diff --git a/results_bvcc/w2v2_base/layer0/run2/model.pt b/results_bvcc/w2v2_base/layer0/run2/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..cbf99855aa8d432e1af8945320a417567206a573 --- /dev/null +++ b/results_bvcc/w2v2_base/layer0/run2/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:39072446dd19190b27a6518cbaded35e7f226989b4b6b1c0c861f8e2983921c1 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer0/run2/model_best.pt b/results_bvcc/w2v2_base/layer0/run2/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..ce59751b1d9681fe62fb2609b2ab8ebfb5d372ea --- /dev/null +++ b/results_bvcc/w2v2_base/layer0/run2/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2479c85d5a78aafa1167c7cbbb43de37648403333e991cb561c4fecdb8a68427 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer0/run2/train.log b/results_bvcc/w2v2_base/layer0/run2/train.log new file mode 100644 index 0000000000000000000000000000000000000000..5e5bfc7df2e65d07af2034cd91812f2bd37517b1 --- /dev/null +++ b/results_bvcc/w2v2_base/layer0/run2/train.log @@ -0,0 +1,163 @@ +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=0.5249429885775615 +INFO:root: +[Valid][0][UTT][ MSE = 0.4273 | LCC = 0.7562 | SRCC = 0.7566 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.3347 | LCC = 0.8364 | SRCC = 0.8395 ] +INFO:root:Average loss=0.30377670434805065 +INFO:root: +[Valid][1][UTT][ MSE = 0.3053 | LCC = 0.8010 | SRCC = 0.7986 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1880 | LCC = 0.8753 | SRCC = 0.8653 ] +INFO:root:Average loss=0.2591542744865784 +INFO:root: +[Valid][2][UTT][ MSE = 0.2888 | LCC = 0.8165 | SRCC = 0.8144 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1695 | LCC = 0.8919 | SRCC = 0.8861 ] +INFO:root:Average loss=0.23760433877125764 +INFO:root: +[Valid][3][UTT][ MSE = 0.2803 | LCC = 0.8266 | SRCC = 0.8273 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1720 | LCC = 0.9016 | SRCC = 0.9013 ] +INFO:root:Average loss=0.21500673259680086 +INFO:root: +[Valid][4][UTT][ MSE = 0.2706 | LCC = 0.8306 | SRCC = 0.8277 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1630 | LCC = 0.9061 | SRCC = 0.8987 ] +INFO:root:Average loss=0.2158757595297618 +INFO:root: +[Valid][5][UTT][ MSE = 0.2761 | LCC = 0.8284 | SRCC = 0.8302 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1438 | LCC = 0.9066 | SRCC = 0.9067 ] +INFO:root:Average loss=0.21459559991191596 +INFO:root: +[Valid][6][UTT][ MSE = 0.2816 | LCC = 0.8356 | SRCC = 0.8333 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1471 | LCC = 0.9174 | SRCC = 0.9117 ] +INFO:root:Average loss=0.1934651486002482 +INFO:root: +[Valid][7][UTT][ MSE = 0.2609 | LCC = 0.8387 | SRCC = 0.8361 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1405 | LCC = 0.9237 | SRCC = 0.9188 ] +INFO:root:Average loss=0.17381173763901758 +INFO:root: +[Valid][8][UTT][ MSE = 0.2560 | LCC = 0.8451 | SRCC = 0.8417 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1467 | LCC = 0.9229 | SRCC = 0.9171 ] +INFO:root:Average loss=0.17481609118672517 +INFO:root: +[Valid][9][UTT][ MSE = 0.2483 | LCC = 0.8449 | SRCC = 0.8430 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1314 | LCC = 0.9251 | SRCC = 0.9197 ] +INFO:root:Average loss=0.17029272182247576 +INFO:root: +[Valid][10][UTT][ MSE = 0.2478 | LCC = 0.8429 | SRCC = 0.8405 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1252 | LCC = 0.9212 | SRCC = 0.9163 ] +INFO:root:Average loss=0.1518475871819716 +INFO:root: +[Valid][11][UTT][ MSE = 0.2435 | LCC = 0.8453 | SRCC = 0.8433 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1191 | LCC = 0.9256 | SRCC = 0.9191 ] +INFO:root:Average loss=0.1431034279939456 +INFO:root: +[Valid][12][UTT][ MSE = 0.2419 | LCC = 0.8462 | SRCC = 0.8440 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1183 | LCC = 0.9252 | SRCC = 0.9196 ] +INFO:root:Average loss=0.13872973315226725 +INFO:root: +[Valid][13][UTT][ MSE = 0.2482 | LCC = 0.8422 | SRCC = 0.8420 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1253 | LCC = 0.9246 | SRCC = 0.9228 ] +INFO:root:Average loss=0.1254938454964222 +INFO:root: +[Valid][14][UTT][ MSE = 0.2690 | LCC = 0.8387 | SRCC = 0.8416 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1587 | LCC = 0.9203 | SRCC = 0.9232 ] +INFO:root:Average loss=0.12176823959900783 +INFO:root: +[Valid][15][UTT][ MSE = 0.2515 | LCC = 0.8395 | SRCC = 0.8416 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1324 | LCC = 0.9226 | SRCC = 0.9240 ] +INFO:root:Average loss=0.11692235389581093 +INFO:root: +[Valid][16][UTT][ MSE = 0.2548 | LCC = 0.8403 | SRCC = 0.8428 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1451 | LCC = 0.9205 | SRCC = 0.9239 ] +INFO:root:Average loss=0.11832601949572563 +INFO:root: +[Valid][17][UTT][ MSE = 0.2401 | LCC = 0.8479 | SRCC = 0.8469 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1205 | LCC = 0.9271 | SRCC = 0.9238 ] +INFO:root:Average loss=0.10400467728957152 +INFO:root: +[Valid][18][UTT][ MSE = 0.2378 | LCC = 0.8469 | SRCC = 0.8464 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1086 | LCC = 0.9307 | SRCC = 0.9273 ] +INFO:root:Average loss=0.10339826460068043 +INFO:root: +[Valid][19][UTT][ MSE = 0.2356 | LCC = 0.8478 | SRCC = 0.8469 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1147 | LCC = 0.9298 | SRCC = 0.9276 ] +INFO:root:Average loss=0.10017874765281494 +INFO:root: +[Valid][20][UTT][ MSE = 0.2394 | LCC = 0.8518 | SRCC = 0.8508 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1305 | LCC = 0.9318 | SRCC = 0.9281 ] +INFO:root:Average loss=0.09424310903518628 +INFO:root: +[Valid][21][UTT][ MSE = 0.2368 | LCC = 0.8460 | SRCC = 0.8457 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1091 | LCC = 0.9294 | SRCC = 0.9279 ] +INFO:root:Average loss=0.09064795965185532 +INFO:root: +[Valid][22][UTT][ MSE = 0.2336 | LCC = 0.8543 | SRCC = 0.8507 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1214 | LCC = 0.9338 | SRCC = 0.9255 ] +INFO:root:Average loss=0.09291106758591457 +INFO:root: +[Valid][23][UTT][ MSE = 0.2494 | LCC = 0.8492 | SRCC = 0.8489 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1401 | LCC = 0.9328 | SRCC = 0.9335 ] +INFO:root:Average loss=0.0867785691546324 +INFO:root: +[Valid][24][UTT][ MSE = 0.2381 | LCC = 0.8477 | SRCC = 0.8477 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1199 | LCC = 0.9325 | SRCC = 0.9303 ] +INFO:root:Average loss=0.08516144384749424 +INFO:root: +[Valid][25][UTT][ MSE = 0.2410 | LCC = 0.8479 | SRCC = 0.8472 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1080 | LCC = 0.9384 | SRCC = 0.9383 ] +INFO:root:Average loss=0.08252089599577281 +INFO:root: +[Valid][26][UTT][ MSE = 0.2298 | LCC = 0.8537 | SRCC = 0.8518 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1116 | LCC = 0.9307 | SRCC = 0.9273 ] +INFO:root:Average loss=0.0788705332730061 +INFO:root: +[Valid][27][UTT][ MSE = 0.2392 | LCC = 0.8526 | SRCC = 0.8505 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1313 | LCC = 0.9333 | SRCC = 0.9303 ] +INFO:root:Average loss=0.07371497244980091 +INFO:root: +[Valid][28][UTT][ MSE = 0.2288 | LCC = 0.8548 | SRCC = 0.8532 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1076 | LCC = 0.9363 | SRCC = 0.9329 ] +INFO:root:Average loss=0.06961088651456894 +INFO:root: +[Valid][29][UTT][ MSE = 0.2316 | LCC = 0.8551 | SRCC = 0.8536 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1062 | LCC = 0.9368 | SRCC = 0.9344 ] +INFO:root:Average loss=0.07108233381922428 +INFO:root: +[Valid][30][UTT][ MSE = 0.2239 | LCC = 0.8575 | SRCC = 0.8542 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1046 | LCC = 0.9353 | SRCC = 0.9271 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2422 | LCC = 0.8542 | SRCC = 0.8538 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1504 | LCC = 0.9074 | SRCC = 0.9035 ] diff --git a/results_bvcc/w2v2_base/layer0/run3/config.gin b/results_bvcc/w2v2_base/layer0/run3/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer0/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 diff --git a/results_bvcc/w2v2_base/layer0/run3/model.pt b/results_bvcc/w2v2_base/layer0/run3/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..772e9e1a52f40718ee3629ad79441b959db34833 --- /dev/null +++ b/results_bvcc/w2v2_base/layer0/run3/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fc4a982063b4ffb202b6ac254c3799359012b89e76327055d54f1c9b1ec838d5 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer0/run3/model_best.pt b/results_bvcc/w2v2_base/layer0/run3/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..b65ccdd2e5c115025cf3d12f2eef40d424852479 --- /dev/null +++ b/results_bvcc/w2v2_base/layer0/run3/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f7621b5451fa3fb81d41231986683d3e6643a5009211efbbae03ed650229ec95 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer0/run3/train.log b/results_bvcc/w2v2_base/layer0/run3/train.log new file mode 100644 index 0000000000000000000000000000000000000000..918a5550a379a47e469c45b01e0288cad1e6840b --- /dev/null +++ b/results_bvcc/w2v2_base/layer0/run3/train.log @@ -0,0 +1,163 @@ +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=0.563079174894553 +INFO:root: +[Valid][0][UTT][ MSE = 0.4337 | LCC = 0.7487 | SRCC = 0.7508 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2867 | LCC = 0.8306 | SRCC = 0.8267 ] +INFO:root:Average loss=0.33041678082484466 +INFO:root: +[Valid][1][UTT][ MSE = 0.3267 | LCC = 0.7845 | SRCC = 0.7847 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1826 | LCC = 0.8734 | SRCC = 0.8705 ] +INFO:root:Average loss=0.27019520084827375 +INFO:root: +[Valid][2][UTT][ MSE = 0.3007 | LCC = 0.8091 | SRCC = 0.8075 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1818 | LCC = 0.8906 | SRCC = 0.8862 ] +INFO:root:Average loss=0.2408250344869418 +INFO:root: +[Valid][3][UTT][ MSE = 0.2780 | LCC = 0.8216 | SRCC = 0.8216 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1474 | LCC = 0.9006 | SRCC = 0.8963 ] +INFO:root:Average loss=0.2187499035245333 +INFO:root: +[Valid][4][UTT][ MSE = 0.2767 | LCC = 0.8258 | SRCC = 0.8234 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1388 | LCC = 0.9111 | SRCC = 0.9036 ] +INFO:root:Average loss=0.2035780111566568 +INFO:root: +[Valid][5][UTT][ MSE = 0.3005 | LCC = 0.8413 | SRCC = 0.8387 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1667 | LCC = 0.9103 | SRCC = 0.9020 ] +INFO:root:Average loss=0.19972144229671893 +INFO:root: +[Valid][6][UTT][ MSE = 0.2593 | LCC = 0.8409 | SRCC = 0.8403 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1387 | LCC = 0.9159 | SRCC = 0.9150 ] +INFO:root:Average loss=0.18598864466334 +INFO:root: +[Valid][7][UTT][ MSE = 0.2857 | LCC = 0.8395 | SRCC = 0.8397 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1415 | LCC = 0.9165 | SRCC = 0.9147 ] +INFO:root:Average loss=0.17202930343456757 +INFO:root: +[Valid][8][UTT][ MSE = 0.2496 | LCC = 0.8409 | SRCC = 0.8413 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1294 | LCC = 0.9146 | SRCC = 0.9129 ] +INFO:root:Average loss=0.1724835539666506 +INFO:root: +[Valid][9][UTT][ MSE = 0.2536 | LCC = 0.8402 | SRCC = 0.8403 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1263 | LCC = 0.9170 | SRCC = 0.9169 ] +INFO:root:Average loss=0.15639462551245323 +INFO:root: +[Valid][10][UTT][ MSE = 0.2747 | LCC = 0.8431 | SRCC = 0.8417 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1342 | LCC = 0.9183 | SRCC = 0.9163 ] +INFO:root:Average loss=0.146468396179187 +INFO:root: +[Valid][11][UTT][ MSE = 0.2483 | LCC = 0.8401 | SRCC = 0.8411 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1328 | LCC = 0.9172 | SRCC = 0.9202 ] +INFO:root:Average loss=0.1407206617295742 +INFO:root: +[Valid][12][UTT][ MSE = 0.2554 | LCC = 0.8406 | SRCC = 0.8417 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1231 | LCC = 0.9119 | SRCC = 0.9137 ] +INFO:root:Average loss=0.1333160166365978 +INFO:root: +[Valid][13][UTT][ MSE = 0.2423 | LCC = 0.8498 | SRCC = 0.8501 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1120 | LCC = 0.9260 | SRCC = 0.9263 ] +INFO:root:Average loss=0.1278129292604251 +INFO:root: +[Valid][14][UTT][ MSE = 0.2507 | LCC = 0.8486 | SRCC = 0.8471 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1249 | LCC = 0.9207 | SRCC = 0.9179 ] +INFO:root:Average loss=0.11769445154529351 +INFO:root: +[Valid][15][UTT][ MSE = 0.2547 | LCC = 0.8402 | SRCC = 0.8384 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1506 | LCC = 0.9111 | SRCC = 0.9063 ] +INFO:root:Average loss=0.11952666766368426 +INFO:root: +[Valid][16][UTT][ MSE = 0.2770 | LCC = 0.8528 | SRCC = 0.8521 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1419 | LCC = 0.9216 | SRCC = 0.9163 ] +INFO:root:Average loss=0.1103925523467553 +INFO:root: +[Valid][17][UTT][ MSE = 0.2377 | LCC = 0.8482 | SRCC = 0.8477 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1205 | LCC = 0.9235 | SRCC = 0.9231 ] +INFO:root:Average loss=0.1063495807063121 +INFO:root: +[Valid][18][UTT][ MSE = 0.2468 | LCC = 0.8488 | SRCC = 0.8468 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1251 | LCC = 0.9173 | SRCC = 0.9115 ] +INFO:root:Average loss=0.09886931179043575 +INFO:root: +[Valid][19][UTT][ MSE = 0.2427 | LCC = 0.8461 | SRCC = 0.8453 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1137 | LCC = 0.9232 | SRCC = 0.9202 ] +INFO:root:Average loss=0.10060006232024768 +INFO:root: +[Valid][20][UTT][ MSE = 0.2573 | LCC = 0.8453 | SRCC = 0.8445 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1289 | LCC = 0.9172 | SRCC = 0.9168 ] +INFO:root:Average loss=0.09440205828883709 +INFO:root: +[Valid][21][UTT][ MSE = 0.2490 | LCC = 0.8523 | SRCC = 0.8508 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1207 | LCC = 0.9225 | SRCC = 0.9174 ] +INFO:root:Average loss=0.08925941424109997 +INFO:root: +[Valid][22][UTT][ MSE = 0.2360 | LCC = 0.8490 | SRCC = 0.8479 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1294 | LCC = 0.9212 | SRCC = 0.9222 ] +INFO:root:Average loss=0.10100043240265968 +INFO:root: +[Valid][23][UTT][ MSE = 0.2638 | LCC = 0.8526 | SRCC = 0.8519 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1328 | LCC = 0.9226 | SRCC = 0.9184 ] +INFO:root:Average loss=0.08822186515690425 +INFO:root: +[Valid][24][UTT][ MSE = 0.2478 | LCC = 0.8504 | SRCC = 0.8506 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1212 | LCC = 0.9195 | SRCC = 0.9163 ] +INFO:root:Average loss=0.08835151099050656 +INFO:root: +[Valid][25][UTT][ MSE = 0.2535 | LCC = 0.8521 | SRCC = 0.8513 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1201 | LCC = 0.9272 | SRCC = 0.9232 ] +INFO:root:Average loss=0.08405076368496968 +INFO:root: +[Valid][26][UTT][ MSE = 0.2364 | LCC = 0.8467 | SRCC = 0.8469 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1268 | LCC = 0.9179 | SRCC = 0.9190 ] +INFO:root:Average loss=0.07794641779783444 +INFO:root: +[Valid][27][UTT][ MSE = 0.2393 | LCC = 0.8504 | SRCC = 0.8501 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1070 | LCC = 0.9272 | SRCC = 0.9239 ] +INFO:root:Average loss=0.07662971938649814 +INFO:root: +[Valid][28][UTT][ MSE = 0.2442 | LCC = 0.8480 | SRCC = 0.8488 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1081 | LCC = 0.9257 | SRCC = 0.9264 ] +INFO:root:Average loss=0.07344848729479007 +INFO:root: +[Valid][29][UTT][ MSE = 0.2342 | LCC = 0.8485 | SRCC = 0.8484 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1131 | LCC = 0.9235 | SRCC = 0.9243 ] +INFO:root:Average loss=0.07641311434025948 +INFO:root: +[Valid][30][UTT][ MSE = 0.2306 | LCC = 0.8542 | SRCC = 0.8536 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1115 | LCC = 0.9230 | SRCC = 0.9198 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2582 | LCC = 0.8525 | SRCC = 0.8507 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1560 | LCC = 0.9164 | SRCC = 0.9162 ] diff --git a/results_bvcc/w2v2_base/layer0/run4/config.gin b/results_bvcc/w2v2_base/layer0/run4/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer0/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 diff --git a/results_bvcc/w2v2_base/layer0/run4/model.pt b/results_bvcc/w2v2_base/layer0/run4/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..132751ef5c408fcc207617e64cde47cef7eb5084 --- /dev/null +++ b/results_bvcc/w2v2_base/layer0/run4/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3b892c4bc00d64d85ad907801d74ae2ed51bae41731b27c3287caa565c085a67 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer0/run4/model_best.pt b/results_bvcc/w2v2_base/layer0/run4/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..7b13744a0f81d721931e9699f6ab1e0d5e29d898 --- /dev/null +++ b/results_bvcc/w2v2_base/layer0/run4/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c5ac614f976d7e5c1af693df380c3438a8015b518b2035d789cb7abb4e5070d2 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer0/run4/train.log b/results_bvcc/w2v2_base/layer0/run4/train.log new file mode 100644 index 0000000000000000000000000000000000000000..9ed88c4a167561aac3e0ed229d4cffd09cf32fea --- /dev/null +++ b/results_bvcc/w2v2_base/layer0/run4/train.log @@ -0,0 +1,163 @@ +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=0.5348865794829833 +INFO:root: +[Valid][0][UTT][ MSE = 0.3870 | LCC = 0.7508 | SRCC = 0.7568 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2511 | LCC = 0.8399 | SRCC = 0.8412 ] +INFO:root:Average loss=0.30528190120672566 +INFO:root: +[Valid][1][UTT][ MSE = 0.3300 | LCC = 0.7935 | SRCC = 0.7971 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1870 | LCC = 0.8798 | SRCC = 0.8809 ] +INFO:root:Average loss=0.27072139905813414 +INFO:root: +[Valid][2][UTT][ MSE = 0.3061 | LCC = 0.8173 | SRCC = 0.8150 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1760 | LCC = 0.8958 | SRCC = 0.8839 ] +INFO:root:Average loss=0.24719594438106585 +INFO:root: +[Valid][3][UTT][ MSE = 0.2898 | LCC = 0.8216 | SRCC = 0.8233 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1825 | LCC = 0.9025 | SRCC = 0.9002 ] +INFO:root:Average loss=0.2230509876822814 +INFO:root: +[Valid][4][UTT][ MSE = 0.2595 | LCC = 0.8320 | SRCC = 0.8328 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1384 | LCC = 0.9110 | SRCC = 0.9070 ] +INFO:root:Average loss=0.21045147875944772 +INFO:root: +[Valid][5][UTT][ MSE = 0.2662 | LCC = 0.8380 | SRCC = 0.8368 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1260 | LCC = 0.9239 | SRCC = 0.9178 ] +INFO:root:Average loss=0.19907851985249764 +INFO:root: +[Valid][6][UTT][ MSE = 0.2598 | LCC = 0.8359 | SRCC = 0.8356 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1449 | LCC = 0.9130 | SRCC = 0.9083 ] +INFO:root:Average loss=0.18605596562608695 +INFO:root: +[Valid][7][UTT][ MSE = 0.2725 | LCC = 0.8359 | SRCC = 0.8348 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1601 | LCC = 0.9201 | SRCC = 0.9151 ] +INFO:root:Average loss=0.17120693194178435 +INFO:root: +[Valid][8][UTT][ MSE = 0.2497 | LCC = 0.8443 | SRCC = 0.8440 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1271 | LCC = 0.9212 | SRCC = 0.9187 ] +INFO:root:Average loss=0.1676345711144117 +INFO:root: +[Valid][9][UTT][ MSE = 0.2577 | LCC = 0.8432 | SRCC = 0.8428 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1316 | LCC = 0.9158 | SRCC = 0.9108 ] +INFO:root:Average loss=0.1591709982126187 +INFO:root: +[Valid][10][UTT][ MSE = 0.2489 | LCC = 0.8445 | SRCC = 0.8425 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1471 | LCC = 0.9161 | SRCC = 0.9117 ] +INFO:root:Average loss=0.1513063931503357 +INFO:root: +[Valid][11][UTT][ MSE = 0.2455 | LCC = 0.8420 | SRCC = 0.8415 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1280 | LCC = 0.9207 | SRCC = 0.9172 ] +INFO:root:Average loss=0.1426447409276779 +INFO:root: +[Valid][12][UTT][ MSE = 0.2440 | LCC = 0.8489 | SRCC = 0.8470 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1391 | LCC = 0.9214 | SRCC = 0.9171 ] +INFO:root:Average loss=0.13574676148784467 +INFO:root: +[Valid][13][UTT][ MSE = 0.2714 | LCC = 0.8429 | SRCC = 0.8428 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1724 | LCC = 0.9209 | SRCC = 0.9193 ] +INFO:root:Average loss=0.12894271820401534 +INFO:root: +[Valid][14][UTT][ MSE = 0.2481 | LCC = 0.8496 | SRCC = 0.8490 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1459 | LCC = 0.9257 | SRCC = 0.9230 ] +INFO:root:Average loss=0.12140956731178822 +INFO:root: +[Valid][15][UTT][ MSE = 0.2455 | LCC = 0.8519 | SRCC = 0.8506 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1155 | LCC = 0.9284 | SRCC = 0.9241 ] +INFO:root:Average loss=0.1175903591972131 +INFO:root: +[Valid][16][UTT][ MSE = 0.2432 | LCC = 0.8484 | SRCC = 0.8483 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1351 | LCC = 0.9256 | SRCC = 0.9244 ] +INFO:root:Average loss=0.12394394878393565 +INFO:root: +[Valid][17][UTT][ MSE = 0.2360 | LCC = 0.8479 | SRCC = 0.8470 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1200 | LCC = 0.9205 | SRCC = 0.9185 ] +INFO:root:Average loss=0.10359991466005643 +INFO:root: +[Valid][18][UTT][ MSE = 0.2292 | LCC = 0.8530 | SRCC = 0.8525 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1115 | LCC = 0.9263 | SRCC = 0.9236 ] +INFO:root:Average loss=0.10143001272510259 +INFO:root: +[Valid][19][UTT][ MSE = 0.2323 | LCC = 0.8522 | SRCC = 0.8506 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1221 | LCC = 0.9221 | SRCC = 0.9172 ] +INFO:root:Average loss=0.09811931160780099 +INFO:root: +[Valid][20][UTT][ MSE = 0.2316 | LCC = 0.8523 | SRCC = 0.8503 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1170 | LCC = 0.9257 | SRCC = 0.9210 ] +INFO:root:Average loss=0.10091435054364876 +INFO:root: +[Valid][21][UTT][ MSE = 0.2368 | LCC = 0.8521 | SRCC = 0.8509 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1222 | LCC = 0.9238 | SRCC = 0.9222 ] +INFO:root:Average loss=0.09523546003187314 +INFO:root: +[Valid][22][UTT][ MSE = 0.2291 | LCC = 0.8537 | SRCC = 0.8521 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1136 | LCC = 0.9258 | SRCC = 0.9201 ] +INFO:root:Average loss=0.09089563381022368 +INFO:root: +[Valid][23][UTT][ MSE = 0.2247 | LCC = 0.8544 | SRCC = 0.8542 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1146 | LCC = 0.9247 | SRCC = 0.9249 ] +INFO:root:Average loss=0.08332412919172874 +INFO:root: +[Valid][24][UTT][ MSE = 0.2285 | LCC = 0.8525 | SRCC = 0.8506 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1080 | LCC = 0.9294 | SRCC = 0.9245 ] +INFO:root:Average loss=0.08792871752610573 +INFO:root: +[Valid][25][UTT][ MSE = 0.2277 | LCC = 0.8567 | SRCC = 0.8546 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1191 | LCC = 0.9283 | SRCC = 0.9248 ] +INFO:root:Average loss=0.08099713615882091 +INFO:root: +[Valid][26][UTT][ MSE = 0.2318 | LCC = 0.8537 | SRCC = 0.8520 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1259 | LCC = 0.9279 | SRCC = 0.9247 ] +INFO:root:Average loss=0.08232074331205624 +INFO:root: +[Valid][27][UTT][ MSE = 0.2305 | LCC = 0.8555 | SRCC = 0.8539 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1268 | LCC = 0.9290 | SRCC = 0.9274 ] +INFO:root:Average loss=0.07562345275894189 +INFO:root: +[Valid][28][UTT][ MSE = 0.2202 | LCC = 0.8610 | SRCC = 0.8587 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1060 | LCC = 0.9394 | SRCC = 0.9350 ] +INFO:root:Average loss=0.07162571507386672 +INFO:root: +[Valid][29][UTT][ MSE = 0.2280 | LCC = 0.8543 | SRCC = 0.8536 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1162 | LCC = 0.9325 | SRCC = 0.9302 ] +INFO:root:Average loss=0.0684171619419104 +INFO:root: +[Valid][30][UTT][ MSE = 0.2380 | LCC = 0.8528 | SRCC = 0.8505 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1342 | LCC = 0.9304 | SRCC = 0.9282 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2461 | LCC = 0.8471 | SRCC = 0.8464 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1242 | LCC = 0.9184 | SRCC = 0.9143 ] diff --git a/results_bvcc/w2v2_base/layer1/run0/config.gin b/results_bvcc/w2v2_base/layer1/run0/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer1/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 diff --git a/results_bvcc/w2v2_base/layer1/run0/model.pt b/results_bvcc/w2v2_base/layer1/run0/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..85b016395593a2f07064685ea55b8896d7c9eb12 --- /dev/null +++ b/results_bvcc/w2v2_base/layer1/run0/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d74b445d94921e01df770e0fca2317b4575a18bcffa59b43f6efcaa870061321 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer1/run0/model_best.pt b/results_bvcc/w2v2_base/layer1/run0/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..e2c58e8477cd65710ce29b3e22f9b7a8f59318b5 --- /dev/null +++ b/results_bvcc/w2v2_base/layer1/run0/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:82325465fff9a7757bdaa487f00ec8bfa4d39ab17fbebef504c32e9a15e747ae +size 2749702 diff --git a/results_bvcc/w2v2_base/layer1/run0/train.log b/results_bvcc/w2v2_base/layer1/run0/train.log new file mode 100644 index 0000000000000000000000000000000000000000..c9b4a020bcdb4a4605bf200a3429ee874bb94658 --- /dev/null +++ b/results_bvcc/w2v2_base/layer1/run0/train.log @@ -0,0 +1,163 @@ +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=0.5303507426228279 +INFO:root: +[Valid][0][UTT][ MSE = 0.3587 | LCC = 0.7716 | SRCC = 0.7740 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2443 | LCC = 0.8536 | SRCC = 0.8533 ] +INFO:root:Average loss=0.273792809400803 +INFO:root: +[Valid][1][UTT][ MSE = 0.2932 | LCC = 0.8067 | SRCC = 0.8129 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1614 | LCC = 0.8929 | SRCC = 0.9015 ] +INFO:root:Average loss=0.2344587573256248 +INFO:root: +[Valid][2][UTT][ MSE = 0.2788 | LCC = 0.8335 | SRCC = 0.8347 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1712 | LCC = 0.9109 | SRCC = 0.9159 ] +INFO:root:Average loss=0.21123353793070868 +INFO:root: +[Valid][3][UTT][ MSE = 0.2564 | LCC = 0.8466 | SRCC = 0.8460 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1470 | LCC = 0.9214 | SRCC = 0.9252 ] +INFO:root:Average loss=0.19582427608279082 +INFO:root: +[Valid][4][UTT][ MSE = 0.2817 | LCC = 0.8441 | SRCC = 0.8424 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1735 | LCC = 0.9215 | SRCC = 0.9235 ] +INFO:root:Average loss=0.18156705643886176 +INFO:root: +[Valid][5][UTT][ MSE = 0.2363 | LCC = 0.8535 | SRCC = 0.8534 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1150 | LCC = 0.9334 | SRCC = 0.9386 ] +INFO:root:Average loss=0.17155700033673874 +INFO:root: +[Valid][6][UTT][ MSE = 0.2387 | LCC = 0.8533 | SRCC = 0.8524 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1198 | LCC = 0.9340 | SRCC = 0.9375 ] +INFO:root:Average loss=0.1611969740344928 +INFO:root: +[Valid][7][UTT][ MSE = 0.2486 | LCC = 0.8581 | SRCC = 0.8560 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1405 | LCC = 0.9351 | SRCC = 0.9363 ] +INFO:root:Average loss=0.14487831963178438 +INFO:root: +[Valid][8][UTT][ MSE = 0.2628 | LCC = 0.8577 | SRCC = 0.8555 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1570 | LCC = 0.9384 | SRCC = 0.9398 ] +INFO:root:Average loss=0.14753570140172273 +INFO:root: +[Valid][9][UTT][ MSE = 0.2528 | LCC = 0.8597 | SRCC = 0.8595 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1479 | LCC = 0.9385 | SRCC = 0.9427 ] +INFO:root:Average loss=0.13028067197555152 +INFO:root: +[Valid][10][UTT][ MSE = 0.2273 | LCC = 0.8649 | SRCC = 0.8641 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1176 | LCC = 0.9425 | SRCC = 0.9470 ] +INFO:root:Average loss=0.11848474179322903 +INFO:root: +[Valid][11][UTT][ MSE = 0.2503 | LCC = 0.8592 | SRCC = 0.8575 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1492 | LCC = 0.9336 | SRCC = 0.9342 ] +INFO:root:Average loss=0.11848746966093014 +INFO:root: +[Valid][12][UTT][ MSE = 0.2657 | LCC = 0.8569 | SRCC = 0.8577 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1625 | LCC = 0.9322 | SRCC = 0.9380 ] +INFO:root:Average loss=0.10621509194756165 +INFO:root: +[Valid][13][UTT][ MSE = 0.2611 | LCC = 0.8605 | SRCC = 0.8571 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1713 | LCC = 0.9290 | SRCC = 0.9259 ] +INFO:root:Average loss=0.0977262431421341 +INFO:root: +[Valid][14][UTT][ MSE = 0.2265 | LCC = 0.8617 | SRCC = 0.8613 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1169 | LCC = 0.9364 | SRCC = 0.9382 ] +INFO:root:Average loss=0.09514909762984668 +INFO:root: +[Valid][15][UTT][ MSE = 0.2794 | LCC = 0.8582 | SRCC = 0.8558 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1889 | LCC = 0.9303 | SRCC = 0.9292 ] +INFO:root:Average loss=0.09148985472245094 +INFO:root: +[Valid][16][UTT][ MSE = 0.2399 | LCC = 0.8646 | SRCC = 0.8634 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1305 | LCC = 0.9424 | SRCC = 0.9436 ] +INFO:root:Average loss=0.09739843593576016 +INFO:root: +[Valid][17][UTT][ MSE = 0.2148 | LCC = 0.8666 | SRCC = 0.8647 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1077 | LCC = 0.9397 | SRCC = 0.9388 ] +INFO:root:Average loss=0.0810291952907275 +INFO:root: +[Valid][18][UTT][ MSE = 0.2443 | LCC = 0.8612 | SRCC = 0.8598 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1418 | LCC = 0.9395 | SRCC = 0.9395 ] +INFO:root:Average loss=0.0809667952454243 +INFO:root: +[Valid][19][UTT][ MSE = 0.2668 | LCC = 0.8560 | SRCC = 0.8556 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1587 | LCC = 0.9368 | SRCC = 0.9400 ] +INFO:root:Average loss=0.07862621292662926 +INFO:root: +[Valid][20][UTT][ MSE = 0.2336 | LCC = 0.8662 | SRCC = 0.8630 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1369 | LCC = 0.9354 | SRCC = 0.9315 ] +INFO:root:Average loss=0.07469513434439133 +INFO:root: +[Valid][21][UTT][ MSE = 0.2299 | LCC = 0.8662 | SRCC = 0.8643 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1314 | LCC = 0.9385 | SRCC = 0.9377 ] +INFO:root:Average loss=0.07152971806816566 +INFO:root: +[Valid][22][UTT][ MSE = 0.2245 | LCC = 0.8649 | SRCC = 0.8630 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1187 | LCC = 0.9403 | SRCC = 0.9398 ] +INFO:root:Average loss=0.06999625805287789 +INFO:root: +[Valid][23][UTT][ MSE = 0.2443 | LCC = 0.8652 | SRCC = 0.8650 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1456 | LCC = 0.9432 | SRCC = 0.9463 ] +INFO:root:Average loss=0.07724339954363994 +INFO:root: +[Valid][24][UTT][ MSE = 0.2206 | LCC = 0.8648 | SRCC = 0.8642 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1147 | LCC = 0.9401 | SRCC = 0.9422 ] +INFO:root:Average loss=0.0698593703026955 +INFO:root: +[Valid][25][UTT][ MSE = 0.2379 | LCC = 0.8625 | SRCC = 0.8616 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1334 | LCC = 0.9405 | SRCC = 0.9405 ] +INFO:root:Average loss=0.06205379327711386 +INFO:root: +[Valid][26][UTT][ MSE = 0.2316 | LCC = 0.8685 | SRCC = 0.8676 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1330 | LCC = 0.9434 | SRCC = 0.9436 ] +INFO:root:Average loss=0.061733570570747055 +INFO:root: +[Valid][27][UTT][ MSE = 0.2087 | LCC = 0.8708 | SRCC = 0.8679 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.0991 | LCC = 0.9410 | SRCC = 0.9373 ] +INFO:root:Average loss=0.061150606554479174 +INFO:root: +[Valid][28][UTT][ MSE = 0.2310 | LCC = 0.8676 | SRCC = 0.8664 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1337 | LCC = 0.9417 | SRCC = 0.9434 ] +INFO:root:Average loss=0.057036525975817286 +INFO:root: +[Valid][29][UTT][ MSE = 0.2331 | LCC = 0.8676 | SRCC = 0.8664 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1354 | LCC = 0.9419 | SRCC = 0.9426 ] +INFO:root:Average loss=0.05510546892690353 +INFO:root: +[Valid][30][UTT][ MSE = 0.2237 | LCC = 0.8658 | SRCC = 0.8650 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1233 | LCC = 0.9387 | SRCC = 0.9414 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2377 | LCC = 0.8587 | SRCC = 0.8551 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1383 | LCC = 0.9224 | SRCC = 0.9203 ] diff --git a/results_bvcc/w2v2_base/layer1/run1/config.gin b/results_bvcc/w2v2_base/layer1/run1/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer1/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 diff --git a/results_bvcc/w2v2_base/layer1/run1/model.pt b/results_bvcc/w2v2_base/layer1/run1/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..2903c63147ba9392cf6ddb0ff5b994b0736a2126 --- /dev/null +++ b/results_bvcc/w2v2_base/layer1/run1/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9557e734a10a18329b6b88241bd5b295f63f75b0530ae47b0ebf1e55a289c366 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer1/run1/model_best.pt b/results_bvcc/w2v2_base/layer1/run1/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..8fca5b0469045a8c677c2062e9d113d2e3001d95 --- /dev/null +++ b/results_bvcc/w2v2_base/layer1/run1/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6707611934c02690335c62cde1fef5d4c5205e932bb7e973f9f62aac74c83690 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer1/run1/train.log b/results_bvcc/w2v2_base/layer1/run1/train.log new file mode 100644 index 0000000000000000000000000000000000000000..9ad143d1792d7be813a1d91d528be99d086ba069 --- /dev/null +++ b/results_bvcc/w2v2_base/layer1/run1/train.log @@ -0,0 +1,163 @@ +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=0.49480017905051893 +INFO:root: +[Valid][0][UTT][ MSE = 0.3707 | LCC = 0.7667 | SRCC = 0.7704 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2482 | LCC = 0.8397 | SRCC = 0.8452 ] +INFO:root:Average loss=0.27845622771061385 +INFO:root: +[Valid][1][UTT][ MSE = 0.2963 | LCC = 0.8141 | SRCC = 0.8191 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1968 | LCC = 0.8806 | SRCC = 0.8873 ] +INFO:root:Average loss=0.2325418186493409 +INFO:root: +[Valid][2][UTT][ MSE = 0.2657 | LCC = 0.8388 | SRCC = 0.8393 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1636 | LCC = 0.9002 | SRCC = 0.9034 ] +INFO:root:Average loss=0.21036842312568274 +INFO:root: +[Valid][3][UTT][ MSE = 0.2611 | LCC = 0.8452 | SRCC = 0.8457 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1507 | LCC = 0.9157 | SRCC = 0.9200 ] +INFO:root:Average loss=0.19588429203782326 +INFO:root: +[Valid][4][UTT][ MSE = 0.2576 | LCC = 0.8492 | SRCC = 0.8464 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1489 | LCC = 0.9182 | SRCC = 0.9138 ] +INFO:root:Average