]
+INFO:root:Average loss=0.18636493563652037
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2931 | LCC = 0.9303 | SRCC = 0.9313 ]
+INFO:root:Average loss=0.17900047713518144
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2693 | LCC = 0.9297 | SRCC = 0.9300 ]
+INFO:root:Average loss=0.16832722479104995
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2727 | LCC = 0.9351 | SRCC = 0.9361 ]
+INFO:root:Average loss=0.16136809945106506
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2506 | LCC = 0.9382 | SRCC = 0.9386 ]
+INFO:root:Average loss=0.1498782126903534
+INFO:root:
+[Valid][11][UTT][ MSE = 0.2281 | LCC = 0.9387 | SRCC = 0.9392 ]
+INFO:root:Average loss=0.13958915346860887
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2393 | LCC = 0.9368 | SRCC = 0.9372 ]
+INFO:root:Average loss=0.13631629371643067
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2315 | LCC = 0.9338 | SRCC = 0.9358 ]
+INFO:root:Average loss=0.13064119237661362
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2396 | LCC = 0.9422 | SRCC = 0.9424 ]
+INFO:root:Average loss=0.1272465716600418
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2452 | LCC = 0.9417 | SRCC = 0.9423 ]
+INFO:root:Average loss=0.12118836542963982
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2440 | LCC = 0.9349 | SRCC = 0.9366 ]
+INFO:root:Average loss=0.11433009663224221
+INFO:root:
+[Valid][17][UTT][ MSE = 0.2635 | LCC = 0.9389 | SRCC = 0.9398 ]
+INFO:root:Average loss=0.11710317665338517
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2165 | LCC = 0.9407 | SRCC = 0.9409 ]
+INFO:root:Average loss=0.10485217961668968
+INFO:root:
+[Valid][19][UTT][ MSE = 0.2331 | LCC = 0.9427 | SRCC = 0.9429 ]
+INFO:root:Average loss=0.1080205430984497
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2403 | LCC = 0.9436 | SRCC = 0.9431 ]
+INFO:root:Average loss=0.10018497738242149
+INFO:root:
+[Valid][21][UTT][ MSE = 0.2411 | LCC = 0.9449 | SRCC = 0.9452 ]
+INFO:root:Average loss=0.09613898831605912
+INFO:root:
+[Valid][22][UTT][ MSE = 0.2187 | LCC = 0.9426 | SRCC = 0.9430 ]
+INFO:root:Average loss=0.09334326723217964
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2130 | LCC = 0.9405 | SRCC = 0.9414 ]
+INFO:root:Average loss=0.09391682940721512
+INFO:root:
+[Valid][24][UTT][ MSE = 0.2603 | LCC = 0.9421 | SRCC = 0.9423 ]
+INFO:root:Average loss=0.08849663078784943
+INFO:root:
+[Valid][25][UTT][ MSE = 0.2087 | LCC = 0.9442 | SRCC = 0.9436 ]
+INFO:root:Average loss=0.08483491101861
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2210 | LCC = 0.9464 | SRCC = 0.9460 ]
+INFO:root:Average loss=0.08490307739377022
+INFO:root:
+[Valid][27][UTT][ MSE = 0.2096 | LCC = 0.9458 | SRCC = 0.9456 ]
+INFO:root:Average loss=0.07878244167566299
+INFO:root:
+[Valid][28][UTT][ MSE = 0.2280 | LCC = 0.9454 | SRCC = 0.9449 ]
+INFO:root:Average loss=0.08274864104390145
+INFO:root:
+[Valid][29][UTT][ MSE = 0.2225 | LCC = 0.9455 | SRCC = 0.9447 ]
+INFO:root:Average loss=0.07389444380998611
+INFO:root:
+[Valid][30][UTT][ MSE = 0.2241 | LCC = 0.9453 | SRCC = 0.9458 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2283 | LCC = 0.9462 | SRCC = 0.9462 ]

results_tencent/w2v2_large_asr_en/layer0/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_large_asr_en/layer0/run3/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cfe4b7e7983d56b8b8372e210b298f660ea2da87d54f68abeb9a82b17460e939
+size 3403339

results_tencent/w2v2_large_asr_en/layer0/run3/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:467a0b3bd794b8d3f9f30addb265f8eba0d9eefe981c70e4bf1fa0890444f6d5
+size 3405062

results_tencent/w2v2_large_asr_en/layer0/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=0.618389678478241
+INFO:root:
+[Valid][0][UTT][ MSE = 0.4107 | LCC = 0.8755 | SRCC = 0.8745 ]
+INFO:root:Average loss=0.35121494472026826
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3838 | LCC = 0.8972 | SRCC = 0.9007 ]
+INFO:root:Average loss=0.2892568988800049
+INFO:root:
+[Valid][2][UTT][ MSE = 0.3334 | LCC = 0.9132 | SRCC = 0.9137 ]
+INFO:root:Average loss=0.2573729250431061
+INFO:root:
+[Valid][3][UTT][ MSE = 0.3814 | LCC = 0.9068 | SRCC = 0.9069 ]
+INFO:root:Average loss=0.23396009427309036
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2855 | LCC = 0.9237 | SRCC = 0.9248 ]
+INFO:root:Average loss=0.2149351971745491
+INFO:root:
+[Valid][5][UTT][ MSE = 0.3038 | LCC = 0.9195 | SRCC = 0.9217 ]
+INFO:root:Average loss=0.20843857795000076
+INFO:root:
+[Valid][6][UTT][ MSE = 0.3273 | LCC = 0.9234 | SRCC = 0.9271 ]
+INFO:root:Average loss=0.19983563405275345
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2876 | LCC = 0.9341 | SRCC = 0.9357 ]
+INFO:root:Average loss=0.18420539343357087
+INFO:root:
+[Valid][8][UTT][ MSE = 0.3130 | LCC = 0.9320 | SRCC = 0.9338 ]
+INFO:root:Average loss=0.17544530963897706
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2366 | LCC = 0.9367 | SRCC = 0.9381 ]
+INFO:root:Average loss=0.15911000841856002
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2164 | LCC = 0.9380 | SRCC = 0.9393 ]
+INFO:root:Average loss=0.14822474956512452
+INFO:root:
+[Valid][11][UTT][ MSE = 0.2618 | LCC = 0.9359 | SRCC = 0.9368 ]
+INFO:root:Average loss=0.15723540180921555
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2825 | LCC = 0.9366 | SRCC = 0.9376 ]
+INFO:root:Average loss=0.13719176137447356
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2436 | LCC = 0.9390 | SRCC = 0.9391 ]
+INFO:root:Average loss=0.13555410206317903
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2749 | LCC = 0.9406 | SRCC = 0.9417 ]
+INFO:root:Average loss=0.12607394224405288
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2703 | LCC = 0.9366 | SRCC = 0.9365 ]
+INFO:root:Average loss=0.1240776018500328
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2650 | LCC = 0.9378 | SRCC = 0.9390 ]
+INFO:root:Average loss=0.11930628028512001
+INFO:root:
+[Valid][17][UTT][ MSE = 0.2459 | LCC = 0.9448 | SRCC = 0.9454 ]
+INFO:root:Average loss=0.117901809155941
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2413 | LCC = 0.9395 | SRCC = 0.9396 ]
+INFO:root:Average loss=0.11178152811527252
+INFO:root:
+[Valid][19][UTT][ MSE = 0.2299 | LCC = 0.9432 | SRCC = 0.9433 ]
+INFO:root:Average loss=0.10334135204553604
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2176 | LCC = 0.9450 | SRCC = 0.9454 ]
+INFO:root:Average loss=0.10154286581277848
+INFO:root:
+[Valid][21][UTT][ MSE = 0.2054 | LCC = 0.9439 | SRCC = 0.9445 ]
+INFO:root:Average loss=0.10280697202682496
+INFO:root:
+[Valid][22][UTT][ MSE = 0.2558 | LCC = 0.9424 | SRCC = 0.9435 ]
+INFO:root:Average loss=0.10026279300451278
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2804 | LCC = 0.9442 | SRCC = 0.9451 ]
+INFO:root:Average loss=0.09706531736254692
+INFO:root:
+[Valid][24][UTT][ MSE = 0.2455 | LCC = 0.9395 | SRCC = 0.9400 ]
+INFO:root:Average loss=0.09057048386335373
+INFO:root:
+[Valid][25][UTT][ MSE = 0.2213 | LCC = 0.9452 | SRCC = 0.9451 ]
+INFO:root:Average loss=0.08575144201517106
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2087 | LCC = 0.9467 | SRCC = 0.9462 ]
+INFO:root:Average loss=0.08538841089606285
+INFO:root:
+[Valid][27][UTT][ MSE = 0.2330 | LCC = 0.9386 | SRCC = 0.9425 ]
+INFO:root:Average loss=0.08338552322983742
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1994 | LCC = 0.9471 | SRCC = 0.9473 ]
+INFO:root:Average loss=0.07937182173132896
+INFO:root:
+[Valid][29][UTT][ MSE = 0.2331 | LCC = 0.9433 | SRCC = 0.9441 ]
+INFO:root:Average loss=0.07781142038106918
+INFO:root:
+[Valid][30][UTT][ MSE = 0.2223 | LCC = 0.9439 | SRCC = 0.9439 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2097 | LCC = 0.9457 | SRCC = 0.9458 ]

results_tencent/w2v2_large_asr_en/layer0/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_large_asr_en/layer0/run4/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:195ac7072f097542df6bce24e23182196ac367d448be61714f0a2e31c7a1a97c
+size 3403339

results_tencent/w2v2_large_asr_en/layer0/run4/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:61ba9166506e4ea57d609855098bb36dfcd5d15f33d37650c212af79c0e446b9
+size 3404998

results_tencent/w2v2_large_asr_en/layer0/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=0.5631957949399948
+INFO:root:
+[Valid][0][UTT][ MSE = 0.4494 | LCC = 0.8740 | SRCC = 0.8730 ]
+INFO:root:Average loss=0.3411657792329788
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3488 | LCC = 0.8964 | SRCC = 0.8966 ]
+INFO:root:Average loss=0.2976254588365555
+INFO:root:
+[Valid][2][UTT][ MSE = 0.3333 | LCC = 0.9059 | SRCC = 0.9046 ]
+INFO:root:Average loss=0.2650909913778305
+INFO:root:
+[Valid][3][UTT][ MSE = 0.3201 | LCC = 0.9120 | SRCC = 0.9129 ]
+INFO:root:Average loss=0.2358146229982376
+INFO:root:
+[Valid][4][UTT][ MSE = 0.3306 | LCC = 0.9116 | SRCC = 0.9145 ]
+INFO:root:Average loss=0.22154437685012818
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2832 | LCC = 0.9230 | SRCC = 0.9233 ]
+INFO:root:Average loss=0.20374881803989411
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2576 | LCC = 0.9301 | SRCC = 0.9307 ]
+INFO:root:Average loss=0.19040218222141267
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2612 | LCC = 0.9320 | SRCC = 0.9325 ]
+INFO:root:Average loss=0.17888089042901992
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2740 | LCC = 0.9285 | SRCC = 0.9289 ]
+INFO:root:Average loss=0.1687595579624176
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2560 | LCC = 0.9289 | SRCC = 0.9297 ]
+INFO:root:Average loss=0.16323179107904434
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2372 | LCC = 0.9362 | SRCC = 0.9374 ]
+INFO:root:Average loss=0.14992485547065734
+INFO:root:
+[Valid][11][UTT][ MSE = 0.2628 | LCC = 0.9367 | SRCC = 0.9375 ]
+INFO:root:Average loss=0.14807665467262268
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2566 | LCC = 0.9340 | SRCC = 0.9343 ]
+INFO:root:Average loss=0.144786834359169
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2536 | LCC = 0.9360 | SRCC = 0.9367 ]
+INFO:root:Average loss=0.13513494366407394
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2272 | LCC = 0.9365 | SRCC = 0.9370 ]
+INFO:root:Average loss=0.1273358070254326
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2196 | LCC = 0.9397 | SRCC = 0.9410 ]
+INFO:root:Average loss=0.12330739915370942
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2180 | LCC = 0.9398 | SRCC = 0.9407 ]
+INFO:root:Average loss=0.11635331761837006
+INFO:root:
+[Valid][17][UTT][ MSE = 0.2562 | LCC = 0.9406 | SRCC = 0.9415 ]
+INFO:root:Average loss=0.11635982328653335
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2257 | LCC = 0.9436 | SRCC = 0.9445 ]
+INFO:root:Average loss=0.11165623599290848
+INFO:root:
+[Valid][19][UTT][ MSE = 0.2017 | LCC = 0.9418 | SRCC = 0.9426 ]
+INFO:root:Average loss=0.1059285136461258
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2069 | LCC = 0.9419 | SRCC = 0.9425 ]
+INFO:root:Average loss=0.10608018425107002
+INFO:root:
+[Valid][21][UTT][ MSE = 0.2171 | LCC = 0.9429 | SRCC = 0.9436 ]
+INFO:root:Average loss=0.10404559701681138
+INFO:root:
+[Valid][22][UTT][ MSE = 0.2162 | LCC = 0.9451 | SRCC = 0.9457 ]
+INFO:root:Average loss=0.09825759163498879
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2109 | LCC = 0.9450 | SRCC = 0.9455 ]
+INFO:root:Average loss=0.09583988213539124
+INFO:root:
+[Valid][24][UTT][ MSE = 0.2023 | LCC = 0.9409 | SRCC = 0.9427 ]
+INFO:root:Average loss=0.0902070954144001
+INFO:root:
+[Valid][25][UTT][ MSE = 0.2103 | LCC = 0.9449 | SRCC = 0.9450 ]
+INFO:root:Average loss=0.09075407809019088
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2025 | LCC = 0.9472 | SRCC = 0.9478 ]
+INFO:root:Average loss=0.08497983774542808
+INFO:root:
+[Valid][27][UTT][ MSE = 0.2004 | LCC = 0.9453 | SRCC = 0.9455 ]
+INFO:root:Average loss=0.08363067823648453
+INFO:root:
+[Valid][28][UTT][ MSE = 0.2146 | LCC = 0.9453 | SRCC = 0.9453 ]
+INFO:root:Average loss=0.08124600887298584
+INFO:root:
+[Valid][29][UTT][ MSE = 0.2302 | LCC = 0.9455 | SRCC = 0.9453 ]
+INFO:root:Average loss=0.08334646984934807
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1921 | LCC = 0.9465 | SRCC = 0.9462 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2072 | LCC = 0.9479 | SRCC = 0.9486 ]