loss=0.18050869554281235 +INFO:root: +[Valid][5][UTT][ MSE = 0.2392 | LCC = 0.8511 | SRCC = 0.8496 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1167 | LCC = 0.9244 | SRCC = 0.9236 ] +INFO:root:Average loss=0.1697243203719457 +INFO:root: +[Valid][6][UTT][ MSE = 0.2389 | LCC = 0.8544 | SRCC = 0.8517 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1242 | LCC = 0.9268 | SRCC = 0.9240 ] +INFO:root:Average loss=0.15253915876532212 +INFO:root: +[Valid][7][UTT][ MSE = 0.2330 | LCC = 0.8590 | SRCC = 0.8571 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1104 | LCC = 0.9312 | SRCC = 0.9303 ] +INFO:root:Average loss=0.14838704428611657 +INFO:root: +[Valid][8][UTT][ MSE = 0.2413 | LCC = 0.8550 | SRCC = 0.8575 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1363 | LCC = 0.9283 | SRCC = 0.9348 ] +INFO:root:Average loss=0.1364711202107943 +INFO:root: +[Valid][9][UTT][ MSE = 0.2387 | LCC = 0.8608 | SRCC = 0.8591 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1134 | LCC = 0.9396 | SRCC = 0.9387 ] +INFO:root:Average loss=0.12884573237254068 +INFO:root: +[Valid][10][UTT][ MSE = 0.2213 | LCC = 0.8658 | SRCC = 0.8639 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1090 | LCC = 0.9361 | SRCC = 0.9352 ] +INFO:root:Average loss=0.13734690319651213 +INFO:root: +[Valid][11][UTT][ MSE = 0.2510 | LCC = 0.8653 | SRCC = 0.8640 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1196 | LCC = 0.9394 | SRCC = 0.9383 ] +INFO:root:Average loss=0.11653993694254985 +INFO:root: +[Valid][12][UTT][ MSE = 0.2371 | LCC = 0.8585 | SRCC = 0.8580 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1264 | LCC = 0.9351 | SRCC = 0.9386 ] +INFO:root:Average loss=0.10805719918929614 +INFO:root: +[Valid][13][UTT][ MSE = 0.2204 | LCC = 0.8631 | SRCC = 0.8619 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.0972 | LCC = 0.9399 | SRCC = 0.9394 ] +INFO:root:Average loss=0.10435576784687164 +INFO:root: +[Valid][14][UTT][ MSE = 0.2360 | LCC = 0.8607 | SRCC = 0.8607 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1297 | LCC = 0.9367 | SRCC = 0.9399 ] +INFO:root:Average loss=0.09637260752228591 +INFO:root: +[Valid][15][UTT][ MSE = 0.2703 | LCC = 0.8568 | SRCC = 0.8589 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1738 | LCC = 0.9327 | SRCC = 0.9391 ] +INFO:root:Average loss=0.10165326985029075 +INFO:root: +[Valid][16][UTT][ MSE = 0.2279 | LCC = 0.8598 | SRCC = 0.8575 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1126 | LCC = 0.9366 | SRCC = 0.9355 ] +INFO:root:Average loss=0.08950381109920832 +INFO:root: +[Valid][17][UTT][ MSE = 0.2257 | LCC = 0.8576 | SRCC = 0.8556 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1056 | LCC = 0.9355 | SRCC = 0.9357 ] +INFO:root:Average loss=0.08791988768065587 +INFO:root: +[Valid][18][UTT][ MSE = 0.2214 | LCC = 0.8607 | SRCC = 0.8588 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1032 | LCC = 0.9386 | SRCC = 0.9374 ] +INFO:root:Average loss=0.0842550451365801 +INFO:root: +[Valid][19][UTT][ MSE = 0.2271 | LCC = 0.8623 | SRCC = 0.8610 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1003 | LCC = 0.9398 | SRCC = 0.9383 ] +INFO:root:Average loss=0.07734839694622235 +INFO:root: +[Valid][20][UTT][ MSE = 0.2193 | LCC = 0.8599 | SRCC = 0.8585 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.0937 | LCC = 0.9395 | SRCC = 0.9408 ] +INFO:root:Average loss=0.07589360111607955 +INFO:root: +[Valid][21][UTT][ MSE = 0.2313 | LCC = 0.8581 | SRCC = 0.8577 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1258 | LCC = 0.9336 | SRCC = 0.9396 ] +INFO:root:Average loss=0.07119909267968093 +INFO:root: +[Valid][22][UTT][ MSE = 0.2225 | LCC = 0.8595 | SRCC = 0.8595 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1092 | LCC = 0.9375 | SRCC = 0.9424 ] +INFO:root:Average loss=0.06933691834983154 +INFO:root: +[Valid][23][UTT][ MSE = 0.2205 | LCC = 0.8625 | SRCC = 0.8595 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1082 | LCC = 0.9367 | SRCC = 0.9368 ] +INFO:root:Average loss=0.06987673411957729 +INFO:root: +[Valid][24][UTT][ MSE = 0.2165 | LCC = 0.8619 | SRCC = 0.8612 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1056 | LCC = 0.9356 | SRCC = 0.9410 ] +INFO:root:Average loss=0.0654885101203735 +INFO:root: +[Valid][25][UTT][ MSE = 0.2249 | LCC = 0.8634 | SRCC = 0.8615 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1198 | LCC = 0.9369 | SRCC = 0.9392 ] +INFO:root:Average loss=0.06392660608085302 +INFO:root: +[Valid][26][UTT][ MSE = 0.2273 | LCC = 0.8618 | SRCC = 0.8602 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1213 | LCC = 0.9369 | SRCC = 0.9393 ] +INFO:root:Average loss=0.060871114858832114 +INFO:root: +[Valid][27][UTT][ MSE = 0.2281 | LCC = 0.8603 | SRCC = 0.8594 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1161 | LCC = 0.9407 | SRCC = 0.9429 ] +INFO:root:Average loss=0.06481427379334584 +INFO:root: +[Valid][28][UTT][ MSE = 0.2182 | LCC = 0.8616 | SRCC = 0.8591 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.0974 | LCC = 0.9408 | SRCC = 0.9411 ] +INFO:root:Average loss=0.05660455673933029 +INFO:root: +[Valid][29][UTT][ MSE = 0.2210 | LCC = 0.8649 | SRCC = 0.8615 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1056 | LCC = 0.9422 | SRCC = 0.9392 ] +INFO:root:Average loss=0.05522246692234125 +INFO:root: +[Valid][30][UTT][ MSE = 0.2146 | LCC = 0.8651 | SRCC = 0.8636 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.0934 | LCC = 0.9415 | SRCC = 0.9418 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2336 | LCC = 0.8647 | SRCC = 0.8619 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1565 | LCC = 0.9132 | SRCC = 0.9094 ] diff --git a/results_bvcc/w2v2_base/layer1/run2/config.gin b/results_bvcc/w2v2_base/layer1/run2/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer1/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 diff --git a/results_bvcc/w2v2_base/layer1/run2/model.pt b/results_bvcc/w2v2_base/layer1/run2/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..8836a681981c00ebbd1dd5b49978f5adf061ae7b --- /dev/null +++ b/results_bvcc/w2v2_base/layer1/run2/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:05ebbc29c7d329bc1ee97f50f3096a73b785e22268be34c95185bffd8b68c7dd +size 2747979 diff --git a/results_bvcc/w2v2_base/layer1/run2/model_best.pt b/results_bvcc/w2v2_base/layer1/run2/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..9cdd422cdad75c21af04041aeff19fd2e102dccd --- /dev/null +++ b/results_bvcc/w2v2_base/layer1/run2/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f573a61f31325e8566236456c2318a88fb7ab56c2c619bd6185de7c66325c33a +size 2749702 diff --git a/results_bvcc/w2v2_base/layer1/run2/train.log b/results_bvcc/w2v2_base/layer1/run2/train.log new file mode 100644 index 0000000000000000000000000000000000000000..b48b14e6ce370958e4b8d0564b8c6b6df0e77055 --- /dev/null +++ b/results_bvcc/w2v2_base/layer1/run2/train.log @@ -0,0 +1,163 @@ +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=0.5799432725478442 +INFO:root: +[Valid][0][UTT][ MSE = 0.3725 | LCC = 0.7729 | SRCC = 0.7788 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2333 | LCC = 0.8492 | SRCC = 0.8574 ] +INFO:root:Average loss=0.2807462093157646 +INFO:root: +[Valid][1][UTT][ MSE = 0.2828 | LCC = 0.8240 | SRCC = 0.8240 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1628 | LCC = 0.9012 | SRCC = 0.9030 ] +INFO:root:Average loss=0.23757183398955908 +INFO:root: +[Valid][2][UTT][ MSE = 0.2619 | LCC = 0.8395 | SRCC = 0.8398 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1473 | LCC = 0.9143 | SRCC = 0.9177 ] +INFO:root:Average loss=0.21484326590330172 +INFO:root: +[Valid][3][UTT][ MSE = 0.2650 | LCC = 0.8500 | SRCC = 0.8505 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1563 | LCC = 0.9237 | SRCC = 0.9270 ] +INFO:root:Average loss=0.20002879679967195 +INFO:root: +[Valid][4][UTT][ MSE = 0.2447 | LCC = 0.8565 | SRCC = 0.8558 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1377 | LCC = 0.9289 | SRCC = 0.9316 ] +INFO:root:Average loss=0.18346711955009362 +INFO:root: +[Valid][5][UTT][ MSE = 0.2279 | LCC = 0.8599 | SRCC = 0.8584 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1060 | LCC = 0.9367 | SRCC = 0.9382 ] +INFO:root:Average loss=0.1678030132674254 +INFO:root: +[Valid][6][UTT][ MSE = 0.2292 | LCC = 0.8561 | SRCC = 0.8585 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1143 | LCC = 0.9311 | SRCC = 0.9361 ] +INFO:root:Average loss=0.15920545428227156 +INFO:root: +[Valid][7][UTT][ MSE = 0.2251 | LCC = 0.8621 | SRCC = 0.8593 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1069 | LCC = 0.9379 | SRCC = 0.9357 ] +INFO:root:Average loss=0.14479620401293802 +INFO:root: +[Valid][8][UTT][ MSE = 0.2212 | LCC = 0.8621 | SRCC = 0.8615 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1146 | LCC = 0.9312 | SRCC = 0.9337 ] +INFO:root:Average loss=0.13811541081239015 +INFO:root: +[Valid][9][UTT][ MSE = 0.2240 | LCC = 0.8674 | SRCC = 0.8632 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1123 | LCC = 0.9409 | SRCC = 0.9389 ] +INFO:root:Average loss=0.13562335083499932 +INFO:root: +[Valid][10][UTT][ MSE = 0.2191 | LCC = 0.8687 | SRCC = 0.8656 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1104 | LCC = 0.9393 | SRCC = 0.9387 ] +INFO:root:Average loss=0.1258917586543621 +INFO:root: +[Valid][11][UTT][ MSE = 0.2320 | LCC = 0.8611 | SRCC = 0.8588 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1264 | LCC = 0.9345 | SRCC = 0.9313 ] +INFO:root:Average loss=0.12100212400158246 +INFO:root: +[Valid][12][UTT][ MSE = 0.2333 | LCC = 0.8706 | SRCC = 0.8672 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1101 | LCC = 0.9452 | SRCC = 0.9416 ] +INFO:root:Average loss=0.11105435886062108 +INFO:root: +[Valid][13][UTT][ MSE = 0.2303 | LCC = 0.8628 | SRCC = 0.8612 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1234 | LCC = 0.9392 | SRCC = 0.9412 ] +INFO:root:Average loss=0.10584970210225154 +INFO:root: +[Valid][14][UTT][ MSE = 0.2299 | LCC = 0.8686 | SRCC = 0.8641 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1114 | LCC = 0.9407 | SRCC = 0.9340 ] +INFO:root:Average loss=0.1046583229341568 +INFO:root: +[Valid][15][UTT][ MSE = 0.2171 | LCC = 0.8656 | SRCC = 0.8622 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1020 | LCC = 0.9402 | SRCC = 0.9382 ] +INFO:root:Average loss=0.09526471244410062 +INFO:root: +[Valid][16][UTT][ MSE = 0.2188 | LCC = 0.8669 | SRCC = 0.8638 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1085 | LCC = 0.9419 | SRCC = 0.9415 ] +INFO:root:Average loss=0.09406070654781964 +INFO:root: +[Valid][17][UTT][ MSE = 0.2255 | LCC = 0.8654 | SRCC = 0.8640 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1177 | LCC = 0.9418 | SRCC = 0.9440 ] +INFO:root:Average loss=0.08924258161240663 +INFO:root: +[Valid][18][UTT][ MSE = 0.2167 | LCC = 0.8688 | SRCC = 0.8649 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.0992 | LCC = 0.9429 | SRCC = 0.9380 ] +INFO:root:Average loss=0.08068299785447426 +INFO:root: +[Valid][19][UTT][ MSE = 0.2161 | LCC = 0.8686 | SRCC = 0.8646 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1112 | LCC = 0.9383 | SRCC = 0.9338 ] +INFO:root:Average loss=0.08465328506934337 +INFO:root: +[Valid][20][UTT][ MSE = 0.2141 | LCC = 0.8689 | SRCC = 0.8651 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1057 | LCC = 0.9440 | SRCC = 0.9409 ] +INFO:root:Average loss=0.07797851294088058 +INFO:root: +[Valid][21][UTT][ MSE = 0.2213 | LCC = 0.8704 | SRCC = 0.8664 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1045 | LCC = 0.9446 | SRCC = 0.9388 ] +INFO:root:Average loss=0.07383806950961933 +INFO:root: +[Valid][22][UTT][ MSE = 0.2054 | LCC = 0.8721 | SRCC = 0.8696 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.0898 | LCC = 0.9471 | SRCC = 0.9455 ] +INFO:root:Average loss=0.06801195805653548 +INFO:root: +[Valid][23][UTT][ MSE = 0.2108 | LCC = 0.8702 | SRCC = 0.8671 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1030 | LCC = 0.9449 | SRCC = 0.9438 ] +INFO:root:Average loss=0.06952695314509746 +INFO:root: +[Valid][24][UTT][ MSE = 0.2072 | LCC = 0.8740 | SRCC = 0.8719 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.0980 | LCC = 0.9481 | SRCC = 0.9477 ] +INFO:root:Average loss=0.06444193609058857 +INFO:root: +[Valid][25][UTT][ MSE = 0.2194 | LCC = 0.8664 | SRCC = 0.8645 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1171 | LCC = 0.9404 | SRCC = 0.9416 ] +INFO:root:Average loss=0.06203614681577071 +INFO:root: +[Valid][26][UTT][ MSE = 0.2200 | LCC = 0.8691 | SRCC = 0.8654 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.0975 | LCC = 0.9463 | SRCC = 0.9413 ] +INFO:root:Average loss=0.05990246645151041 +INFO:root: +[Valid][27][UTT][ MSE = 0.2192 | LCC = 0.8689 | SRCC = 0.8656 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1137 | LCC = 0.9417 | SRCC = 0.9394 ] +INFO:root:Average loss=0.060253849921700284 +INFO:root: +[Valid][28][UTT][ MSE = 0.2255 | LCC = 0.8686 | SRCC = 0.8647 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1256 | LCC = 0.9428 | SRCC = 0.9410 ] +INFO:root:Average loss=0.057994768262291566 +INFO:root: +[Valid][29][UTT][ MSE = 0.2181 | LCC = 0.8666 | SRCC = 0.8649 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1148 | LCC = 0.9399 | SRCC = 0.9399 ] +INFO:root:Average loss=0.05246185893431688 +INFO:root: +[Valid][30][UTT][ MSE = 0.2124 | LCC = 0.8670 | SRCC = 0.8639 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1001 | LCC = 0.9424 | SRCC = 0.9408 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2362 | LCC = 0.8620 | SRCC = 0.8610 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1418 | LCC = 0.9148 | SRCC = 0.9130 ] diff --git a/results_bvcc/w2v2_base/layer1/run3/config.gin b/results_bvcc/w2v2_base/layer1/run3/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer1/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 diff --git a/results_bvcc/w2v2_base/layer1/run3/model.pt b/results_bvcc/w2v2_base/layer1/run3/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..0603093f575676674e036c7adf7fda1edf352ab9 --- /dev/null +++ b/results_bvcc/w2v2_base/layer1/run3/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:346cbada120b6a68fa4def54564ec2d0494a6dd82cdfd147b8e54b0d8823730a +size 2747979 diff --git a/results_bvcc/w2v2_base/layer1/run3/model_best.pt b/results_bvcc/w2v2_base/layer1/run3/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..1382aa081329ffa689ea18d73c53609077941a3e --- /dev/null +++ b/results_bvcc/w2v2_base/layer1/run3/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:88c82d0e77fb4254ac1593d055299fa629b2f0f064e853c1b6feb98fb96b3754 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer1/run3/train.log b/results_bvcc/w2v2_base/layer1/run3/train.log new file mode 100644 index 0000000000000000000000000000000000000000..4918b105fd1cfd63c61c46f3154a28c8c5613e5d --- /dev/null +++ b/results_bvcc/w2v2_base/layer1/run3/train.log @@ -0,0 +1,163 @@ +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=0.46871751355819213 +INFO:root: +[Valid][0][UTT][ MSE = 0.3610 | LCC = 0.7796 | SRCC = 0.7904 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2642 | LCC = 0.8517 | SRCC = 0.8624 ] +INFO:root:Average loss=0.27671181295926756 +INFO:root: +[Valid][1][UTT][ MSE = 0.2842 | LCC = 0.8195 | SRCC = 0.8230 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1693 | LCC = 0.8907 | SRCC = 0.8959 ] +INFO:root:Average loss=0.23440087357392678 +INFO:root: +[Valid][2][UTT][ MSE = 0.2656 | LCC = 0.8478 | SRCC = 0.8472 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1619 | LCC = 0.9165 | SRCC = 0.9190 ] +INFO:root:Average loss=0.2157593487929075 +INFO:root: +[Valid][3][UTT][ MSE = 0.2531 | LCC = 0.8485 | SRCC = 0.8497 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1396 | LCC = 0.9237 | SRCC = 0.9290 ] +INFO:root:Average loss=0.19040684105876166 +INFO:root: +[Valid][4][UTT][ MSE = 0.2466 | LCC = 0.8519 | SRCC = 0.8542 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1470 | LCC = 0.9214 | SRCC = 0.9277 ] +INFO:root:Average loss=0.18986343038387787 +INFO:root: +[Valid][5][UTT][ MSE = 0.2325 | LCC = 0.8606 | SRCC = 0.8614 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1093 | LCC = 0.9342 | SRCC = 0.9375 ] +INFO:root:Average loss=0.16716703103903013 +INFO:root: +[Valid][6][UTT][ MSE = 0.2298 | LCC = 0.8626 | SRCC = 0.8629 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1131 | LCC = 0.9310 | SRCC = 0.9332 ] +INFO:root:Average loss=0.15025553078605577 +INFO:root: +[Valid][7][UTT][ MSE = 0.2353 | LCC = 0.8638 | SRCC = 0.8606 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1247 | LCC = 0.9331 | SRCC = 0.9293 ] +INFO:root:Average loss=0.14692852034782752 +INFO:root: +[Valid][8][UTT][ MSE = 0.2680 | LCC = 0.8597 | SRCC = 0.8601 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1739 | LCC = 0.9308 | SRCC = 0.9349 ] +INFO:root:Average loss=0.13832987290926468 +INFO:root: +[Valid][9][UTT][ MSE = 0.2308 | LCC = 0.8610 | SRCC = 0.8612 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1122 | LCC = 0.9356 | SRCC = 0.9380 ] +INFO:root:Average loss=0.12529149489142957 +INFO:root: +[Valid][10][UTT][ MSE = 0.2126 | LCC = 0.8691 | SRCC = 0.8693 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1073 | LCC = 0.9379 | SRCC = 0.9403 ] +INFO:root:Average loss=0.12105241398780774 +INFO:root: +[Valid][11][UTT][ MSE = 0.2209 | LCC = 0.8679 | SRCC = 0.8677 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1138 | LCC = 0.9410 | SRCC = 0.9420 ] +INFO:root:Average loss=0.11520517350007327 +INFO:root: +[Valid][12][UTT][ MSE = 0.2131 | LCC = 0.8694 | SRCC = 0.8693 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1143 | LCC = 0.9357 | SRCC = 0.9355 ] +INFO:root:Average loss=0.10659859214837734 +INFO:root: +[Valid][13][UTT][ MSE = 0.2217 | LCC = 0.8688 | SRCC = 0.8675 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1192 | LCC = 0.9406 | SRCC = 0.9401 ] +INFO:root:Average loss=0.10968755309780438 +INFO:root: +[Valid][14][UTT][ MSE = 0.2105 | LCC = 0.8711 | SRCC = 0.8689 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1039 | LCC = 0.9395 | SRCC = 0.9362 ] +INFO:root:Average loss=0.10449104106579071 +INFO:root: +[Valid][15][UTT][ MSE = 0.2135 | LCC = 0.8707 | SRCC = 0.8703 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1221 | LCC = 0.9335 | SRCC = 0.9342 ] +INFO:root:Average loss=0.0918716401912463 +INFO:root: +[Valid][16][UTT][ MSE = 0.2455 | LCC = 0.8659 | SRCC = 0.8691 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1538 | LCC = 0.9333 | SRCC = 0.9413 ] +INFO:root:Average loss=0.0942793258298666 +INFO:root: +[Valid][17][UTT][ MSE = 0.2137 | LCC = 0.8720 | SRCC = 0.8689 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1149 | LCC = 0.9356 | SRCC = 0.9312 ] +INFO:root:Average loss=0.08522394886956765 +INFO:root: +[Valid][18][UTT][ MSE = 0.2344 | LCC = 0.8721 | SRCC = 0.8702 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1425 | LCC = 0.9390 | SRCC = 0.9406 ] +INFO:root:Average loss=0.08684245191323452 +INFO:root: +[Valid][19][UTT][ MSE = 0.2264 | LCC = 0.8685 | SRCC = 0.8688 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1366 | LCC = 0.9331 | SRCC = 0.9367 ] +INFO:root:Average loss=0.07973687728055012 +INFO:root: +[Valid][20][UTT][ MSE = 0.2107 | LCC = 0.8706 | SRCC = 0.8691 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1134 | LCC = 0.9364 | SRCC = 0.9359 ] +INFO:root:Average loss=0.07529464426139991 +INFO:root: +[Valid][21][UTT][ MSE = 0.2128 | LCC = 0.8692 | SRCC = 0.8677 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1149 | LCC = 0.9344 | SRCC = 0.9338 ] +INFO:root:Average loss=0.07538936134332265 +INFO:root: +[Valid][22][UTT][ MSE = 0.2157 | LCC = 0.8723 | SRCC = 0.8697 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1235 | LCC = 0.9354 | SRCC = 0.9309 ] +INFO:root:Average loss=0.07405028368035953 +INFO:root: +[Valid][23][UTT][ MSE = 0.2426 | LCC = 0.8700 | SRCC = 0.8697 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1588 | LCC = 0.9342 | SRCC = 0.9355 ] +INFO:root:Average loss=0.06866880921790233 +INFO:root: +[Valid][24][UTT][ MSE = 0.2103 | LCC = 0.8714 | SRCC = 0.8695 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1108 | LCC = 0.9367 | SRCC = 0.9354 ] +INFO:root:Average loss=0.0702828078602369 +INFO:root: +[Valid][25][UTT][ MSE = 0.2039 | LCC = 0.8734 | SRCC = 0.8705 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1016 | LCC = 0.9377 | SRCC = 0.9323 ] +INFO:root:Average loss=0.06925542175005643 +INFO:root: +[Valid][26][UTT][ MSE = 0.2048 | LCC = 0.8728 | SRCC = 0.8711 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.0996 | LCC = 0.9356 | SRCC = 0.9303 ] +INFO:root:Average loss=0.06134156978283173 +INFO:root: +[Valid][27][UTT][ MSE = 0.2206 | LCC = 0.8677 | SRCC = 0.8649 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1209 | LCC = 0.9371 | SRCC = 0.9334 ] +INFO:root:Average loss=0.058670274125268825 +INFO:root: +[Valid][28][UTT][ MSE = 0.2128 | LCC = 0.8692 | SRCC = 0.8667 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1019 | LCC = 0.9381 | SRCC = 0.9330 ] +INFO:root:Average loss=0.05878604814792291 +INFO:root: +[Valid][29][UTT][ MSE = 0.2116 | LCC = 0.8700 | SRCC = 0.8661 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1094 | LCC = 0.9363 | SRCC = 0.9316 ] +INFO:root:Average loss=0.055968338528122656 +INFO:root: +[Valid][30][UTT][ MSE = 0.2131 | LCC = 0.8720 | SRCC = 0.8678 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1195 | LCC = 0.9362 | SRCC = 0.9300 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2366 | LCC = 0.8587 | SRCC = 0.8593 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1257 | LCC = 0.9256 | SRCC = 0.9279 ] diff --git a/results_bvcc/w2v2_base/layer1/run4/config.gin b/results_bvcc/w2v2_base/layer1/run4/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer1/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 diff --git a/results_bvcc/w2v2_base/layer1/run4/model.pt b/results_bvcc/w2v2_base/layer1/run4/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..f6766b17b467a76df8960929b86085a4c9163f0a --- /dev/null +++ b/results_bvcc/w2v2_base/layer1/run4/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:47e12b81faf37172040cce1a322306ec6c1fc2972e1f5316d571952d17534f46 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer1/run4/model_best.pt b/results_bvcc/w2v2_base/layer1/run4/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..7a59e0bf19197aeb5cd70075cb4163ab505c5a42 --- /dev/null +++ b/results_bvcc/w2v2_base/layer1/run4/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:10b7f6630f6e8cd9082f912d75b22a7d2699b33fd9d4ffa7be02629920573271 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer1/run4/train.log b/results_bvcc/w2v2_base/layer1/run4/train.log new file mode 100644 index 0000000000000000000000000000000000000000..0fbbca3f83df16d06d89106ef8084adc46441b34 --- /dev/null +++ b/results_bvcc/w2v2_base/layer1/run4/train.log @@ -0,0 +1,163 @@ +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=0.5067888184999808 +INFO:root: +[Valid][0][UTT][ MSE = 0.3787 | LCC = 0.7676 | SRCC = 0.7716 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2580 | LCC = 0.8547 | SRCC = 0.8602 ] +INFO:root:Average loss=0.28694182959122533 +INFO:root: +[Valid][1][UTT][ MSE = 0.3054 | LCC = 0.8105 | SRCC = 0.8177 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1863 | LCC = 0.8930 | SRCC = 0.9010 ] +INFO:root:Average loss=0.23920302990919504 +INFO:root: +[Valid][2][UTT][ MSE = 0.2741 | LCC = 0.8386 | SRCC = 0.8393 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1467 | LCC = 0.9210 | SRCC = 0.9235 ] +INFO:root:Average loss=0.21217203235779053 +INFO:root: +[Valid][3][UTT][ MSE = 0.2616 | LCC = 0.8402 | SRCC = 0.8408 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1357 | LCC = 0.9121 | SRCC = 0.9131 ] +INFO:root:Average loss=0.19956507419164365 +INFO:root: +[Valid][4][UTT][ MSE = 0.2505 | LCC = 0.8506 | SRCC = 0.8512 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1288 | LCC = 0.9267 | SRCC = 0.9300 ] +INFO:root:Average loss=0.18455939013988543 +INFO:root: +[Valid][5][UTT][ MSE = 0.2528 | LCC = 0.8536 | SRCC = 0.8531 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1231 | LCC = 0.9293 | SRCC = 0.9300 ] +INFO:root:Average loss=0.16596660868097574 +INFO:root: +[Valid][6][UTT][ MSE = 0.2391 | LCC = 0.8592 | SRCC = 0.8570 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1131 | LCC = 0.9329 | SRCC = 0.9302 ] +INFO:root:Average loss=0.15587880462408066 +INFO:root: +[Valid][7][UTT][ MSE = 0.2430 | LCC = 0.8558 | SRCC = 0.8537 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1304 | LCC = 0.9287 | SRCC = 0.9277 ] +INFO:root:Average loss=0.14549451025250631 +INFO:root: +[Valid][8][UTT][ MSE = 0.2537 | LCC = 0.8568 | SRCC = 0.8572 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1547 | LCC = 0.9273 | SRCC = 0.9312 ] +INFO:root:Average loss=0.14089427315271819 +INFO:root: +[Valid][9][UTT][ MSE = 0.2450 | LCC = 0.8516 | SRCC = 0.8518 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1301 | LCC = 0.9300 | SRCC = 0.9345 ] +INFO:root:Average loss=0.1309406508046847 +INFO:root: +[Valid][10][UTT][ MSE = 0.2291 | LCC = 0.8593 | SRCC = 0.8578 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1124 | LCC = 0.9358 | SRCC = 0.9358 ] +INFO:root:Average loss=0.12291629870350544 +INFO:root: +[Valid][11][UTT][ MSE = 0.2327 | LCC = 0.8623 | SRCC = 0.8601 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1056 | LCC = 0.9389 | SRCC = 0.9350 ] +INFO:root:Average loss=0.1144241709739734 +INFO:root: +[Valid][12][UTT][ MSE = 0.2246 | LCC = 0.8619 | SRCC = 0.8598 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1116 | LCC = 0.9373 | SRCC = 0.9369 ] +INFO:root:Average loss=0.10850251150818971 +INFO:root: +[Valid][13][UTT][ MSE = 0.2227 | LCC = 0.8616 | SRCC = 0.8594 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1055 | LCC = 0.9379 | SRCC = 0.9376 ] +INFO:root:Average loss=0.10566861149019156 +INFO:root: +[Valid][14][UTT][ MSE = 0.2212 | LCC = 0.8608 | SRCC = 0.8593 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1054 | LCC = 0.9364 | SRCC = 0.9392 ] +INFO:root:Average loss=0.10105009959676327 +INFO:root: +[Valid][15][UTT][ MSE = 0.2293 | LCC = 0.8554 | SRCC = 0.8543 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1115 | LCC = 0.9358 | SRCC = 0.9401 ] +INFO:root:Average loss=0.09794910175678058 +INFO:root: +[Valid][16][UTT][ MSE = 0.2247 | LCC = 0.8605 | SRCC = 0.8583 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1109 | LCC = 0.9350 | SRCC = 0.9357 ] +INFO:root:Average loss=0.0921257061836047 +INFO:root: +[Valid][17][UTT][ MSE = 0.2215 | LCC = 0.8593 | SRCC = 0.8574 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.0989 | LCC = 0.9380 | SRCC = 0.9397 ] +INFO:root:Average loss=0.08327317982912064 +INFO:root: +[Valid][18][UTT][ MSE = 0.2166 | LCC = 0.8626 | SRCC = 0.8613 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.0964 | LCC = 0.9357 | SRCC = 0.9363 ] +INFO:root:Average loss=0.09554384051798245 +INFO:root: +[Valid][19][UTT][ MSE = 0.2440 | LCC = 0.8582 | SRCC = 0.8562 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1127 | LCC = 0.9306 | SRCC = 0.9285 ] +INFO:root:Average loss=0.0855937705685695 +INFO:root: +[Valid][20][UTT][ MSE = 0.2158 | LCC = 0.8632 | SRCC = 0.8613 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.0965 | LCC = 0.9388 | SRCC = 0.9368 ] +INFO:root:Average loss=0.07785610897609821 +INFO:root: +[Valid][21][UTT][ MSE = 0.2248 | LCC = 0.8624 | SRCC = 0.8590 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1069 | LCC = 0.9421 | SRCC = 0.9383 ] +INFO:root:Average loss=0.07071381960159694 +INFO:root: +[Valid][22][UTT][ MSE = 0.2162 | LCC = 0.8618 | SRCC = 0.8584 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.0971 | LCC = 0.9382 | SRCC = 0.9361 ] +INFO:root:Average loss=0.0724777755064842 +INFO:root: +[Valid][23][UTT][ MSE = 0.2325 | LCC = 0.8600 | SRCC = 0.8571 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1164 | LCC = 0.9416 | SRCC = 0.9406 ] +INFO:root:Average loss=0.06505997803730842 +INFO:root: +[Valid][24][UTT][ MSE = 0.2212 | LCC = 0.8638 | SRCC = 0.8613 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.0981 | LCC = 0.9424 | SRCC = 0.9394 ] +INFO:root:Average loss=0.0653465433189502 +INFO:root: +[Valid][25][UTT][ MSE = 0.2362 | LCC = 0.8579 | SRCC = 0.8555 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1219 | LCC = 0.9391 | SRCC = 0.9373 ] +INFO:root:Average loss=0.06491684961395386 +INFO:root: +[Valid][26][UTT][ MSE = 0.2331 | LCC = 0.8584 | SRCC = 0.8556 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1030 | LCC = 0.9373 | SRCC = 0.9331 ] +INFO:root:Average loss=0.0655498340821419 +INFO:root: +[Valid][27][UTT][ MSE = 0.2169 | LCC = 0.8609 | SRCC = 0.8594 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.0975 | LCC = 0.9371 | SRCC = 0.9358 ] +INFO:root:Average loss=0.05822523506597067 +INFO:root: +[Valid][28][UTT][ MSE = 0.2225 | LCC = 0.8625 | SRCC = 0.8595 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1022 | LCC = 0.9419 | SRCC = 0.9394 ] +INFO:root:Average loss=0.06262223728192158 +INFO:root: +[Valid][29][UTT][ MSE = 0.2209 | LCC = 0.8590 | SRCC = 0.8579 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.0941 | LCC = 0.9341 | SRCC = 0.9334 ] +INFO:root:Average loss=0.061830226809550554 +INFO:root: +[Valid][30][UTT][ MSE = 0.2208 | LCC = 0.8628 | SRCC = 0.8607 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.0992 | LCC = 0.9370 | SRCC = 0.9355 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2542 | LCC = 0.8541 | SRCC = 0.8542 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1683 | LCC = 0.9123 | SRCC = 0.9148 ] diff --git a/results_bvcc/w2v2_base/layer10/run0/config.gin b/results_bvcc/w2v2_base/layer10/run0/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer10/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 diff --git a/results_bvcc/w2v2_base/layer10/run0/model.pt b/results_bvcc/w2v2_base/layer10/run0/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..b3bffc51638583bffa931719c4e01d56c16974e3 --- /dev/null +++ b/results_bvcc/w2v2_base/layer10/run0/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:34c1a63ec83a1430692c77ecf5066f37ea69f3e10b10914faf5dfa5611595741 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer10/run0/model_best.pt b/results_bvcc/w2v2_base/layer10/run0/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..9d2eb101b84c56bb8a6625cfa9ff8b475f009337 --- /dev/null +++ b/results_bvcc/w2v2_base/layer10/run0/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1a2c5e2c1fc190799f3088ee8b609abb21a34b65276a7504346f38a469024ab1 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer10/run0/train.log b/results_bvcc/w2v2_base/layer10/run0/train.log new file mode 100644 index 0000000000000000000000000000000000000000..d91dda727d61957cca3284f6242f22b710c9c86b --- /dev/null +++ b/results_bvcc/w2v2_base/layer10/run0/train.log @@ -0,0 +1,163 @@ +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=0.6340566608004081 +INFO:root: +[Valid][0][UTT][ MSE = 0.4106 | LCC = 0.7514 | SRCC = 0.7600 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.3009 | LCC = 0.8284 | SRCC = 0.8318 ] +INFO:root:Average loss=0.3847933440254285 +INFO:root: +[Valid][1][UTT][ MSE = 0.3723 | LCC = 0.7784 | SRCC = 0.7856 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2704 | LCC = 0.8500 | SRCC = 0.8501 ] +INFO:root:Average loss=0.30829281417223126 +INFO:root: +[Valid][2][UTT][ MSE = 0.3684 | LCC = 0.7963 | SRCC = 0.7985 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.2706 | LCC = 0.8699 | SRCC = 0.8680 ] +INFO:root:Average loss=0.31271022462692016 +INFO:root: +[Valid][3][UTT][ MSE = 0.3253 | LCC = 0.8152 | SRCC = 0.8146 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1964 | LCC = 0.8825 | SRCC = 0.8798 ] +INFO:root:Average loss=0.302239019710284 +INFO:root: +[Valid][4][UTT][ MSE = 0.3614 | LCC = 0.8170 | SRCC = 0.8186 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.2740 | LCC = 0.8858 | SRCC = 0.8825 ] +INFO:root:Average loss=0.3026625733727064 +INFO:root: +[Valid][5][UTT][ MSE = 0.3027 | LCC = 0.8235 | SRCC = 0.8254 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.2049 | LCC = 0.8904 | SRCC = 0.8884 ] +INFO:root:Average loss=0.27604097318954957 +INFO:root: +[Valid][6][UTT][ MSE = 0.3244 | LCC = 0.8330 | SRCC = 0.8298 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.2295 | LCC = 0.8938 | SRCC = 0.8855 ] +INFO:root:Average loss=0.26417100658783543 +INFO:root: +[Valid][7][UTT][ MSE = 0.3696 | LCC = 0.8294 | SRCC = 0.8307 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.2218 | LCC = 0.8951 | SRCC = 0.8915 ] +INFO:root:Average loss=0.2875201409825912 +INFO:root: +[Valid][8][UTT][ MSE = 0.3298 | LCC = 0.8324 | SRCC = 0.8360 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.2464 | LCC = 0.8945 | SRCC = 0.8941 ] +INFO:root:Average loss=0.24525141200193992 +INFO:root: +[Valid][9][UTT][ MSE = 0.2832 | LCC = 0.8442 | SRCC = 0.8422 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1743 | LCC = 0.9060 | SRCC = 0.9009 ] +INFO:root:Average loss=0.2364092094775958 +INFO:root: +[Valid][10][UTT][ MSE = 0.3405 | LCC = 0.8430 | SRCC = 0.8429 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.2558 | LCC = 0.9111 | SRCC = 0.9070 ] +INFO:root:Average loss=0.25019912135142547 +INFO:root: +[Valid][11][UTT][ MSE = 0.2921 | LCC = 0.8451 | SRCC = 0.8446 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1648 | LCC = 0.9099 | SRCC = 0.9045 ] +INFO:root:Average loss=0.27862557539573085 +INFO:root: +[Valid][12][UTT][ MSE = 0.2895 | LCC = 0.8432 | SRCC = 0.8413 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1873 | LCC = 0.9077 | SRCC = 0.9023 ] +INFO:root:Average loss=0.23084685531182167 +INFO:root: +[Valid][13][UTT][ MSE = 0.2827 | LCC = 0.8465 | SRCC = 0.8464 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1910 | LCC = 0.9125 | SRCC = 0.9102 ] +INFO:root:Average loss=0.22213796965586832 +INFO:root: +[Valid][14][UTT][ MSE = 0.3125 | LCC = 0.8457 | SRCC = 0.8463 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.2268 | LCC = 0.9111 | SRCC = 0.9076 ] +INFO:root:Average