results_tencent/w2v2_large_asr_en/layer10/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_large_asr_en/layer10/run0/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8dacd0896b3b6f8f490a4c27412fbe80d584c6a7f473778a120d3611ec75d410
+size 3403339

results_tencent/w2v2_large_asr_en/layer10/run0/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:09756ea7c63f4192036f30c43e302f9dda1faf63ac1aa8f66461a96c5243fd5d
+size 3404998

results_tencent/w2v2_large_asr_en/layer10/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=0.5010222245454788
+INFO:root:
+[Valid][0][UTT][ MSE = 0.3356 | LCC = 0.9079 | SRCC = 0.9067 ]
+INFO:root:Average loss=0.22929055041074753
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2595 | LCC = 0.9292 | SRCC = 0.9272 ]
+INFO:root:Average loss=0.18144736593961716
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2243 | LCC = 0.9390 | SRCC = 0.9375 ]
+INFO:root:Average loss=0.15326694977283478
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2462 | LCC = 0.9455 | SRCC = 0.9449 ]
+INFO:root:Average loss=0.13040729987621308
+INFO:root:
+[Valid][4][UTT][ MSE = 0.1936 | LCC = 0.9489 | SRCC = 0.9490 ]
+INFO:root:Average loss=0.11421453300118446
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2101 | LCC = 0.9508 | SRCC = 0.9506 ]
+INFO:root:Average loss=0.10502200680971145
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2050 | LCC = 0.9525 | SRCC = 0.9523 ]
+INFO:root:Average loss=0.10170439013838768
+INFO:root:
+[Valid][7][UTT][ MSE = 0.1755 | LCC = 0.9533 | SRCC = 0.9527 ]
+INFO:root:Average loss=0.100304323554039
+INFO:root:
+[Valid][8][UTT][ MSE = 0.1908 | LCC = 0.9543 | SRCC = 0.9541 ]
+INFO:root:Average loss=0.08370878720283509
+INFO:root:
+[Valid][9][UTT][ MSE = 0.1980 | LCC = 0.9562 | SRCC = 0.9555 ]
+INFO:root:Average loss=0.07891840344667435
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2028 | LCC = 0.9559 | SRCC = 0.9552 ]
+INFO:root:Average loss=0.07109805217385293
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1710 | LCC = 0.9553 | SRCC = 0.9546 ]
+INFO:root:Average loss=0.07400908750295639
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2114 | LCC = 0.9551 | SRCC = 0.9550 ]
+INFO:root:Average loss=0.06603842422366142
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2193 | LCC = 0.9560 | SRCC = 0.9554 ]
+INFO:root:Average loss=0.06004586634039879
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1790 | LCC = 0.9560 | SRCC = 0.9555 ]
+INFO:root:Average loss=0.061918595537543294
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2059 | LCC = 0.9556 | SRCC = 0.9553 ]
+INFO:root:Average loss=0.05104477721452713
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1862 | LCC = 0.9563 | SRCC = 0.9552 ]
+INFO:root:Average loss=0.04904086169600487
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1856 | LCC = 0.9561 | SRCC = 0.9550 ]
+INFO:root:Average loss=0.04913599762320518
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2109 | LCC = 0.9555 | SRCC = 0.9548 ]
+INFO:root:Average loss=0.04368538872897625
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1904 | LCC = 0.9568 | SRCC = 0.9551 ]
+INFO:root:Average loss=0.04146124598383903
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1693 | LCC = 0.9564 | SRCC = 0.9548 ]
+INFO:root:Average loss=0.04018029852211475
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1828 | LCC = 0.9564 | SRCC = 0.9550 ]
+INFO:root:Average loss=0.03878097864985466
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1932 | LCC = 0.9556 | SRCC = 0.9541 ]
+INFO:root:Average loss=0.03755505308508873
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1820 | LCC = 0.9572 | SRCC = 0.9559 ]
+INFO:root:Average loss=0.03594144396483898
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1679 | LCC = 0.9556 | SRCC = 0.9544 ]
+INFO:root:Average loss=0.03358615885674954
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1846 | LCC = 0.9563 | SRCC = 0.9552 ]
+INFO:root:Average loss=0.031770835474133494
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1805 | LCC = 0.9578 | SRCC = 0.9562 ]
+INFO:root:Average loss=0.030430072009563446
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1853 | LCC = 0.9567 | SRCC = 0.9553 ]
+INFO:root:Average loss=0.028494312539696694
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1808 | LCC = 0.9569 | SRCC = 0.9555 ]
+INFO:root:Average loss=0.02948863671720028
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1704 | LCC = 0.9568 | SRCC = 0.9554 ]
+INFO:root:Average loss=0.027734526216983794
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1943 | LCC = 0.9565 | SRCC = 0.9553 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1827 | LCC = 0.9590 | SRCC = 0.9593 ]

results_tencent/w2v2_large_asr_en/layer10/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_large_asr_en/layer10/run1/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2b8c0529f25555f4e3e5dc2f51ea5aaf5d0c12e07f5d98269b201f7327ae20ac
+size 3403339