loss=0.21562972512000647 +INFO:root: +[Valid][15][UTT][ MSE = 0.2733 | LCC = 0.8510 | SRCC = 0.8493 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1689 | LCC = 0.9194 | SRCC = 0.9141 ] +INFO:root:Average loss=0.20544885232662544 +INFO:root: +[Valid][16][UTT][ MSE = 0.2582 | LCC = 0.8532 | SRCC = 0.8531 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1508 | LCC = 0.9153 | SRCC = 0.9110 ] +INFO:root:Average loss=0.19289059908344194 +INFO:root: +[Valid][17][UTT][ MSE = 0.2610 | LCC = 0.8567 | SRCC = 0.8573 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1694 | LCC = 0.9142 | SRCC = 0.9118 ] +INFO:root:Average loss=0.1999117200000164 +INFO:root: +[Valid][18][UTT][ MSE = 0.3502 | LCC = 0.8483 | SRCC = 0.8489 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.2647 | LCC = 0.9167 | SRCC = 0.9136 ] +INFO:root:Average loss=0.21447260677814484 +INFO:root: +[Valid][19][UTT][ MSE = 0.2960 | LCC = 0.8518 | SRCC = 0.8516 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1730 | LCC = 0.9164 | SRCC = 0.9078 ] +INFO:root:Average loss=0.20319852691430312 +INFO:root: +[Valid][20][UTT][ MSE = 0.2693 | LCC = 0.8538 | SRCC = 0.8543 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1738 | LCC = 0.9145 | SRCC = 0.9124 ] +INFO:root:Average loss=0.21281205280086932 +INFO:root: +[Valid][21][UTT][ MSE = 0.2741 | LCC = 0.8537 | SRCC = 0.8529 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1787 | LCC = 0.9133 | SRCC = 0.9075 ] +INFO:root:Average loss=0.222677637082644 +INFO:root: +[Valid][22][UTT][ MSE = 0.2668 | LCC = 0.8554 | SRCC = 0.8565 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1624 | LCC = 0.9186 | SRCC = 0.9168 ] +INFO:root:Average loss=0.18749733603535554 +INFO:root: +[Valid][23][UTT][ MSE = 0.2773 | LCC = 0.8512 | SRCC = 0.8497 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1769 | LCC = 0.9171 | SRCC = 0.9100 ] +INFO:root:Average loss=0.21209082150688538 +INFO:root: +[Valid][24][UTT][ MSE = 0.2507 | LCC = 0.8562 | SRCC = 0.8569 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1468 | LCC = 0.9200 | SRCC = 0.9166 ] +INFO:root:Average loss=0.17649115985020614 +INFO:root: +[Valid][25][UTT][ MSE = 0.2750 | LCC = 0.8531 | SRCC = 0.8521 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1826 | LCC = 0.9181 | SRCC = 0.9133 ] +INFO:root:Average loss=0.1752329561381768 +INFO:root: +[Valid][26][UTT][ MSE = 0.2695 | LCC = 0.8559 | SRCC = 0.8585 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1472 | LCC = 0.9228 | SRCC = 0.9195 ] +INFO:root:Average loss=0.18399202670806494 +INFO:root: +[Valid][27][UTT][ MSE = 0.2607 | LCC = 0.8577 | SRCC = 0.8576 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1629 | LCC = 0.9278 | SRCC = 0.9257 ] +INFO:root:Average loss=0.1969409548701384 +INFO:root: +[Valid][28][UTT][ MSE = 0.3153 | LCC = 0.8558 | SRCC = 0.8557 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.2016 | LCC = 0.9231 | SRCC = 0.9166 ] +INFO:root:Average loss=0.16963752693472764 +INFO:root: +[Valid][29][UTT][ MSE = 0.2594 | LCC = 0.8517 | SRCC = 0.8516 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1607 | LCC = 0.9219 | SRCC = 0.9184 ] +INFO:root:Average loss=0.15732398762916908 +INFO:root: +[Valid][30][UTT][ MSE = 0.2864 | LCC = 0.8545 | SRCC = 0.8538 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1970 | LCC = 0.9199 | SRCC = 0.9155 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2476 | LCC = 0.8550 | SRCC = 0.8542 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1155 | LCC = 0.9270 | SRCC = 0.9283 ] diff --git a/results_bvcc/w2v2_base/layer10/run1/config.gin b/results_bvcc/w2v2_base/layer10/run1/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer10/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 diff --git a/results_bvcc/w2v2_base/layer10/run1/model.pt b/results_bvcc/w2v2_base/layer10/run1/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..60c202c243963c55f7424273d5f1f07e5cad604b --- /dev/null +++ b/results_bvcc/w2v2_base/layer10/run1/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:598dc0012f60cdfc06d5253279853e5fc3e9ab77df774cfa9e1e3ed384fce21b +size 2747979 diff --git a/results_bvcc/w2v2_base/layer10/run1/model_best.pt b/results_bvcc/w2v2_base/layer10/run1/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..ce5c21f7a049264198244845e28704b9ff55c565 --- /dev/null +++ b/results_bvcc/w2v2_base/layer10/run1/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:306949bbbe0fb3d5003ae9772e2142cba578554ebc0cb3f80988304a6e3fdbc6 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer10/run1/train.log b/results_bvcc/w2v2_base/layer10/run1/train.log new file mode 100644 index 0000000000000000000000000000000000000000..d358aedf02b6a65aea3ee8ff6597a32a4ac39b56 --- /dev/null +++ b/results_bvcc/w2v2_base/layer10/run1/train.log @@ -0,0 +1,163 @@ +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=0.7796100844175388 +INFO:root: +[Valid][0][UTT][ MSE = 0.6900 | LCC = 0.6318 | SRCC = 0.6236 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.5890 | LCC = 0.7349 | SRCC = 0.7272 ] +INFO:root:Average loss=0.43842379137491566 +INFO:root: +[Valid][1][UTT][ MSE = 0.4299 | LCC = 0.7606 | SRCC = 0.7773 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.3243 | LCC = 0.8389 | SRCC = 0.8462 ] +INFO:root:Average loss=0.33721469266292375 +INFO:root: +[Valid][2][UTT][ MSE = 0.3480 | LCC = 0.7960 | SRCC = 0.7999 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.2464 | LCC = 0.8636 | SRCC = 0.8620 ] +INFO:root:Average loss=0.3431662714634186 +INFO:root: +[Valid][3][UTT][ MSE = 0.3195 | LCC = 0.8054 | SRCC = 0.8097 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.2148 | LCC = 0.8697 | SRCC = 0.8674 ] +INFO:root:Average loss=0.30712308906591856 +INFO:root: +[Valid][4][UTT][ MSE = 0.3144 | LCC = 0.8185 | SRCC = 0.8199 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1908 | LCC = 0.8827 | SRCC = 0.8823 ] +INFO:root:Average loss=0.28962959884068906 +INFO:root: +[Valid][5][UTT][ MSE = 0.3502 | LCC = 0.8183 | SRCC = 0.8205 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.2134 | LCC = 0.8749 | SRCC = 0.8772 ] +INFO:root:Average loss=0.28130568888707036 +INFO:root: +[Valid][6][UTT][ MSE = 0.3032 | LCC = 0.8288 | SRCC = 0.8286 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1951 | LCC = 0.8926 | SRCC = 0.8887 ] +INFO:root:Average loss=0.2795686335899891 +INFO:root: +[Valid][7][UTT][ MSE = 0.3075 | LCC = 0.8311 | SRCC = 0.8295 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.2042 | LCC = 0.8928 | SRCC = 0.8890 ] +INFO:root:Average loss=0.25468874741823244 +INFO:root: +[Valid][8][UTT][ MSE = 0.3200 | LCC = 0.8369 | SRCC = 0.8390 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1840 | LCC = 0.9005 | SRCC = 0.9005 ] +INFO:root:Average loss=0.2632343329680272 +INFO:root: +[Valid][9][UTT][ MSE = 0.3394 | LCC = 0.8340 | SRCC = 0.8335 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.2438 | LCC = 0.8982 | SRCC = 0.8920 ] +INFO:root:Average loss=0.2561314589320085 +INFO:root: +[Valid][10][UTT][ MSE = 0.2806 | LCC = 0.8409 | SRCC = 0.8399 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1698 | LCC = 0.9036 | SRCC = 0.8982 ] +INFO:root:Average loss=0.2590295391587111 +INFO:root: +[Valid][11][UTT][ MSE = 0.3220 | LCC = 0.8449 | SRCC = 0.8445 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1911 | LCC = 0.9085 | SRCC = 0.9055 ] +INFO:root:Average loss=0.24746226251889497 +INFO:root: +[Valid][12][UTT][ MSE = 0.2911 | LCC = 0.8403 | SRCC = 0.8417 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1961 | LCC = 0.9034 | SRCC = 0.9010 ] +INFO:root:Average loss=0.2464088023855136 +INFO:root: +[Valid][13][UTT][ MSE = 0.2756 | LCC = 0.8400 | SRCC = 0.8411 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1819 | LCC = 0.9006 | SRCC = 0.8981 ] +INFO:root:Average loss=0.2302155883457416 +INFO:root: +[Valid][14][UTT][ MSE = 0.2737 | LCC = 0.8459 | SRCC = 0.8459 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1546 | LCC = 0.9050 | SRCC = 0.9027 ] +INFO:root:Average loss=0.2331328264031655 +INFO:root: +[Valid][15][UTT][ MSE = 0.2586 | LCC = 0.8418 | SRCC = 0.8408 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1446 | LCC = 0.9063 | SRCC = 0.9006 ] +INFO:root:Average loss=0.21601504622361598 +INFO:root: +[Valid][16][UTT][ MSE = 0.2791 | LCC = 0.8456 | SRCC = 0.8441 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1559 | LCC = 0.9105 | SRCC = 0.9069 ] +INFO:root:Average loss=0.21259737740724516 +INFO:root: +[Valid][17][UTT][ MSE = 0.3048 | LCC = 0.8435 | SRCC = 0.8427 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.2081 | LCC = 0.9063 | SRCC = 0.9020 ] +INFO:root:Average loss=0.2054756549306405 +INFO:root: +[Valid][18][UTT][ MSE = 0.2682 | LCC = 0.8466 | SRCC = 0.8461 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1679 | LCC = 0.9086 | SRCC = 0.9032 ] +INFO:root:Average loss=0.19966396890007532 +INFO:root: +[Valid][19][UTT][ MSE = 0.2665 | LCC = 0.8492 | SRCC = 0.8481 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1388 | LCC = 0.9182 | SRCC = 0.9130 ] +INFO:root:Average loss=0.20426208172471094 +INFO:root: +[Valid][20][UTT][ MSE = 0.3294 | LCC = 0.8432 | SRCC = 0.8414 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.2330 | LCC = 0.9149 | SRCC = 0.9120 ] +INFO:root:Average loss=0.1957240331058319 +INFO:root: +[Valid][21][UTT][ MSE = 0.2638 | LCC = 0.8438 | SRCC = 0.8424 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1370 | LCC = 0.9156 | SRCC = 0.9094 ] +INFO:root:Average loss=0.18055324447460663 +INFO:root: +[Valid][22][UTT][ MSE = 0.3058 | LCC = 0.8398 | SRCC = 0.8420 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.2248 | LCC = 0.9039 | SRCC = 0.9026 ] +INFO:root:Average loss=0.20639980890047857 +INFO:root: +[Valid][23][UTT][ MSE = 0.2672 | LCC = 0.8460 | SRCC = 0.8458 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1678 | LCC = 0.9135 | SRCC = 0.9100 ] +INFO:root:Average loss=0.19168832009801498 +INFO:root: +[Valid][24][UTT][ MSE = 0.2609 | LCC = 0.8521 | SRCC = 0.8501 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1466 | LCC = 0.9210 | SRCC = 0.9134 ] +INFO:root:Average loss=0.18861015102802178 +INFO:root: +[Valid][25][UTT][ MSE = 0.2697 | LCC = 0.8510 | SRCC = 0.8511 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1404 | LCC = 0.9210 | SRCC = 0.9152 ] +INFO:root:Average loss=0.17494672307601342 +INFO:root: +[Valid][26][UTT][ MSE = 0.2967 | LCC = 0.8508 | SRCC = 0.8511 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1617 | LCC = 0.9181 | SRCC = 0.9122 ] +INFO:root:Average loss=0.19273283638251135 +INFO:root: +[Valid][27][UTT][ MSE = 0.3081 | LCC = 0.8413 | SRCC = 0.8424 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.2173 | LCC = 0.9092 | SRCC = 0.9051 ] +INFO:root:Average loss=0.16662715948545015 +INFO:root: +[Valid][28][UTT][ MSE = 0.2756 | LCC = 0.8517 | SRCC = 0.8503 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1756 | LCC = 0.9223 | SRCC = 0.9152 ] +INFO:root:Average loss=0.15800289943432197 +INFO:root: +[Valid][29][UTT][ MSE = 0.2639 | LCC = 0.8480 | SRCC = 0.8473 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1819 | LCC = 0.9090 | SRCC = 0.9048 ] +INFO:root:Average loss=0.15854160783764643 +INFO:root: +[Valid][30][UTT][ MSE = 0.2453 | LCC = 0.8465 | SRCC = 0.8471 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1411 | LCC = 0.9131 | SRCC = 0.9123 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2646 | LCC = 0.8492 | SRCC = 0.8482 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1480 | LCC = 0.9267 | SRCC = 0.9284 ] diff --git a/results_bvcc/w2v2_base/layer10/run2/config.gin b/results_bvcc/w2v2_base/layer10/run2/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer10/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 diff --git a/results_bvcc/w2v2_base/layer10/run2/model.pt b/results_bvcc/w2v2_base/layer10/run2/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..2ce0ceee68573313b456f27938a3fdd2b27e1036 --- /dev/null +++ b/results_bvcc/w2v2_base/layer10/run2/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e33ada00c9dfc1300d134ec03d9ddded0705ab909eb5090ab0661e2e5e3aa644 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer10/run2/model_best.pt b/results_bvcc/w2v2_base/layer10/run2/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..b114bb7d3f0a6beb38a47dcc6f4567751072e4f6 --- /dev/null +++ b/results_bvcc/w2v2_base/layer10/run2/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1e9d94a3abab797f3aee2f89678a15c9349c120ecbbd8a45c6e6891c822cff0f +size 2749702 diff --git a/results_bvcc/w2v2_base/layer10/run2/train.log b/results_bvcc/w2v2_base/layer10/run2/train.log new file mode 100644 index 0000000000000000000000000000000000000000..6347588966685c8221a7a11bc8bb38dfbf9699c2 --- /dev/null +++ b/results_bvcc/w2v2_base/layer10/run2/train.log @@ -0,0 +1,163 @@ +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=0.7398962053732995 +INFO:root: +[Valid][0][UTT][ MSE = 0.5114 | LCC = 0.6916 | SRCC = 0.6911 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.4090 | LCC = 0.7632 | SRCC = 0.7634 ] +INFO:root:Average loss=0.37356871232772487 +INFO:root: +[Valid][1][UTT][ MSE = 0.3672 | LCC = 0.7884 | SRCC = 0.7919 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2653 | LCC = 0.8618 | SRCC = 0.8618 ] +INFO:root:Average loss=0.3334188083043465 +INFO:root: +[Valid][2][UTT][ MSE = 0.3171 | LCC = 0.8014 | SRCC = 0.8030 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.2022 | LCC = 0.8692 | SRCC = 0.8707 ] +INFO:root:Average loss=0.3181392959295175 +INFO:root: +[Valid][3][UTT][ MSE = 0.3506 | LCC = 0.8125 | SRCC = 0.8092 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.2285 | LCC = 0.8834 | SRCC = 0.8795 ] +INFO:root:Average loss=0.2956486004285323 +INFO:root: +[Valid][4][UTT][ MSE = 0.3246 | LCC = 0.8153 | SRCC = 0.8148 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.2220 | LCC = 0.8831 | SRCC = 0.8801 ] +INFO:root:Average loss=0.2900301647874025 +INFO:root: +[Valid][5][UTT][ MSE = 0.3632 | LCC = 0.8221 | SRCC = 0.8200 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.2366 | LCC = 0.8955 | SRCC = 0.8903 ] +INFO:root:Average loss=0.27746068762663084 +INFO:root: +[Valid][6][UTT][ MSE = 0.3168 | LCC = 0.8305 | SRCC = 0.8286 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.2027 | LCC = 0.8982 | SRCC = 0.8937 ] +INFO:root:Average loss=0.2560170080799323 +INFO:root: +[Valid][7][UTT][ MSE = 0.2914 | LCC = 0.8352 | SRCC = 0.8338 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1847 | LCC = 0.9009 | SRCC = 0.8981 ] +INFO:root:Average loss=0.2501956919829051 +INFO:root: +[Valid][8][UTT][ MSE = 0.4080 | LCC = 0.8264 | SRCC = 0.8285 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.3328 | LCC = 0.8892 | SRCC = 0.8856 ] +INFO:root:Average loss=0.24749419876398185 +INFO:root: +[Valid][9][UTT][ MSE = 0.4722 | LCC = 0.8236 | SRCC = 0.8263 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.4021 | LCC = 0.8900 | SRCC = 0.8886 ] +INFO:root:Average loss=0.24561113271957788 +INFO:root: +[Valid][10][UTT][ MSE = 0.3281 | LCC = 0.8385 | SRCC = 0.8402 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.2398 | LCC = 0.9049 | SRCC = 0.9064 ] +INFO:root:Average loss=0.2400172662276488 +INFO:root: +[Valid][11][UTT][ MSE = 0.3102 | LCC = 0.8428 | SRCC = 0.8429 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.2185 | LCC = 0.9077 | SRCC = 0.9060 ] +INFO:root:Average loss=0.23402776301671296 +INFO:root: +[Valid][12][UTT][ MSE = 0.3086 | LCC = 0.8419 | SRCC = 0.8401 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.2166 | LCC = 0.9024 | SRCC = 0.8960 ] +INFO:root:Average loss=0.21971334545658186 +INFO:root: +[Valid][13][UTT][ MSE = 0.3033 | LCC = 0.8444 | SRCC = 0.8451 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.2220 | LCC = 0.9051 | SRCC = 0.9021 ] +INFO:root:Average loss=0.23659904224750322 +INFO:root: +[Valid][14][UTT][ MSE = 0.3431 | LCC = 0.8426 | SRCC = 0.8446 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.2683 | LCC = 0.9018 | SRCC = 0.9008 ] +INFO:root:Average loss=0.23464051614969206 +INFO:root: +[Valid][15][UTT][ MSE = 0.2890 | LCC = 0.8491 | SRCC = 0.8482 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1912 | LCC = 0.9092 | SRCC = 0.9054 ] +INFO:root:Average loss=0.21507561799043265 +INFO:root: +[Valid][16][UTT][ MSE = 0.2793 | LCC = 0.8515 | SRCC = 0.8499 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1891 | LCC = 0.9127 | SRCC = 0.9065 ] +INFO:root:Average loss=0.21147131709716258 +INFO:root: +[Valid][17][UTT][ MSE = 0.2529 | LCC = 0.8499 | SRCC = 0.8496 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1409 | LCC = 0.9103 | SRCC = 0.9051 ] +INFO:root:Average loss=0.21192121945130518 +INFO:root: +[Valid][18][UTT][ MSE = 0.2614 | LCC = 0.8458 | SRCC = 0.8458 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1423 | LCC = 0.9126 | SRCC = 0.9092 ] +INFO:root:Average loss=0.19656659729587725 +INFO:root: +[Valid][19][UTT][ MSE = 0.2669 | LCC = 0.8525 | SRCC = 0.8516 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1671 | LCC = 0.9199 | SRCC = 0.9171 ] +INFO:root:Average loss=0.2072015079932335 +INFO:root: +[Valid][20][UTT][ MSE = 0.2580 | LCC = 0.8494 | SRCC = 0.8492 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1512 | LCC = 0.9145 | SRCC = 0.9120 ] +INFO:root:Average loss=0.18787010644490904 +INFO:root: +[Valid][21][UTT][ MSE = 0.2558 | LCC = 0.8537 | SRCC = 0.8530 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1398 | LCC = 0.9229 | SRCC = 0.9196 ] +INFO:root:Average loss=0.18262528609006834 +INFO:root: +[Valid][22][UTT][ MSE = 0.3109 | LCC = 0.8497 | SRCC = 0.8504 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.2268 | LCC = 0.9188 | SRCC = 0.9193 ] +INFO:root:Average loss=0.18379567716366205 +INFO:root: +[Valid][23][UTT][ MSE = 0.2671 | LCC = 0.8502 | SRCC = 0.8511 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1596 | LCC = 0.9184 | SRCC = 0.9178 ] +INFO:root:Average loss=0.16958561491889831 +INFO:root: +[Valid][24][UTT][ MSE = 0.3290 | LCC = 0.8524 | SRCC = 0.8527 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.2510 | LCC = 0.9205 | SRCC = 0.9204 ] +INFO:root:Average loss=0.17480431143672037 +INFO:root: +[Valid][25][UTT][ MSE = 0.2704 | LCC = 0.8578 | SRCC = 0.8573 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1756 | LCC = 0.9267 | SRCC = 0.9243 ] +INFO:root:Average loss=0.16138773172711715 +INFO:root: +[Valid][26][UTT][ MSE = 0.2760 | LCC = 0.8530 | SRCC = 0.8534 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1867 | LCC = 0.9202 | SRCC = 0.9194 ] +INFO:root:Average loss=0.17517884572347006 +INFO:root: +[Valid][27][UTT][ MSE = 0.2612 | LCC = 0.8536 | SRCC = 0.8533 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1487 | LCC = 0.9164 | SRCC = 0.9122 ] +INFO:root:Average loss=0.1903034653992225 +INFO:root: +[Valid][28][UTT][ MSE = 0.2716 | LCC = 0.8538 | SRCC = 0.8544 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1490 | LCC = 0.9221 | SRCC = 0.9188 ] +INFO:root:Average loss=0.16389165293329802 +INFO:root: +[Valid][29][UTT][ MSE = 0.3065 | LCC = 0.8502 | SRCC = 0.8510 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.2208 | LCC = 0.9189 | SRCC = 0.9184 ] +INFO:root:Average loss=0.15112435368773264 +INFO:root: +[Valid][30][UTT][ MSE = 0.2571 | LCC = 0.8564 | SRCC = 0.8551 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1691 | LCC = 0.9202 | SRCC = 0.9159 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2634 | LCC = 0.8483 | SRCC = 0.8472 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1228 | LCC = 0.9229 | SRCC = 0.9240 ] diff --git a/results_bvcc/w2v2_base/layer10/run3/config.gin b/results_bvcc/w2v2_base/layer10/run3/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer10/run3/config.gin @@ -0,0 +1,31 @@ +# Training config on BVCC dataset. + +# ------------------------------ +# Architecture +# ------------------------------ +ProjectionHead: + conv_channels = (32, 32) + dense_neurons = (64, 32, 1) + use_poolings = (True, True) + kernel_size = 5 + pool_size = 5 + +# ------------------------------ +# Dataset +# ------------------------------ +Bvcc: + debug = False + +get_dataloader: + num_workers = 12 + shuffle = True + +# ------------------------------ +# Training loop +# ------------------------------ +TrainingLoop: + num_epochs = 30 + learning_rate = 1e-4 + dataset_cls = @Bvcc + weight_decay = 0.0 + batch_size_train = 64 diff --git a/results_bvcc/w2v2_base/layer10/run3/model.pt b/results_bvcc/w2v2_base/layer10/run3/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..88d2ce5a3b5368e6815c8399a2616337c0867f40 --- /dev/null +++ b/results_bvcc/w2v2_base/layer10/run3/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9e1f35fa31a5b15a56dd2b81dfe4855db0a5288f1d261075e354090f21f607f6 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer10/run3/model_best.pt b/results_bvcc/w2v2_base/layer10/run3/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..2d21b3224b08f6926b3b9b3ab459385e4585020c --- /dev/null +++ b/results_bvcc/w2v2_base/layer10/run3/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:de2c91977e3a45e25c04d693e1efdad18f4a4ebf787fc846577d2a8ffc126cba +size 2749638 diff --git a/results_bvcc/w2v2_base/layer10/run3/train.log b/results_bvcc/w2v2_base/layer10/run3/train.log new file mode 100644 index 0000000000000000000000000000000000000000..c7b6d89fa4b7bc0a9cfedfe61e51778b47b2ed9c --- /dev/null +++ b/results_bvcc/w2v2_base/layer10/run3/train.log @@ -0,0 +1,163 @@ +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=0.6151072016129127 +INFO:root: +[Valid][0][UTT][ MSE = 0.4204 | LCC = 0.7370 | SRCC = 0.7415 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2999 | LCC = 0.8171 | SRCC = 0.8177 ] +INFO:root:Average loss=0.38759309473710185 +INFO:root: +[Valid][1][UTT][ MSE = 0.3523 | LCC = 0.7804 | SRCC = 0.7795 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2224 | LCC = 0.8444 | SRCC = 0.8431 ] +INFO:root:Average loss=0.3734013302586017 +INFO:root: +[Valid][2][UTT][ MSE = 0.3707 | LCC = 0.7873 | SRCC = 0.7909 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.2624 | LCC = 0.8608 | SRCC = 0.8615 ] +INFO:root:Average loss=0.3029930610687305 +INFO:root: +[Valid][3][UTT][ MSE = 0.3446 | LCC = 0.8111 | SRCC = 0.8094 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.2205 | LCC = 0.8816 | SRCC = 0.8775 ] +INFO:root:Average loss=0.3115337605659778 +INFO:root: +[Valid][4][UTT][ MSE = 0.4950 | LCC = 0.7943 | SRCC = 0.8059 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.4047 | LCC = 0.8674 | SRCC = 0.8729 ] +INFO:root:Average loss=0.2675825441494966 +INFO:root: +[Valid][5][UTT][ MSE = 0.2955 | LCC = 0.8212 | SRCC = 0.8203 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1821 | LCC = 0.8854 | SRCC = 0.8803 ] +INFO:root:Average loss=0.2922194231397066 +INFO:root: +[Valid][6][UTT][ MSE = 0.3622 | LCC = 0.8166 | SRCC = 0.8248 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.2780 | LCC = 0.8809 | SRCC = 0.8855 ] +INFO:root:Average loss=0.2934482196966807 +INFO:root: +[Valid][7][UTT][ MSE = 0.4435 | LCC = 0.8111 | SRCC = 0.8185 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.3622 | LCC = 0.8790 | SRCC = 0.8799 ] +INFO:root:Average loss=0.2818548060380496 +INFO:root: +[Valid][8][UTT][ MSE = 0.3128 | LCC = 0.8260 | SRCC = 0.8290 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.2179 | LCC = 0.8915 | SRCC = 0.8921 ] +INFO:root:Average loss=0.24064893084458816 +INFO:root: +[Valid][9][UTT][ MSE = 0.2917 | LCC = 0.8292 | SRCC = 0.8278 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1780 | LCC = 0.8979 | SRCC = 0.8924 ] +INFO:root:Average loss=0.242812563593571 +INFO:root: +[Valid][10][UTT][ MSE = 0.2794 | LCC = 0.8348 | SRCC = 0.8349 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1536 | LCC = 0.9036 | SRCC = 0.8996 ] +INFO:root:Average loss=0.23124572883049646 +INFO:root: +[Valid][11][UTT][ MSE = 0.3585 | LCC = 0.8292 | SRCC = 0.8331 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.2724 | LCC = 0.8929 | SRCC = 0.8921 ] +INFO:root:Average loss=0.23634144434562096 +INFO:root: +[Valid][12][UTT][ MSE = 0.3049 | LCC = 0.8325 | SRCC = 0.8369 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.2208 | LCC = 0.8946 | SRCC = 0.8953 ] +INFO:root:Average loss=0.25324304974996126 +INFO:root: +[Valid][13][UTT][ MSE = 0.3333 | LCC = 0.8284 | SRCC = 0.8334 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.2473 | LCC = 0.8942 | SRCC = 0.8948 ] +INFO:root:Average loss=0.22296481732374582 +INFO:root: +[Valid][14][UTT][ MSE = 0.3288 | LCC = 0.8310 | SRCC = 0.8329 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.2406 | LCC = 0.8992 | SRCC = 0.8979 ] +INFO:root:Average loss=0.2211699033012757 +INFO:root: +[Valid][15][UTT][ MSE = 0.3366 | LCC = 0.8352 | SRCC = 0.8372 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.2561 | LCC = 0.8978 | SRCC = 0.8955 ] +INFO:root:Average loss=0.22075015917802468 +INFO:root: +[Valid][16][UTT][ MSE = 0.2678 | LCC = 0.8435 | SRCC = 0.8434 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1686 | LCC = 0.9088 | SRCC = 0.9047 ] +INFO:root:Average loss=0.20845631261666617 +INFO:root: +[Valid][17][UTT][ MSE = 0.2795 | LCC = 0.8428 | SRCC = 0.8424 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1774 | LCC = 0.9096 | SRCC = 0.9050 ] +INFO:root:Average loss=0.21116518544463012 +INFO:root: +[Valid][18][UTT][ MSE = 0.2620 | LCC = 0.8462 | SRCC = 0.8460 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1636 | LCC = 0.9117 | SRCC = 0.9080 ] +INFO:root:Average loss=0.19843919689838702 +INFO:root: +[Valid][19][UTT][ MSE = 0.2834 | LCC = 0.8428 | SRCC = 0.8439 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1841 | LCC = 0.9136 | SRCC = 0.9119 ] +INFO:root:Average loss=0.19134420423935622 +INFO:root: +[Valid][20][UTT][ MSE = 0.2904 | LCC = 0.8457 | SRCC = 0.8455 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1989 | LCC = 0.9120 | SRCC = 0.9076 ] +INFO:root:Average loss=0.20944310189821783 +INFO:root: +[Valid][21][UTT][ MSE = 0.2646 | LCC = 0.8465 | SRCC = 0.8490 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1666 | LCC = 0.9136 | SRCC = 0.9139 ] +INFO:root:Average loss=0.19671830009573546 +INFO:root: +[Valid][22][UTT][ MSE = 0.2879 | LCC = 0.8445 | SRCC = 0.8469 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.2027 | LCC = 0.9099 | SRCC = 0.9105 ] +INFO:root:Average loss=0.1823623925447464 +INFO:root: +[Valid][23][UTT][ MSE = 0.2827 | LCC = 0.8465 | SRCC = 0.8485 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1454 | LCC = 0.9202 | SRCC = 0.9189 ] +INFO:root:Average loss=0.19276750651307595 +INFO:root: +[Valid][24][UTT][ MSE = 0.2823 | LCC = 0.8405 | SRCC = 0.8389 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1512 | LCC = 0.9130 | SRCC = 0.9092 ] +INFO:root:Average loss=0.17636762196436906 +INFO:root: +[Valid][25][UTT][ MSE = 0.2556 | LCC = 0.8436 | SRCC = 0.8448 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1602 | LCC = 0.9090 | SRCC = 0.9076 ] +INFO:root:Average loss=0.20817641866130707 +INFO:root: +[Valid][26][UTT][ MSE = 0.2682 | LCC = 0.8417 | SRCC = 0.8437 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1759 | LCC = 0.9074 | SRCC = 0.9057 ] +INFO:root:Average loss=0.16617436869404253 +INFO:root: +[Valid][27][UTT][ MSE = 0.2511 | LCC = 0.8464 | SRCC = 0.8461 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1193 | LCC = 0.9229 | SRCC = 0.9178 ] +INFO:root:Average loss=0.16740939957209122 +INFO:root: +[Valid][28][UTT][ MSE = 0.2690 | LCC = 0.8440 | SRCC = 0.8437 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1364 | LCC = 0.9144 | SRCC = 0.9107 ] +INFO:root:Average loss=0.16532027176939523 +INFO:root: +[Valid][29][UTT][ MSE = 0.3082 | LCC = 0.8402 | SRCC = 0.8406 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.2247 | LCC = 0.9109 | SRCC = 0.9082 ] +INFO:root:Average loss=0.15212263835546297 +INFO:root: +[Valid][30][UTT][ MSE = 0.2527 | LCC = 0.8465 | SRCC = 0.8445 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1419 | LCC = 0.9131 | SRCC = 0.9048 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2515 | LCC = 0.8470 | SRCC = 0.8473 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1390 | LCC = 0.9138 | SRCC = 0.9160 ] diff --git a/results_bvcc/w2v2_base/layer10/run4/config.gin b/results_bvcc/w2v2_base/layer10/run4/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer10/run4/config.gin @@ -0,0 +1,31 @@ +# Training config on BVCC dataset. + +# ------------------------------ +# Architecture +# ------------------------------ +ProjectionHead: + conv_channels = (32, 32) + dense_neurons = (64, 32, 1) + use_poolings = (True, True) + kernel_size = 5 + pool_size = 5 + +# ------------------------------ +# Dataset +# ------------------------------ +Bvcc: + debug = False + +get_dataloader: + num_workers = 12 + shuffle = True + +# ------------------------------ +# Training loop +# ------------------------------ +TrainingLoop: + num_epochs = 30 + learning_rate = 1e-4 + dataset_cls = @Bvcc + weight_decay = 0.0 + batch_size_train = 64 diff --git a/results_bvcc/w2v2_base/layer10/run4/model.pt b/results_bvcc/w2v2_base/layer10/run4/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..fe1e507c4f14ef5777a26af56904bf4b61065afa --- /dev/null +++ b/results_bvcc/w2v2_base/layer10/run4/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8233559288f0e7cce5c389b0d98eff5da9230888a799b95da497236000e1a818 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer10/run4/model_best.pt b/results_bvcc/w2v2_base/layer10/run4/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..fe59bcb5feb5a29dc82cf1de0e73332959440335 --- /dev/null +++ b/results_bvcc/w2v2_base/layer10/run4/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4568757eef5a6b5eea5b8fcbb5aac40f903c07142d42636a57d741fcc45125b1 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer10/run4/train.log b/results_bvcc/w2v2_base/layer10/run4/train.log new file mode 100644 index 0000000000000000000000000000000000000000..14d139d9c5f2765730a9d5c9a62de44c5d99c39c --- /dev/null +++ b/results_bvcc/w2v2_base/layer10/run4/train.log @@ -0,0 +1,163 @@ +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=0.6545072178810071 +INFO:root: +[Valid][0][UTT][ MSE = 0.4145 | LCC = 0.7394 | SRCC = 0.7458 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2926 | LCC = 0.8168 | SRCC = 0.8205 ] +INFO:root:Average loss=0.355220785125708 +INFO:root: +[Valid][1][UTT][ MSE = 0.3932 | LCC = 0.7778 | SRCC = 0.7817 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2915 | LCC = 0.8502 | SRCC = 0.8524 ] +INFO:root:Average loss=0.3223578440837371 +INFO:root: +[Valid][2][UTT][ MSE = 0.3430 | LCC = 0.8016 | SRCC = 0.8013 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.2191 | LCC = 0.8739 | SRCC = 0.8727 ] +INFO:root:Average loss=0.2996859038487459 +INFO:root: +[Valid][3][UTT][ MSE = 0.3301 | LCC = 0.8054 | SRCC = 0.8083 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.2334 | LCC = 0.8701 | SRCC = 0.8720 ] +INFO:root:Average loss=0.2845711891467755 +INFO:root: +[Valid][4][UTT][ MSE = 0.3250 | LCC = 0.8211 | SRCC = 0.8189 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.2040 | LCC = 0.8849 | SRCC = 0.8811 ] +INFO:root:Average loss=0.27655357638230693 +INFO:root: +[Valid][5][UTT][ MSE = 0.3029 | LCC = 0.8262 | SRCC = 0.8253 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1902 | LCC = 0.8937 | SRCC = 0.8892 ] +INFO:root:Average loss=0.25849863504752135 +INFO:root: +[Valid][6][UTT][ MSE = 0.2964 | LCC = 0.8321 | SRCC = 0.8329 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.2004 | LCC = 0.8976 | SRCC = 0.8939 ] +INFO:root:Average loss=0.2515467290694897 +INFO:root: +[Valid][7][UTT][ MSE = 0.3240 | LCC = 0.8310 | SRCC = 0.8324 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1842 | LCC = 0.9034 | SRCC = 0.8998 ] +INFO:root:Average loss=0.27844069592463666 +INFO:root: +[Valid][8][UTT][ MSE = 0.4449 | LCC = 0.8209 | SRCC = 0.8246 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.3579 | LCC = 0.8906 | SRCC = 0.8906 ] +INFO:root:Average loss=0.26968440126914245 +INFO:root: +[Valid][9][UTT][ MSE = 0.2838 | LCC = 0.8295 | SRCC = 0.8282 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1730 | LCC = 0.8944 | SRCC = 0.8902 ] +INFO:root:Average loss=0.2712944417427748 +INFO:root: +[Valid][10][UTT][ MSE = 0.2849 | LCC = 0.8366 | SRCC = 0.8354 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1620 | LCC = 0.9042 | SRCC = 0.8981 ] +INFO:root:Average loss=0.2607968611977039 +INFO:root: +[Valid][11][UTT][ MSE = 0.3064 | LCC = 0.8345 | SRCC = 0.8328 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.2161 | LCC = 0.8940 | SRCC = 0.8878 ] +INFO:root:Average loss=0.23532921314621583 +INFO:root: +[Valid][12][UTT][ MSE = 0.3120 | LCC = 0.8419 | SRCC = 0.8409 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1901 | LCC = 0.9129 | SRCC = 0.9088 ] +INFO:root:Average loss=0.2828981878283696 +INFO:root: +[Valid][13][UTT][ MSE = 0.3866 | LCC = 0.8293 | SRCC = 0.8299 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.2966 | LCC = 0.8977 | SRCC = 0.8952 ] +INFO:root:Average loss=0.2335819055636724 +INFO:root: +[Valid][14][UTT][ MSE = 0.3715 | LCC = 0.8360 | SRCC = 0.8363 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.2870 | LCC = 0.9040 | SRCC = 0.9009 ] +INFO:root:Average loss=0.23957400100353438 +INFO:root: +[Valid][15][UTT][ MSE = 0.3176 | LCC = 0.8405 | SRCC = 0.8403 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.2366 | LCC = 0.9023 | SRCC = 0.8987 ] +INFO:root:Average loss=0.24155533466583642 +INFO:root: +[Valid][16][UTT][ MSE = 0.3208 | LCC = 0.8441 | SRCC = 0.8428 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1898 | LCC = 0.9131 | SRCC = 0.9075 ] +INFO:root:Average loss=0.24156073385324234 +INFO:root: +[Valid][17][UTT][ MSE = 0.2841 | LCC = 0.8425 | SRCC = 0.8423 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1892 | LCC = 0.9048 | SRCC = 0.9042 ] +INFO:root:Average loss=0.20834966548360312 +INFO:root: +[Valid][18][UTT][ MSE = 0.3095 | LCC = 0.8416 | SRCC = 0.8422 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.2186 | LCC = 0.9105 | SRCC = 0.9073 ] +INFO:root:Average