results_tencent/w2v2_large_asr_en/layer10/run1/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:01ef34f7d3abfae5e3448480f86dead58af5f7b54b983043b4c53021e0d5e6f4
+size 3405062

results_tencent/w2v2_large_asr_en/layer10/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=0.5603787733316421
+INFO:root:
+[Valid][0][UTT][ MSE = 0.3758 | LCC = 0.8986 | SRCC = 0.9004 ]
+INFO:root:Average loss=0.2266037258505821
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3077 | LCC = 0.9323 | SRCC = 0.9312 ]
+INFO:root:Average loss=0.17661147236824035
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2388 | LCC = 0.9411 | SRCC = 0.9413 ]
+INFO:root:Average loss=0.16322550082206727
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2356 | LCC = 0.9469 | SRCC = 0.9468 ]
+INFO:root:Average loss=0.146523366689682
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2567 | LCC = 0.9492 | SRCC = 0.9496 ]
+INFO:root:Average loss=0.12082720899581909
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2178 | LCC = 0.9527 | SRCC = 0.9522 ]
+INFO:root:Average loss=0.10759995889663697
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2165 | LCC = 0.9545 | SRCC = 0.9543 ]
+INFO:root:Average loss=0.0963390297293663
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2414 | LCC = 0.9537 | SRCC = 0.9538 ]
+INFO:root:Average loss=0.0931014162003994
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2232 | LCC = 0.9546 | SRCC = 0.9547 ]
+INFO:root:Average loss=0.08377202424407006
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2424 | LCC = 0.9552 | SRCC = 0.9549 ]
+INFO:root:Average loss=0.0782404323220253
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2017 | LCC = 0.9566 | SRCC = 0.9560 ]
+INFO:root:Average loss=0.07125890260934829
+INFO:root:
+[Valid][11][UTT][ MSE = 0.2306 | LCC = 0.9543 | SRCC = 0.9542 ]
+INFO:root:Average loss=0.06637917655706406
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2030 | LCC = 0.9560 | SRCC = 0.9553 ]
+INFO:root:Average loss=0.06531040970981121
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2353 | LCC = 0.9560 | SRCC = 0.9559 ]
+INFO:root:Average loss=0.05905053097009659
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2284 | LCC = 0.9565 | SRCC = 0.9567 ]
+INFO:root:Average loss=0.05572305777668953
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2050 | LCC = 0.9569 | SRCC = 0.9565 ]
+INFO:root:Average loss=0.049080815181136134
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2475 | LCC = 0.9548 | SRCC = 0.9554 ]
+INFO:root:Average loss=0.049781987801194194
+INFO:root:
+[Valid][17][UTT][ MSE = 0.2452 | LCC = 0.9548 | SRCC = 0.9554 ]
+INFO:root:Average loss=0.04713604381680488
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2155 | LCC = 0.9533 | SRCC = 0.9526 ]
+INFO:root:Average loss=0.04531936103105545
+INFO:root:
+[Valid][19][UTT][ MSE = 0.2101 | LCC = 0.9552 | SRCC = 0.9546 ]
+INFO:root:Average loss=0.04381908024847508
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2090 | LCC = 0.9558 | SRCC = 0.9552 ]
+INFO:root:Average loss=0.03895128692686558
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1920 | LCC = 0.9547 | SRCC = 0.9539 ]
+INFO:root:Average loss=0.03864530585706234
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1911 | LCC = 0.9552 | SRCC = 0.9540 ]
+INFO:root:Average loss=0.03801547911763191
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2114 | LCC = 0.9566 | SRCC = 0.9558 ]
+INFO:root:Average loss=0.03450804188847542
+INFO:root:
+[Valid][24][UTT][ MSE = 0.2195 | LCC = 0.9553 | SRCC = 0.9553 ]
+INFO:root:Average loss=0.03415686471760273
+INFO:root:
+[Valid][25][UTT][ MSE = 0.2021 | LCC = 0.9558 | SRCC = 0.9552 ]
+INFO:root:Average loss=0.031128952831029892
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2291 | LCC = 0.9552 | SRCC = 0.9555 ]
+INFO:root:Average loss=0.03010747820138931
+INFO:root:
+[Valid][27][UTT][ MSE = 0.2379 | LCC = 0.9548 | SRCC = 0.9549 ]
+INFO:root:Average loss=0.02996332250535488
+INFO:root:
+[Valid][28][UTT][ MSE = 0.2181 | LCC = 0.9561 | SRCC = 0.9556 ]
+INFO:root:Average loss=0.028923279009759426
+INFO:root:
+[Valid][29][UTT][ MSE = 0.2238 | LCC = 0.9542 | SRCC = 0.9536 ]
+INFO:root:Average loss=0.027694006398320196
+INFO:root:
+[Valid][30][UTT][ MSE = 0.2069 | LCC = 0.9549 | SRCC = 0.9540 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2182 | LCC = 0.9548 | SRCC = 0.9564 ]

results_tencent/w2v2_large_asr_en/layer10/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_large_asr_en/layer10/run2/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a315d1cd6b47fc559bd25de3646cf62c9727503bb12777188169988a2611c87a
+size 3403339

results_tencent/w2v2_large_asr_en/layer10/run2/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:786764fac91eeb10e5844efd1460952eb5771e8190986ef1c5a29511412e9d83
+size 3405062

results_tencent/w2v2_large_asr_en/layer10/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=0.5990666986703873
+INFO:root:
+[Valid][0][UTT][ MSE = 0.3596 | LCC = 0.8934 | SRCC = 0.8964 ]
+INFO:root:Average loss=0.23174694460630416
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2969 | LCC = 0.9247 | SRCC = 0.9234 ]
+INFO:root:Average loss=0.19153335118293763
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2587 | LCC = 0.9346 | SRCC = 0.9355 ]
+INFO:root:Average loss=0.16517070412635804
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2289 | LCC = 0.9430 | SRCC = 0.9434 ]
+INFO:root:Average loss=0.13948925870656967
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2573 | LCC = 0.9450 | SRCC = 0.9447 ]
+INFO:root:Average loss=0.13110521811246872
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2340 | LCC = 0.9502 | SRCC = 0.9499 ]
+INFO:root:Average loss=0.11405591505765915
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2594 | LCC = 0.9498 | SRCC = 0.9502 ]
+INFO:root:Average loss=0.10888171818852424
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2659 | LCC = 0.9497 | SRCC = 0.9500 ]
+INFO:root:Average loss=0.09537407386302948
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2092 | LCC = 0.9542 | SRCC = 0.9540 ]
+INFO:root:Average loss=0.08707308340072632
+INFO:root:
+[Valid][9][UTT][ MSE = 0.1970 | LCC = 0.9549 | SRCC = 0.9543 ]
+INFO:root:Average loss=0.08054021772742272
+INFO:root:
+[Valid][10][UTT][ MSE = 0.1781 | LCC = 0.9552 | SRCC = 0.9555 ]
+INFO:root:Average loss=0.07421033498644829
+INFO:root:
+[Valid][11][UTT][ MSE = 0.2133 | LCC = 0.9555 | SRCC = 0.9553 ]
+INFO:root:Average loss=0.06622179347276688
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2521 | LCC = 0.9535 | SRCC = 0.9537 ]
+INFO:root:Average loss=0.07030860358476639
+INFO:root:
+[Valid][13][UTT][ MSE = 0.1998 | LCC = 0.9570 | SRCC = 0.9557 ]
+INFO:root:Average loss=0.05837066516280174
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1780 | LCC = 0.9545 | SRCC = 0.9544 ]
+INFO:root:Average loss=0.05676400099694729
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2248 | LCC = 0.9560 | SRCC = 0.9555 ]
+INFO:root:Average loss=0.0513066251128912
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1889 | LCC = 0.9573 | SRCC = 0.9563 ]
+INFO:root:Average loss=0.04992554013431072
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1918 | LCC = 0.9578 | SRCC = 0.9568 ]
+INFO:root:Average loss=0.04731315691769123
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2096 | LCC = 0.9567 | SRCC = 0.9549 ]
+INFO:root:Average loss=0.04341712997853756
+INFO:root:
+[Valid][19][UTT][ MSE = 0.2004 | LCC = 0.9579 | SRCC = 0.9564 ]
+INFO:root:Average loss=0.04224585399031639
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1944 | LCC = 0.9563 | SRCC = 0.9545 ]
+INFO:root:Average loss=0.040629407063126566
+INFO:root:
+[Valid][21][UTT][ MSE = 0.2280 | LCC = 0.9552 | SRCC = 0.9545 ]
+INFO:root:Average loss=0.039427185893058776
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1903 | LCC = 0.9572 | SRCC = 0.9561 ]
+INFO:root:Average loss=0.03544106629490852
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2200 | LCC = 0.9564 | SRCC = 0.9553 ]
+INFO:root:Average loss=0.03559505261480808
+INFO:root:
+[Valid][24][UTT][ MSE = 0.2227 | LCC = 0.9568 | SRCC = 0.9558 ]
+INFO:root:Average loss=0.03225844207406044
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1825 | LCC = 0.9565 | SRCC = 0.9557 ]
+INFO:root:Average loss=0.03080828219652176
+INFO:root:
+[Valid][26][UTT][ MSE = 0.2094 | LCC = 0.9573 | SRCC = 0.9561 ]
+INFO:root:Average loss=0.029174505949020386
+INFO:root:
+[Valid][27][UTT][ MSE = 0.2377 | LCC = 0.9557 | SRCC = 0.9553 ]
+INFO:root:Average loss=0.027514540180563925
+INFO:root:
+[Valid][28][UTT][ MSE = 0.2186 | LCC = 0.9571 | SRCC = 0.9561 ]
+INFO:root:Average loss=0.027990632250905038
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1997 | LCC = 0.9578 | SRCC = 0.9565 ]
+INFO:root:Average loss=0.02687372672557831
+INFO:root:
+[Valid][30][UTT][ MSE = 0.2045 | LCC = 0.9567 | SRCC = 0.9554 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2121 | LCC = 0.9568 | SRCC = 0.9573 ]

results_tencent/w2v2_large_asr_en/layer10/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_large_asr_en/layer10/run3/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c5c62df18135890ce298a17c9a644675cb85598a8942b349728aabcc94887a24
+size 3403339

results_tencent/w2v2_large_asr_en/layer10/run3/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e729122522432f0835838f26d2cad3fa76b7925d6f28f007f20fe56e3c1ea3e4
+size 3405062

results_tencent/w2v2_large_asr_en/layer10/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=0.5285066267251969
+INFO:root:
+[Valid][0][UTT][ MSE = 0.3495 | LCC = 0.9046 | SRCC = 0.9025 ]
+INFO:root:Average loss=0.22524885666370392
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2786 | LCC = 0.9305 | SRCC = 0.9289 ]
+INFO:root:Average loss=0.1791326133608818
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2578 | LCC = 0.9395 | SRCC = 0.9389 ]
+INFO:root:Average loss=0.15077698367834091
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2336 | LCC = 0.9419 | SRCC = 0.9421 ]
+INFO:root:Average loss=0.12690383332967758
+INFO:root:
+[Valid][4][UTT][ MSE = 0.1791 | LCC = 0.9497 | SRCC = 0.9504 ]
+INFO:root:Average loss=0.11462375044822692
+INFO:root:
+[Valid][5][UTT][ MSE = 0.1806 | LCC = 0.9521 | SRCC = 0.9532 ]
+INFO:root:Average loss=0.10862916389107705
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2203 | LCC = 0.9544 | SRCC = 0.9539 ]
+INFO:root:Average loss=0.10128800648450852
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2116 | LCC = 0.9529 | SRCC = 0.9530 ]
+INFO:root:Average loss=0.0923602003455162
+INFO:root:
+[Valid][8][UTT][ MSE = 0.1858 | LCC = 0.9569 | SRCC = 0.9568 ]
+INFO:root:Average loss=0.08683293190598487
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2051 | LCC = 0.9565 | SRCC = 0.9563 ]
+INFO:root:Average loss=0.07982811743021011
+INFO:root:
+[Valid][10][UTT][ MSE = 0.2036 | LCC = 0.9551 | SRCC = 0.9543 ]
+INFO:root:Average loss=0.07268274888396263
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1692 | LCC = 0.9566 | SRCC = 0.9562 ]
+INFO:root:Average loss=0.06928026416897774
+INFO:root:
+[Valid][12][UTT][ MSE = 0.1884 | LCC = 0.9563 | SRCC = 0.9557 ]
+INFO:root:Average loss=0.06979656729102135
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2117 | LCC = 0.9575 | SRCC = 0.9568 ]
+INFO:root:Average loss=0.062166826754808426
+INFO:root:
+[Valid][14][UTT][ MSE = 0.1802 | LCC = 0.9578 | SRCC = 0.9569 ]
+INFO:root:Average loss=0.05576982533931732
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1588 | LCC = 0.9577 | SRCC = 0.9569 ]
+INFO:root:Average loss=0.05502957463264466
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1907 | LCC = 0.9581 | SRCC = 0.9568 ]
+INFO:root:Average loss=0.05059075519442558
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1882 | LCC = 0.9579 | SRCC = 0.9565 ]
+INFO:root:Average loss=0.04845566868782043
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1956 | LCC = 0.9557 | SRCC = 0.9551 ]
+INFO:root:Average loss=0.04369436579942703
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1805 | LCC = 0.9541 | SRCC = 0.9543 ]
+INFO:root:Average loss=0.04274707092344761
+INFO:root:
+[Valid][20][UTT][ MSE = 0.1729 | LCC = 0.9543 | SRCC = 0.9549 ]
+INFO:root:Average loss=0.04194863259792328
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1857 | LCC = 0.9552 | SRCC = 0.9548 ]
+INFO:root:Average loss=0.03765071985125542
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1634 | LCC = 0.9571 | SRCC = 0.9564 ]
+INFO:root:Average loss=0.036161151051521304
+INFO:root:
+[Valid][23][UTT][ MSE = 0.1778 | LCC = 0.9563 | SRCC = 0.9553 ]
+INFO:root:Average loss=0.03502218876779079
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1776 | LCC = 0.9573 | SRCC = 0.9566 ]
+INFO:root:Average loss=0.03258236289024353
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1839 | LCC = 0.9563 | SRCC = 0.9556 ]
+INFO:root:Average loss=0.0320270359814167
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1739 | LCC = 0.9572 | SRCC = 0.9563 ]
+INFO:root:Average loss=0.03372133381664753
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1898 | LCC = 0.9562 | SRCC = 0.9556 ]
+INFO:root:Average loss=0.029788642928004264
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1704 | LCC = 0.9580 | SRCC = 0.9566 ]
+INFO:root:Average loss=0.027675464630126954
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1706 | LCC = 0.9574 | SRCC = 0.9562 ]
+INFO:root:Average loss=0.02931845810264349
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1761 | LCC = 0.9568 | SRCC = 0.9559 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2010 | LCC = 0.9559 | SRCC = 0.9561 ]

results_tencent/w2v2_large_asr_en/layer10/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_large_asr_en/layer10/run4/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7d89fb5b7ee874d8f2b2329bfdaf00206a69a2c6d50fcffb28fcc7018a85f686
+size 3403339

results_tencent/w2v2_large_asr_en/layer10/run4/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b29ccc4dd14e6d74d110e0d069399158ed91be6cbc2d0a38471482ba971c2f0b
+size 3405062

results_tencent/w2v2_large_asr_en/layer10/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=0.44579941058158873
+INFO:root:
+[Valid][0][UTT][ MSE = 0.3057 | LCC = 0.9118 | SRCC = 0.9096 ]
+INFO:root:Average loss=0.21829465001821519
+INFO:root:
+[Valid][1][UTT][ MSE = 0.2358 | LCC = 0.9341 | SRCC = 0.9331 ]
+INFO:root:Average loss=0.17866651034355163
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2394 | LCC = 0.9425 | SRCC = 0.9419 ]
+INFO:root:Average loss=0.14484544485807418
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2758 | LCC = 0.9447 | SRCC = 0.9464 ]
+INFO:root:Average loss=0.12815326994657517
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2093 | LCC = 0.9523 | SRCC = 0.9521 ]
+INFO:root:Average loss=0.12252556091547012
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2464 | LCC = 0.9509 | SRCC = 0.9519 ]
+INFO:root:Average loss=0.10439216762781144
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2176 | LCC = 0.9536 | SRCC = 0.9537 ]
+INFO:root:Average loss=0.09497732418775559
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2263 | LCC = 0.9553 | SRCC = 0.9552 ]
+INFO:root:Average loss=0.08847856071591377
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2399 | LCC = 0.9537 | SRCC = 0.9539 ]
+INFO:root:Average loss=0.08257314336299897
+INFO:root:
+[Valid][9][UTT][ MSE = 0.1794 | LCC = 0.9571 | SRCC = 0.9569 ]
+INFO:root:Average loss=0.07822251987457275
+INFO:root:
+[Valid][10][UTT][ MSE = 0.1992 | LCC = 0.9567 | SRCC = 0.9561 ]
+INFO:root:Average loss=0.06896637162566185
+INFO:root:
+[Valid][11][UTT][ MSE = 0.2219 | LCC = 0.9544 | SRCC = 0.9556 ]
+INFO:root:Average loss=0.06737528908252716
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2053 | LCC = 0.9568 | SRCC = 0.9565 ]
+INFO:root:Average loss=0.060266792118549344
+INFO:root:
+[Valid][13][UTT][ MSE = 0.1941 | LCC = 0.9572 | SRCC = 0.9562 ]
+INFO:root:Average loss=0.05566529126465321
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2073 | LCC = 0.9576 | SRCC = 0.9571 ]
+INFO:root:Average loss=0.052824598386883734
+INFO:root:
+[Valid][15][UTT][ MSE = 0.1887 | LCC = 0.9592 | SRCC = 0.9582 ]
+INFO:root:Average loss=0.049190616264939306
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2054 | LCC = 0.9580 | SRCC = 0.9580 ]
+INFO:root:Average loss=0.04600373490154743
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1944 | LCC = 0.9581 | SRCC = 0.9573 ]
+INFO:root:Average loss=0.04593441557884216
+INFO:root:
+[Valid][18][UTT][ MSE = 0.1909 | LCC = 0.9585 | SRCC = 0.9576 ]
+INFO:root:Average loss=0.04096190817654133
+INFO:root:
+[Valid][19][UTT][ MSE = 0.2361 | LCC = 0.9555 | SRCC = 0.9565 ]
+INFO:root:Average loss=0.0421585553586483
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2020 | LCC = 0.9573 | SRCC = 0.9567 ]
+INFO:root:Average loss=0.039209578692913055
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1755 | LCC = 0.9584 | SRCC = 0.9572 ]
+INFO:root:Average loss=0.035189966961741447
+INFO:root:
+[Valid][22][UTT][ MSE = 0.2024 | LCC = 0.9577 | SRCC = 0.9573 ]
+INFO:root:Average loss=0.035884251922369007
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2137 | LCC = 0.9573 | SRCC = 0.9571 ]
+INFO:root:Average loss=0.03409900195896626
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1867 | LCC = 0.9573 | SRCC = 0.9562 ]
+INFO:root:Average loss=0.03412789665162563
+INFO:root:
+[Valid][25][UTT][ MSE = 0.2245 | LCC = 0.9564 | SRCC = 0.9563 ]
+INFO:root:Average loss=0.030063528805971147
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1909 | LCC = 0.9580 | SRCC = 0.9576 ]
+INFO:root:Average loss=0.02850383374094963
+INFO:root:
+[Valid][27][UTT][ MSE = 0.2043 | LCC = 0.9570 | SRCC = 0.9568 ]
+INFO:root:Average loss=0.026539348259568215
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1923 | LCC = 0.9572 | SRCC = 0.9564 ]
+INFO:root:Average loss=0.027131718397140504
+INFO:root:
+[Valid][29][UTT][ MSE = 0.2032 | LCC = 0.9571 | SRCC = 0.9573 ]
+INFO:root:Average loss=0.02667437396198511
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1941 | LCC = 0.9578 | SRCC = 0.9576 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.1959 | LCC = 0.9581 | SRCC = 0.9585 ]

results_tencent/w2v2_large_asr_en/layer12/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_large_asr_en/layer12/run0/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8f5490d3050280dd1d0dcd42aa4f25880b59d9ef1a4683b60a8b09daaf1744b2
+size 3403339

results_tencent/w2v2_large_asr_en/layer12/run0/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:37d30204eb352a8f128f1e245eaf9dccf89799781a279e4dbbbc3b2439a1437e
+size 3404998

results_tencent/w2v2_large_asr_en/layer12/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=0.45763952028751376
+INFO:root:
+[Valid][0][UTT][ MSE = 0.3437 | LCC = 0.9045 | SRCC = 0.9028 ]
+INFO:root:Average loss=0.24739523291587828
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3047 | LCC = 0.9260 | SRCC = 0.9258 ]
+INFO:root:Average loss=0.18281020152568817
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2355 | LCC = 0.9383 | SRCC = 0.9379 ]
+INFO:root:Average loss=0.15608748096227645
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2260 | LCC = 0.9424 | SRCC = 0.9421 ]
+INFO:root:Average loss=0.1361794403195381
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2345 | LCC = 0.9475 | SRCC = 0.9471 ]
+INFO:root:Average loss=0.12376651817560196
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2251 | LCC = 0.9479 | SRCC = 0.9481 ]
+INFO:root:Average loss=0.11315364986658097
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2690 | LCC = 0.9493 | SRCC = 0.9497 ]
+INFO:root:Average loss=0.1021771268248558
+INFO:root:
+[Valid][7][UTT][ MSE = 0.2839 | LCC = 0.9500 | SRCC = 0.9500 ]
+INFO:root:Average loss=0.09653241819143295
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2269 | LCC = 0.9508 | SRCC = 0.9509 ]
+INFO:root:Average loss=0.0871372711956501
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2007 | LCC = 0.9517 | SRCC = 0.9512 ]
+INFO:root:Average loss=0.08229440996050835
+INFO:root:
+[Valid][10][UTT][ MSE = 0.1970 | LCC = 0.9513 | SRCC = 0.9499 ]
+INFO:root:Average loss=0.07431881454586983
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1926 | LCC = 0.9520 | SRCC = 0.9512 ]
+INFO:root:Average loss=0.07261369603872299
+INFO:root:
+[Valid][12][UTT][ MSE = 0.2162 | LCC = 0.9520 | SRCC = 0.9511 ]
+INFO:root:Average loss=0.06527467966079711
+INFO:root:
+[Valid][13][UTT][ MSE = 0.2009 | LCC = 0.9535 | SRCC = 0.9523 ]
+INFO:root:Average loss=0.05860449504852295
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2104 | LCC = 0.9516 | SRCC = 0.9515 ]
+INFO:root:Average loss=0.057510033637285234
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2019 | LCC = 0.9529 | SRCC = 0.9523 ]
+INFO:root:Average loss=0.055362048029899594
+INFO:root:
+[Valid][16][UTT][ MSE = 0.1898 | LCC = 0.9530 | SRCC = 0.9514 ]
+INFO:root:Average loss=0.05099905371665955
+INFO:root:
+[Valid][17][UTT][ MSE = 0.1937 | LCC = 0.9522 | SRCC = 0.9506 ]
+INFO:root:Average loss=0.04922864617407322
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2055 | LCC = 0.9525 | SRCC = 0.9510 ]
+INFO:root:Average loss=0.04761053936183453
+INFO:root:
+[Valid][19][UTT][ MSE = 0.1911 | LCC = 0.9522 | SRCC = 0.9506 ]
+INFO:root:Average loss=0.043699118316173555
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2161 | LCC = 0.9512 | SRCC = 0.9502 ]
+INFO:root:Average loss=0.04208632671833038
+INFO:root:
+[Valid][21][UTT][ MSE = 0.2412 | LCC = 0.9494 | SRCC = 0.9483 ]
+INFO:root:Average loss=0.038514207944273945
+INFO:root:
+[Valid][22][UTT][ MSE = 0.2358 | LCC = 0.9523 | SRCC = 0.9519 ]
+INFO:root:Average loss=0.036645598217844966
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2005 | LCC = 0.9529 | SRCC = 0.9517 ]
+INFO:root:Average loss=0.03616778099536896
+INFO:root:
+[Valid][24][UTT][ MSE = 0.1920 | LCC = 0.9535 | SRCC = 0.9520 ]
+INFO:root:Average loss=0.03495833389461041
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1988 | LCC = 0.9514 | SRCC = 0.9511 ]
+INFO:root:Average loss=0.03206825006008148
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1873 | LCC = 0.9535 | SRCC = 0.9519 ]
+INFO:root:Average loss=0.030375116378068924
+INFO:root:
+[Valid][27][UTT][ MSE = 0.1942 | LCC = 0.9518 | SRCC = 0.9505 ]
+INFO:root:Average loss=0.030753920525312423
+INFO:root:
+[Valid][28][UTT][ MSE = 0.1980 | LCC = 0.9509 | SRCC = 0.9502 ]
+INFO:root:Average loss=0.028391597509384154
+INFO:root:
+[Valid][29][UTT][ MSE = 0.1949 | LCC = 0.9528 | SRCC = 0.9513 ]
+INFO:root:Average loss=0.026957260347902776
+INFO:root:
+[Valid][30][UTT][ MSE = 0.1927 | LCC = 0.9515 | SRCC = 0.9500 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2032 | LCC = 0.9542 | SRCC = 0.9545 ]

results_tencent/w2v2_large_asr_en/layer12/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_large_asr_en/layer12/run1/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:068f827b4e26008dac4f4ad6a7df63ec4066478904352478d90efbd4c561646f
+size 3403339

results_tencent/w2v2_large_asr_en/layer12/run1/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:abf5f71f20517f37cd1f136856396f91a6a0c6e2061cf47e29efbd942b4c05df
+size 3405062

results_tencent/w2v2_large_asr_en/layer12/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=0.5242924990653992
+INFO:root:
+[Valid][0][UTT][ MSE = 0.3788 | LCC = 0.8942 | SRCC = 0.8931 ]
+INFO:root:Average loss=0.24546742510795594
+INFO:root:
+[Valid][1][UTT][ MSE = 0.3027 | LCC = 0.9209 | SRCC = 0.9223 ]
+INFO:root:Average loss=0.18577863663434982
+INFO:root:
+[Valid][2][UTT][ MSE = 0.2813 | LCC = 0.9356 | SRCC = 0.9340 ]
+INFO:root:Average loss=0.16184018909931183
+INFO:root:
+[Valid][3][UTT][ MSE = 0.2496 | LCC = 0.9412 | SRCC = 0.9415 ]
+INFO:root:Average loss=0.13380484807491302
+INFO:root:
+[Valid][4][UTT][ MSE = 0.2195 | LCC = 0.9447 | SRCC = 0.9466 ]
+INFO:root:Average loss=0.12330664277076721
+INFO:root:
+[Valid][5][UTT][ MSE = 0.2310 | LCC = 0.9502 | SRCC = 0.9500 ]
+INFO:root:Average loss=0.10776367819309235
+INFO:root:
+[Valid][6][UTT][ MSE = 0.2348 | LCC = 0.9504 | SRCC = 0.9501 ]
+INFO:root:Average loss=0.10152350184321403
+INFO:root:
+[Valid][7][UTT][ MSE = 0.1826 | LCC = 0.9526 | SRCC = 0.9525 ]
+INFO:root:Average loss=0.09220429676771164
+INFO:root:
+[Valid][8][UTT][ MSE = 0.2323 | LCC = 0.9520 | SRCC = 0.9519 ]
+INFO:root:Average loss=0.08813998591899871
+INFO:root:
+[Valid][9][UTT][ MSE = 0.2221 | LCC = 0.9508 | SRCC = 0.9503 ]
+INFO:root:Average loss=0.08187764972448348
+INFO:root:
+[Valid][10][UTT][ MSE = 0.1963 | LCC = 0.9533 | SRCC = 0.9525 ]
+INFO:root:Average loss=0.07334526661038399
+INFO:root:
+[Valid][11][UTT][ MSE = 0.1770 | LCC = 0.9530 | SRCC = 0.9522 ]
+INFO:root:Average loss=0.06764051532745362
+INFO:root:
+[Valid][12][UTT][ MSE = 0.1863 | LCC = 0.9535 | SRCC = 0.9526 ]
+INFO:root:Average loss=0.06379980775713921
+INFO:root:
+[Valid][13][UTT][ MSE = 0.1914 | LCC = 0.9536 | SRCC = 0.9529 ]
+INFO:root:Average loss=0.057914644062519074
+INFO:root:
+[Valid][14][UTT][ MSE = 0.2259 | LCC = 0.9540 | SRCC = 0.9534 ]
+INFO:root:Average loss=0.05512319526076317
+INFO:root:
+[Valid][15][UTT][ MSE = 0.2205 | LCC = 0.9510 | SRCC = 0.9500 ]
+INFO:root:Average loss=0.05528501798212528
+INFO:root:
+[Valid][16][UTT][ MSE = 0.2199 | LCC = 0.9544 | SRCC = 0.9541 ]
+INFO:root:Average loss=0.049439133182168006
+INFO:root:
+[Valid][17][UTT][ MSE = 0.2171 | LCC = 0.9530 | SRCC = 0.9525 ]
+INFO:root:Average loss=0.049794646456837656
+INFO:root:
+[Valid][18][UTT][ MSE = 0.2210 | LCC = 0.9545 | SRCC = 0.9539 ]
+INFO:root:Average loss=0.045176528945565224
+INFO:root:
+[Valid][19][UTT][ MSE = 0.2093 | LCC = 0.9544 | SRCC = 0.9537 ]
+INFO:root:Average loss=0.040274559304118156
+INFO:root:
+[Valid][20][UTT][ MSE = 0.2322 | LCC = 0.9528 | SRCC = 0.9519 ]
+INFO:root:Average loss=0.037963664829730986
+INFO:root:
+[Valid][21][UTT][ MSE = 0.1920 | LCC = 0.9537 | SRCC = 0.9527 ]
+INFO:root:Average loss=0.03684345783293247
+INFO:root:
+[Valid][22][UTT][ MSE = 0.1989 | LCC = 0.9549 | SRCC = 0.9538 ]
+INFO:root:Average loss=0.037093067303299906
+INFO:root:
+[Valid][23][UTT][ MSE = 0.2083 | LCC = 0.9552 | SRCC = 0.9541 ]
+INFO:root:Average loss=0.03685079918801785
+INFO:root:
+[Valid][24][UTT][ MSE = 0.2049 | LCC = 0.9553 | SRCC = 0.9544 ]
+INFO:root:Average loss=0.032861173123121264
+INFO:root:
+[Valid][25][UTT][ MSE = 0.1961 | LCC = 0.9542 | SRCC = 0.9538 ]
+INFO:root:Average loss=0.030662636265158652
+INFO:root:
+[Valid][26][UTT][ MSE = 0.1996 | LCC = 0.9541 | SRCC = 0.9537 ]
+INFO:root:Average loss=0.029129821971058845
+INFO:root:
+[Valid][27][UTT][ MSE = 0.2035 | LCC = 0.9540 | SRCC = 0.9536 ]
+INFO:root:Average loss=0.029114353954792022
+INFO:root:
+[Valid][28][UTT][ MSE = 0.2031 | LCC = 0.9548 | SRCC = 0.9543 ]
+INFO:root:Average loss=0.0274615458548069
+INFO:root:
+[Valid][29][UTT][ MSE = 0.2411 | LCC = 0.9537 | SRCC = 0.9531 ]
+INFO:root:Average loss=0.026666629649698733
+INFO:root:
+[Valid][30][UTT][ MSE = 0.2095 | LCC = 0.9546 | SRCC = 0.9535 ]
+INFO:root:
+[Test][31][UTT][ MSE = 0.2065 | LCC = 0.9542 | SRCC = 0.9548 ]

results_tencent/w2v2_large_asr_en/layer12/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_large_asr_en/layer12/run2/model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6e4ac779f0ae2adf79ab1e5c41fedf75f8cdec951fbba9079c3e19c2aa006334
+size 3403339

results_tencent/w2v2_large_asr_en/layer12/run2/model_best.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c18347eb9d88284d63e91c24739e6e1a6703c195440327d404f064c14e959731
+size 3405062