loss=0.21217099806437126 +INFO:root: +[Valid][19][UTT][ MSE = 0.2720 | LCC = 0.8501 | SRCC = 0.8484 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1652 | LCC = 0.9162 | SRCC = 0.9114 ] +INFO:root:Average loss=0.19776207074905053 +INFO:root: +[Valid][20][UTT][ MSE = 0.2636 | LCC = 0.8447 | SRCC = 0.8453 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1664 | LCC = 0.9097 | SRCC = 0.9068 ] +INFO:root:Average loss=0.19794282622826406 +INFO:root: +[Valid][21][UTT][ MSE = 0.2411 | LCC = 0.8514 | SRCC = 0.8515 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1362 | LCC = 0.9154 | SRCC = 0.9123 ] +INFO:root:Average loss=0.1933632479646267 +INFO:root: +[Valid][22][UTT][ MSE = 0.2746 | LCC = 0.8498 | SRCC = 0.8500 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1695 | LCC = 0.9207 | SRCC = 0.9184 ] +INFO:root:Average loss=0.1782150381268599 +INFO:root: +[Valid][23][UTT][ MSE = 0.2668 | LCC = 0.8488 | SRCC = 0.8503 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1449 | LCC = 0.9202 | SRCC = 0.9170 ] +INFO:root:Average loss=0.18513291052136666 +INFO:root: +[Valid][24][UTT][ MSE = 0.3103 | LCC = 0.8454 | SRCC = 0.8476 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1695 | LCC = 0.9138 | SRCC = 0.9128 ] +INFO:root:Average loss=0.18012897880413595 +INFO:root: +[Valid][25][UTT][ MSE = 0.2771 | LCC = 0.8489 | SRCC = 0.8477 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1834 | LCC = 0.9111 | SRCC = 0.9067 ] +INFO:root:Average loss=0.20076192113069388 +INFO:root: +[Valid][26][UTT][ MSE = 0.2845 | LCC = 0.8477 | SRCC = 0.8475 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1971 | LCC = 0.9096 | SRCC = 0.9059 ] +INFO:root:Average loss=0.17811192266452006 +INFO:root: +[Valid][27][UTT][ MSE = 0.3167 | LCC = 0.8465 | SRCC = 0.8468 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.2296 | LCC = 0.9130 | SRCC = 0.9091 ] +INFO:root:Average loss=0.17272290339072546 +INFO:root: +[Valid][28][UTT][ MSE = 0.2849 | LCC = 0.8521 | SRCC = 0.8519 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1649 | LCC = 0.9216 | SRCC = 0.9194 ] +INFO:root:Average loss=0.16984143423346373 +INFO:root: +[Valid][29][UTT][ MSE = 0.2588 | LCC = 0.8487 | SRCC = 0.8484 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1560 | LCC = 0.9123 | SRCC = 0.9062 ] +INFO:root:Average loss=0.15709087529625648 +INFO:root: +[Valid][30][UTT][ MSE = 0.2657 | LCC = 0.8477 | SRCC = 0.8492 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1605 | LCC = 0.9181 | SRCC = 0.9184 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2939 | LCC = 0.8437 | SRCC = 0.8424 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1769 | LCC = 0.9285 | SRCC = 0.9301 ] diff --git a/results_bvcc/w2v2_base/layer11/run0/config.gin b/results_bvcc/w2v2_base/layer11/run0/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer11/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 diff --git a/results_bvcc/w2v2_base/layer11/run0/model.pt b/results_bvcc/w2v2_base/layer11/run0/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..80fe0d8c6f1187a71df973ffbd74d60d44e53872 --- /dev/null +++ b/results_bvcc/w2v2_base/layer11/run0/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dfc11f80c30a82b85cb4b4bdfcb97edbbfa133c881922ae9ba98706b8b0297e6 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer11/run0/model_best.pt b/results_bvcc/w2v2_base/layer11/run0/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..1b923bad2cd4f715141c60b3095a22b92eef3ff8 --- /dev/null +++ b/results_bvcc/w2v2_base/layer11/run0/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ad52517bf6eb877a028312970a4ce86574839f1c3ecc9de81d76699ab96f37bc +size 2749702 diff --git a/results_bvcc/w2v2_base/layer11/run0/train.log b/results_bvcc/w2v2_base/layer11/run0/train.log new file mode 100644 index 0000000000000000000000000000000000000000..dc0bd0dc27152495a19b0d2306bd4139db021e1a --- /dev/null +++ b/results_bvcc/w2v2_base/layer11/run0/train.log @@ -0,0 +1,163 @@ +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=0.6455222926078699 +INFO:root: +[Valid][0][UTT][ MSE = 0.5647 | LCC = 0.6759 | SRCC = 0.7119 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.4597 | LCC = 0.7481 | SRCC = 0.7608 ] +INFO:root:Average loss=0.40357579787572223 +INFO:root: +[Valid][1][UTT][ MSE = 0.4272 | LCC = 0.7687 | SRCC = 0.7667 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2970 | LCC = 0.8387 | SRCC = 0.8310 ] +INFO:root:Average loss=0.39697744773748594 +INFO:root: +[Valid][2][UTT][ MSE = 0.4527 | LCC = 0.7757 | SRCC = 0.7818 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.3565 | LCC = 0.8446 | SRCC = 0.8423 ] +INFO:root:Average loss=0.3299111532859313 +INFO:root: +[Valid][3][UTT][ MSE = 0.5422 | LCC = 0.7850 | SRCC = 0.7971 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.4597 | LCC = 0.8536 | SRCC = 0.8605 ] +INFO:root:Average loss=0.31106368318582195 +INFO:root: +[Valid][4][UTT][ MSE = 0.3591 | LCC = 0.8123 | SRCC = 0.8108 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.2491 | LCC = 0.8777 | SRCC = 0.8750 ] +INFO:root:Average loss=0.32614133583429533 +INFO:root: +[Valid][5][UTT][ MSE = 0.3149 | LCC = 0.8091 | SRCC = 0.8065 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.2032 | LCC = 0.8766 | SRCC = 0.8740 ] +INFO:root:Average loss=0.2989775559458977 +INFO:root: +[Valid][6][UTT][ MSE = 0.4583 | LCC = 0.8003 | SRCC = 0.8025 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.3748 | LCC = 0.8690 | SRCC = 0.8686 ] +INFO:root:Average loss=0.2956921005478272 +INFO:root: +[Valid][7][UTT][ MSE = 0.3242 | LCC = 0.8107 | SRCC = 0.8095 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.2181 | LCC = 0.8710 | SRCC = 0.8697 ] +INFO:root:Average loss=0.2990660447722826 +INFO:root: +[Valid][8][UTT][ MSE = 0.3365 | LCC = 0.8165 | SRCC = 0.8128 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.2320 | LCC = 0.8804 | SRCC = 0.8774 ] +INFO:root:Average loss=0.29290045702304596 +INFO:root: +[Valid][9][UTT][ MSE = 0.3527 | LCC = 0.8114 | SRCC = 0.8117 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.2550 | LCC = 0.8804 | SRCC = 0.8779 ] +INFO:root:Average loss=0.28881488606715816 +INFO:root: +[Valid][10][UTT][ MSE = 0.3209 | LCC = 0.8259 | SRCC = 0.8227 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.2150 | LCC = 0.8904 | SRCC = 0.8853 ] +INFO:root:Average loss=0.26799667072601807 +INFO:root: +[Valid][11][UTT][ MSE = 0.3062 | LCC = 0.8212 | SRCC = 0.8230 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.2023 | LCC = 0.8892 | SRCC = 0.8882 ] +INFO:root:Average loss=0.26787725988871014 +INFO:root: +[Valid][12][UTT][ MSE = 0.2852 | LCC = 0.8253 | SRCC = 0.8229 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1774 | LCC = 0.8877 | SRCC = 0.8857 ] +INFO:root:Average loss=0.2579400602441568 +INFO:root: +[Valid][13][UTT][ MSE = 0.3570 | LCC = 0.8197 | SRCC = 0.8169 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.2613 | LCC = 0.8890 | SRCC = 0.8840 ] +INFO:root:Average loss=0.2424776370708759 +INFO:root: +[Valid][14][UTT][ MSE = 0.3429 | LCC = 0.8199 | SRCC = 0.8163 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.2379 | LCC = 0.8902 | SRCC = 0.8829 ] +INFO:root:Average loss=0.23430171819069448 +INFO:root: +[Valid][15][UTT][ MSE = 0.4064 | LCC = 0.8138 | SRCC = 0.8131 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.3059 | LCC = 0.8898 | SRCC = 0.8835 ] +INFO:root:Average loss=0.23789403893244573 +INFO:root: +[Valid][16][UTT][ MSE = 0.3125 | LCC = 0.8208 | SRCC = 0.8172 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.2108 | LCC = 0.8861 | SRCC = 0.8795 ] +INFO:root:Average loss=0.23053006713206953 +INFO:root: +[Valid][17][UTT][ MSE = 0.3741 | LCC = 0.8181 | SRCC = 0.8211 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.2951 | LCC = 0.8825 | SRCC = 0.8796 ] +INFO:root:Average loss=0.25142463602316684 +INFO:root: +[Valid][18][UTT][ MSE = 0.4135 | LCC = 0.8214 | SRCC = 0.8191 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.3286 | LCC = 0.8891 | SRCC = 0.8810 ] +INFO:root:Average loss=0.22007848140903008 +INFO:root: +[Valid][19][UTT][ MSE = 0.2813 | LCC = 0.8310 | SRCC = 0.8301 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1659 | LCC = 0.9020 | SRCC = 0.8989 ] +INFO:root:Average loss=0.21565520534148583 +INFO:root: +[Valid][20][UTT][ MSE = 0.3021 | LCC = 0.8334 | SRCC = 0.8321 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1809 | LCC = 0.9008 | SRCC = 0.8960 ] +INFO:root:Average loss=0.21774061788351107 +INFO:root: +[Valid][21][UTT][ MSE = 0.2965 | LCC = 0.8323 | SRCC = 0.8291 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1958 | LCC = 0.8977 | SRCC = 0.8927 ] +INFO:root:Average loss=0.21372454011669526 +INFO:root: +[Valid][22][UTT][ MSE = 0.3308 | LCC = 0.8298 | SRCC = 0.8241 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.2081 | LCC = 0.8992 | SRCC = 0.8880 ] +INFO:root:Average loss=0.2222265633635032 +INFO:root: +[Valid][23][UTT][ MSE = 0.2914 | LCC = 0.8400 | SRCC = 0.8372 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1932 | LCC = 0.9076 | SRCC = 0.9017 ] +INFO:root:Average loss=0.22320872621658522 +INFO:root: +[Valid][24][UTT][ MSE = 0.3454 | LCC = 0.8277 | SRCC = 0.8262 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.2540 | LCC = 0.8985 | SRCC = 0.8919 ] +INFO:root:Average loss=0.21909576444289622 +INFO:root: +[Valid][25][UTT][ MSE = 0.2585 | LCC = 0.8413 | SRCC = 0.8399 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1501 | LCC = 0.9039 | SRCC = 0.8987 ] +INFO:root:Average loss=0.23766303731080812 +INFO:root: +[Valid][26][UTT][ MSE = 0.2841 | LCC = 0.8333 | SRCC = 0.8326 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1888 | LCC = 0.8965 | SRCC = 0.8938 ] +INFO:root:Average loss=0.19886974168893617 +INFO:root: +[Valid][27][UTT][ MSE = 0.2978 | LCC = 0.8314 | SRCC = 0.8269 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1885 | LCC = 0.8945 | SRCC = 0.8863 ] +INFO:root:Average loss=0.19182698858472017 +INFO:root: +[Valid][28][UTT][ MSE = 0.2951 | LCC = 0.8328 | SRCC = 0.8277 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1793 | LCC = 0.9046 | SRCC = 0.8945 ] +INFO:root:Average loss=0.18665764146508315 +INFO:root: +[Valid][29][UTT][ MSE = 0.2899 | LCC = 0.8379 | SRCC = 0.8334 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1939 | LCC = 0.9061 | SRCC = 0.8967 ] +INFO:root:Average loss=0.18604436191992882 +INFO:root: +[Valid][30][UTT][ MSE = 0.3017 | LCC = 0.8313 | SRCC = 0.8254 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.2016 | LCC = 0.9000 | SRCC = 0.8884 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2745 | LCC = 0.8364 | SRCC = 0.8363 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1628 | LCC = 0.9073 | SRCC = 0.9092 ] diff --git a/results_bvcc/w2v2_base/layer11/run1/config.gin b/results_bvcc/w2v2_base/layer11/run1/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer11/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 diff --git a/results_bvcc/w2v2_base/layer11/run1/model.pt b/results_bvcc/w2v2_base/layer11/run1/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..f68b7424b7f8127b6406162a064c1d77f781e4e8 --- /dev/null +++ b/results_bvcc/w2v2_base/layer11/run1/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ec8f46c457db3e704e5cbca6e59fa55c37174a910473172478ccf686ee201df1 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer11/run1/model_best.pt b/results_bvcc/w2v2_base/layer11/run1/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..bf6e9960e7bb9fa0c25691f4c7260455181748af --- /dev/null +++ b/results_bvcc/w2v2_base/layer11/run1/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:36a9eeebbabf95f7ccaa1933d3c6a43c755c99621905ac481d64b73787dd7179 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer11/run1/train.log b/results_bvcc/w2v2_base/layer11/run1/train.log new file mode 100644 index 0000000000000000000000000000000000000000..3a21c16bf69966b5c631839caecc377b0ff66010 --- /dev/null +++ b/results_bvcc/w2v2_base/layer11/run1/train.log @@ -0,0 +1,163 @@ +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=0.6777121298588239 +INFO:root: +[Valid][0][UTT][ MSE = 0.5208 | LCC = 0.7150 | SRCC = 0.7291 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.4165 | LCC = 0.7770 | SRCC = 0.7842 ] +INFO:root:Average loss=0.41207178242695636 +INFO:root: +[Valid][1][UTT][ MSE = 0.5329 | LCC = 0.7174 | SRCC = 0.7447 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.4299 | LCC = 0.7978 | SRCC = 0.8202 ] +INFO:root:Average loss=0.3891752799734091 +INFO:root: +[Valid][2][UTT][ MSE = 0.3544 | LCC = 0.7856 | SRCC = 0.7873 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.2358 | LCC = 0.8463 | SRCC = 0.8498 ] +INFO:root:Average loss=0.36639245389363706 +INFO:root: +[Valid][3][UTT][ MSE = 0.3719 | LCC = 0.7815 | SRCC = 0.7807 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.2601 | LCC = 0.8495 | SRCC = 0.8464 ] +INFO:root:Average loss=0.32774755855401355 +INFO:root: +[Valid][4][UTT][ MSE = 0.3613 | LCC = 0.7930 | SRCC = 0.7925 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.2549 | LCC = 0.8583 | SRCC = 0.8555 ] +INFO:root:Average loss=0.30906936583610684 +INFO:root: +[Valid][5][UTT][ MSE = 0.3296 | LCC = 0.8100 | SRCC = 0.8077 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.2179 | LCC = 0.8734 | SRCC = 0.8695 ] +INFO:root:Average loss=0.3046371056101261 +INFO:root: +[Valid][6][UTT][ MSE = 0.3526 | LCC = 0.8069 | SRCC = 0.8058 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.2528 | LCC = 0.8687 | SRCC = 0.8679 ] +INFO:root:Average loss=0.2846126778003497 +INFO:root: +[Valid][7][UTT][ MSE = 0.3285 | LCC = 0.8181 | SRCC = 0.8151 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.2170 | LCC = 0.8794 | SRCC = 0.8759 ] +INFO:root:Average loss=0.2865399620853938 +INFO:root: +[Valid][8][UTT][ MSE = 0.3877 | LCC = 0.8014 | SRCC = 0.8063 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.2862 | LCC = 0.8720 | SRCC = 0.8754 ] +INFO:root:Average loss=0.264769526437307 +INFO:root: +[Valid][9][UTT][ MSE = 0.3275 | LCC = 0.8228 | SRCC = 0.8195 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.2051 | LCC = 0.8820 | SRCC = 0.8772 ] +INFO:root:Average loss=0.2721201202426201 +INFO:root: +[Valid][10][UTT][ MSE = 0.3243 | LCC = 0.8214 | SRCC = 0.8176 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.2141 | LCC = 0.8804 | SRCC = 0.8742 ] +INFO:root:Average loss=0.27247229237587023 +INFO:root: +[Valid][11][UTT][ MSE = 0.4476 | LCC = 0.7903 | SRCC = 0.7961 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.3478 | LCC = 0.8667 | SRCC = 0.8664 ] +INFO:root:Average loss=0.25402150914455074 +INFO:root: +[Valid][12][UTT][ MSE = 0.3202 | LCC = 0.8181 | SRCC = 0.8123 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1962 | LCC = 0.8902 | SRCC = 0.8825 ] +INFO:root:Average loss=0.25222226690787536 +INFO:root: +[Valid][13][UTT][ MSE = 0.3210 | LCC = 0.8316 | SRCC = 0.8308 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1919 | LCC = 0.8901 | SRCC = 0.8872 ] +INFO:root:Average loss=0.2519393419990173 +INFO:root: +[Valid][14][UTT][ MSE = 0.2873 | LCC = 0.8269 | SRCC = 0.8227 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1733 | LCC = 0.8884 | SRCC = 0.8820 ] +INFO:root:Average loss=0.23889287064472833 +INFO:root: +[Valid][15][UTT][ MSE = 0.2837 | LCC = 0.8258 | SRCC = 0.8232 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1780 | LCC = 0.8913 | SRCC = 0.8858 ] +INFO:root:Average loss=0.23880851880098 +INFO:root: +[Valid][16][UTT][ MSE = 0.3039 | LCC = 0.8251 | SRCC = 0.8191 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1981 | LCC = 0.8865 | SRCC = 0.8756 ] +INFO:root:Average loss=0.23545555942333662 +INFO:root: +[Valid][17][UTT][ MSE = 0.2894 | LCC = 0.8264 | SRCC = 0.8211 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1755 | LCC = 0.8922 | SRCC = 0.8817 ] +INFO:root:Average loss=0.23744415530027488 +INFO:root: +[Valid][18][UTT][ MSE = 0.3457 | LCC = 0.8198 | SRCC = 0.8164 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.2371 | LCC = 0.8952 | SRCC = 0.8878 ] +INFO:root:Average loss=0.23250375477931437 +INFO:root: +[Valid][19][UTT][ MSE = 0.2920 | LCC = 0.8290 | SRCC = 0.8241 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1779 | LCC = 0.8944 | SRCC = 0.8829 ] +INFO:root:Average loss=0.22382532327603072 +INFO:root: +[Valid][20][UTT][ MSE = 0.3448 | LCC = 0.8138 | SRCC = 0.8088 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.2440 | LCC = 0.8793 | SRCC = 0.8676 ] +INFO:root:Average loss=0.23441962324655974 +INFO:root: +[Valid][21][UTT][ MSE = 0.2683 | LCC = 0.8385 | SRCC = 0.8365 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1597 | LCC = 0.9025 | SRCC = 0.8959 ] +INFO:root:Average loss=0.21088194598754248 +INFO:root: +[Valid][22][UTT][ MSE = 0.2623 | LCC = 0.8352 | SRCC = 0.8333 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1488 | LCC = 0.8955 | SRCC = 0.8902 ] +INFO:root:Average loss=0.2082187430216716 +INFO:root: +[Valid][23][UTT][ MSE = 0.2659 | LCC = 0.8346 | SRCC = 0.8308 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1578 | LCC = 0.8952 | SRCC = 0.8854 ] +INFO:root:Average loss=0.2000242100121119 +INFO:root: +[Valid][24][UTT][ MSE = 0.3160 | LCC = 0.8235 | SRCC = 0.8200 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.2155 | LCC = 0.8939 | SRCC = 0.8859 ] +INFO:root:Average loss=0.20128546349513224 +INFO:root: +[Valid][25][UTT][ MSE = 0.2892 | LCC = 0.8389 | SRCC = 0.8367 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1661 | LCC = 0.9058 | SRCC = 0.8982 ] +INFO:root:Average loss=0.18621900734993127 +INFO:root: +[Valid][26][UTT][ MSE = 0.3234 | LCC = 0.8287 | SRCC = 0.8266 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.2166 | LCC = 0.8978 | SRCC = 0.8908 ] +INFO:root:Average loss=0.21298517439609918 +INFO:root: +[Valid][27][UTT][ MSE = 0.3378 | LCC = 0.8320 | SRCC = 0.8346 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1914 | LCC = 0.8987 | SRCC = 0.8907 ] +INFO:root:Average loss=0.18291082156774324 +INFO:root: +[Valid][28][UTT][ MSE = 0.2814 | LCC = 0.8298 | SRCC = 0.8276 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1782 | LCC = 0.8973 | SRCC = 0.8925 ] +INFO:root:Average loss=0.1762896945270208 +INFO:root: +[Valid][29][UTT][ MSE = 0.3062 | LCC = 0.8319 | SRCC = 0.8299 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.2148 | LCC = 0.8946 | SRCC = 0.8869 ] +INFO:root:Average loss=0.17860180750871316 +INFO:root: +[Valid][30][UTT][ MSE = 0.2800 | LCC = 0.8348 | SRCC = 0.8323 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1669 | LCC = 0.9055 | SRCC = 0.8957 ] +INFO:root: +[Test][31][UTT][ MSE = 0.3215 | LCC = 0.8348 | SRCC = 0.8323 ] +INFO:root: +[Test][31][SYS][ MSE = 0.2133 | LCC = 0.9161 | SRCC = 0.9142 ] diff --git a/results_bvcc/w2v2_base/layer11/run2/config.gin b/results_bvcc/w2v2_base/layer11/run2/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer11/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 diff --git a/results_bvcc/w2v2_base/layer11/run2/model.pt b/results_bvcc/w2v2_base/layer11/run2/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..ff7c93b86dd733ed51512e138f06e315227adb5b --- /dev/null +++ b/results_bvcc/w2v2_base/layer11/run2/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b804aef5164a80567d4ce0c2d99af6258a756a98b876fe8942690eb11ae10324 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer11/run2/model_best.pt b/results_bvcc/w2v2_base/layer11/run2/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..71839808f43fdce8d67f2081dfc332b72e53c305 --- /dev/null +++ b/results_bvcc/w2v2_base/layer11/run2/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:01a0c6cbcfa1bb44c3cfba4d3b3977fcdaefbeb74836e93f7712184aa6166979 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer11/run2/train.log b/results_bvcc/w2v2_base/layer11/run2/train.log new file mode 100644 index 0000000000000000000000000000000000000000..0efb11ee5e92ac2a25336adc1a783b3cfadd5ab6 --- /dev/null +++ b/results_bvcc/w2v2_base/layer11/run2/train.log @@ -0,0 +1,163 @@ +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=0.6003104799833053 +INFO:root: +[Valid][0][UTT][ MSE = 0.4680 | LCC = 0.7318 | SRCC = 0.7361 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.3594 | LCC = 0.8059 | SRCC = 0.8062 ] +INFO:root:Average loss=0.3892658226765119 +INFO:root: +[Valid][1][UTT][ MSE = 0.3872 | LCC = 0.7728 | SRCC = 0.7727 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2698 | LCC = 0.8371 | SRCC = 0.8376 ] +INFO:root:Average loss=0.34596275385373676 +INFO:root: +[Valid][2][UTT][ MSE = 0.3512 | LCC = 0.7984 | SRCC = 0.7979 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.2387 | LCC = 0.8583 | SRCC = 0.8555 ] +INFO:root:Average loss=0.3237921830553275 +INFO:root: +[Valid][3][UTT][ MSE = 0.3430 | LCC = 0.8074 | SRCC = 0.8044 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.2409 | LCC = 0.8655 | SRCC = 0.8606 ] +INFO:root:Average loss=0.3099866583943367 +INFO:root: +[Valid][4][UTT][ MSE = 0.4142 | LCC = 0.8046 | SRCC = 0.8063 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.3186 | LCC = 0.8699 | SRCC = 0.8688 ] +INFO:root:Average loss=0.3186366760577911 +INFO:root: +[Valid][5][UTT][ MSE = 0.3651 | LCC = 0.8053 | SRCC = 0.8063 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.2601 | LCC = 0.8717 | SRCC = 0.8709 ] +INFO:root:Average loss=0.29510455024548066 +INFO:root: +[Valid][6][UTT][ MSE = 0.3430 | LCC = 0.8205 | SRCC = 0.8175 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.2247 | LCC = 0.8836 | SRCC = 0.8797 ] +INFO:root:Average loss=0.29405762885625547 +INFO:root: +[Valid][7][UTT][ MSE = 0.3464 | LCC = 0.8253 | SRCC = 0.8217 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.2272 | LCC = 0.8900 | SRCC = 0.8868 ] +INFO:root:Average loss=0.2764970312515895 +INFO:root: +[Valid][8][UTT][ MSE = 0.3751 | LCC = 0.8126 | SRCC = 0.8115 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.2666 | LCC = 0.8831 | SRCC = 0.8801 ] +INFO:root:Average loss=0.2914778310327958 +INFO:root: +[Valid][9][UTT][ MSE = 0.3969 | LCC = 0.8187 | SRCC = 0.8201 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.2542 | LCC = 0.8824 | SRCC = 0.8814 ] +INFO:root:Average loss=0.2988517676026393 +INFO:root: +[Valid][10][UTT][ MSE = 0.3473 | LCC = 0.8121 | SRCC = 0.8118 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.2348 | LCC = 0.8817 | SRCC = 0.8807 ] +INFO:root:Average loss=0.27727187004608983 +INFO:root: +[Valid][11][UTT][ MSE = 0.3326 | LCC = 0.8251 | SRCC = 0.8228 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.2117 | LCC = 0.8857 | SRCC = 0.8818 ] +INFO:root:Average loss=0.27046331190145934 +INFO:root: +[Valid][12][UTT][ MSE = 0.3154 | LCC = 0.8211 | SRCC = 0.8192 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1901 | LCC = 0.8878 | SRCC = 0.8853 ] +INFO:root:Average loss=0.2739945594698955 +INFO:root: +[Valid][13][UTT][ MSE = 0.3231 | LCC = 0.8264 | SRCC = 0.8220 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.2072 | LCC = 0.8896 | SRCC = 0.8832 ] +INFO:root:Average loss=0.258020879366459 +INFO:root: +[Valid][14][UTT][ MSE = 0.3184 | LCC = 0.8261 | SRCC = 0.8223 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.2033 | LCC = 0.8825 | SRCC = 0.8778 ] +INFO:root:Average loss=0.27254069997714114 +INFO:root: +[Valid][15][UTT][ MSE = 0.3430 | LCC = 0.8225 | SRCC = 0.8193 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.2314 | LCC = 0.8881 | SRCC = 0.8798 ] +INFO:root:Average loss=0.2408681788887733 +INFO:root: +[Valid][16][UTT][ MSE = 0.2906 | LCC = 0.8254 | SRCC = 0.8238 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1831 | LCC = 0.8917 | SRCC = 0.8872 ] +INFO:root:Average loss=0.23906497179697722 +INFO:root: +[Valid][17][UTT][ MSE = 0.2966 | LCC = 0.8233 | SRCC = 0.8202 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1882 | LCC = 0.8881 | SRCC = 0.8809 ] +INFO:root:Average loss=0.24480519080773377 +INFO:root: +[Valid][18][UTT][ MSE = 0.3642 | LCC = 0.8174 | SRCC = 0.8194 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.2574 | LCC = 0.8907 | SRCC = 0.8882 ] +INFO:root:Average loss=0.2429893903243236 +INFO:root: +[Valid][19][UTT][ MSE = 0.4137 | LCC = 0.8100 | SRCC = 0.8148 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.3132 | LCC = 0.8835 | SRCC = 0.8819 ] +INFO:root:Average loss=0.260135826201011 +INFO:root: +[Valid][20][UTT][ MSE = 0.2907 | LCC = 0.8286 | SRCC = 0.8261 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1859 | LCC = 0.8925 | SRCC = 0.8868 ] +INFO:root:Average loss=0.22067575175792742 +INFO:root: +[Valid][21][UTT][ MSE = 0.3179 | LCC = 0.8335 | SRCC = 0.8294 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.2053 | LCC = 0.8991 | SRCC = 0.8932 ] +INFO:root:Average loss=0.21902744949628145 +INFO:root: +[Valid][22][UTT][ MSE = 0.2907 | LCC = 0.8362 | SRCC = 0.8327 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1786 | LCC = 0.8920 | SRCC = 0.8878 ] +INFO:root:Average loss=0.24296566996818933 +INFO:root: +[Valid][23][UTT][ MSE = 0.2824 | LCC = 0.8310 | SRCC = 0.8293 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1784 | LCC = 0.8956 | SRCC = 0.8910 ] +INFO:root:Average loss=0.21568279407727411 +INFO:root: +[Valid][24][UTT][ MSE = 0.3220 | LCC = 0.8248 | SRCC = 0.8266 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.2199 | LCC = 0.8956 | SRCC = 0.8933 ] +INFO:root:Average loss=0.2115793648438576 +INFO:root: +[Valid][25][UTT][ MSE = 0.3106 | LCC = 0.8246 | SRCC = 0.8189 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1950 | LCC = 0.8962 | SRCC = 0.8871 ] +INFO:root:Average loss=0.21290304225224715 +INFO:root: +[Valid][26][UTT][ MSE = 0.2655 | LCC = 0.8369 | SRCC = 0.8357 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1538 | LCC = 0.8970 | SRCC = 0.8951 ] +INFO:root:Average loss=0.2119023553453959 +INFO:root: +[Valid][27][UTT][ MSE = 0.2836 | LCC = 0.8367 | SRCC = 0.8330 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1773 | LCC = 0.9019 | SRCC = 0.8930 ] +INFO:root:Average loss=0.21173850943644842 +INFO:root: +[Valid][28][UTT][ MSE = 0.2990 | LCC = 0.8336 | SRCC = 0.8301 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1755 | LCC = 0.8988 | SRCC = 0.8922 ] +INFO:root:Average loss=0.20261248659629089 +INFO:root: +[Valid][29][UTT][ MSE = 0.3009 | LCC = 0.8321 | SRCC = 0.8297 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.2014 | LCC = 0.8978 | SRCC = 0.8916 ] +INFO:root:Average loss=0.19815457650484183 +INFO:root: +[Valid][30][UTT][ MSE = 0.2975 | LCC = 0.8295 | SRCC = 0.8257 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1976 | LCC = 0.8912 | SRCC = 0.8832 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2877 | LCC = 0.8364 | SRCC = 0.8431 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1739 | LCC = 0.9090 | SRCC = 0.9128 ] diff --git a/results_bvcc/w2v2_base/layer11/run3/config.gin b/results_bvcc/w2v2_base/layer11/run3/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer11/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 diff --git a/results_bvcc/w2v2_base/layer11/run3/model.pt b/results_bvcc/w2v2_base/layer11/run3/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..c8fc4966860d651c3c61f5f673f18893a7400c5e --- /dev/null +++ b/results_bvcc/w2v2_base/layer11/run3/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:74580fe053e800a41750ca6168d3fdf8d9f5ed55c60ef36a91e699735b8a049a +size 2747979 diff --git a/results_bvcc/w2v2_base/layer11/run3/model_best.pt b/results_bvcc/w2v2_base/layer11/run3/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..d1d3eb3ad8df7f09f8ec6758990b351b63645601 --- /dev/null +++ b/results_bvcc/w2v2_base/layer11/run3/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ba0bc686031b105ca842e5c9877be3b0b5634fdb678d692f79b53acf69adc125 +size 2749638 diff --git a/results_bvcc/w2v2_base/layer11/run3/train.log b/results_bvcc/w2v2_base/layer11/run3/train.log new file mode 100644 index 0000000000000000000000000000000000000000..6745de04c67c79646e98da7ee3ee9a074e021259 --- /dev/null +++ b/results_bvcc/w2v2_base/layer11/run3/train.log @@ -0,0 +1,163 @@ +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=0.6673062390241867 +INFO:root: +[Valid][0][UTT][ MSE = 0.4497 | LCC = 0.7222 | SRCC = 0.7272 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.3693 | LCC = 0.7712 | SRCC = 0.7743 ] +INFO:root:Average loss=0.3851980106570782 +INFO:root: +[Valid][1][UTT][ MSE = 0.3850 | LCC = 0.7646 | SRCC = 0.7653 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2616 | LCC = 0.8378 | SRCC = 0.8371 ] +INFO:root:Average loss=0.38525681369579756 +INFO:root: +[Valid][2][UTT][ MSE = 0.3943 | LCC = 0.7871 | SRCC = 0.7841 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.2975 | LCC = 0.8474 | SRCC = 0.8419 ] +INFO:root:Average loss=0.34646142885471004 +INFO:root: +[Valid][3][UTT][ MSE = 0.3691 | LCC = 0.7985 | SRCC = 0.7956 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.2537 | LCC = 0.8611 | SRCC = 0.8570 ] +INFO:root:Average loss=0.3217357878501599 +INFO:root: +[Valid][4][UTT][ MSE = 0.3353 | LCC = 0.8013 | SRCC = 0.7986 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.2372 | LCC = 0.8601 | SRCC = 0.8581 ] +INFO:root:Average loss=0.2999670543731787 +INFO:root: +[Valid][5][UTT][ MSE = 0.3242 | LCC = 0.8132 | SRCC = 0.8095 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.2184 | LCC = 0.8769 | SRCC = 0.8735 ] +INFO:root:Average loss=0.2896061759346571 +INFO:root: +[Valid][6][UTT][ MSE = 0.3496 | LCC = 0.8111 | SRCC = 0.8075 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.2321 | LCC = 0.8769 | SRCC = 0.8715 ] +INFO:root:Average loss=0.31383106609185535 +INFO:root: +[Valid][7][UTT][ MSE = 0.3840 | LCC = 0.8064 | SRCC = 0.8069 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.2934 | LCC = 0.8692 | SRCC = 0.8672 ] +INFO:root:Average loss=0.2928306609392166 +INFO:root: +[Valid][8][UTT][ MSE = 0.3586 | LCC = 0.8154 | SRCC = 0.8100 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.2683 | LCC = 0.8751 | SRCC = 0.8692 ] +INFO:root:Average loss=0.280482394190935 +INFO:root: +[Valid][9][UTT][ MSE = 0.3762 | LCC = 0.8194 | SRCC = 0.8243 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.2476 | LCC = 0.8879 | SRCC = 0.8890 ] +INFO:root:Average loss=0.2833488344764098 +INFO:root: +[Valid][10][UTT][ MSE = 0.3155 | LCC = 0.8236 | SRCC = 0.8208 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.2140 | LCC = 0.8853 | SRCC = 0.8816 ] +INFO:root:Average loss=0.2735230911236543 +INFO:root: +[Valid][11][UTT][ MSE = 0.3046 | LCC = 0.8259 | SRCC = 0.8233 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1966 | LCC = 0.8873 | SRCC = 0.8850 ] +INFO:root:Average loss=0.248969737917949 +INFO:root: +[Valid][12][UTT][ MSE = 0.3347 | LCC = 0.8164 | SRCC = 0.8163 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.2394 | LCC = 0.8831 | SRCC = 0.8817 ] +INFO:root:Average loss=0.24772374484783563 +INFO:root: +[Valid][13][UTT][ MSE = 0.5375 | LCC = 0.7997 | SRCC = 0.8076 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.4513 | LCC = 0.8728 | SRCC = 0.8733 ] +INFO:root:Average loss=0.26829605797926587 +INFO:root: +[Valid][14][UTT][ MSE = 0.4204 | LCC = 0.8071 | SRCC = 0.8138 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.3252 | LCC = 0.8799 | SRCC = 0.8816 ] +INFO:root:Average loss=0.25674715905617446 +INFO:root: +[Valid][15][UTT][ MSE = 0.2978 | LCC = 0.8261 | SRCC = 0.8240 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1927 | LCC = 0.8950 | SRCC = 0.8907 ] +INFO:root:Average loss=0.2518398566887929 +INFO:root: +[Valid][16][UTT][ MSE = 0.4337 | LCC = 0.8097 | SRCC = 0.8115 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.3396 | LCC = 0.8837 | SRCC = 0.8790 ] +INFO:root:Average loss=0.23362450683728242 +INFO:root: +[Valid][17][UTT][ MSE = 0.2853 | LCC = 0.8266 | SRCC = 0.8267 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1901 | LCC = 0.8885 | SRCC = 0.8890 ] +INFO:root:Average loss=0.2333294852421834 +INFO:root: +[Valid][18][UTT][ MSE = 0.4068 | LCC = 0.8173 | SRCC = 0.8179 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.3106 | LCC = 0.8922 | SRCC = 0.8893 ] +INFO:root:Average loss=0.2209210529541358 +INFO:root: +[Valid][19][UTT][ MSE = 0.3067 | LCC = 0.8329 | SRCC = 0.8281 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1872 | LCC = 0.9048 | SRCC = 0.8986 ] +INFO:root:Average loss=0.22563823618185827 +INFO:root: +[Valid][20][UTT][ MSE = 0.2855 | LCC = 0.8339 | SRCC = 0.8323 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1823 | LCC = 0.9006 | SRCC = 0.8972 ] +INFO:root:Average loss=0.22429958558999574 +INFO:root: +[Valid][21][UTT][ MSE = 0.2899 | LCC = 0.8240 | SRCC = 0.8223 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1807 | LCC = 0.8839 | SRCC = 0.8825 ] +INFO:root:Average loss=0.22609921659414584 +INFO:root: +[Valid][22][UTT][ MSE = 0.2869 | LCC = 0.8288 | SRCC = 0.8238 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1908 | LCC = 0.8856 | SRCC = 0.8785 ] +INFO:root:Average loss=0.2121251737460112 +INFO:root: +[Valid][23][UTT][ MSE = 0.3567 | LCC = 0.8274 | SRCC = 0.8259 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.2589 | LCC = 0.9052 | SRCC = 0.8997 ] +INFO:root:Average loss=0.21041092372093445 +INFO:root: +[Valid][24][UTT][ MSE = 0.3075 | LCC = 0.8325 | SRCC = 0.8307 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.2119 | LCC = 0.9009 | SRCC = 0.8942 ] +INFO:root:Average loss=0.1975307013743963 +INFO:root: +[Valid][25][UTT][ MSE = 0.2909 | LCC = 0.8325 | SRCC = 0.8287 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1831 | LCC = 0.8994 | SRCC = 0.8932 ] +INFO:root:Average loss=0.19424283943879297 +INFO:root: +[Valid][26][UTT][ MSE = 0.2812 | LCC = 0.8316 | SRCC = 0.8262 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1708 | LCC = 0.9006 | SRCC = 0.8919 ] +INFO:root:Average loss=0.18941571630346468 +INFO:root: +[Valid][27][UTT][ MSE = 0.3145 | LCC = 0.8265 | SRCC = 0.8240 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.2128 | LCC = 0.8991 | SRCC = 0.8906 ] +INFO:root:Average loss=0.20809702632518914 +INFO:root: +[Valid][28][UTT][ MSE = 0.3558 | LCC = 0.8252 | SRCC = 0.8223 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.2603 | LCC = 0.8992 | SRCC = 0.8922 ] +INFO:root:Average loss=0.19708576053380966 +INFO:root: +[Valid][29][UTT][ MSE = 0.3216 | LCC = 0.8276 | SRCC = 0.8261 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.2197 | LCC = 0.9015 | SRCC = 0.8983 ] +INFO:root:Average loss=0.18196845866548708 +INFO:root: +[Valid][30][UTT][ MSE = 0.3658 | LCC = 0.8177 | SRCC = 0.8162 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.2639 | LCC = 0.8996 | SRCC = 0.8954 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2613 | LCC = 0.8433 | SRCC = 0.8427 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1338 | LCC = 0.9130 | SRCC = 0.9148 ] diff --git a/results_bvcc/w2v2_base/layer11/run4/config.gin b/results_bvcc/w2v2_base/layer11/run4/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer11/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 diff --git a/results_bvcc/w2v2_base/layer11/run4/model.pt b/results_bvcc/w2v2_base/layer11/run4/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..cf6ae537ff8f71730247eaba02ab3666df3f4957 --- /dev/null +++ b/results_bvcc/w2v2_base/layer11/run4/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3efefbf0f2069fd705e339c4574f9e23220fb188db2ac62bb28313df7a1126ae +size 2747979 diff --git a/results_bvcc/w2v2_base/layer11/run4/model_best.pt b/results_bvcc/w2v2_base/layer11/run4/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..63322e617e8b0564e90d72cefef50d6aa58ae419 --- /dev/null +++ b/results_bvcc/w2v2_base/layer11/run4/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f503bfe9b9756328340e763c66027dbb696f22c33abfc870d553669ee3e156cd +size 2749638 diff --git a/results_bvcc/w2v2_base/layer11/run4/train.log b/results_bvcc/w2v2_base/layer11/run4/train.log new file mode 100644 index 0000000000000000000000000000000000000000..5f1a884cfbdcd31f10f9a28434d764454bf62eae --- /dev/null +++ b/results_bvcc/w2v2_base/layer11/run4/train.log @@ -0,0 +1,163 @@ +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=0.6826842091022394 +INFO:root: +[Valid][0][UTT][ MSE = 0.5820 | LCC = 0.6435 | SRCC = 0.6466 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.4546 | LCC = 0.7377 | SRCC = 0.7277 ] +INFO:root:Average loss=0.4120133725496439 +INFO:root: +[Valid][1][UTT][ MSE = 0.3882 | LCC = 0.7657 | SRCC = 0.7657 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2755 | LCC = 0.8284 | SRCC = 0.8259 ] +INFO:root:Average loss=0.34557417512704164 +INFO:root: +[Valid][2][UTT][ MSE = 0.3622 | LCC = 0.7899 | SRCC = 0.7869 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.2423 | LCC = 0.8561 | SRCC = 0.8555 ] +INFO:root:Average loss=0.33199873184546447 +INFO:root: +[Valid][3][UTT][ MSE = 0.3485 | LCC = 0.7948 | SRCC = 0.7903 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.2306 | LCC = 0.8638 | SRCC = 0.8596 ] +INFO:root:Average loss=0.30769283515520585 +INFO:root: +[Valid][4][UTT][ MSE = 0.3431 | LCC = 0.8125 | SRCC = 0.8098 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.2186 | LCC = 0.8758 | SRCC = 0.8720 ] +INFO:root:Average loss=0.30204098518842304 +INFO:root: +[Valid][5][UTT][ MSE = 0.3392 | LCC = 0.8140 | SRCC = 0.8100 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.2240 | LCC = 0.8794 | SRCC = 0.8752 ] +INFO:root:Average loss=0.3005861930358104 +INFO:root: +[Valid][6][UTT][ MSE = 0.3586 | LCC = 0.8036 | SRCC = 0.8033 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.2551 | LCC = 0.8733 | SRCC = 0.8734 ] +INFO:root:Average loss=0.2986571742938115 +INFO:root: +[Valid][7][UTT][ MSE = 0.3210 | LCC = 0.8085 | SRCC = 0.8096 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.2200 | LCC = 0.8717 | SRCC = 0.8710 ] +INFO:root:Average loss=0.3032072772964453 +INFO:root: +[Valid][8][UTT][ MSE = 0.3505 | LCC = 0.8009 | SRCC = 0.7979 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.2400 | LCC = 0.8710 | SRCC = 0.8694 ] +INFO:root:Average loss=0.2738923659691444 +INFO:root: +[Valid][9][UTT][ MSE = 0.3376 | LCC = 0.8130 | SRCC = 0.8097 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.2334 | LCC = 0.8789 | SRCC = 0.8760 ] +INFO:root:Average loss=0.28141820965669095 +INFO:root: +[Valid][10][UTT][ MSE = 0.3324 | LCC = 0.8221 | SRCC = 0.8208 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.2053 | LCC = 0.8859 | SRCC = 0.8817 ] +INFO:root:Average loss=0.26851738549959964 +INFO:root: +[Valid][11][UTT][ MSE = 0.3217 | LCC = 0.8217 | SRCC = 0.8186 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1921 | LCC = 0.8942 | SRCC = 0.8888 ] +INFO:root:Average loss=0.257943457708909 +INFO:root: +[Valid][12][UTT][ MSE = 0.3132 | LCC = 0.8275 | SRCC = 0.8258 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1852 | LCC = 0.8936 | SRCC = 0.8881 ] +INFO:root:Average loss=0.2578059950700173 +INFO:root: +[Valid][13][UTT][ MSE = 0.3135 | LCC = 0.8307 | SRCC = 0.8287 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1932 | LCC = 0.8986 | SRCC = 0.8929 ] +INFO:root:Average loss=0.24126039216151604 +INFO:root: +[Valid][14][UTT][ MSE = 0.2871 | LCC = 0.8256 | SRCC = 0.8219 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1708 | LCC = 0.8946 | SRCC = 0.8891 ] +INFO:root:Average loss=0.24108283737531075 +INFO:root: +[Valid][15][UTT][ MSE = 0.2890 | LCC = 0.8282 | SRCC = 0.8281 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1777 | LCC = 0.8995 | SRCC = 0.9011 ] +INFO:root:Average loss=0.2567561503786307 +INFO:root: +[Valid][16][UTT][ MSE = 0.3144 | LCC = 0.8204 | SRCC = 0.8153 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.2043 | LCC = 0.8862 | SRCC = 0.8806 ] +INFO:root:Average loss=0.23038973621068856 +INFO:root: +[Valid][17][UTT][ MSE = 0.3740 | LCC = 0.8193 | SRCC = 0.8153 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.2712 | LCC = 0.8917 | SRCC = 0.8831 ] +INFO:root:Average loss=0.23056428965467674 +INFO:root: +[Valid][18][UTT][ MSE = 0.3573 | LCC = 0.8124 | SRCC = 0.8144 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.2534 | LCC = 0.8869 | SRCC = 0.8875 ] +INFO:root:Average loss=0.2271525175907673 +INFO:root: +[Valid][19][UTT][ MSE = 0.2838 | LCC = 0.8303 | SRCC = 0.8253 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1656 | LCC = 0.9015 | SRCC = 0.8941 ] +INFO:root:Average loss=0.23410523167023292 +INFO:root: +[Valid][20][UTT][ MSE = 0.2998 | LCC = 0.8248 | SRCC = 0.8224 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1924 | LCC = 0.8968 | SRCC = 0.8914 ] +INFO:root:Average loss=0.23758421265161955 +INFO:root: +[Valid][21][UTT][ MSE = 0.2806 | LCC = 0.8379 | SRCC = 0.8338 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1620 | LCC = 0.9031 | SRCC = 0.8957 ] +INFO:root:Average loss=0.21090209598724657 +INFO:root: +[Valid][22][UTT][ MSE = 0.2987 | LCC = 0.8284 | SRCC = 0.8222 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1841 | LCC = 0.9012 | SRCC = 0.8921 ] +INFO:root:Average loss=0.20552279761968514 +INFO:root: +[Valid][23][UTT][ MSE = 0.2847 | LCC = 0.8311 | SRCC = 0.8274 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1777 | LCC = 0.9015 | SRCC = 0.8963 ] +INFO:root:Average loss=0.214265919648684 +INFO:root: +[Valid][24][UTT][ MSE = 0.3319 | LCC = 0.8228 | SRCC = 0.8239 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.2312 | LCC = 0.8936 | SRCC = 0.8925 ] +INFO:root:Average loss=0.21972282058917558 +INFO:root: +[Valid][25][UTT][ MSE = 0.3411 | LCC = 0.8249 | SRCC = 0.8246 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.2411 | LCC = 0.8977 | SRCC = 0.8941 ] +INFO:root:Average loss=0.21510682923671526 +INFO:root: +[Valid][26][UTT][ MSE = 0.4188 | LCC = 0.8211 | SRCC = 0.8189 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.3261 | LCC = 0.8982 | SRCC = 0.8921 ] +INFO:root:Average loss=0.20609653741121292 +INFO:root: +[Valid][27][UTT][ MSE = 0.2706 | LCC = 0.8390 | SRCC = 0.8345 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1505 | LCC = 0.9124 | SRCC = 0.9062 ] +INFO:root:Average loss=0.19128478614565653 +INFO:root: +[Valid][28][UTT][ MSE = 0.2918 | LCC = 0.8367 | SRCC = 0.8327 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1806 | LCC = 0.9133 | SRCC = 0.9041 ] +INFO:root:Average loss=0.18737139782080284 +INFO:root: +[Valid][29][UTT][ MSE = 0.2696 | LCC = 0.8385 | SRCC = 0.8357 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1604 | LCC = 0.9130 | SRCC = 0.9089 ] +INFO:root:Average loss=0.1908378616357461 +INFO:root: +[Valid][30][UTT][ MSE = 0.2758 | LCC = 0.8305 | SRCC = 0.8282 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1664 | LCC = 0.9008 | SRCC = 0.8977 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2707 | LCC = 0.8424 | SRCC = 0.8398 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1542 | LCC = 0.9174 | SRCC = 0.9194 ] diff --git a/results_bvcc/w2v2_base/layer2/run0/config.gin b/results_bvcc/w2v2_base/layer2/run0/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer2/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 diff --git a/results_bvcc/w2v2_base/layer2/run0/model.pt b/results_bvcc/w2v2_base/layer2/run0/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..03373211c9be036331f3648864078403f4573059 --- /dev/null +++ b/results_bvcc/w2v2_base/layer2/run0/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a87172d8205d3ba509819605657245c181076dc059002beffcde77ddb18d7b58 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer2/run0/model_best.pt b/results_bvcc/w2v2_base/layer2/run0/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..e3dc2edbbefaa0fdf3703df9551a89e4a978c548 --- /dev/null +++ b/results_bvcc/w2v2_base/layer2/run0/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4dd6b62774dd1d04828d7282661daae5612e857abaa7065c83f159dfae251d40 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer2/run0/train.log b/results_bvcc/w2v2_base/layer2/run0/train.log new file mode 100644 index 0000000000000000000000000000000000000000..48ff22c6e857468b1f9687594f46a1a20da59f53 --- /dev/null +++ b/results_bvcc/w2v2_base/layer2/run0/train.log @@ -0,0 +1,163 @@ +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=0.482954451862054 +INFO:root: +[Valid][0][UTT][ MSE = 0.3220 | LCC = 0.8048 | SRCC = 0.8073 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2228 | LCC = 0.8679 | SRCC = 0.8721 ] +INFO:root:Average loss=0.27200799607313597 +INFO:root: +[Valid][1][UTT][ MSE = 0.3449 | LCC = 0.8315 | SRCC = 0.8384 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2669 | LCC = 0.8933 | SRCC = 0.9026 ] +INFO:root:Average loss=0.22906939265055534 +INFO:root: +[Valid][2][UTT][ MSE = 0.2732 | LCC = 0.8474 | SRCC = 0.8498 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1962 | LCC = 0.8998 | SRCC = 0.9026 ] +INFO:root:Average loss=0.20422833766310644 +INFO:root: +[Valid][3][UTT][ MSE = 0.2503 | LCC = 0.8597 | SRCC = 0.8610 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1651 | LCC = 0.9203 | SRCC = 0.9233 ] +INFO:root:Average loss=0.18800765944597048 +INFO:root: +[Valid][4][UTT][ MSE = 0.2322 | LCC = 0.8654 | SRCC = 0.8635 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1337 | LCC = 0.9287 | SRCC = 0.9267 ] +INFO:root:Average loss=0.16873418931395578 +INFO:root: +[Valid][5][UTT][ MSE = 0.2379 | LCC = 0.8683 | SRCC = 0.8672 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1463 | LCC = 0.9318 | SRCC = 0.9312 ] +INFO:root:Average loss=0.1603218610279071 +INFO:root: +[Valid][6][UTT][ MSE = 0.2351 | LCC = 0.8699 | SRCC = 0.8689 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1421 | LCC = 0.9356 | SRCC = 0.9361 ] +INFO:root:Average loss=0.14570359560923699 +INFO:root: +[Valid][7][UTT][ MSE = 0.2265 | LCC = 0.8685 | SRCC = 0.8683 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1369 | LCC = 0.9300 | SRCC = 0.9317 ] +INFO:root:Average loss=0.14478182687591284 +INFO:root: +[Valid][8][UTT][ MSE = 0.2130 | LCC = 0.8719 | SRCC = 0.8707 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1234 | LCC = 0.9290 | SRCC = 0.9292 ] +INFO:root:Average loss=0.13394135389572534 +INFO:root: +[Valid][9][UTT][ MSE = 0.2425 | LCC = 0.8673 | SRCC = 0.8678 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1554 | LCC = 0.9316 | SRCC = 0.9324 ] +INFO:root:Average loss=0.13410376289334053 +INFO:root: +[Valid][10][UTT][ MSE = 0.2146 | LCC = 0.8740 | SRCC = 0.8727 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1236 | LCC = 0.9341 | SRCC = 0.9316 ] +INFO:root:Average loss=0.11399285858258223 +INFO:root: +[Valid][11][UTT][ MSE = 0.2250 | LCC = 0.8718 | SRCC = 0.8714 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1358 | LCC = 0.9333 | SRCC = 0.9344 ] +INFO:root:Average loss=0.11316550475282547 +INFO:root: +[Valid][12][UTT][ MSE = 0.2149 | LCC = 0.8736 | SRCC = 0.8721 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1218 | LCC = 0.9338 | SRCC = 0.9320 ] +INFO:root:Average loss=0.10255196776527625 +INFO:root: +[Valid][13][UTT][ MSE = 0.2094 | LCC = 0.8764 | SRCC = 0.8753 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1070 | LCC = 0.9410 | SRCC = 0.9390 ] +INFO:root:Average loss=0.09897774954636891 +INFO:root: +[Valid][14][UTT][ MSE = 0.2842 | LCC = 0.8652 | SRCC = 0.8679 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.2061 | LCC = 0.9293 | SRCC = 0.9330 ] +INFO:root:Average loss=0.09658107161521912 +INFO:root: +[Valid][15][UTT][ MSE = 0.2040 | LCC = 0.8759 | SRCC = 0.8751 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.0996 | LCC = 0.9365 | SRCC = 0.9342 ] +INFO:root:Average loss=0.08910296293787467 +INFO:root: +[Valid][16][UTT][ MSE = 0.2313 | LCC = 0.8712 | SRCC = 0.8697 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1462 | LCC = 0.9331 | SRCC = 0.9288 ] +INFO:root:Average loss=0.08440985568822959 +INFO:root: +[Valid][17][UTT][ MSE = 0.2168 | LCC = 0.8740 | SRCC = 0.8728 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1312 | LCC = 0.9339 | SRCC = 0.9335 ] +INFO:root:Average loss=0.07786295754023087 +INFO:root: +[Valid][18][UTT][ MSE = 0.2173 | LCC = 0.8732 | SRCC = 0.8717 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1311 | LCC = 0.9326 | SRCC = 0.9309 ] +INFO:root:Average loss=0.07497032913259971 +INFO:root: +[Valid][19][UTT][ MSE = 0.2093 | LCC = 0.8726 | SRCC = 0.8717 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1112 | LCC = 0.9415 | SRCC = 0.9399 ] +INFO:root:Average loss=0.07791718300909568 +INFO:root: +[Valid][20][UTT][ MSE = 0.2306 | LCC = 0.8713 | SRCC = 0.8704 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1433 | LCC = 0.9379 | SRCC = 0.9357 ] +INFO:root:Average loss=0.06942635335219212 +INFO:root: +[Valid][21][UTT][ MSE = 0.2157 | LCC = 0.8691 | SRCC = 0.8673 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1162 | LCC = 0.9368 | SRCC = 0.9340 ] +INFO:root:Average loss=0.07016778393433644 +INFO:root: +[Valid][22][UTT][ MSE = 0.2352 | LCC = 0.8669 | SRCC = 0.8646 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1467 | LCC = 0.9349 | SRCC = 0.9304 ] +INFO:root:Average loss=0.06461190370221932 +INFO:root: +[Valid][23][UTT][ MSE = 0.2409 | LCC = 0.8687 | SRCC = 0.8673 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1565 | LCC = 0.9362 | SRCC = 0.9344 ] +INFO:root:Average loss=0.06316893170468318 +INFO:root: +[Valid][24][UTT][ MSE = 0.2233 | LCC = 0.8704 | SRCC = 0.8683 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1370 | LCC = 0.9321 | SRCC = 0.9273 ] +INFO:root:Average loss=0.05942078526967611 +INFO:root: +[Valid][25][UTT][ MSE = 0.2347 | LCC = 0.8688 | SRCC = 0.8668 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1519 | LCC = 0.9322 | SRCC = 0.9277 ] +INFO:root:Average loss=0.05671664800208349 +INFO:root: +[Valid][26][UTT][ MSE = 0.2135 | LCC = 0.8691 | SRCC = 0.8678 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1149 | LCC = 0.9376 | SRCC = 0.9353 ] +INFO:root:Average loss=0.05771111439053829 +INFO:root: +[Valid][27][UTT][ MSE = 0.2166 | LCC = 0.8700 | SRCC = 0.8673 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1201 | LCC = 0.9336 | SRCC = 0.9269 ] +INFO:root:Average loss=0.05265827018481035 +INFO:root: +[Valid][28][UTT][ MSE = 0.2263 | LCC = 0.8681 | SRCC = 0.8665 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1297 | LCC = 0.9362 | SRCC = 0.9349 ] +INFO:root:Average loss=0.05084032777887888 +INFO:root: +[Valid][29][UTT][ MSE = 0.2232 | LCC = 0.8697 | SRCC = 0.8679 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1226 | LCC = 0.9404 | SRCC = 0.9356 ] +INFO:root:Average loss=0.05115930165331333 +INFO:root: +[Valid][30][UTT][ MSE = 0.2095 | LCC = 0.8715 | SRCC = 0.8692 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1064 | LCC = 0.9365 | SRCC = 0.9306 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2230 | LCC = 0.8723 | SRCC = 0.8720 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1486 | LCC = 0.9162 | SRCC = 0.9163 ] diff --git a/results_bvcc/w2v2_base/layer2/run1/config.gin b/results_bvcc/w2v2_base/layer2/run1/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer2/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 diff --git a/results_bvcc/w2v2_base/layer2/run1/model.pt b/results_bvcc/w2v2_base/layer2/run1/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..50ba78d401efd1f56ce08dec08f1cd4edeab0369 --- /dev/null +++ b/results_bvcc/w2v2_base/layer2/run1/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f8f7254246ff87700be63a345c9c0ee81fad3e1f797f39cd3a31202a5bcad6b3 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer2/run1/model_best.pt b/results_bvcc/w2v2_base/layer2/run1/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..d82edb7a819f3c7bf25bff68d608dc27cc68700d --- /dev/null +++ b/results_bvcc/w2v2_base/layer2/run1/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8fb5e484b58299ce462b6d7c231c9e06af935822c720260c61f882a3c67b9971 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer2/run1/train.log b/results_bvcc/w2v2_base/layer2/run1/train.log new file mode 100644 index 0000000000000000000000000000000000000000..e1abb9bbde3b1b7d4f516084fd78325c105b3e8a --- /dev/null +++ b/results_bvcc/w2v2_base/layer2/run1/train.log @@ -0,0 +1,163 @@ +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=0.6088308644218322 +INFO:root: +[Valid][0][UTT][ MSE = 0.3502 | LCC = 0.7788 | SRCC = 0.7889 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2564 | LCC = 0.8382 | SRCC = 0.8525 ] +INFO:root:Average loss=0.27738805210743195 +INFO:root: +[Valid][1][UTT][ MSE = 0.2760 | LCC = 0.8394 | SRCC = 0.8404 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1710 | LCC = 0.8979 | SRCC = 0.9011 ] +INFO:root:Average loss=0.2269272389702308 +INFO:root: +[Valid][2][UTT][ MSE = 0.2783 | LCC = 0.8550 | SRCC = 0.8569 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1886 | LCC = 0.9164 | SRCC = 0.9219 ] +INFO:root:Average loss=0.19760936116560912 +INFO:root: +[Valid][3][UTT][ MSE = 0.2357 | LCC = 0.8666 | SRCC = 0.8656 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1393 | LCC = 0.9268 | SRCC = 0.9267 ] +INFO:root:Average loss=0.17750048541869873 +INFO:root: +[Valid][4][UTT][ MSE = 0.2451 | LCC = 0.8636 | SRCC = 0.8634 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1520 | LCC = 0.9265 | SRCC = 0.9283 ] +INFO:root:Average loss=0.16420816907133812 +INFO:root: +[Valid][5][UTT][ MSE = 0.2240 | LCC = 0.8719 | SRCC = 0.8745 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1363 | LCC = 0.9329 | SRCC = 0.9374 ] +INFO:root:Average loss=0.15668228402351722 +INFO:root: +[Valid][6][UTT][ MSE = 0.2177 | LCC = 0.8731 | SRCC = 0.8732 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1242 | LCC = 0.9348 | SRCC = 0.9384 ] +INFO:root:Average loss=0.142537188071471 +INFO:root: +[Valid][7][UTT][ MSE = 0.2166 | LCC = 0.8753 | SRCC = 0.8743 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1217 | LCC = 0.9381 | SRCC = 0.9395 ] +INFO:root:Average loss=0.13764455007054868 +INFO:root: +[Valid][8][UTT][ MSE = 0.2125 | LCC = 0.8732 | SRCC = 0.8736 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1173 | LCC = 0.9364 | SRCC = 0.9391 ] +INFO:root:Average loss=0.12069327412889554 +INFO:root: +[Valid][9][UTT][ MSE = 0.2123 | LCC = 0.8767 | SRCC = 0.8772 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1225 | LCC = 0.9395 | SRCC = 0.9435 ] +INFO:root:Average loss=0.11904759408953862 +INFO:root: +[Valid][10][UTT][ MSE = 0.2135 | LCC = 0.8777 | SRCC = 0.8768 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1299 | LCC = 0.9371 | SRCC = 0.9379 ] +INFO:root:Average loss=0.11248624358230676 +INFO:root: +[Valid][11][UTT][ MSE = 0.2000 | LCC = 0.8788 | SRCC = 0.8783 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1053 | LCC = 0.9412 | SRCC = 0.9412 ] +INFO:root:Average loss=0.10821921344942008 +INFO:root: +[Valid][12][UTT][ MSE = 0.2054 | LCC = 0.8781 | SRCC = 0.8771 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1153 | LCC = 0.9407 | SRCC = 0.9414 ] +INFO:root:Average loss=0.09464836254333839 +INFO:root: +[Valid][13][UTT][ MSE = 0.2163 | LCC = 0.8763 | SRCC = 0.8758 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1290 | LCC = 0.9395 | SRCC = 0.9436 ] +INFO:root:Average loss=0.09158457261629593 +INFO:root: +[Valid][14][UTT][ MSE = 0.2170 | LCC = 0.8732 | SRCC = 0.8743 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1269 | LCC = 0.9370 | SRCC = 0.9424 ] +INFO:root:Average loss=0.0874979652177829 +INFO:root: +[Valid][15][UTT][ MSE = 0.2571 | LCC = 0.8703 | SRCC = 0.8718 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1811 | LCC = 0.9331 | SRCC = 0.9391 ] +INFO:root:Average loss=0.0838656835258007 +INFO:root: +[Valid][16][UTT][ MSE = 0.2166 | LCC = 0.8754 | SRCC = 0.8751 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1303 | LCC = 0.9389 | SRCC = 0.9382 ] +INFO:root:Average loss=0.0768170833396606 +INFO:root: +[Valid][17][UTT][ MSE = 0.1963 | LCC = 0.8762 | SRCC = 0.8761 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.0940 | LCC = 0.9439 | SRCC = 0.9447 ] +INFO:root:Average loss=0.07279678166676791 +INFO:root: +[Valid][18][UTT][ MSE = 0.2068 | LCC = 0.8726 | SRCC = 0.8719 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1103 | LCC = 0.9369 | SRCC = 0.9387 ] +INFO:root:Average loss=0.0753249945357824 +INFO:root: +[Valid][19][UTT][ MSE = 0.2332 | LCC = 0.8714 | SRCC = 0.8719 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1504 | LCC = 0.9370 | SRCC = 0.9418 ] +INFO:root:Average loss=0.06767097402077454 +INFO:root: +[Valid][20][UTT][ MSE = 0.2148 | LCC = 0.8742 | SRCC = 0.8726 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1214 | LCC = 0.9406 | SRCC = 0.9399 ] +INFO:root:Average loss=0.07097760659570877 +INFO:root: +[Valid][21][UTT][ MSE = 0.2100 | LCC = 0.8731 | SRCC = 0.8718 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1109 | LCC = 0.9404 | SRCC = 0.9383 ] +INFO:root:Average loss=0.06349302465334916 +INFO:root: +[Valid][22][UTT][ MSE = 0.2215 | LCC = 0.8714 | SRCC = 0.8703 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1303 | LCC = 0.9388 | SRCC = 0.9385 ] +INFO:root:Average loss=0.06413250368757126 +INFO:root: +[Valid][23][UTT][ MSE = 0.2209 | LCC = 0.8713 | SRCC = 0.8713 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1258 | LCC = 0.9414 | SRCC = 0.9441 ] +INFO:root:Average loss=0.057331459620633185 +INFO:root: +[Valid][24][UTT][ MSE = 0.2516 | LCC = 0.8716 | SRCC = 0.8714 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1699 | LCC = 0.9378 | SRCC = 0.9406 ] +INFO:root:Average loss=0.05777340133984884 +INFO:root: +[Valid][25][UTT][ MSE = 0.2118 | LCC = 0.8715 | SRCC = 0.8702 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1156 | LCC = 0.9388 | SRCC = 0.9369 ] +INFO:root:Average loss=0.05594903440811695 +INFO:root: +[Valid][26][UTT][ MSE = 0.2089 | LCC = 0.8742 | SRCC = 0.8734 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1111 | LCC = 0.9422 | SRCC = 0.9440 ] +INFO:root:Average loss=0.05032864819543484 +INFO:root: +[Valid][27][UTT][ MSE = 0.2078 | LCC = 0.8730 | SRCC = 0.8713 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1056 | LCC = 0.9435 | SRCC = 0.9429 ] +INFO:root:Average loss=0.04933541982124249 +INFO:root: +[Valid][28][UTT][ MSE = 0.2160 | LCC = 0.8709 | SRCC = 0.8697 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1189 | LCC = 0.9407 | SRCC = 0.9408 ] +INFO:root:Average loss=0.04820012315534628 +INFO:root: +[Valid][29][UTT][ MSE = 0.2271 | LCC = 0.8729 | SRCC = 0.8724 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1376 | LCC = 0.9419 | SRCC = 0.9428 ] +INFO:root:Average loss=0.04829420927816477 +INFO:root: +[Valid][30][UTT][ MSE = 0.2172 | LCC = 0.8736 | SRCC = 0.8723 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1241 | LCC = 0.9432 | SRCC = 0.9434 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2189 | LCC = 0.8686 | SRCC = 0.8681 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1261 | LCC = 0.9256 | SRCC = 0.9257 ] diff --git a/results_bvcc/w2v2_base/layer2/run2/config.gin b/results_bvcc/w2v2_base/layer2/run2/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer2/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 diff --git a/results_bvcc/w2v2_base/layer2/run2/model.pt b/results_bvcc/w2v2_base/layer2/run2/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..b02367a12df911e677bd32ec933c7783bf41ab8b --- /dev/null +++ b/results_bvcc/w2v2_base/layer2/run2/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2a83befebc6c6240d2820a173bfbcbad63e4f5eadf60f353819bcaddf1a2cbc9 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer2/run2/model_best.pt b/results_bvcc/w2v2_base/layer2/run2/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..20d2e54bd09332ac2c8f1029850f320638f41305 --- /dev/null +++ b/results_bvcc/w2v2_base/layer2/run2/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c61f22c4d11d65f5ecf455513dfab454e1195e56a0fc2201f880ea3c9d8aaf93 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer2/run2/train.log b/results_bvcc/w2v2_base/layer2/run2/train.log new file mode 100644 index 0000000000000000000000000000000000000000..2139474e02143332b8baf8662983089ac74b15ee --- /dev/null +++ b/results_bvcc/w2v2_base/layer2/run2/train.log @@ -0,0 +1,163 @@ +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=0.5705115963251163 +INFO:root: +[Valid][0][UTT][ MSE = 0.3555 | LCC = 0.7787 | SRCC = 0.7899 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2513 | LCC = 0.8477 | SRCC = 0.8612 ] +INFO:root:Average loss=0.2832266852641717 +INFO:root: +[Valid][1][UTT][ MSE = 0.3086 | LCC = 0.8289 | SRCC = 0.8360 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2188 | LCC = 0.8928 | SRCC = 0.9000 ] +INFO:root:Average loss=0.2279693358219587 +INFO:root: +[Valid][2][UTT][ MSE = 0.2631 | LCC = 0.8508 | SRCC = 0.8516 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1646 | LCC = 0.9124 | SRCC = 0.9110 ] +INFO:root:Average loss=0.20434632477087852 +INFO:root: +[Valid][3][UTT][ MSE = 0.2406 | LCC = 0.8592 | SRCC = 0.8611 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1447 | LCC = 0.9234 | SRCC = 0.9239 ] +INFO:root:Average loss=0.1874878452374385 +INFO:root: +[Valid][4][UTT][ MSE = 0.2336 | LCC = 0.8638 | SRCC = 0.8640 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1298 | LCC = 0.9226 | SRCC = 0.9210 ] +INFO:root:Average loss=0.17548125561995384 +INFO:root: +[Valid][5][UTT][ MSE = 0.2240 | LCC = 0.8642 | SRCC = 0.8651 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1298 | LCC = 0.9237 | SRCC = 0.9229 ] +INFO:root:Average loss=0.16438788786912575 +INFO:root: +[Valid][6][UTT][ MSE = 0.2569 | LCC = 0.8640 | SRCC = 0.8682 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1746 | LCC = 0.9253 | SRCC = 0.9287 ] +INFO:root:Average loss=0.1527534862741446 +INFO:root: +[Valid][7][UTT][ MSE = 0.2308 | LCC = 0.8638 | SRCC = 0.8664 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1375 | LCC = 0.9284 | SRCC = 0.9315 ] +INFO:root:Average loss=0.13682211677615458 +INFO:root: +[Valid][8][UTT][ MSE = 0.2261 | LCC = 0.8654 | SRCC = 0.8651 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1325 | LCC = 0.9274 | SRCC = 0.9268 ] +INFO:root:Average loss=0.1315191204731281 +INFO:root: +[Valid][9][UTT][ MSE = 0.2022 | LCC = 0.8708 | SRCC = 0.8720 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.0921 | LCC = 0.9373 | SRCC = 0.9381 ] +INFO:root:Average loss=0.1235604960566912 +INFO:root: +[Valid][10][UTT][ MSE = 0.2412 | LCC = 0.8683 | SRCC = 0.8691 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1521 | LCC = 0.9352 | SRCC = 0.9353 ] +INFO:root:Average loss=0.12201854185416149 +INFO:root: +[Valid][11][UTT][ MSE = 0.2180 | LCC = 0.8717 | SRCC = 0.8712 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1202 | LCC = 0.9372 | SRCC = 0.9342 ] +INFO:root:Average loss=0.1139524788237535 +INFO:root: +[Valid][12][UTT][ MSE = 0.2120 | LCC = 0.8702 | SRCC = 0.8711 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1211 | LCC = 0.9305 | SRCC = 0.9301 ] +INFO:root:Average loss=0.10202818449873191 +INFO:root: +[Valid][13][UTT][ MSE = 0.2379 | LCC = 0.8699 | SRCC = 0.8704 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1544 | LCC = 0.9341 | SRCC = 0.9336 ] +INFO:root:Average loss=0.09498395121250397 +INFO:root: +[Valid][14][UTT][ MSE = 0.2465 | LCC = 0.8680 | SRCC = 0.8694 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1572 | LCC = 0.9340 | SRCC = 0.9344 ] +INFO:root:Average loss=0.09195501581789592 +INFO:root: +[Valid][15][UTT][ MSE = 0.2034 | LCC = 0.8716 | SRCC = 0.8715 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.0975 | LCC = 0.9355 | SRCC = 0.9348 ] +INFO:root:Average loss=0.09475782599586707 +INFO:root: +[Valid][16][UTT][ MSE = 0.2314 | LCC = 0.8688 | SRCC = 0.8706 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1387 | LCC = 0.9367 | SRCC = 0.9406 ] +INFO:root:Average loss=0.08204915374517441 +INFO:root: +[Valid][17][UTT][ MSE = 0.2397 | LCC = 0.8668 | SRCC = 0.8682 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1515 | LCC = 0.9314 | SRCC = 0.9335 ] +INFO:root:Average loss=0.08633479012701756 +INFO:root: +[Valid][18][UTT][ MSE = 0.2600 | LCC = 0.8659 | SRCC = 0.8693 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1776 | LCC = 0.9305 | SRCC = 0.9321 ] +INFO:root:Average loss=0.07447785931902054 +INFO:root: +[Valid][19][UTT][ MSE = 0.2205 | LCC = 0.8718 | SRCC = 0.8732 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1327 | LCC = 0.9341 | SRCC = 0.9324 ] +INFO:root:Average loss=0.07263165206099167 +INFO:root: +[Valid][20][UTT][ MSE = 0.2134 | LCC = 0.8725 | SRCC = 0.8723 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1169 | LCC = 0.9385 | SRCC = 0.9368 ] +INFO:root:Average loss=0.07442939759064944 +INFO:root: +[Valid][21][UTT][ MSE = 0.2279 | LCC = 0.8710 | SRCC = 0.8726 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1314 | LCC = 0.9411 | SRCC = 0.9405 ] +INFO:root:Average loss=0.0702505098321499 +INFO:root: +[Valid][22][UTT][ MSE = 0.2374 | LCC = 0.8680 | SRCC = 0.8702 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1462 | LCC = 0.9361 | SRCC = 0.9351 ] +INFO:root:Average loss=0.06415436440744461 +INFO:root: +[Valid][23][UTT][ MSE = 0.2244 | LCC = 0.8708 | SRCC = 0.8718 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1321 | LCC = 0.9386 | SRCC = 0.9376 ] +INFO:root:Average loss=0.060042525617740095 +INFO:root: +[Valid][24][UTT][ MSE = 0.2325 | LCC = 0.8713 | SRCC = 0.8720 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1396 | LCC = 0.9399 | SRCC = 0.9395 ] +INFO:root:Average loss=0.05847985058640822 +INFO:root: +[Valid][25][UTT][ MSE = 0.2324 | LCC = 0.8691 | SRCC = 0.8691 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1474 | LCC = 0.9334 | SRCC = 0.9308 ] +INFO:root:Average loss=0.06009320652064604 +INFO:root: +[Valid][26][UTT][ MSE = 0.2519 | LCC = 0.8682 | SRCC = 0.8689 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1704 | LCC = 0.9316 | SRCC = 0.9301 ] +INFO:root:Average loss=0.057863487265048884 +INFO:root: +[Valid][27][UTT][ MSE = 0.2431 | LCC = 0.8683 | SRCC = 0.8704 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1568 | LCC = 0.9340 | SRCC = 0.9338 ] +INFO:root:Average loss=0.052555808367637485 +INFO:root: +[Valid][28][UTT][ MSE = 0.2270 | LCC = 0.8705 | SRCC = 0.8709 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1371 | LCC = 0.9349 | SRCC = 0.9337 ] +INFO:root:Average loss=0.0516127003595615 +INFO:root: +[Valid][29][UTT][ MSE = 0.2462 | LCC = 0.8657 | SRCC = 0.8672 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1608 | LCC = 0.9319 | SRCC = 0.9328 ] +INFO:root:Average loss=0.05194547691215307 +INFO:root: +[Valid][30][UTT][ MSE = 0.2335 | LCC = 0.8696 | SRCC = 0.8701 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1434 | LCC = 0.9373 | SRCC = 0.9362 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2146 | LCC = 0.8673 | SRCC = 0.8661 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1150 | LCC = 0.9175 | SRCC = 0.9156 ] diff --git a/results_bvcc/w2v2_base/layer2/run3/config.gin b/results_bvcc/w2v2_base/layer2/run3/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer2/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 diff --git a/results_bvcc/w2v2_base/layer2/run3/model.pt b/results_bvcc/w2v2_base/layer2/run3/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..175b4a891cd136a9061f489e168c70a046cd7ca6 --- /dev/null +++ b/results_bvcc/w2v2_base/layer2/run3/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:44d00882ce9bdb27572e9e7e029b746274167f0fee356d1fca384ce6bd2ff4e8 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer2/run3/model_best.pt b/results_bvcc/w2v2_base/layer2/run3/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..cdb39b4e92edb4f62dbf6aedb2fcd39e45eede24 --- /dev/null +++ b/results_bvcc/w2v2_base/layer2/run3/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c53d6202dbac59f3eb1366167bc18d7dbc8917af22dfcc319114143583d4e009 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer2/run3/train.log b/results_bvcc/w2v2_base/layer2/run3/train.log new file mode 100644 index 0000000000000000000000000000000000000000..d65c4f25c5b3c8949acbb17f8d3c249fd7fa9315 --- /dev/null +++ b/results_bvcc/w2v2_base/layer2/run3/train.log @@ -0,0 +1,163 @@ +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=0.47186055855873305 +INFO:root: +[Valid][0][UTT][ MSE = 0.3180 | LCC = 0.7977 | SRCC = 0.8064 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2032 | LCC = 0.8637 | SRCC = 0.8745 ] +INFO:root:Average loss=0.2624582029305972 +INFO:root: +[Valid][1][UTT][ MSE = 0.2600 | LCC = 0.8367 | SRCC = 0.8406 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1522 | LCC = 0.9006 | SRCC = 0.9075 ] +INFO:root:Average loss=0.22028402334604508 +INFO:root: +[Valid][2][UTT][ MSE = 0.2417 | LCC = 0.8501 | SRCC = 0.8539 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1450 | LCC = 0.9101 | SRCC = 0.9179 ] +INFO:root:Average loss=0.20250543436178794 +INFO:root: +[Valid][3][UTT][ MSE = 0.2395 | LCC = 0.8538 | SRCC = 0.8580 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1448 | LCC = 0.9157 | SRCC = 0.9245 ] +INFO:root:Average loss=0.19198323528353983 +INFO:root: +[Valid][4][UTT][ MSE = 0.2402 | LCC = 0.8610 | SRCC = 0.8629 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1486 | LCC = 0.9241 | SRCC = 0.9273 ] +INFO:root:Average loss=0.17301098352823502 +INFO:root: +[Valid][5][UTT][ MSE = 0.2192 | LCC = 0.8605 | SRCC = 0.8630 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1110 | LCC = 0.9246 | SRCC = 0.9310 ] +INFO:root:Average loss=0.1627328746402875 +INFO:root: +[Valid][6][UTT][ MSE = 0.2415 | LCC = 0.8638 | SRCC = 0.8650 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1565 | LCC = 0.9237 | SRCC = 0.9268 ] +INFO:root:Average loss=0.14924141181967196 +INFO:root: +[Valid][7][UTT][ MSE = 0.2151 | LCC = 0.8671 | SRCC = 0.8671 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1167 | LCC = 0.9301 | SRCC = 0.9314 ] +INFO:root:Average loss=0.1419510532839176 +INFO:root: +[Valid][8][UTT][ MSE = 0.2448 | LCC = 0.8668 | SRCC = 0.8699 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1651 | LCC = 0.9260 | SRCC = 0.9295 ] +INFO:root:Average loss=0.13182480852955428 +INFO:root: +[Valid][9][UTT][ MSE = 0.2100 | LCC = 0.8712 | SRCC = 0.8705 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1153 | LCC = 0.9305 | SRCC = 0.9300 ] +INFO:root:Average loss=0.11972101118702155 +INFO:root: +[Valid][10][UTT][ MSE = 0.2134 | LCC = 0.8671 | SRCC = 0.8671 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1172 | LCC = 0.9302 | SRCC = 0.9305 ] +INFO:root:Average loss=0.10903176283225036 +INFO:root: +[Valid][11][UTT][ MSE = 0.2125 | LCC = 0.8675 | SRCC = 0.8681 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1010 | LCC = 0.9333 | SRCC = 0.9340 ] +INFO:root:Average loss=0.1088200886375629 +INFO:root: +[Valid][12][UTT][ MSE = 0.2181 | LCC = 0.8687 | SRCC = 0.8691 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1264 | LCC = 0.9324 | SRCC = 0.9327 ] +INFO:root:Average loss=0.10160787862080795 +INFO:root: +[Valid][13][UTT][ MSE = 0.2122 | LCC = 0.8684 | SRCC = 0.8678 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1059 | LCC = 0.9353 | SRCC = 0.9360 ] +INFO:root:Average loss=0.09808076445299846 +INFO:root: +[Valid][14][UTT][ MSE = 0.2123 | LCC = 0.8684 | SRCC = 0.8684 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1020 | LCC = 0.9383 | SRCC = 0.9410 ] +INFO:root:Average loss=0.09452442691112176 +INFO:root: +[Valid][15][UTT][ MSE = 0.2089 | LCC = 0.8705 | SRCC = 0.8714 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1124 | LCC = 0.9355 | SRCC = 0.9373 ] +INFO:root:Average loss=0.08577243210031436 +INFO:root: +[Valid][16][UTT][ MSE = 0.2181 | LCC = 0.8696 | SRCC = 0.8696 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1266 | LCC = 0.9344 | SRCC = 0.9336 ] +INFO:root:Average loss=0.08414324425542966 +INFO:root: +[Valid][17][UTT][ MSE = 0.2276 | LCC = 0.8678 | SRCC = 0.8695 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1354 | LCC = 0.9345 | SRCC = 0.9397 ] +INFO:root:Average loss=0.0833642983283752 +INFO:root: +[Valid][18][UTT][ MSE = 0.2019 | LCC = 0.8723 | SRCC = 0.8727 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1021 | LCC = 0.9362 | SRCC = 0.9374 ] +INFO:root:Average loss=0.07738380731107333 +INFO:root: +[Valid][19][UTT][ MSE = 0.2095 | LCC = 0.8680 | SRCC = 0.8696 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1053 | LCC = 0.9352 | SRCC = 0.9393 ] +INFO:root:Average loss=0.0696348740408818 +INFO:root: +[Valid][20][UTT][ MSE = 0.2151 | LCC = 0.8691 | SRCC = 0.8680 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1035 | LCC = 0.9360 | SRCC = 0.9349 ] +INFO:root:Average loss=0.06817801893712619 +INFO:root: +[Valid][21][UTT][ MSE = 0.2034 | LCC = 0.8697 | SRCC = 0.8698 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.0986 | LCC = 0.9355 | SRCC = 0.9388 ] +INFO:root:Average loss=0.06503444012158956 +INFO:root: +[Valid][22][UTT][ MSE = 0.2067 | LCC = 0.8717 | SRCC = 0.8712 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1006 | LCC = 0.9365 | SRCC = 0.9365 ] +INFO:root:Average loss=0.06267922534010349 +INFO:root: +[Valid][23][UTT][ MSE = 0.2133 | LCC = 0.8693 | SRCC = 0.8687 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1000 | LCC = 0.9355 | SRCC = 0.9339 ] +INFO:root:Average loss=0.06198704605683302 +INFO:root: +[Valid][24][UTT][ MSE = 0.2206 | LCC = 0.8658 | SRCC = 0.8675 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1266 | LCC = 0.9322 | SRCC = 0.9366 ] +INFO:root:Average loss=0.05823039750640209 +INFO:root: +[Valid][25][UTT][ MSE = 0.2054 | LCC = 0.8711 | SRCC = 0.8695 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1019 | LCC = 0.9391 | SRCC = 0.9364 ] +INFO:root:Average loss=0.05994454604119827 +INFO:root: +[Valid][26][UTT][ MSE = 0.2051 | LCC = 0.8711 | SRCC = 0.8704 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1023 | LCC = 0.9388 | SRCC = 0.9402 ] +INFO:root:Average loss=0.05661682106363468 +INFO:root: +[Valid][27][UTT][ MSE = 0.2098 | LCC = 0.8706 | SRCC = 0.8708 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1087 | LCC = 0.9374 | SRCC = 0.9392 ] +INFO:root:Average loss=0.057367766801363386 +INFO:root: +[Valid][28][UTT][ MSE = 0.2236 | LCC = 0.8713 | SRCC = 0.8702 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1024 | LCC = 0.9401 | SRCC = 0.9382 ] +INFO:root:Average loss=0.05500842788471626 +INFO:root: +[Valid][29][UTT][ MSE = 0.2143 | LCC = 0.8686 | SRCC = 0.8681 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1079 | LCC = 0.9409 | SRCC = 0.9411 ] +INFO:root:Average loss=0.058606830783761464 +INFO:root: +[Valid][30][UTT][ MSE = 0.2091 | LCC = 0.8709 | SRCC = 0.8711 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1049 | LCC = 0.9415 | SRCC = 0.9435 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2206 | LCC = 0.8645 | SRCC = 0.8635 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1282 | LCC = 0.9192 | SRCC = 0.9178 ] diff --git a/results_bvcc/w2v2_base/layer2/run4/config.gin b/results_bvcc/w2v2_base/layer2/run4/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer2/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 diff --git a/results_bvcc/w2v2_base/layer2/run4/model.pt b/results_bvcc/w2v2_base/layer2/run4/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..6d38d7cba4756495c815ae49a676f5c177944579 --- /dev/null +++ b/results_bvcc/w2v2_base/layer2/run4/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0738fb49d851f769bffa7bfaa45e3f7c74fdc8c2fad61d0a4d2d78baca564579 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer2/run4/model_best.pt b/results_bvcc/w2v2_base/layer2/run4/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..dc6c707f1454edd1d624e92882096cd96f0d8f7f --- /dev/null +++ b/results_bvcc/w2v2_base/layer2/run4/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b688bb54318876871aab5ba587ccfe10e10741c85d1269d689ea9efd84e94a9d +size 2749702 diff --git a/results_bvcc/w2v2_base/layer2/run4/train.log b/results_bvcc/w2v2_base/layer2/run4/train.log new file mode 100644 index 0000000000000000000000000000000000000000..50d7564c2548ce51e9fc241dc69d941273e40281 --- /dev/null +++ b/results_bvcc/w2v2_base/layer2/run4/train.log @@ -0,0 +1,163 @@ +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=0.546703078234807 +INFO:root: +[Valid][0][UTT][ MSE = 0.3588 | LCC = 0.7844 | SRCC = 0.7916 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2672 | LCC = 0.8472 | SRCC = 0.8528 ] +INFO:root:Average loss=0.27466181971323794 +INFO:root: +[Valid][1][UTT][ MSE = 0.2624 | LCC = 0.8375 | SRCC = 0.8399 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1535 | LCC = 0.8992 | SRCC = 0.9027 ] +INFO:root:Average loss=0.2284278623186625 +INFO:root: +[Valid][2][UTT][ MSE = 0.2567 | LCC = 0.8502 | SRCC = 0.8542 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1584 | LCC = 0.9142 | SRCC = 0.9215 ] +INFO:root:Average loss=0.20395248440595773 +INFO:root: +[Valid][3][UTT][ MSE = 0.2272 | LCC = 0.8599 | SRCC = 0.8633 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1299 | LCC = 0.9235 | SRCC = 0.9304 ] +INFO:root:Average loss=0.18677303787225333 +INFO:root: +[Valid][4][UTT][ MSE = 0.2344 | LCC = 0.8626 | SRCC = 0.8628 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1206 | LCC = 0.9261 | SRCC = 0.9246 ] +INFO:root:Average loss=0.1753699792883335 +INFO:root: +[Valid][5][UTT][ MSE = 0.2259 | LCC = 0.8636 | SRCC = 0.8647 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1159 | LCC = 0.9275 | SRCC = 0.9288 ] +INFO:root:Average loss=0.16694142325566366 +INFO:root: +[Valid][6][UTT][ MSE = 0.2452 | LCC = 0.8695 | SRCC = 0.8696 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1228 | LCC = 0.9352 | SRCC = 0.9356 ] +INFO:root:Average loss=0.15656071939529517 +INFO:root: +[Valid][7][UTT][ MSE = 0.2183 | LCC = 0.8695 | SRCC = 0.8707 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1166 | LCC = 0.9307 | SRCC = 0.9325 ] +INFO:root:Average loss=0.1479238289862107 +INFO:root: +[Valid][8][UTT][ MSE = 0.2018 | LCC = 0.8737 | SRCC = 0.8749 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1021 | LCC = 0.9335 | SRCC = 0.9331 ] +INFO:root:Average loss=0.1357195248397497 +INFO:root: +[Valid][9][UTT][ MSE = 0.2023 | LCC = 0.8718 | SRCC = 0.8728 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1068 | LCC = 0.9290 | SRCC = 0.9300 ] +INFO:root:Average loss=0.12724740583545122 +INFO:root: +[Valid][10][UTT][ MSE = 0.2130 | LCC = 0.8738 | SRCC = 0.8751 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1247 | LCC = 0.9293 | SRCC = 0.9297 ] +INFO:root:Average loss=0.1182455844604052 +INFO:root: +[Valid][11][UTT][ MSE = 0.2131 | LCC = 0.8734 | SRCC = 0.8751 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1272 | LCC = 0.9316 | SRCC = 0.9338 ] +INFO:root:Average loss=0.11133041013127719 +INFO:root: +[Valid][12][UTT][ MSE = 0.2046 | LCC = 0.8744 | SRCC = 0.8752 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1176 | LCC = 0.9309 | SRCC = 0.9313 ] +INFO:root:Average loss=0.1087566380126354 +INFO:root: +[Valid][13][UTT][ MSE = 0.2021 | LCC = 0.8737 | SRCC = 0.8736 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1041 | LCC = 0.9321 | SRCC = 0.9325 ] +INFO:root:Average loss=0.09840855422692421 +INFO:root: +[Valid][14][UTT][ MSE = 0.1995 | LCC = 0.8747 | SRCC = 0.8749 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1036 | LCC = 0.9355 | SRCC = 0.9339 ] +INFO:root:Average loss=0.09325690815846126 +INFO:root: +[Valid][15][UTT][ MSE = 0.2025 | LCC = 0.8732 | SRCC = 0.8722 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.0990 | LCC = 0.9353 | SRCC = 0.9337 ] +INFO:root:Average loss=0.0917385072948841 +INFO:root: +[Valid][16][UTT][ MSE = 0.2108 | LCC = 0.8708 | SRCC = 0.8730 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1210 | LCC = 0.9337 | SRCC = 0.9377 ] +INFO:root:Average loss=0.08631781135232021 +INFO:root: +[Valid][17][UTT][ MSE = 0.2089 | LCC = 0.8745 | SRCC = 0.8755 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1246 | LCC = 0.9344 | SRCC = 0.9348 ] +INFO:root:Average loss=0.07788006894481488 +INFO:root: +[Valid][18][UTT][ MSE = 0.2055 | LCC = 0.8723 | SRCC = 0.8735 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1197 | LCC = 0.9306 | SRCC = 0.9318 ] +INFO:root:Average loss=0.07628878841224389 +INFO:root: +[Valid][19][UTT][ MSE = 0.1982 | LCC = 0.8747 | SRCC = 0.8751 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1043 | LCC = 0.9326 | SRCC = 0.9313 ] +INFO:root:Average loss=0.0727895686450677 +INFO:root: +[Valid][20][UTT][ MSE = 0.2037 | LCC = 0.8721 | SRCC = 0.8732 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1109 | LCC = 0.9310 | SRCC = 0.9329 ] +INFO:root:Average loss=0.07348711019716202 +INFO:root: +[Valid][21][UTT][ MSE = 0.2059 | LCC = 0.8751 | SRCC = 0.8758 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.0983 | LCC = 0.9367 | SRCC = 0.9387 ] +INFO:root:Average loss=0.07345395544782662 +INFO:root: +[Valid][22][UTT][ MSE = 0.2100 | LCC = 0.8740 | SRCC = 0.8743 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1350 | LCC = 0.9273 | SRCC = 0.9275 ] +INFO:root:Average loss=0.07317100283809197 +INFO:root: +[Valid][23][UTT][ MSE = 0.2008 | LCC = 0.8747 | SRCC = 0.8756 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1090 | LCC = 0.9336 | SRCC = 0.9368 ] +INFO:root:Average loss=0.06463421093156704 +INFO:root: +[Valid][24][UTT][ MSE = 0.2050 | LCC = 0.8722 | SRCC = 0.8728 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1031 | LCC = 0.9338 | SRCC = 0.9356 ] +INFO:root:Average loss=0.06862837128723279 +INFO:root: +[Valid][25][UTT][ MSE = 0.2059 | LCC = 0.8722 | SRCC = 0.8726 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1041 | LCC = 0.9307 | SRCC = 0.9321 ] +INFO:root:Average loss=0.057527173429918595 +INFO:root: +[Valid][26][UTT][ MSE = 0.1972 | LCC = 0.8741 | SRCC = 0.8746 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.0985 | LCC = 0.9359 | SRCC = 0.9390 ] +INFO:root:Average loss=0.05614383848240742 +INFO:root: +[Valid][27][UTT][ MSE = 0.2055 | LCC = 0.8738 | SRCC = 0.8746 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1148 | LCC = 0.9351 | SRCC = 0.9359 ] +INFO:root:Average loss=0.05448296217200083 +INFO:root: +[Valid][28][UTT][ MSE = 0.2062 | LCC = 0.8759 | SRCC = 0.8782 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1188 | LCC = 0.9361 | SRCC = 0.9403 ] +INFO:root:Average loss=0.05338920180041056 +INFO:root: +[Valid][29][UTT][ MSE = 0.2300 | LCC = 0.8759 | SRCC = 0.8757 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1112 | LCC = 0.9356 | SRCC = 0.9340 ] +INFO:root:Average loss=0.053298711633453004 +INFO:root: +[Valid][30][UTT][ MSE = 0.1984 | LCC = 0.8753 | SRCC = 0.8757 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.0938 | LCC = 0.9384 | SRCC = 0.9399 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2209 | LCC = 0.8641 | SRCC = 0.8606 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1045 | LCC = 0.9294 | SRCC = 0.9249 ] diff --git a/results_bvcc/w2v2_base/layer3/run0/config.gin b/results_bvcc/w2v2_base/layer3/run0/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer3/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 diff --git a/results_bvcc/w2v2_base/layer3/run0/model.pt b/results_bvcc/w2v2_base/layer3/run0/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..5a898bc4e2a85a06b39c868cb1f6d1ef9f209ec4 --- /dev/null +++ b/results_bvcc/w2v2_base/layer3/run0/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fdf33753a6e81026b64b783810bac0171dc3aaf056eda0b468165f4c2d86021c +size 2747979 diff --git a/results_bvcc/w2v2_base/layer3/run0/model_best.pt b/results_bvcc/w2v2_base/layer3/run0/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..88059701e38161a4cd301d4d416ee20688da94b2 --- /dev/null +++ b/results_bvcc/w2v2_base/layer3/run0/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1bee814c8ea8e7b55732a510b1f8ce0774d0c4ba469fa862ffeaef38178c72d4 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer3/run0/train.log b/results_bvcc/w2v2_base/layer3/run0/train.log new file mode 100644 index 0000000000000000000000000000000000000000..01f3b840e899c0ccfb72b091d794e9d190c5034f --- /dev/null +++ b/results_bvcc/w2v2_base/layer3/run0/train.log @@ -0,0 +1,163 @@ +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=0.5388963071581645 +INFO:root: +[Valid][0][UTT][ MSE = 0.3512 | LCC = 0.7939 | SRCC = 0.7945 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2600 | LCC = 0.8442 | SRCC = 0.8438 ] +INFO:root:Average loss=0.27322102185243213 +INFO:root: +[Valid][1][UTT][ MSE = 0.2811 | LCC = 0.8311 | SRCC = 0.8322 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1753 | LCC = 0.8854 | SRCC = 0.8881 ] +INFO:root:Average loss=0.2318616206638324 +INFO:root: +[Valid][2][UTT][ MSE = 0.2779 | LCC = 0.8477 | SRCC = 0.8471 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1908 | LCC = 0.9080 | SRCC = 0.9095 ] +INFO:root:Average loss=0.20415112949334657 +INFO:root: +[Valid][3][UTT][ MSE = 0.2638 | LCC = 0.8523 | SRCC = 0.8499 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1818 | LCC = 0.9087 | SRCC = 0.9041 ] +INFO:root:Average loss=0.18547061362709755 +INFO:root: +[Valid][4][UTT][ MSE = 0.2861 | LCC = 0.8565 | SRCC = 0.8546 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.2072 | LCC = 0.9171 | SRCC = 0.9145 ] +INFO:root:Average loss=0.17083025150574171 +INFO:root: +[Valid][5][UTT][ MSE = 0.2423 | LCC = 0.8596 | SRCC = 0.8577 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1564 | LCC = 0.9205 | SRCC = 0.9173 ] +INFO:root:Average loss=0.16075854204021967 +INFO:root: +[Valid][6][UTT][ MSE = 0.2366 | LCC = 0.8606 | SRCC = 0.8573 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1433 | LCC = 0.9225 | SRCC = 0.9165 ] +INFO:root:Average loss=0.1513323341615689 +INFO:root: +[Valid][7][UTT][ MSE = 0.2528 | LCC = 0.8620 | SRCC = 0.8582 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1722 | LCC = 0.9219 | SRCC = 0.9148 ] +INFO:root:Average loss=0.14375656250959787 +INFO:root: +[Valid][8][UTT][ MSE = 0.2534 | LCC = 0.8642 | SRCC = 0.8617 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1582 | LCC = 0.9275 | SRCC = 0.9214 ] +INFO:root:Average loss=0.13486209139227867 +INFO:root: +[Valid][9][UTT][ MSE = 0.2702 | LCC = 0.8623 | SRCC = 0.8596 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1906 | LCC = 0.9249 | SRCC = 0.9197 ] +INFO:root:Average loss=0.12303440234599969 +INFO:root: +[Valid][10][UTT][ MSE = 0.2664 | LCC = 0.8629 | SRCC = 0.8597 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1781 | LCC = 0.9325 | SRCC = 0.9274 ] +INFO:root:Average loss=0.11504628881812096 +INFO:root: +[Valid][11][UTT][ MSE = 0.2496 | LCC = 0.8590 | SRCC = 0.8558 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1585 | LCC = 0.9256 | SRCC = 0.9208 ] +INFO:root:Average loss=0.11191778477185811 +INFO:root: +[Valid][12][UTT][ MSE = 0.2513 | LCC = 0.8614 | SRCC = 0.8580 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1519 | LCC = 0.9338 | SRCC = 0.9299 ] +INFO:root:Average loss=0.11290468418827423 +INFO:root: +[Valid][13][UTT][ MSE = 0.2294 | LCC = 0.8622 | SRCC = 0.8584 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1273 | LCC = 0.9318 | SRCC = 0.9254 ] +INFO:root:Average loss=0.09664473429513283 +INFO:root: +[Valid][14][UTT][ MSE = 0.2275 | LCC = 0.8593 | SRCC = 0.8565 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1063 | LCC = 0.9294 | SRCC = 0.9243 ] +INFO:root:Average loss=0.09712230867873399 +INFO:root: +[Valid][15][UTT][ MSE = 0.2296 | LCC = 0.8617 | SRCC = 0.8581 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1272 | LCC = 0.9334 | SRCC = 0.9283 ] +INFO:root:Average loss=0.09159498513700107 +INFO:root: +[Valid][16][UTT][ MSE = 0.2242 | LCC = 0.8610 | SRCC = 0.8574 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1112 | LCC = 0.9345 | SRCC = 0.9280 ] +INFO:root:Average loss=0.08877285192601192 +INFO:root: +[Valid][17][UTT][ MSE = 0.2437 | LCC = 0.8588 | SRCC = 0.8555 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1415 | LCC = 0.9320 | SRCC = 0.9288 ] +INFO:root:Average loss=0.08063112065578118 +INFO:root: +[Valid][18][UTT][ MSE = 0.2455 | LCC = 0.8561 | SRCC = 0.8526 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1498 | LCC = 0.9254 | SRCC = 0.9209 ] +INFO:root:Average loss=0.07690747359242195 +INFO:root: +[Valid][19][UTT][ MSE = 0.2671 | LCC = 0.8548 | SRCC = 0.8527 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1799 | LCC = 0.9232 | SRCC = 0.9177 ] +INFO:root:Average loss=0.07784985402264656 +INFO:root: +[Valid][20][UTT][ MSE = 0.2575 | LCC = 0.8543 | SRCC = 0.8528 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1639 | LCC = 0.9265 | SRCC = 0.9245 ] +INFO:root:Average loss=0.0726441254791541 +INFO:root: +[Valid][21][UTT][ MSE = 0.2362 | LCC = 0.8570 | SRCC = 0.8532 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1341 | LCC = 0.9280 | SRCC = 0.9217 ] +INFO:root:Average loss=0.06858268786126222 +INFO:root: +[Valid][22][UTT][ MSE = 0.2390 | LCC = 0.8561 | SRCC = 0.8534 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1356 | LCC = 0.9288 | SRCC = 0.9254 ] +INFO:root:Average loss=0.06611003927313365 +INFO:root: +[Valid][23][UTT][ MSE = 0.2279 | LCC = 0.8582 | SRCC = 0.8548 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1188 | LCC = 0.9321 | SRCC = 0.9256 ] +INFO:root:Average loss=0.06680632244126919 +INFO:root: +[Valid][24][UTT][ MSE = 0.2328 | LCC = 0.8572 | SRCC = 0.8546 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1228 | LCC = 0.9319 | SRCC = 0.9282 ] +INFO:root:Average loss=0.06613343476484983 +INFO:root: +[Valid][25][UTT][ MSE = 0.2402 | LCC = 0.8571 | SRCC = 0.8548 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1372 | LCC = 0.9334 | SRCC = 0.9297 ] +INFO:root:Average loss=0.06038058568269779 +INFO:root: +[Valid][26][UTT][ MSE = 0.2448 | LCC = 0.8563 | SRCC = 0.8534 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1497 | LCC = 0.9279 | SRCC = 0.9240 ] +INFO:root:Average loss=0.06123598622014889 +INFO:root: +[Valid][27][UTT][ MSE = 0.2293 | LCC = 0.8583 | SRCC = 0.8563 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1218 | LCC = 0.9303 | SRCC = 0.9247 ] +INFO:root:Average loss=0.05458702681920467 +INFO:root: +[Valid][28][UTT][ MSE = 0.2293 | LCC = 0.8579 | SRCC = 0.8550 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1114 | LCC = 0.9385 | SRCC = 0.9341 ] +INFO:root:Average loss=0.05240754589724999 +INFO:root: +[Valid][29][UTT][ MSE = 0.2390 | LCC = 0.8566 | SRCC = 0.8549 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1378 | LCC = 0.9309 | SRCC = 0.9286 ] +INFO:root:Average loss=0.05166725773746387 +INFO:root: +[Valid][30][UTT][ MSE = 0.2476 | LCC = 0.8566 | SRCC = 0.8539 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1460 | LCC = 0.9331 | SRCC = 0.9267 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2576 | LCC = 0.8631 | SRCC = 0.8653 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1635 | LCC = 0.9126 | SRCC = 0.9165 ] diff --git a/results_bvcc/w2v2_base/layer3/run1/config.gin b/results_bvcc/w2v2_base/layer3/run1/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer3/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 diff --git a/results_bvcc/w2v2_base/layer3/run1/model.pt b/results_bvcc/w2v2_base/layer3/run1/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..4187dd632cac14cfb9ba85f171e54d008f5b7a80 --- /dev/null +++ b/results_bvcc/w2v2_base/layer3/run1/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a39f217dd1dedc053997a1bdc2d01d17a99666a2387eb9ea6f4eaddf07ef8384 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer3/run1/model_best.pt b/results_bvcc/w2v2_base/layer3/run1/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..1360ead87cfe8374804af7149c46422f28bab3d2 --- /dev/null +++ b/results_bvcc/w2v2_base/layer3/run1/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dd74199e0dba7775ba74536db0aecba9f4dcb34bd45b38ec0af680d8db2e6739 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer3/run1/train.log b/results_bvcc/w2v2_base/layer3/run1/train.log new file mode 100644 index 0000000000000000000000000000000000000000..3c64a5dd5213d19ae62a1b56038bd9f593f28411 --- /dev/null +++ b/results_bvcc/w2v2_base/layer3/run1/train.log @@ -0,0 +1,163 @@ +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=0.6045276709856131 +INFO:root: +[Valid][0][UTT][ MSE = 0.3528 | LCC = 0.7753 | SRCC = 0.7834 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2512 | LCC = 0.8350 | SRCC = 0.8419 ] +INFO:root:Average loss=0.276482614951256 +INFO:root: +[Valid][1][UTT][ MSE = 0.2983 | LCC = 0.8376 | SRCC = 0.8397 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1809 | LCC = 0.8952 | SRCC = 0.8956 ] +INFO:root:Average loss=0.23211595110404185 +INFO:root: +[Valid][2][UTT][ MSE = 0.2653 | LCC = 0.8551 | SRCC = 0.8543 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1613 | LCC = 0.9139 | SRCC = 0.9107 ] +INFO:root:Average loss=0.2051768224590864 +INFO:root: +[Valid][3][UTT][ MSE = 0.2569 | LCC = 0.8638 | SRCC = 0.8628 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1611 | LCC = 0.9238 | SRCC = 0.9218 ] +INFO:root:Average loss=0.18358159428223586 +INFO:root: +[Valid][4][UTT][ MSE = 0.2485 | LCC = 0.8652 | SRCC = 0.8636 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1632 | LCC = 0.9231 | SRCC = 0.9207 ] +INFO:root:Average loss=0.17219183870997185 +INFO:root: +[Valid][5][UTT][ MSE = 0.2531 | LCC = 0.8688 | SRCC = 0.8673 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1419 | LCC = 0.9308 | SRCC = 0.9245 ] +INFO:root:Average loss=0.16229827138475883 +INFO:root: +[Valid][6][UTT][ MSE = 0.2436 | LCC = 0.8715 | SRCC = 0.8705 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1630 | LCC = 0.9330 | SRCC = 0.9304 ] +INFO:root:Average loss=0.14533317518921998 +INFO:root: +[Valid][7][UTT][ MSE = 0.2222 | LCC = 0.8700 | SRCC = 0.8685 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1263 | LCC = 0.9304 | SRCC = 0.9264 ] +INFO:root:Average loss=0.1511033176420591 +INFO:root: +[Valid][8][UTT][ MSE = 0.2502 | LCC = 0.8694 | SRCC = 0.8680 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1442 | LCC = 0.9287 | SRCC = 0.9218 ] +INFO:root:Average loss=0.13616069969840539 +INFO:root: +[Valid][9][UTT][ MSE = 0.2253 | LCC = 0.8718 | SRCC = 0.8700 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1267 | LCC = 0.9365 | SRCC = 0.9305 ] +INFO:root:Average loss=0.12982463636077368 +INFO:root: +[Valid][10][UTT][ MSE = 0.2278 | LCC = 0.8714 | SRCC = 0.8701 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1230 | LCC = 0.9399 | SRCC = 0.9347 ] +INFO:root:Average loss=0.12225270586518142 +INFO:root: +[Valid][11][UTT][ MSE = 0.2335 | LCC = 0.8698 | SRCC = 0.8692 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1279 | LCC = 0.9349 | SRCC = 0.9288 ] +INFO:root:Average loss=0.1111568048214301 +INFO:root: +[Valid][12][UTT][ MSE = 0.2213 | LCC = 0.8720 | SRCC = 0.8700 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1300 | LCC = 0.9358 | SRCC = 0.9302 ] +INFO:root:Average loss=0.09923397931150901 +INFO:root: +[Valid][13][UTT][ MSE = 0.2504 | LCC = 0.8699 | SRCC = 0.8688 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1655 | LCC = 0.9372 | SRCC = 0.9333 ] +INFO:root:Average loss=0.098825214860531 +INFO:root: +[Valid][14][UTT][ MSE = 0.2265 | LCC = 0.8690 | SRCC = 0.8683 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1179 | LCC = 0.9354 | SRCC = 0.9310 ] +INFO:root:Average loss=0.09269962283090138 +INFO:root: +[Valid][15][UTT][ MSE = 0.2205 | LCC = 0.8721 | SRCC = 0.8710 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1237 | LCC = 0.9359 | SRCC = 0.9305 ] +INFO:root:Average loss=0.08557081676255433 +INFO:root: +[Valid][16][UTT][ MSE = 0.2120 | LCC = 0.8721 | SRCC = 0.8708 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1124 | LCC = 0.9358 | SRCC = 0.9310 ] +INFO:root:Average loss=0.08366892195473878 +INFO:root: +[Valid][17][UTT][ MSE = 0.2337 | LCC = 0.8683 | SRCC = 0.8665 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1298 | LCC = 0.9318 | SRCC = 0.9260 ] +INFO:root:Average loss=0.07955511253422652 +INFO:root: +[Valid][18][UTT][ MSE = 0.2187 | LCC = 0.8687 | SRCC = 0.8674 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1149 | LCC = 0.9334 | SRCC = 0.9271 ] +INFO:root:Average loss=0.07645396310358475 +INFO:root: +[Valid][19][UTT][ MSE = 0.2162 | LCC = 0.8672 | SRCC = 0.8660 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1113 | LCC = 0.9364 | SRCC = 0.9307 ] +INFO:root:Average loss=0.07043788954615593 +INFO:root: +[Valid][20][UTT][ MSE = 0.2195 | LCC = 0.8682 | SRCC = 0.8681 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1075 | LCC = 0.9383 | SRCC = 0.9344 ] +INFO:root:Average loss=0.06872585272559753 +INFO:root: +[Valid][21][UTT][ MSE = 0.2158 | LCC = 0.8680 | SRCC = 0.8669 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1111 | LCC = 0.9388 | SRCC = 0.9353 ] +INFO:root:Average loss=0.0641730479322947 +INFO:root: +[Valid][22][UTT][ MSE = 0.2187 | LCC = 0.8652 | SRCC = 0.8644 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1028 | LCC = 0.9389 | SRCC = 0.9364 ] +INFO:root:Average loss=0.06244218024687889 +INFO:root: +[Valid][23][UTT][ MSE = 0.2303 | LCC = 0.8649 | SRCC = 0.8637 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1155 | LCC = 0.9361 | SRCC = 0.9292 ] +INFO:root:Average loss=0.06570703416871719 +INFO:root: +[Valid][24][UTT][ MSE = 0.2219 | LCC = 0.8659 | SRCC = 0.8637 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1218 | LCC = 0.9327 | SRCC = 0.9259 ] +INFO:root:Average loss=0.05749523018797239 +INFO:root: +[Valid][25][UTT][ MSE = 0.2209 | LCC = 0.8642 | SRCC = 0.8629 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1084 | LCC = 0.9337 | SRCC = 0.9275 ] +INFO:root:Average loss=0.054782983870842517 +INFO:root: +[Valid][26][UTT][ MSE = 0.2314 | LCC = 0.8639 | SRCC = 0.8620 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1302 | LCC = 0.9340 | SRCC = 0.9260 ] +INFO:root:Average loss=0.05425381350020567 +INFO:root: +[Valid][27][UTT][ MSE = 0.2338 | LCC = 0.8635 | SRCC = 0.8615 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1350 | LCC = 0.9358 | SRCC = 0.9312 ] +INFO:root:Average loss=0.05194462983845136 +INFO:root: +[Valid][28][UTT][ MSE = 0.2383 | LCC = 0.8637 | SRCC = 0.8603 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1332 | LCC = 0.9367 | SRCC = 0.9305 ] +INFO:root:Average loss=0.049382959277584 +INFO:root: +[Valid][29][UTT][ MSE = 0.2317 | LCC = 0.8654 | SRCC = 0.8633 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1310 | LCC = 0.9387 | SRCC = 0.9334 ] +INFO:root:Average loss=0.046610791116761856 +INFO:root: +[Valid][30][UTT][ MSE = 0.2229 | LCC = 0.8633 | SRCC = 0.8613 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1128 | LCC = 0.9322 | SRCC = 0.9270 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2338 | LCC = 0.8633 | SRCC = 0.8631 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1369 | LCC = 0.9161 | SRCC = 0.9144 ] diff --git a/results_bvcc/w2v2_base/layer3/run2/config.gin b/results_bvcc/w2v2_base/layer3/run2/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer3/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 diff --git a/results_bvcc/w2v2_base/layer3/run2/model.pt b/results_bvcc/w2v2_base/layer3/run2/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..b9fe81b8945c5dfee931a45189d6ba1e9db39ad2 --- /dev/null +++ b/results_bvcc/w2v2_base/layer3/run2/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e1b75942f542b02fb8dc639bcd569b1a99484bfbfb6fd38cc340956fd97d089d +size 2747979 diff --git a/results_bvcc/w2v2_base/layer3/run2/model_best.pt b/results_bvcc/w2v2_base/layer3/run2/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..2523e68d738004730a96b98e1a4ab7e5fe492425 --- /dev/null +++ b/results_bvcc/w2v2_base/layer3/run2/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1c78e22a00126378d4d417793e6b04df4f64ab0a124727486ccddc694db03314 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer3/run2/train.log b/results_bvcc/w2v2_base/layer3/run2/train.log new file mode 100644 index 0000000000000000000000000000000000000000..5f4119cb201cba1185bbaf8e4e5d6a0f9177b6da --- /dev/null +++ b/results_bvcc/w2v2_base/layer3/run2/train.log @@ -0,0 +1,163 @@ +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=0.618041337109529 +INFO:root: +[Valid][0][UTT][ MSE = 0.3437 | LCC = 0.7714 | SRCC = 0.7771 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2400 | LCC = 0.8256 | SRCC = 0.8296 ] +INFO:root:Average loss=0.2772234756595049 +INFO:root: +[Valid][1][UTT][ MSE = 0.2928 | LCC = 0.8326 | SRCC = 0.8343 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2050 | LCC = 0.8881 | SRCC = 0.8924 ] +INFO:root:Average loss=0.22710150289229858 +INFO:root: +[Valid][2][UTT][ MSE = 0.2608 | LCC = 0.8505 | SRCC = 0.8490 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1733 | LCC = 0.9061 | SRCC = 0.9044 ] +INFO:root:Average loss=0.2087478186839666 +INFO:root: +[Valid][3][UTT][ MSE = 0.2380 | LCC = 0.8567 | SRCC = 0.8561 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1516 | LCC = 0.9108 | SRCC = 0.9095 ] +INFO:root:Average loss=0.1865193887780874 +INFO:root: +[Valid][4][UTT][ MSE = 0.2831 | LCC = 0.8544 | SRCC = 0.8559 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.2103 | LCC = 0.9157 | SRCC = 0.9219 ] +INFO:root:Average loss=0.17166964795726997 +INFO:root: +[Valid][5][UTT][ MSE = 0.2424 | LCC = 0.8626 | SRCC = 0.8591 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1621 | LCC = 0.9188 | SRCC = 0.9157 ] +INFO:root:Average loss=0.1617496408140048 +INFO:root: +[Valid][6][UTT][ MSE = 0.2678 | LCC = 0.8641 | SRCC = 0.8640 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1971 | LCC = 0.9220 | SRCC = 0.9257 ] +INFO:root:Average loss=0.1450485422825202 +INFO:root: +[Valid][7][UTT][ MSE = 0.2475 | LCC = 0.8645 | SRCC = 0.8629 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1803 | LCC = 0.9193 | SRCC = 0.9202 ] +INFO:root:Average loss=0.1357928833518273 +INFO:root: +[Valid][8][UTT][ MSE = 0.2334 | LCC = 0.8654 | SRCC = 0.8633 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1588 | LCC = 0.9215 | SRCC = 0.9209 ] +INFO:root:Average loss=0.12330357300547454 +INFO:root: +[Valid][9][UTT][ MSE = 0.2168 | LCC = 0.8692 | SRCC = 0.8656 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1258 | LCC = 0.9305 | SRCC = 0.9286 ] +INFO:root:Average loss=0.11874586467941602 +INFO:root: +[Valid][10][UTT][ MSE = 0.2850 | LCC = 0.8641 | SRCC = 0.8622 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.2185 | LCC = 0.9264 | SRCC = 0.9285 ] +INFO:root:Average loss=0.1149993587571841 +INFO:root: +[Valid][11][UTT][ MSE = 0.2829 | LCC = 0.8640 | SRCC = 0.8623 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.2235 | LCC = 0.9246 | SRCC = 0.9269 ] +INFO:root:Average loss=0.10477700199072178 +INFO:root: +[Valid][12][UTT][ MSE = 0.2421 | LCC = 0.8628 | SRCC = 0.8616 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1622 | LCC = 0.9286 | SRCC = 0.9296 ] +INFO:root:Average loss=0.10028586307397255 +INFO:root: +[Valid][13][UTT][ MSE = 0.2405 | LCC = 0.8625 | SRCC = 0.8603 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1517 | LCC = 0.9335 | SRCC = 0.9345 ] +INFO:root:Average loss=0.08870371980353808 +INFO:root: +[Valid][14][UTT][ MSE = 0.2286 | LCC = 0.8608 | SRCC = 0.8580 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1326 | LCC = 0.9324 | SRCC = 0.9295 ] +INFO:root:Average loss=0.09085492293039958 +INFO:root: +[Valid][15][UTT][ MSE = 0.2530 | LCC = 0.8599 | SRCC = 0.8576 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1677 | LCC = 0.9289 | SRCC = 0.9268 ] +INFO:root:Average loss=0.0872667694512086 +INFO:root: +[Valid][16][UTT][ MSE = 0.2214 | LCC = 0.8600 | SRCC = 0.8571 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1130 | LCC = 0.9293 | SRCC = 0.9240 ] +INFO:root:Average loss=0.0894361056196384 +INFO:root: +[Valid][17][UTT][ MSE = 0.2320 | LCC = 0.8593 | SRCC = 0.8550 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1214 | LCC = 0.9305 | SRCC = 0.9272 ] +INFO:root:Average loss=0.07583379124601682 +INFO:root: +[Valid][18][UTT][ MSE = 0.2417 | LCC = 0.8584 | SRCC = 0.8570 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1514 | LCC = 0.9303 | SRCC = 0.9319 ] +INFO:root:Average loss=0.07111639109177467 +INFO:root: +[Valid][19][UTT][ MSE = 0.2731 | LCC = 0.8588 | SRCC = 0.8573 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.2008 | LCC = 0.9275 | SRCC = 0.9285 ] +INFO:root:Average loss=0.07006627545715907 +INFO:root: +[Valid][20][UTT][ MSE = 0.2273 | LCC = 0.8601 | SRCC = 0.8573 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1297 | LCC = 0.9308 | SRCC = 0.9294 ] +INFO:root:Average loss=0.06404362069681668 +INFO:root: +[Valid][21][UTT][ MSE = 0.2183 | LCC = 0.8609 | SRCC = 0.8578 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1197 | LCC = 0.9261 | SRCC = 0.9246 ] +INFO:root:Average loss=0.06599630224399078 +INFO:root: +[Valid][22][UTT][ MSE = 0.2211 | LCC = 0.8622 | SRCC = 0.8593 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1181 | LCC = 0.9331 | SRCC = 0.9294 ] +INFO:root:Average loss=0.06542258260723872 +INFO:root: +[Valid][23][UTT][ MSE = 0.2451 | LCC = 0.8611 | SRCC = 0.8586 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1695 | LCC = 0.9252 | SRCC = 0.9249 ] +INFO:root:Average loss=0.0584898883333573 +INFO:root: +[Valid][24][UTT][ MSE = 0.2561 | LCC = 0.8568 | SRCC = 0.8553 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1757 | LCC = 0.9245 | SRCC = 0.9251 ] +INFO:root:Average loss=0.056313919404951424 +INFO:root: +[Valid][25][UTT][ MSE = 0.2289 | LCC = 0.8596 | SRCC = 0.8569 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1350 | LCC = 0.9270 | SRCC = 0.9238 ] +INFO:root:Average loss=0.057840629743459895 +INFO:root: +[Valid][26][UTT][ MSE = 0.2219 | LCC = 0.8587 | SRCC = 0.8558 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1207 | LCC = 0.9274 | SRCC = 0.9231 ] +INFO:root:Average loss=0.0556940395767108 +INFO:root: +[Valid][27][UTT][ MSE = 0.2501 | LCC = 0.8580 | SRCC = 0.8562 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1715 | LCC = 0.9229 | SRCC = 0.9221 ] +INFO:root:Average loss=0.0545005032267326 +INFO:root: +[Valid][28][UTT][ MSE = 0.2296 | LCC = 0.8601 | SRCC = 0.8567 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1405 | LCC = 0.9251 | SRCC = 0.9213 ] +INFO:root:Average loss=0.048206741945483744 +INFO:root: +[Valid][29][UTT][ MSE = 0.2366 | LCC = 0.8605 | SRCC = 0.8582 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1535 | LCC = 0.9264 | SRCC = 0.9227 ] +INFO:root:Average loss=0.04664461219157928 +INFO:root: +[Valid][30][UTT][ MSE = 0.2589 | LCC = 0.8573 | SRCC = 0.8562 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1841 | LCC = 0.9222 | SRCC = 0.9247 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2174 | LCC = 0.8691 | SRCC = 0.8677 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1332 | LCC = 0.9097 | SRCC = 0.9066 ] diff --git a/results_bvcc/w2v2_base/layer3/run3/config.gin b/results_bvcc/w2v2_base/layer3/run3/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer3/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 diff --git a/results_bvcc/w2v2_base/layer3/run3/model.pt b/results_bvcc/w2v2_base/layer3/run3/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..bfe4b357189e4ed990c4647992c85bb81a781209 --- /dev/null +++ b/results_bvcc/w2v2_base/layer3/run3/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c91dde9323ca8e5a3f6899fb2bbecddef878e9904798a7cda6cf67d924c2c68 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer3/run3/model_best.pt b/results_bvcc/w2v2_base/layer3/run3/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..ab5979f55d93f4dff2f8e3272ddfb2443f3ad005 --- /dev/null +++ b/results_bvcc/w2v2_base/layer3/run3/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1f548a2ee9de747087936de24cd72f6a43bd78c3de57097e60d2799c9229af5b +size 2749702 diff --git a/results_bvcc/w2v2_base/layer3/run3/train.log b/results_bvcc/w2v2_base/layer3/run3/train.log new file mode 100644 index 0000000000000000000000000000000000000000..f44470a4bd6b559c61c72f2e376a3cc5b5869a25 --- /dev/null +++ b/results_bvcc/w2v2_base/layer3/run3/train.log @@ -0,0 +1,163 @@ +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=0.5180282392180883 +INFO:root: +[Valid][0][UTT][ MSE = 0.3457 | LCC = 0.8063 | SRCC = 0.8103 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2580 | LCC = 0.8510 | SRCC = 0.8568 ] +INFO:root:Average loss=0.2746687919283525 +INFO:root: +[Valid][1][UTT][ MSE = 0.3072 | LCC = 0.8393 | SRCC = 0.8444 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2292 | LCC = 0.8944 | SRCC = 0.9021 ] +INFO:root:Average loss=0.22535095306543204 +INFO:root: +[Valid][2][UTT][ MSE = 0.2846 | LCC = 0.8533 | SRCC = 0.8537 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.2128 | LCC = 0.9096 | SRCC = 0.9118 ] +INFO:root:Average loss=0.19969538713877016 +INFO:root: +[Valid][3][UTT][ MSE = 0.2755 | LCC = 0.8591 | SRCC = 0.8611 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1914 | LCC = 0.9251 | SRCC = 0.9300 ] +INFO:root:Average loss=0.18626682383891863 +INFO:root: +[Valid][4][UTT][ MSE = 0.2442 | LCC = 0.8708 | SRCC = 0.8691 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1613 | LCC = 0.9328 | SRCC = 0.9311 ] +INFO:root:Average loss=0.1728838854111158 +INFO:root: +[Valid][5][UTT][ MSE = 0.2233 | LCC = 0.8677 | SRCC = 0.8661 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1286 | LCC = 0.9306 | SRCC = 0.9284 ] +INFO:root:Average loss=0.15506736141366836 +INFO:root: +[Valid][6][UTT][ MSE = 0.2231 | LCC = 0.8670 | SRCC = 0.8652 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1305 | LCC = 0.9319 | SRCC = 0.9271 ] +INFO:root:Average loss=0.14330583008436057 +INFO:root: +[Valid][7][UTT][ MSE = 0.2280 | LCC = 0.8716 | SRCC = 0.8691 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1338 | LCC = 0.9408 | SRCC = 0.9357 ] +INFO:root:Average loss=0.13510834196439156 +INFO:root: +[Valid][8][UTT][ MSE = 0.2143 | LCC = 0.8703 | SRCC = 0.8687 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1066 | LCC = 0.9417 | SRCC = 0.9355 ] +INFO:root:Average loss=0.12729691274655172 +INFO:root: +[Valid][9][UTT][ MSE = 0.2293 | LCC = 0.8690 | SRCC = 0.8674 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1369 | LCC = 0.9373 | SRCC = 0.9339 ] +INFO:root:Average loss=0.12344187708237232 +INFO:root: +[Valid][10][UTT][ MSE = 0.2122 | LCC = 0.8686 | SRCC = 0.8669 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1049 | LCC = 0.9415 | SRCC = 0.9373 ] +INFO:root:Average loss=0.11213754470913838 +INFO:root: +[Valid][11][UTT][ MSE = 0.2174 | LCC = 0.8684 | SRCC = 0.8674 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1257 | LCC = 0.9343 | SRCC = 0.9304 ] +INFO:root:Average loss=0.10677822230335994 +INFO:root: +[Valid][12][UTT][ MSE = 0.2204 | LCC = 0.8686 | SRCC = 0.8684 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1258 | LCC = 0.9390 | SRCC = 0.9357 ] +INFO:root:Average loss=0.0998713538910334 +INFO:root: +[Valid][13][UTT][ MSE = 0.2194 | LCC = 0.8674 | SRCC = 0.8651 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1176 | LCC = 0.9374 | SRCC = 0.9329 ] +INFO:root:Average loss=0.09522502835935508 +INFO:root: +[Valid][14][UTT][ MSE = 0.2305 | LCC = 0.8661 | SRCC = 0.8670 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1339 | LCC = 0.9403 | SRCC = 0.9411 ] +INFO:root:Average loss=0.09259714778417195 +INFO:root: +[Valid][15][UTT][ MSE = 0.2171 | LCC = 0.8684 | SRCC = 0.8666 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1072 | LCC = 0.9413 | SRCC = 0.9373 ] +INFO:root:Average loss=0.08832140801808773 +INFO:root: +[Valid][16][UTT][ MSE = 0.2125 | LCC = 0.8658 | SRCC = 0.8657 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.0971 | LCC = 0.9418 | SRCC = 0.9380 ] +INFO:root:Average loss=0.08183478320447299 +INFO:root: +[Valid][17][UTT][ MSE = 0.2093 | LCC = 0.8662 | SRCC = 0.8665 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.0991 | LCC = 0.9384 | SRCC = 0.9360 ] +INFO:root:Average loss=0.08206631973958932 +INFO:root: +[Valid][18][UTT][ MSE = 0.2076 | LCC = 0.8677 | SRCC = 0.8680 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.0931 | LCC = 0.9434 | SRCC = 0.9413 ] +INFO:root:Average loss=0.07576178214870967 +INFO:root: +[Valid][19][UTT][ MSE = 0.2179 | LCC = 0.8671 | SRCC = 0.8665 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1018 | LCC = 0.9477 | SRCC = 0.9458 ] +INFO:root:Average loss=0.07124239153777942 +INFO:root: +[Valid][20][UTT][ MSE = 0.2340 | LCC = 0.8652 | SRCC = 0.8649 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1318 | LCC = 0.9423 | SRCC = 0.9406 ] +INFO:root:Average loss=0.06956854190390843 +INFO:root: +[Valid][21][UTT][ MSE = 0.2196 | LCC = 0.8642 | SRCC = 0.8630 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1130 | LCC = 0.9397 | SRCC = 0.9369 ] +INFO:root:Average loss=0.06603325253877884 +INFO:root: +[Valid][22][UTT][ MSE = 0.2188 | LCC = 0.8647 | SRCC = 0.8642 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1052 | LCC = 0.9436 | SRCC = 0.9413 ] +INFO:root:Average loss=0.05982308340473817 +INFO:root: +[Valid][23][UTT][ MSE = 0.2171 | LCC = 0.8613 | SRCC = 0.8611 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.0919 | LCC = 0.9409 | SRCC = 0.9387 ] +INFO:root:Average loss=0.06440985943071353 +INFO:root: +[Valid][24][UTT][ MSE = 0.2329 | LCC = 0.8622 | SRCC = 0.8614 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1214 | LCC = 0.9458 | SRCC = 0.9457 ] +INFO:root:Average loss=0.058364116706145115 +INFO:root: +[Valid][25][UTT][ MSE = 0.2226 | LCC = 0.8635 | SRCC = 0.8618 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1101 | LCC = 0.9448 | SRCC = 0.9423 ] +INFO:root:Average loss=0.05689558596947254 +INFO:root: +[Valid][26][UTT][ MSE = 0.2267 | LCC = 0.8585 | SRCC = 0.8584 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1100 | LCC = 0.9412 | SRCC = 0.9386 ] +INFO:root:Average loss=0.05729831917545734 +INFO:root: +[Valid][27][UTT][ MSE = 0.2450 | LCC = 0.8591 | SRCC = 0.8597 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1480 | LCC = 0.9374 | SRCC = 0.9377 ] +INFO:root:Average loss=0.05534830479285656 +INFO:root: +[Valid][28][UTT][ MSE = 0.2206 | LCC = 0.8614 | SRCC = 0.8612 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1068 | LCC = 0.9403 | SRCC = 0.9371 ] +INFO:root:Average loss=0.050912948564077035 +INFO:root: +[Valid][29][UTT][ MSE = 0.2216 | LCC = 0.8599 | SRCC = 0.8604 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.0974 | LCC = 0.9462 | SRCC = 0.9460 ] +INFO:root:Average loss=0.04927240398067694 +INFO:root: +[Valid][30][UTT][ MSE = 0.2297 | LCC = 0.8634 | SRCC = 0.8635 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1263 | LCC = 0.9415 | SRCC = 0.9400 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2242 | LCC = 0.8702 | SRCC = 0.8680 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1269 | LCC = 0.9155 | SRCC = 0.9142 ] diff --git a/results_bvcc/w2v2_base/layer3/run4/config.gin b/results_bvcc/w2v2_base/layer3/run4/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer3/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 diff --git a/results_bvcc/w2v2_base/layer3/run4/model.pt b/results_bvcc/w2v2_base/layer3/run4/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..f3bfb9c0df1eb8ee9e5608a5c46d2b2f05b2bd5b --- /dev/null +++ b/results_bvcc/w2v2_base/layer3/run4/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4a50ac265e9bef6d83e6ba452030c574c8080de1e7cdb51e618c154b981a7e0b +size 2747979 diff --git a/results_bvcc/w2v2_base/layer3/run4/model_best.pt b/results_bvcc/w2v2_base/layer3/run4/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..94bd13e1055d35a1b2895beed824530f9326a409 --- /dev/null +++ b/results_bvcc/w2v2_base/layer3/run4/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1d3ce52b63da32180eb3a59106930d5183dd6d9a445a22bda7d62cea846def31 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer3/run4/train.log b/results_bvcc/w2v2_base/layer3/run4/train.log new file mode 100644 index 0000000000000000000000000000000000000000..bb2248a48ea72b050d2781ddf30b689693361a1d --- /dev/null +++ b/results_bvcc/w2v2_base/layer3/run4/train.log @@ -0,0 +1,163 @@ +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=0.6844065410968585 +INFO:root: +[Valid][0][UTT][ MSE = 0.3566 | LCC = 0.7676 | SRCC = 0.7714 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2404 | LCC = 0.8304 | SRCC = 0.8402 ] +INFO:root:Average loss=0.286894916341855 +INFO:root: +[Valid][1][UTT][ MSE = 0.2938 | LCC = 0.8334 | SRCC = 0.8340 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2155 | LCC = 0.8865 | SRCC = 0.8920 ] +INFO:root:Average loss=0.23104553478650558 +INFO:root: +[Valid][2][UTT][ MSE = 0.2693 | LCC = 0.8531 | SRCC = 0.8510 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1777 | LCC = 0.9123 | SRCC = 0.9111 ] +INFO:root:Average loss=0.2135846028343225 +INFO:root: +[Valid][3][UTT][ MSE = 0.2386 | LCC = 0.8586 | SRCC = 0.8570 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1430 | LCC = 0.9177 | SRCC = 0.9138 ] +INFO:root:Average loss=0.18364031125719732 +INFO:root: +[Valid][4][UTT][ MSE = 0.2330 | LCC = 0.8647 | SRCC = 0.8629 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1338 | LCC = 0.9295 | SRCC = 0.9234 ] +INFO:root:Average loss=0.17280855029821396 +INFO:root: +[Valid][5][UTT][ MSE = 0.2456 | LCC = 0.8674 | SRCC = 0.8658 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1606 | LCC = 0.9320 | SRCC = 0.9289 ] +INFO:root:Average loss=0.16010016842912406 +INFO:root: +[Valid][6][UTT][ MSE = 0.2302 | LCC = 0.8685 | SRCC = 0.8670 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1238 | LCC = 0.9397 | SRCC = 0.9335 ] +INFO:root:Average loss=0.1615201665613896 +INFO:root: +[Valid][7][UTT][ MSE = 0.2545 | LCC = 0.8669 | SRCC = 0.8644 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1608 | LCC = 0.9364 | SRCC = 0.9303 ] +INFO:root:Average loss=0.13865628360937804 +INFO:root: +[Valid][8][UTT][ MSE = 0.2326 | LCC = 0.8726 | SRCC = 0.8711 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1471 | LCC = 0.9385 | SRCC = 0.9339 ] +INFO:root:Average loss=0.12536052795938957 +INFO:root: +[Valid][9][UTT][ MSE = 0.2182 | LCC = 0.8707 | SRCC = 0.8689 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1266 | LCC = 0.9341 | SRCC = 0.9272 ] +INFO:root:Average loss=0.12356129002112609 +INFO:root: +[Valid][10][UTT][ MSE = 0.2251 | LCC = 0.8698 | SRCC = 0.8679 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1254 | LCC = 0.9384 | SRCC = 0.9329 ] +INFO:root:Average loss=0.11592093587685855 +INFO:root: +[Valid][11][UTT][ MSE = 0.2464 | LCC = 0.8687 | SRCC = 0.8650 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1586 | LCC = 0.9358 | SRCC = 0.9288 ] +INFO:root:Average loss=0.1086466150979201 +INFO:root: +[Valid][12][UTT][ MSE = 0.2341 | LCC = 0.8669 | SRCC = 0.8661 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1194 | LCC = 0.9375 | SRCC = 0.9307 ] +INFO:root:Average loss=0.10961012513591693 +INFO:root: +[Valid][13][UTT][ MSE = 0.2418 | LCC = 0.8702 | SRCC = 0.8677 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1554 | LCC = 0.9374 | SRCC = 0.9317 ] +INFO:root:Average loss=0.09485020942221849 +INFO:root: +[Valid][14][UTT][ MSE = 0.2220 | LCC = 0.8672 | SRCC = 0.8659 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1103 | LCC = 0.9357 | SRCC = 0.9277 ] +INFO:root:Average loss=0.09101185985864738 +INFO:root: +[Valid][15][UTT][ MSE = 0.2195 | LCC = 0.8676 | SRCC = 0.8655 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1238 | LCC = 0.9326 | SRCC = 0.9256 ] +INFO:root:Average loss=0.08243084746675614 +INFO:root: +[Valid][16][UTT][ MSE = 0.2277 | LCC = 0.8680 | SRCC = 0.8660 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1294 | LCC = 0.9386 | SRCC = 0.9335 ] +INFO:root:Average loss=0.0799717598905166 +INFO:root: +[Valid][17][UTT][ MSE = 0.2227 | LCC = 0.8665 | SRCC = 0.8647 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1274 | LCC = 0.9341 | SRCC = 0.9270 ] +INFO:root:Average loss=0.07730969509635216 +INFO:root: +[Valid][18][UTT][ MSE = 0.2269 | LCC = 0.8686 | SRCC = 0.8660 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1301 | LCC = 0.9378 | SRCC = 0.9313 ] +INFO:root:Average loss=0.07513376339696921 +INFO:root: +[Valid][19][UTT][ MSE = 0.2499 | LCC = 0.8664 | SRCC = 0.8654 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1611 | LCC = 0.9362 | SRCC = 0.9303 ] +INFO:root:Average loss=0.06604956873716453 +INFO:root: +[Valid][20][UTT][ MSE = 0.2237 | LCC = 0.8674 | SRCC = 0.8656 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1230 | LCC = 0.9414 | SRCC = 0.9346 ] +INFO:root:Average loss=0.06491754146722648 +INFO:root: +[Valid][21][UTT][ MSE = 0.2123 | LCC = 0.8667 | SRCC = 0.8657 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.0967 | LCC = 0.9440 | SRCC = 0.9380 ] +INFO:root:Average loss=0.06417986741050696 +INFO:root: +[Valid][22][UTT][ MSE = 0.2329 | LCC = 0.8662 | SRCC = 0.8643 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1355 | LCC = 0.9393 | SRCC = 0.9333 ] +INFO:root:Average loss=0.06302707222027656 +INFO:root: +[Valid][23][UTT][ MSE = 0.2315 | LCC = 0.8681 | SRCC = 0.8656 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1305 | LCC = 0.9423 | SRCC = 0.9337 ] +INFO:root:Average loss=0.059602705761790276 +INFO:root: +[Valid][24][UTT][ MSE = 0.2239 | LCC = 0.8651 | SRCC = 0.8633 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1056 | LCC = 0.9387 | SRCC = 0.9289 ] +INFO:root:Average loss=0.06030423046113589 +INFO:root: +[Valid][25][UTT][ MSE = 0.2355 | LCC = 0.8674 | SRCC = 0.8644 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1456 | LCC = 0.9367 | SRCC = 0.9284 ] +INFO:root:Average loss=0.053316498843905255 +INFO:root: +[Valid][26][UTT][ MSE = 0.2394 | LCC = 0.8653 | SRCC = 0.8627 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1421 | LCC = 0.9385 | SRCC = 0.9306 ] +INFO:root:Average loss=0.0535005309069768 +INFO:root: +[Valid][27][UTT][ MSE = 0.2440 | LCC = 0.8670 | SRCC = 0.8635 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1501 | LCC = 0.9380 | SRCC = 0.9267 ] +INFO:root:Average loss=0.05456977363866873 +INFO:root: +[Valid][28][UTT][ MSE = 0.2624 | LCC = 0.8672 | SRCC = 0.8644 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1744 | LCC = 0.9382 | SRCC = 0.9279 ] +INFO:root:Average loss=0.054004807789356284 +INFO:root: +[Valid][29][UTT][ MSE = 0.2196 | LCC = 0.8670 | SRCC = 0.8637 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1167 | LCC = 0.9375 | SRCC = 0.9270 ] +INFO:root:Average loss=0.04918697146842113 +INFO:root: +[Valid][30][UTT][ MSE = 0.2193 | LCC = 0.8692 | SRCC = 0.8654 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1164 | LCC = 0.9419 | SRCC = 0.9302 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2255 | LCC = 0.8668 | SRCC = 0.8655 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1173 | LCC = 0.9179 | SRCC = 0.9143 ] diff --git a/results_bvcc/w2v2_base/layer4/run0/config.gin b/results_bvcc/w2v2_base/layer4/run0/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/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 diff --git a/results_bvcc/w2v2_base/layer4/run0/model.pt b/results_bvcc/w2v2_base/layer4/run0/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..6563c13247d72fd6235bfb062a416c5ebc252824 --- /dev/null +++ b/results_bvcc/w2v2_base/layer4/run0/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9e16d8d1fcbaa1ec4fcd806f1a66badc96c523b255f76faf65d0092f8f6e70d6 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer4/run0/model_best.pt b/results_bvcc/w2v2_base/layer4/run0/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..cca7b2ed32226abda13de5a0b1f67a5395e2295e --- /dev/null +++ b/results_bvcc/w2v2_base/layer4/run0/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:384c4afe4f14828b60db6447df70766acbfb715953fd6058662bae57c0de2820 +size 2749638 diff --git a/results_bvcc/w2v2_base/layer4/run0/train.log b/results_bvcc/w2v2_base/layer4/run0/train.log new file mode 100644 index 0000000000000000000000000000000000000000..7c167d3a771119e4d0d6bd828dc655709a6dfab5 --- /dev/null +++ b/results_bvcc/w2v2_base/layer4/run0/train.log @@ -0,0 +1,163 @@ +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=0.6518440366937563 +INFO:root: +[Valid][0][UTT][ MSE = 0.3691 | LCC = 0.7673 | SRCC = 0.7700 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2678 | LCC = 0.8392 | SRCC = 0.8478 ] +INFO:root:Average loss=0.2980204480580794 +INFO:root: +[Valid][1][UTT][ MSE = 0.3228 | LCC = 0.8205 | SRCC = 0.8216 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2290 | LCC = 0.8915 | SRCC = 0.8931 ] +INFO:root:Average loss=0.23852598533416405 +INFO:root: +[Valid][2][UTT][ MSE = 0.2998 | LCC = 0.8344 | SRCC = 0.8336 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.2089 | LCC = 0.9050 | SRCC = 0.9043 ] +INFO:root:Average loss=0.2203069405677991 +INFO:root: +[Valid][3][UTT][ MSE = 0.2549 | LCC = 0.8477 | SRCC = 0.8454 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1512 | LCC = 0.9134 | SRCC = 0.9111 ] +INFO:root:Average loss=0.19342320144940645 +INFO:root: +[Valid][4][UTT][ MSE = 0.2427 | LCC = 0.8533 | SRCC = 0.8517 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1298 | LCC = 0.9254 | SRCC = 0.9226 ] +INFO:root:Average loss=0.17977268735949808 +INFO:root: +[Valid][5][UTT][ MSE = 0.2396 | LCC = 0.8559 | SRCC = 0.8529 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1316 | LCC = 0.9277 | SRCC = 0.9231 ] +INFO:root:Average loss=0.16202573420909736 +INFO:root: +[Valid][6][UTT][ MSE = 0.3117 | LCC = 0.8517 | SRCC = 0.8504 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.2418 | LCC = 0.9191 | SRCC = 0.9192 ] +INFO:root:Average loss=0.15009707021407592 +INFO:root: +[Valid][7][UTT][ MSE = 0.2360 | LCC = 0.8547 | SRCC = 0.8512 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1203 | LCC = 0.9317 | SRCC = 0.9285 ] +INFO:root:Average loss=0.13682709433711493 +INFO:root: +[Valid][8][UTT][ MSE = 0.2529 | LCC = 0.8533 | SRCC = 0.8522 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1584 | LCC = 0.9260 | SRCC = 0.9242 ] +INFO:root:Average loss=0.12654114161164332 +INFO:root: +[Valid][9][UTT][ MSE = 0.2748 | LCC = 0.8533 | SRCC = 0.8505 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1801 | LCC = 0.9304 | SRCC = 0.9283 ] +INFO:root:Average loss=0.11697492557458389 +INFO:root: +[Valid][10][UTT][ MSE = 0.2399 | LCC = 0.8539 | SRCC = 0.8520 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1388 | LCC = 0.9274 | SRCC = 0.9230 ] +INFO:root:Average loss=0.11015717818950996 +INFO:root: +[Valid][11][UTT][ MSE = 0.2474 | LCC = 0.8497 | SRCC = 0.8471 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1392 | LCC = 0.9276 | SRCC = 0.9244 ] +INFO:root:Average loss=0.09738092019389837 +INFO:root: +[Valid][12][UTT][ MSE = 0.2394 | LCC = 0.8506 | SRCC = 0.8469 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1335 | LCC = 0.9239 | SRCC = 0.9151 ] +INFO:root:Average loss=0.09146557222956266 +INFO:root: +[Valid][13][UTT][ MSE = 0.2361 | LCC = 0.8504 | SRCC = 0.8478 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1282 | LCC = 0.9217 | SRCC = 0.9179 ] +INFO:root:Average loss=0.09438327733331765 +INFO:root: +[Valid][14][UTT][ MSE = 0.2411 | LCC = 0.8509 | SRCC = 0.8481 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1312 | LCC = 0.9301 | SRCC = 0.9248 ] +INFO:root:Average loss=0.08172654083523995 +INFO:root: +[Valid][15][UTT][ MSE = 0.2294 | LCC = 0.8518 | SRCC = 0.8511 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1099 | LCC = 0.9300 | SRCC = 0.9286 ] +INFO:root:Average loss=0.08050136019786198 +INFO:root: +[Valid][16][UTT][ MSE = 0.2551 | LCC = 0.8488 | SRCC = 0.8456 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1625 | LCC = 0.9214 | SRCC = 0.9152 ] +INFO:root:Average loss=0.07946349900120343 +INFO:root: +[Valid][17][UTT][ MSE = 0.2467 | LCC = 0.8484 | SRCC = 0.8470 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1447 | LCC = 0.9244 | SRCC = 0.9204 ] +INFO:root:Average loss=0.0735778076430926 +INFO:root: +[Valid][18][UTT][ MSE = 0.2599 | LCC = 0.8496 | SRCC = 0.8481 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1685 | LCC = 0.9259 | SRCC = 0.9250 ] +INFO:root:Average loss=0.06909938591221969 +INFO:root: +[Valid][19][UTT][ MSE = 0.2482 | LCC = 0.8458 | SRCC = 0.8449 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1166 | LCC = 0.9336 | SRCC = 0.9301 ] +INFO:root:Average loss=0.06664102113781831 +INFO:root: +[Valid][20][UTT][ MSE = 0.2371 | LCC = 0.8468 | SRCC = 0.8448 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1176 | LCC = 0.9266 | SRCC = 0.9249 ] +INFO:root:Average loss=0.06509335539661922 +INFO:root: +[Valid][21][UTT][ MSE = 0.2520 | LCC = 0.8480 | SRCC = 0.8455 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1569 | LCC = 0.9216 | SRCC = 0.9186 ] +INFO:root:Average loss=0.06702007200473394 +INFO:root: +[Valid][22][UTT][ MSE = 0.2406 | LCC = 0.8476 | SRCC = 0.8461 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1340 | LCC = 0.9220 | SRCC = 0.9189 ] +INFO:root:Average loss=0.05538847426382395 +INFO:root: +[Valid][23][UTT][ MSE = 0.2402 | LCC = 0.8504 | SRCC = 0.8493 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1334 | LCC = 0.9279 | SRCC = 0.9257 ] +INFO:root:Average loss=0.058735111441749796 +INFO:root: +[Valid][24][UTT][ MSE = 0.2368 | LCC = 0.8484 | SRCC = 0.8471 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1213 | LCC = 0.9282 | SRCC = 0.9241 ] +INFO:root:Average loss=0.053573257934588656 +INFO:root: +[Valid][25][UTT][ MSE = 0.2500 | LCC = 0.8482 | SRCC = 0.8464 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1489 | LCC = 0.9262 | SRCC = 0.9241 ] +INFO:root:Average loss=0.05971602560617985 +INFO:root: +[Valid][26][UTT][ MSE = 0.2564 | LCC = 0.8484 | SRCC = 0.8468 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1655 | LCC = 0.9218 | SRCC = 0.9181 ] +INFO:root:Average loss=0.05014622359512708 +INFO:root: +[Valid][27][UTT][ MSE = 0.2422 | LCC = 0.8479 | SRCC = 0.8475 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1325 | LCC = 0.9279 | SRCC = 0.9260 ] +INFO:root:Average loss=0.047758462671668105 +INFO:root: +[Valid][28][UTT][ MSE = 0.2437 | LCC = 0.8472 | SRCC = 0.8460 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1315 | LCC = 0.9281 | SRCC = 0.9232 ] +INFO:root:Average loss=0.0468454816354773 +INFO:root: +[Valid][29][UTT][ MSE = 0.2499 | LCC = 0.8479 | SRCC = 0.8468 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1519 | LCC = 0.9249 | SRCC = 0.9221 ] +INFO:root:Average loss=0.04511046369010822 +INFO:root: +[Valid][30][UTT][ MSE = 0.2394 | LCC = 0.8464 | SRCC = 0.8454 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1292 | LCC = 0.9216 | SRCC = 0.9168 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2346 | LCC = 0.8601 | SRCC = 0.8619 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1405 | LCC = 0.9114 | SRCC = 0.9137 ] diff --git a/results_bvcc/w2v2_base/layer4/run1/config.gin b/results_bvcc/w2v2_base/layer4/run1/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/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 diff --git a/results_bvcc/w2v2_base/layer4/run1/model.pt b/results_bvcc/w2v2_base/layer4/run1/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..badc15bd1ebbfc76eb167f4ed57d8f31318728b1 --- /dev/null +++ b/results_bvcc/w2v2_base/layer4/run1/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8b64bed13b53a8acecc5e5f6cd1a7ee76a937ab880f4942285342e2482be9f6f +size 2747979 diff --git a/results_bvcc/w2v2_base/layer4/run1/model_best.pt b/results_bvcc/w2v2_base/layer4/run1/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..283ee3761e37a691529536b5815dc93b0c4face3 --- /dev/null +++ b/results_bvcc/w2v2_base/layer4/run1/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:af54ce931c21d643e402e64498cea74ec1119fe465b5468ed08f14572ca16dee +size 2749638 diff --git a/results_bvcc/w2v2_base/layer4/run1/train.log b/results_bvcc/w2v2_base/layer4/run1/train.log new file mode 100644 index 0000000000000000000000000000000000000000..612c0d34602b36df41a7d76a45f50fe69b519d92 --- /dev/null +++ b/results_bvcc/w2v2_base/layer4/run1/train.log @@ -0,0 +1,163 @@ +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=0.6995040017824906 +INFO:root: +[Valid][0][UTT][ MSE = 0.3744 | LCC = 0.7721 | SRCC = 0.7779 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2619 | LCC = 0.8271 | SRCC = 0.8370 ] +INFO:root:Average loss=0.292100517031474 +INFO:root: +[Valid][1][UTT][ MSE = 0.2846 | LCC = 0.8336 | SRCC = 0.8354 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1936 | LCC = 0.8953 | SRCC = 0.8977 ] +INFO:root:Average loss=0.23422882151909363 +INFO:root: +[Valid][2][UTT][ MSE = 0.2706 | LCC = 0.8480 | SRCC = 0.8468 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1625 | LCC = 0.9109 | SRCC = 0.9089 ] +INFO:root:Average loss=0.20164333417629585 +INFO:root: +[Valid][3][UTT][ MSE = 0.2545 | LCC = 0.8531 | SRCC = 0.8517 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1497 | LCC = 0.9191 | SRCC = 0.9216 ] +INFO:root:Average loss=0.18401004240298882 +INFO:root: +[Valid][4][UTT][ MSE = 0.2540 | LCC = 0.8524 | SRCC = 0.8497 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1433 | LCC = 0.9229 | SRCC = 0.9206 ] +INFO:root:Average loss=0.16525128856301308 +INFO:root: +[Valid][5][UTT][ MSE = 0.2439 | LCC = 0.8540 | SRCC = 0.8522 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1334 | LCC = 0.9196 | SRCC = 0.9151 ] +INFO:root:Average loss=0.1554804721321815 +INFO:root: +[Valid][6][UTT][ MSE = 0.2294 | LCC = 0.8584 | SRCC = 0.8572 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1303 | LCC = 0.9219 | SRCC = 0.9212 ] +INFO:root:Average loss=0.14740966241329145 +INFO:root: +[Valid][7][UTT][ MSE = 0.2452 | LCC = 0.8558 | SRCC = 0.8537 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1225 | LCC = 0.9270 | SRCC = 0.9225 ] +INFO:root:Average loss=0.1314019387922226 +INFO:root: +[Valid][8][UTT][ MSE = 0.2454 | LCC = 0.8547 | SRCC = 0.8539 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1244 | LCC = 0.9239 | SRCC = 0.9246 ] +INFO:root:Average loss=0.13122159213973925 +INFO:root: +[Valid][9][UTT][ MSE = 0.2378 | LCC = 0.8575 | SRCC = 0.8558 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1297 | LCC = 0.9269 | SRCC = 0.9252 ] +INFO:root:Average loss=0.11824162017840606 +INFO:root: +[Valid][10][UTT][ MSE = 0.2381 | LCC = 0.8516 | SRCC = 0.8513 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1447 | LCC = 0.9167 | SRCC = 0.9156 ] +INFO:root:Average loss=0.10952124381676698 +INFO:root: +[Valid][11][UTT][ MSE = 0.2369 | LCC = 0.8548 | SRCC = 0.8528 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1199 | LCC = 0.9260 | SRCC = 0.9207 ] +INFO:root:Average loss=0.10048425531922242 +INFO:root: +[Valid][12][UTT][ MSE = 0.2401 | LCC = 0.8538 | SRCC = 0.8524 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1287 | LCC = 0.9218 | SRCC = 0.9187 ] +INFO:root:Average loss=0.09728018652934295 +INFO:root: +[Valid][13][UTT][ MSE = 0.2352 | LCC = 0.8536 | SRCC = 0.8520 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1306 | LCC = 0.9232 | SRCC = 0.9180 ] +INFO:root:Average loss=0.08805536120556869 +INFO:root: +[Valid][14][UTT][ MSE = 0.2387 | LCC = 0.8520 | SRCC = 0.8522 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1147 | LCC = 0.9214 | SRCC = 0.9184 ] +INFO:root:Average loss=0.08045589847442432 +INFO:root: +[Valid][15][UTT][ MSE = 0.2470 | LCC = 0.8494 | SRCC = 0.8480 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1281 | LCC = 0.9218 | SRCC = 0.9162 ] +INFO:root:Average loss=0.07845703942271379 +INFO:root: +[Valid][16][UTT][ MSE = 0.2419 | LCC = 0.8512 | SRCC = 0.8504 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1153 | LCC = 0.9257 | SRCC = 0.9207 ] +INFO:root:Average loss=0.07501401604177096 +INFO:root: +[Valid][17][UTT][ MSE = 0.2468 | LCC = 0.8499 | SRCC = 0.8489 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1451 | LCC = 0.9258 | SRCC = 0.9229 ] +INFO:root:Average loss=0.07132870985720402 +INFO:root: +[Valid][18][UTT][ MSE = 0.2370 | LCC = 0.8547 | SRCC = 0.8543 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1124 | LCC = 0.9311 | SRCC = 0.9297 ] +INFO:root:Average loss=0.06897051646732368 +INFO:root: +[Valid][19][UTT][ MSE = 0.2635 | LCC = 0.8472 | SRCC = 0.8484 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1764 | LCC = 0.9192 | SRCC = 0.9188 ] +INFO:root:Average loss=0.06110574868627083 +INFO:root: +[Valid][20][UTT][ MSE = 0.2393 | LCC = 0.8488 | SRCC = 0.8474 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1373 | LCC = 0.9193 | SRCC = 0.9175 ] +INFO:root:Average loss=0.0640004864678933 +INFO:root: +[Valid][21][UTT][ MSE = 0.2394 | LCC = 0.8531 | SRCC = 0.8533 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1161 | LCC = 0.9221 | SRCC = 0.9212 ] +INFO:root:Average loss=0.06125358525568094 +INFO:root: +[Valid][22][UTT][ MSE = 0.2377 | LCC = 0.8533 | SRCC = 0.8528 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1249 | LCC = 0.9215 | SRCC = 0.9168 ] +INFO:root:Average loss=0.05879857443655149 +INFO:root: +[Valid][23][UTT][ MSE = 0.2330 | LCC = 0.8510 | SRCC = 0.8504 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1189 | LCC = 0.9223 | SRCC = 0.9204 ] +INFO:root:Average loss=0.06154265242795914 +INFO:root: +[Valid][24][UTT][ MSE = 0.2387 | LCC = 0.8523 | SRCC = 0.8518 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1255 | LCC = 0.9255 | SRCC = 0.9220 ] +INFO:root:Average loss=0.05104140994640497 +INFO:root: +[Valid][25][UTT][ MSE = 0.2389 | LCC = 0.8509 | SRCC = 0.8498 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1270 | LCC = 0.9213 | SRCC = 0.9182 ] +INFO:root:Average loss=0.04806591773357911 +INFO:root: +[Valid][26][UTT][ MSE = 0.2402 | LCC = 0.8508 | SRCC = 0.8501 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1171 | LCC = 0.9258 | SRCC = 0.9242 ] +INFO:root:Average loss=0.049024469075867765 +INFO:root: +[Valid][27][UTT][ MSE = 0.2377 | LCC = 0.8527 | SRCC = 0.8519 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1131 | LCC = 0.9276 | SRCC = 0.9263 ] +INFO:root:Average loss=0.05045450175515352 +INFO:root: +[Valid][28][UTT][ MSE = 0.2526 | LCC = 0.8527 | SRCC = 0.8534 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1151 | LCC = 0.9257 | SRCC = 0.9221 ] +INFO:root:Average loss=0.052950735084521465 +INFO:root: +[Valid][29][UTT][ MSE = 0.2377 | LCC = 0.8504 | SRCC = 0.8493 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1297 | LCC = 0.9197 | SRCC = 0.9133 ] +INFO:root:Average loss=0.04517366875631687 +INFO:root: +[Valid][30][UTT][ MSE = 0.2447 | LCC = 0.8494 | SRCC = 0.8492 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1225 | LCC = 0.9263 | SRCC = 0.9245 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2286 | LCC = 0.8611 | SRCC = 0.8614 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1296 | LCC = 0.9158 | SRCC = 0.9195 ] diff --git a/results_bvcc/w2v2_base/layer4/run2/config.gin b/results_bvcc/w2v2_base/layer4/run2/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/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 diff --git a/results_bvcc/w2v2_base/layer4/run2/model.pt b/results_bvcc/w2v2_base/layer4/run2/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..137a4a5a441e0c64472ba4180ec7bc10cc7d19ea --- /dev/null +++ b/results_bvcc/w2v2_base/layer4/run2/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:793bd76d2a24860b2ff46340efd0395827fae209fe2efceeccc43688377a97cd +size 2747979 diff --git a/results_bvcc/w2v2_base/layer4/run2/model_best.pt b/results_bvcc/w2v2_base/layer4/run2/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..187b38f3e30b00fc678858b33014f0005fb7ab26 --- /dev/null +++ b/results_bvcc/w2v2_base/layer4/run2/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f031b23f79cfd25939fdd60db1a82be7f94c6562580c8981b5e4dbbacf0b1238 +size 2749638 diff --git a/results_bvcc/w2v2_base/layer4/run2/train.log b/results_bvcc/w2v2_base/layer4/run2/train.log new file mode 100644 index 0000000000000000000000000000000000000000..c89a79d3323bbc5a9dee65097917399d78d47af7 --- /dev/null +++ b/results_bvcc/w2v2_base/layer4/run2/train.log @@ -0,0 +1,163 @@ +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=0.5343131203299913 +INFO:root: +[Valid][0][UTT][ MSE = 0.3462 | LCC = 0.8062 | SRCC = 0.8092 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2436 | LCC = 0.8694 | SRCC = 0.8717 ] +INFO:root:Average loss=0.27175192019114125 +INFO:root: +[Valid][1][UTT][ MSE = 0.2863 | LCC = 0.8349 | SRCC = 0.8361 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1729 | LCC = 0.9034 | SRCC = 0.9050 ] +INFO:root:Average loss=0.2316355369029901 +INFO:root: +[Valid][2][UTT][ MSE = 0.2653 | LCC = 0.8448 | SRCC = 0.8445 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1581 | LCC = 0.9109 | SRCC = 0.9082 ] +INFO:root:Average loss=0.2039906156177704 +INFO:root: +[Valid][3][UTT][ MSE = 0.2471 | LCC = 0.8509 | SRCC = 0.8507 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1419 | LCC = 0.9147 | SRCC = 0.9137 ] +INFO:root:Average loss=0.18347042149458176 +INFO:root: +[Valid][4][UTT][ MSE = 0.2402 | LCC = 0.8526 | SRCC = 0.8512 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1165 | LCC = 0.9279 | SRCC = 0.9265 ] +INFO:root:Average loss=0.17145636314765 +INFO:root: +[Valid][5][UTT][ MSE = 0.2396 | LCC = 0.8510 | SRCC = 0.8493 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1257 | LCC = 0.9232 | SRCC = 0.9208 ] +INFO:root:Average loss=0.15737829061272818 +INFO:root: +[Valid][6][UTT][ MSE = 0.2479 | LCC = 0.8517 | SRCC = 0.8488 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1373 | LCC = 0.9241 | SRCC = 0.9218 ] +INFO:root:Average loss=0.15031299014122057 +INFO:root: +[Valid][7][UTT][ MSE = 0.2758 | LCC = 0.8518 | SRCC = 0.8486 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1275 | LCC = 0.9372 | SRCC = 0.9325 ] +INFO:root:Average loss=0.13447498081204218 +INFO:root: +[Valid][8][UTT][ MSE = 0.2399 | LCC = 0.8532 | SRCC = 0.8506 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1168 | LCC = 0.9328 | SRCC = 0.9304 ] +INFO:root:Average loss=0.11924519437627915 +INFO:root: +[Valid][9][UTT][ MSE = 0.2407 | LCC = 0.8550 | SRCC = 0.8527 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1138 | LCC = 0.9334 | SRCC = 0.9315 ] +INFO:root:Average loss=0.11490686906453891 +INFO:root: +[Valid][10][UTT][ MSE = 0.2476 | LCC = 0.8494 | SRCC = 0.8494 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1390 | LCC = 0.9307 | SRCC = 0.9319 ] +INFO:root:Average loss=0.11408099121390244 +INFO:root: +[Valid][11][UTT][ MSE = 0.2453 | LCC = 0.8527 | SRCC = 0.8504 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1177 | LCC = 0.9331 | SRCC = 0.9319 ] +INFO:root:Average loss=0.10160443177207923 +INFO:root: +[Valid][12][UTT][ MSE = 0.2535 | LCC = 0.8479 | SRCC = 0.8448 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1439 | LCC = 0.9288 | SRCC = 0.9250 ] +INFO:root:Average loss=0.09523698162191953 +INFO:root: +[Valid][13][UTT][ MSE = 0.2378 | LCC = 0.8518 | SRCC = 0.8493 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1087 | LCC = 0.9303 | SRCC = 0.9291 ] +INFO:root:Average loss=0.09384617987924661 +INFO:root: +[Valid][14][UTT][ MSE = 0.2370 | LCC = 0.8507 | SRCC = 0.8485 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1139 | LCC = 0.9320 | SRCC = 0.9297 ] +INFO:root:Average loss=0.09187376570816223 +INFO:root: +[Valid][15][UTT][ MSE = 0.2517 | LCC = 0.8479 | SRCC = 0.8447 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1392 | LCC = 0.9300 | SRCC = 0.9248 ] +INFO:root:Average loss=0.08201310315575355 +INFO:root: +[Valid][16][UTT][ MSE = 0.2378 | LCC = 0.8463 | SRCC = 0.8445 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1096 | LCC = 0.9282 | SRCC = 0.9250 ] +INFO:root:Average loss=0.07383349213080528 +INFO:root: +[Valid][17][UTT][ MSE = 0.2428 | LCC = 0.8463 | SRCC = 0.8429 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1146 | LCC = 0.9320 | SRCC = 0.9267 ] +INFO:root:Average loss=0.07333376893821435 +INFO:root: +[Valid][18][UTT][ MSE = 0.2563 | LCC = 0.8400 | SRCC = 0.8358 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1352 | LCC = 0.9258 | SRCC = 0.9179 ] +INFO:root:Average loss=0.06984003021931037 +INFO:root: +[Valid][19][UTT][ MSE = 0.2454 | LCC = 0.8440 | SRCC = 0.8412 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1247 | LCC = 0.9266 | SRCC = 0.9223 ] +INFO:root:Average loss=0.06705957226073131 +INFO:root: +[Valid][20][UTT][ MSE = 0.2617 | LCC = 0.8406 | SRCC = 0.8373 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1524 | LCC = 0.9239 | SRCC = 0.9200 ] +INFO:root:Average loss=0.061417680472517625 +INFO:root: +[Valid][21][UTT][ MSE = 0.2540 | LCC = 0.8447 | SRCC = 0.8421 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1334 | LCC = 0.9348 | SRCC = 0.9325 ] +INFO:root:Average loss=0.0598450720978853 +INFO:root: +[Valid][22][UTT][ MSE = 0.2533 | LCC = 0.8436 | SRCC = 0.8432 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1373 | LCC = 0.9318 | SRCC = 0.9322 ] +INFO:root:Average loss=0.058676363876423776 +INFO:root: +[Valid][23][UTT][ MSE = 0.2433 | LCC = 0.8443 | SRCC = 0.8406 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1091 | LCC = 0.9288 | SRCC = 0.9225 ] +INFO:root:Average loss=0.06033519416665419 +INFO:root: +[Valid][24][UTT][ MSE = 0.2405 | LCC = 0.8474 | SRCC = 0.8454 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1140 | LCC = 0.9291 | SRCC = 0.9235 ] +INFO:root:Average loss=0.050451572554615826 +INFO:root: +[Valid][25][UTT][ MSE = 0.2442 | LCC = 0.8453 | SRCC = 0.8436 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1317 | LCC = 0.9262 | SRCC = 0.9206 ] +INFO:root:Average loss=0.05087349609209177 +INFO:root: +[Valid][26][UTT][ MSE = 0.2446 | LCC = 0.8433 | SRCC = 0.8408 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1150 | LCC = 0.9295 | SRCC = 0.9245 ] +INFO:root:Average loss=0.04974463123541612 +INFO:root: +[Valid][27][UTT][ MSE = 0.2527 | LCC = 0.8420 | SRCC = 0.8402 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1397 | LCC = 0.9258 | SRCC = 0.9200 ] +INFO:root:Average loss=0.04788704111407965 +INFO:root: +[Valid][28][UTT][ MSE = 0.2505 | LCC = 0.8432 | SRCC = 0.8398 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1198 | LCC = 0.9272 | SRCC = 0.9203 ] +INFO:root:Average loss=0.04461211735048355 +INFO:root: +[Valid][29][UTT][ MSE = 0.2511 | LCC = 0.8434 | SRCC = 0.8389 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1209 | LCC = 0.9270 | SRCC = 0.9176 ] +INFO:root:Average loss=0.04255519038400589 +INFO:root: +[Valid][30][UTT][ MSE = 0.2407 | LCC = 0.8453 | SRCC = 0.8425 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1166 | LCC = 0.9281 | SRCC = 0.9210 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2429 | LCC = 0.8605 | SRCC = 0.8570 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1607 | LCC = 0.9140 | SRCC = 0.9134 ] diff --git a/results_bvcc/w2v2_base/layer4/run3/config.gin b/results_bvcc/w2v2_base/layer4/run3/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/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 diff --git a/results_bvcc/w2v2_base/layer4/run3/model.pt b/results_bvcc/w2v2_base/layer4/run3/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..1efdce14c294e5dfb45d50f2d9fa2932e99d77b9 --- /dev/null +++ b/results_bvcc/w2v2_base/layer4/run3/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8fe868479c254c44e5783dab4ff47d7e645a3b222cc1e8cf3dd269491a5b208a +size 2747979 diff --git a/results_bvcc/w2v2_base/layer4/run3/model_best.pt b/results_bvcc/w2v2_base/layer4/run3/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..80c3c34d9a3d7fcd6ddaed051e0b2fdc930946b9 --- /dev/null +++ b/results_bvcc/w2v2_base/layer4/run3/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:faf17fc7763f699a84177e24815be4a6eaa06ecbae11609e8c609b1238d03286 +size 2749638 diff --git a/results_bvcc/w2v2_base/layer4/run3/train.log b/results_bvcc/w2v2_base/layer4/run3/train.log new file mode 100644 index 0000000000000000000000000000000000000000..60fb7be88b9fcebe62f53eb343bce1a0b9a86736 --- /dev/null +++ b/results_bvcc/w2v2_base/layer4/run3/train.log @@ -0,0 +1,163 @@ +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=0.5085462654630343 +INFO:root: +[Valid][0][UTT][ MSE = 0.3282 | LCC = 0.7951 | SRCC = 0.7976 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2068 | LCC = 0.8645 | SRCC = 0.8663 ] +INFO:root:Average loss=0.2736496201310402 +INFO:root: +[Valid][1][UTT][ MSE = 0.2974 | LCC = 0.8266 | SRCC = 0.8256 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1848 | LCC = 0.8973 | SRCC = 0.8972 ] +INFO:root:Average loss=0.23099989520433623 +INFO:root: +[Valid][2][UTT][ MSE = 0.2775 | LCC = 0.8375 | SRCC = 0.8342 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1581 | LCC = 0.9121 | SRCC = 0.9060 ] +INFO:root:Average loss=0.21685103766429117 +INFO:root: +[Valid][3][UTT][ MSE = 0.2648 | LCC = 0.8463 | SRCC = 0.8444 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1362 | LCC = 0.9247 | SRCC = 0.9214 ] +INFO:root:Average loss=0.18828245386099204 +INFO:root: +[Valid][4][UTT][ MSE = 0.2554 | LCC = 0.8475 | SRCC = 0.8461 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1486 | LCC = 0.9180 | SRCC = 0.9146 ] +INFO:root:Average loss=0.1756950860413221 +INFO:root: +[Valid][5][UTT][ MSE = 0.2549 | LCC = 0.8457 | SRCC = 0.8444 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1456 | LCC = 0.9175 | SRCC = 0.9147 ] +INFO:root:Average loss=0.16025372403554428 +INFO:root: +[Valid][6][UTT][ MSE = 0.2729 | LCC = 0.8476 | SRCC = 0.8472 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1313 | LCC = 0.9236 | SRCC = 0.9219 ] +INFO:root:Average loss=0.15151278464457926 +INFO:root: +[Valid][7][UTT][ MSE = 0.2621 | LCC = 0.8462 | SRCC = 0.8458 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1612 | LCC = 0.9191 | SRCC = 0.9219 ] +INFO:root:Average loss=0.14629306873449913 +INFO:root: +[Valid][8][UTT][ MSE = 0.2474 | LCC = 0.8495 | SRCC = 0.8484 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1201 | LCC = 0.9252 | SRCC = 0.9235 ] +INFO:root:Average loss=0.13070591968985704 +INFO:root: +[Valid][9][UTT][ MSE = 0.2506 | LCC = 0.8510 | SRCC = 0.8508 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1181 | LCC = 0.9257 | SRCC = 0.9223 ] +INFO:root:Average loss=0.12174523822390117 +INFO:root: +[Valid][10][UTT][ MSE = 0.2421 | LCC = 0.8478 | SRCC = 0.8474 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1216 | LCC = 0.9240 | SRCC = 0.9229 ] +INFO:root:Average loss=0.11293161641328762 +INFO:root: +[Valid][11][UTT][ MSE = 0.2400 | LCC = 0.8531 | SRCC = 0.8525 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1149 | LCC = 0.9283 | SRCC = 0.9268 ] +INFO:root:Average loss=0.11384149487966146 +INFO:root: +[Valid][12][UTT][ MSE = 0.2410 | LCC = 0.8514 | SRCC = 0.8525 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1396 | LCC = 0.9216 | SRCC = 0.9244 ] +INFO:root:Average loss=0.10897560723316975 +INFO:root: +[Valid][13][UTT][ MSE = 0.2391 | LCC = 0.8499 | SRCC = 0.8493 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1168 | LCC = 0.9301 | SRCC = 0.9291 ] +INFO:root:Average loss=0.09427648849594288 +INFO:root: +[Valid][14][UTT][ MSE = 0.2592 | LCC = 0.8488 | SRCC = 0.8487 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1220 | LCC = 0.9215 | SRCC = 0.9182 ] +INFO:root:Average loss=0.08931664425211075 +INFO:root: +[Valid][15][UTT][ MSE = 0.2695 | LCC = 0.8424 | SRCC = 0.8411 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1210 | LCC = 0.9287 | SRCC = 0.9299 ] +INFO:root:Average loss=0.08630830622636355 +INFO:root: +[Valid][16][UTT][ MSE = 0.2433 | LCC = 0.8457 | SRCC = 0.8455 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1283 | LCC = 0.9237 | SRCC = 0.9232 ] +INFO:root:Average loss=0.08225185681994145 +INFO:root: +[Valid][17][UTT][ MSE = 0.2457 | LCC = 0.8462 | SRCC = 0.8470 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1203 | LCC = 0.9233 | SRCC = 0.9244 ] +INFO:root:Average loss=0.07563946792521538 +INFO:root: +[Valid][18][UTT][ MSE = 0.2550 | LCC = 0.8435 | SRCC = 0.8448 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1439 | LCC = 0.9235 | SRCC = 0.9258 ] +INFO:root:Average loss=0.07579196258806266 +INFO:root: +[Valid][19][UTT][ MSE = 0.2552 | LCC = 0.8409 | SRCC = 0.8409 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1348 | LCC = 0.9220 | SRCC = 0.9201 ] +INFO:root:Average loss=0.07054669085221413 +INFO:root: +[Valid][20][UTT][ MSE = 0.2596 | LCC = 0.8438 | SRCC = 0.8448 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1157 | LCC = 0.9252 | SRCC = 0.9230 ] +INFO:root:Average loss=0.06740382051047607 +INFO:root: +[Valid][21][UTT][ MSE = 0.2585 | LCC = 0.8462 | SRCC = 0.8470 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1122 | LCC = 0.9265 | SRCC = 0.9268 ] +INFO:root:Average loss=0.06237097313770881 +INFO:root: +[Valid][22][UTT][ MSE = 0.2516 | LCC = 0.8390 | SRCC = 0.8390 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1359 | LCC = 0.9196 | SRCC = 0.9166 ] +INFO:root:Average loss=0.06210563231546145 +INFO:root: +[Valid][23][UTT][ MSE = 0.2502 | LCC = 0.8419 | SRCC = 0.8426 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1186 | LCC = 0.9272 | SRCC = 0.9307 ] +INFO:root:Average loss=0.061679850451839276 +INFO:root: +[Valid][24][UTT][ MSE = 0.2583 | LCC = 0.8429 | SRCC = 0.8435 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1146 | LCC = 0.9250 | SRCC = 0.9234 ] +INFO:root:Average loss=0.05688843985971732 +INFO:root: +[Valid][25][UTT][ MSE = 0.2570 | LCC = 0.8439 | SRCC = 0.8441 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1131 | LCC = 0.9267 | SRCC = 0.9255 ] +INFO:root:Average loss=0.060697861588918246 +INFO:root: +[Valid][26][UTT][ MSE = 0.2408 | LCC = 0.8470 | SRCC = 0.8481 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1145 | LCC = 0.9214 | SRCC = 0.9233 ] +INFO:root:Average loss=0.050061373422161125 +INFO:root: +[Valid][27][UTT][ MSE = 0.2491 | LCC = 0.8479 | SRCC = 0.8492 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1099 | LCC = 0.9272 | SRCC = 0.9300 ] +INFO:root:Average loss=0.04986120075082932 +INFO:root: +[Valid][28][UTT][ MSE = 0.2520 | LCC = 0.8453 | SRCC = 0.8462 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1091 | LCC = 0.9278 | SRCC = 0.9270 ] +INFO:root:Average loss=0.05015027785721498 +INFO:root: +[Valid][29][UTT][ MSE = 0.2409 | LCC = 0.8459 | SRCC = 0.8458 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1227 | LCC = 0.9257 | SRCC = 0.9261 ] +INFO:root:Average loss=0.04641652630212215 +INFO:root: +[Valid][30][UTT][ MSE = 0.2401 | LCC = 0.8453 | SRCC = 0.8453 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1248 | LCC = 0.9181 | SRCC = 0.9159 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2458 | LCC = 0.8566 | SRCC = 0.8560 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1512 | LCC = 0.9160 | SRCC = 0.9161 ] diff --git a/results_bvcc/w2v2_base/layer4/run4/config.gin b/results_bvcc/w2v2_base/layer4/run4/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/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 diff --git a/results_bvcc/w2v2_base/layer4/run4/model.pt b/results_bvcc/w2v2_base/layer4/run4/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..522b40709bce55263c8a507e21c32681374c1c8a --- /dev/null +++ b/results_bvcc/w2v2_base/layer4/run4/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9bcf2511ab2f27703daead23c08dadae388ce812ff44dd68b1ffef61be3e14b4 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer4/run4/model_best.pt b/results_bvcc/w2v2_base/layer4/run4/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..ec0c4ed7a4a9f96a67c092f40b9b3d387c8bf874 --- /dev/null +++ b/results_bvcc/w2v2_base/layer4/run4/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:114422f0c0dc347753b632f73912412dec137d6c0de9e88229a8a4ab54d0b741 +size 2749638 diff --git a/results_bvcc/w2v2_base/layer4/run4/train.log b/results_bvcc/w2v2_base/layer4/run4/train.log new file mode 100644 index 0000000000000000000000000000000000000000..d9b0860d213726b975680f9d47675550fd965a35 --- /dev/null +++ b/results_bvcc/w2v2_base/layer4/run4/train.log @@ -0,0 +1,163 @@ +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=0.5468503362857379 +INFO:root: +[Valid][0][UTT][ MSE = 0.3448 | LCC = 0.7943 | SRCC = 0.8006 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2443 | LCC = 0.8535 | SRCC = 0.8566 ] +INFO:root:Average loss=0.27077613637233394 +INFO:root: +[Valid][1][UTT][ MSE = 0.2820 | LCC = 0.8264 | SRCC = 0.8264 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1600 | LCC = 0.8961 | SRCC = 0.8944 ] +INFO:root:Average loss=0.22810082997267062 +INFO:root: +[Valid][2][UTT][ MSE = 0.2672 | LCC = 0.8386 | SRCC = 0.8367 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1558 | LCC = 0.9067 | SRCC = 0.9022 ] +INFO:root:Average loss=0.20190775021910667 +INFO:root: +[Valid][3][UTT][ MSE = 0.2598 | LCC = 0.8466 | SRCC = 0.8470 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1672 | LCC = 0.9131 | SRCC = 0.9101 ] +INFO:root:Average loss=0.18852522033147323 +INFO:root: +[Valid][4][UTT][ MSE = 0.2565 | LCC = 0.8505 | SRCC = 0.8500 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1397 | LCC = 0.9297 | SRCC = 0.9281 ] +INFO:root:Average loss=0.17150766545763382 +INFO:root: +[Valid][5][UTT][ MSE = 0.2497 | LCC = 0.8486 | SRCC = 0.8474 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1356 | LCC = 0.9263 | SRCC = 0.9212 ] +INFO:root:Average loss=0.15543746690337473 +INFO:root: +[Valid][6][UTT][ MSE = 0.3014 | LCC = 0.8476 | SRCC = 0.8467 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.2182 | LCC = 0.9209 | SRCC = 0.9172 ] +INFO:root:Average loss=0.14557785464402956 +INFO:root: +[Valid][7][UTT][ MSE = 0.2356 | LCC = 0.8514 | SRCC = 0.8501 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1150 | LCC = 0.9310 | SRCC = 0.9292 ] +INFO:root:Average loss=0.139043383873426 +INFO:root: +[Valid][8][UTT][ MSE = 0.2399 | LCC = 0.8554 | SRCC = 0.8542 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1350 | LCC = 0.9298 | SRCC = 0.9280 ] +INFO:root:Average loss=0.12835628338731253 +INFO:root: +[Valid][9][UTT][ MSE = 0.2367 | LCC = 0.8529 | SRCC = 0.8513 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1302 | LCC = 0.9268 | SRCC = 0.9229 ] +INFO:root:Average loss=0.11538153485609935 +INFO:root: +[Valid][10][UTT][ MSE = 0.2492 | LCC = 0.8485 | SRCC = 0.8461 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1211 | LCC = 0.9309 | SRCC = 0.9269 ] +INFO:root:Average loss=0.11206608275190377 +INFO:root: +[Valid][11][UTT][ MSE = 0.2642 | LCC = 0.8449 | SRCC = 0.8436 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1211 | LCC = 0.9255 | SRCC = 0.9221 ] +INFO:root:Average loss=0.10401512520053448 +INFO:root: +[Valid][12][UTT][ MSE = 0.2515 | LCC = 0.8466 | SRCC = 0.8443 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1171 | LCC = 0.9276 | SRCC = 0.9234 ] +INFO:root:Average loss=0.09643900934129189 +INFO:root: +[Valid][13][UTT][ MSE = 0.2451 | LCC = 0.8463 | SRCC = 0.8448 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1193 | LCC = 0.9295 | SRCC = 0.9251 ] +INFO:root:Average loss=0.0941178413728873 +INFO:root: +[Valid][14][UTT][ MSE = 0.2492 | LCC = 0.8506 | SRCC = 0.8495 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1571 | LCC = 0.9236 | SRCC = 0.9235 ] +INFO:root:Average loss=0.09112726806257015 +INFO:root: +[Valid][15][UTT][ MSE = 0.2383 | LCC = 0.8477 | SRCC = 0.8462 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1113 | LCC = 0.9281 | SRCC = 0.9240 ] +INFO:root:Average loss=0.08408225915179803 +INFO:root: +[Valid][16][UTT][ MSE = 0.2437 | LCC = 0.8481 | SRCC = 0.8458 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1224 | LCC = 0.9282 | SRCC = 0.9266 ] +INFO:root:Average loss=0.0798466527977815 +INFO:root: +[Valid][17][UTT][ MSE = 0.2685 | LCC = 0.8447 | SRCC = 0.8434 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1808 | LCC = 0.9174 | SRCC = 0.9165 ] +INFO:root:Average loss=0.07477873582870533 +INFO:root: +[Valid][18][UTT][ MSE = 0.2417 | LCC = 0.8455 | SRCC = 0.8439 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1195 | LCC = 0.9247 | SRCC = 0.9220 ] +INFO:root:Average loss=0.07728042964560863 +INFO:root: +[Valid][19][UTT][ MSE = 0.2514 | LCC = 0.8435 | SRCC = 0.8421 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1397 | LCC = 0.9269 | SRCC = 0.9279 ] +INFO:root:Average loss=0.07046954491390632 +INFO:root: +[Valid][20][UTT][ MSE = 0.2380 | LCC = 0.8487 | SRCC = 0.8473 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1200 | LCC = 0.9283 | SRCC = 0.9263 ] +INFO:root:Average loss=0.06266084585625392 +INFO:root: +[Valid][21][UTT][ MSE = 0.2428 | LCC = 0.8446 | SRCC = 0.8432 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1167 | LCC = 0.9258 | SRCC = 0.9222 ] +INFO:root:Average loss=0.06497646395403606 +INFO:root: +[Valid][22][UTT][ MSE = 0.2423 | LCC = 0.8477 | SRCC = 0.8465 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1315 | LCC = 0.9248 | SRCC = 0.9224 ] +INFO:root:Average loss=0.05855887244718197 +INFO:root: +[Valid][23][UTT][ MSE = 0.2397 | LCC = 0.8444 | SRCC = 0.8431 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1163 | LCC = 0.9239 | SRCC = 0.9213 ] +INFO:root:Average loss=0.058047509585053496 +INFO:root: +[Valid][24][UTT][ MSE = 0.2619 | LCC = 0.8435 | SRCC = 0.8427 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1131 | LCC = 0.9279 | SRCC = 0.9261 ] +INFO:root:Average loss=0.05773174597953375 +INFO:root: +[Valid][25][UTT][ MSE = 0.2454 | LCC = 0.8445 | SRCC = 0.8426 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1199 | LCC = 0.9227 | SRCC = 0.9193 ] +INFO:root:Average loss=0.05239785586794218 +INFO:root: +[Valid][26][UTT][ MSE = 0.2471 | LCC = 0.8418 | SRCC = 0.8396 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1224 | LCC = 0.9268 | SRCC = 0.9246 ] +INFO:root:Average loss=0.058313708465832934 +INFO:root: +[Valid][27][UTT][ MSE = 0.2479 | LCC = 0.8432 | SRCC = 0.8412 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1444 | LCC = 0.9182 | SRCC = 0.9171 ] +INFO:root:Average loss=0.0516114946311483 +INFO:root: +[Valid][28][UTT][ MSE = 0.2501 | LCC = 0.8400 | SRCC = 0.8384 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1224 | LCC = 0.9213 | SRCC = 0.9171 ] +INFO:root:Average loss=0.048133330419659615 +INFO:root: +[Valid][29][UTT][ MSE = 0.2488 | LCC = 0.8419 | SRCC = 0.8403 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1127 | LCC = 0.9239 | SRCC = 0.9176 ] +INFO:root:Average loss=0.04460425904163948 +INFO:root: +[Valid][30][UTT][ MSE = 0.2531 | LCC = 0.8384 | SRCC = 0.8362 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1125 | LCC = 0.9271 | SRCC = 0.9241 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2336 | LCC = 0.8569 | SRCC = 0.8560 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1321 | LCC = 0.9084 | SRCC = 0.9090 ] diff --git a/results_bvcc/w2v2_base/layer5/run0/config.gin b/results_bvcc/w2v2_base/layer5/run0/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer5/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 diff --git a/results_bvcc/w2v2_base/layer5/run0/model.pt b/results_bvcc/w2v2_base/layer5/run0/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..7066b34e1eed608e3bf514a9c178f63cd535b95f --- /dev/null +++ b/results_bvcc/w2v2_base/layer5/run0/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:522eef7b2a6679a6164abe42b860eac1b70c7107af1af590793f74b8e5a94fa8 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer5/run0/model_best.pt b/results_bvcc/w2v2_base/layer5/run0/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..34461ce1160b97fb96c3baac2b1552e899196e20 --- /dev/null +++ b/results_bvcc/w2v2_base/layer5/run0/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:118981f23c705dc8de4a7e4cd6790d5fd33f823b0ebf305b4d9b760de3466867 +size 2749638 diff --git a/results_bvcc/w2v2_base/layer5/run0/train.log b/results_bvcc/w2v2_base/layer5/run0/train.log new file mode 100644 index 0000000000000000000000000000000000000000..c13a8002949e45a6c12cc839e393580b3db5f25b --- /dev/null +++ b/results_bvcc/w2v2_base/layer5/run0/train.log @@ -0,0 +1,163 @@ +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=0.5386151056259106 +INFO:root: +[Valid][0][UTT][ MSE = 0.3590 | LCC = 0.7954 | SRCC = 0.8000 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2577 | LCC = 0.8553 | SRCC = 0.8563 ] +INFO:root:Average loss=0.277979993667358 +INFO:root: +[Valid][1][UTT][ MSE = 0.2879 | LCC = 0.8313 | SRCC = 0.8321 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1763 | LCC = 0.8954 | SRCC = 0.8947 ] +INFO:root:Average loss=0.24505285383799139 +INFO:root: +[Valid][2][UTT][ MSE = 0.3084 | LCC = 0.8342 | SRCC = 0.8389 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.2267 | LCC = 0.8945 | SRCC = 0.8966 ] +INFO:root:Average loss=0.21206361131790358 +INFO:root: +[Valid][3][UTT][ MSE = 0.2693 | LCC = 0.8395 | SRCC = 0.8398 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1485 | LCC = 0.9078 | SRCC = 0.9077 ] +INFO:root:Average loss=0.1959761346761997 +INFO:root: +[Valid][4][UTT][ MSE = 0.2659 | LCC = 0.8468 | SRCC = 0.8495 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1599 | LCC = 0.9174 | SRCC = 0.9203 ] +INFO:root:Average loss=0.17937187334665886 +INFO:root: +[Valid][5][UTT][ MSE = 0.2580 | LCC = 0.8474 | SRCC = 0.8468 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1492 | LCC = 0.9136 | SRCC = 0.9100 ] +INFO:root:Average loss=0.15804126342901817 +INFO:root: +[Valid][6][UTT][ MSE = 0.2943 | LCC = 0.8447 | SRCC = 0.8461 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.2036 | LCC = 0.9160 | SRCC = 0.9173 ] +INFO:root:Average loss=0.15064545453358918 +INFO:root: +[Valid][7][UTT][ MSE = 0.2534 | LCC = 0.8502 | SRCC = 0.8489 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1436 | LCC = 0.9251 | SRCC = 0.9240 ] +INFO:root:Average loss=0.13708414557652596 +INFO:root: +[Valid][8][UTT][ MSE = 0.2500 | LCC = 0.8492 | SRCC = 0.8474 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1346 | LCC = 0.9187 | SRCC = 0.9129 ] +INFO:root:Average loss=0.12941377180126998 +INFO:root: +[Valid][9][UTT][ MSE = 0.2543 | LCC = 0.8445 | SRCC = 0.8449 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1561 | LCC = 0.9115 | SRCC = 0.9101 ] +INFO:root:Average loss=0.12873738325941256 +INFO:root: +[Valid][10][UTT][ MSE = 0.2482 | LCC = 0.8451 | SRCC = 0.8440 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1304 | LCC = 0.9129 | SRCC = 0.9082 ] +INFO:root:Average loss=0.10760594780246417 +INFO:root: +[Valid][11][UTT][ MSE = 0.2493 | LCC = 0.8447 | SRCC = 0.8438 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1350 | LCC = 0.9158 | SRCC = 0.9121 ] +INFO:root:Average loss=0.10614026471590385 +INFO:root: +[Valid][12][UTT][ MSE = 0.2687 | LCC = 0.8409 | SRCC = 0.8406 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1764 | LCC = 0.9098 | SRCC = 0.9068 ] +INFO:root:Average loss=0.09480574306769249 +INFO:root: +[Valid][13][UTT][ MSE = 0.2555 | LCC = 0.8457 | SRCC = 0.8444 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1551 | LCC = 0.9197 | SRCC = 0.9178 ] +INFO:root:Average loss=0.09497289440761773 +INFO:root: +[Valid][14][UTT][ MSE = 0.2614 | LCC = 0.8423 | SRCC = 0.8428 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1600 | LCC = 0.9161 | SRCC = 0.9166 ] +INFO:root:Average loss=0.08831849675147961 +INFO:root: +[Valid][15][UTT][ MSE = 0.2481 | LCC = 0.8449 | SRCC = 0.8438 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1357 | LCC = 0.9196 | SRCC = 0.9156 ] +INFO:root:Average loss=0.08021045896487358 +INFO:root: +[Valid][16][UTT][ MSE = 0.2531 | LCC = 0.8379 | SRCC = 0.8371 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1308 | LCC = 0.9158 | SRCC = 0.9119 ] +INFO:root:Average loss=0.0786331050002422 +INFO:root: +[Valid][17][UTT][ MSE = 0.2760 | LCC = 0.8366 | SRCC = 0.8372 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1704 | LCC = 0.9153 | SRCC = 0.9129 ] +INFO:root:Average loss=0.07629947918347824 +INFO:root: +[Valid][18][UTT][ MSE = 0.2511 | LCC = 0.8413 | SRCC = 0.8416 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1434 | LCC = 0.9147 | SRCC = 0.9116 ] +INFO:root:Average loss=0.06719083396288064 +INFO:root: +[Valid][19][UTT][ MSE = 0.2622 | LCC = 0.8356 | SRCC = 0.8345 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1463 | LCC = 0.9144 | SRCC = 0.9103 ] +INFO:root:Average loss=0.0656877676359354 +INFO:root: +[Valid][20][UTT][ MSE = 0.2610 | LCC = 0.8381 | SRCC = 0.8380 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1460 | LCC = 0.9211 | SRCC = 0.9203 ] +INFO:root:Average loss=0.06476138116648564 +INFO:root: +[Valid][21][UTT][ MSE = 0.2557 | LCC = 0.8380 | SRCC = 0.8366 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1383 | LCC = 0.9151 | SRCC = 0.9099 ] +INFO:root:Average loss=0.07254046626771107 +INFO:root: +[Valid][22][UTT][ MSE = 0.2552 | LCC = 0.8407 | SRCC = 0.8397 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1439 | LCC = 0.9158 | SRCC = 0.9121 ] +INFO:root:Average loss=0.05787904063860575 +INFO:root: +[Valid][23][UTT][ MSE = 0.2576 | LCC = 0.8353 | SRCC = 0.8331 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1358 | LCC = 0.9116 | SRCC = 0.9046 ] +INFO:root:Average loss=0.05635955633643346 +INFO:root: +[Valid][24][UTT][ MSE = 0.2532 | LCC = 0.8379 | SRCC = 0.8370 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1379 | LCC = 0.9133 | SRCC = 0.9107 ] +INFO:root:Average loss=0.055401684334262825 +INFO:root: +[Valid][25][UTT][ MSE = 0.2457 | LCC = 0.8414 | SRCC = 0.8397 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1258 | LCC = 0.9183 | SRCC = 0.9158 ] +INFO:root:Average loss=0.053006258458854295 +INFO:root: +[Valid][26][UTT][ MSE = 0.2873 | LCC = 0.8379 | SRCC = 0.8377 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1413 | LCC = 0.9132 | SRCC = 0.9101 ] +INFO:root:Average loss=0.05385311547284707 +INFO:root: +[Valid][27][UTT][ MSE = 0.2561 | LCC = 0.8374 | SRCC = 0.8357 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1311 | LCC = 0.9165 | SRCC = 0.9135 ] +INFO:root:Average loss=0.04797640824929262 +INFO:root: +[Valid][28][UTT][ MSE = 0.2506 | LCC = 0.8426 | SRCC = 0.8409 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1290 | LCC = 0.9184 | SRCC = 0.9155 ] +INFO:root:Average loss=0.04571760445833206 +INFO:root: +[Valid][29][UTT][ MSE = 0.2500 | LCC = 0.8396 | SRCC = 0.8384 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1310 | LCC = 0.9149 | SRCC = 0.9116 ] +INFO:root:Average loss=0.04728874014929319 +INFO:root: +[Valid][30][UTT][ MSE = 0.2644 | LCC = 0.8401 | SRCC = 0.8371 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1592 | LCC = 0.9129 | SRCC = 0.9064 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2392 | LCC = 0.8561 | SRCC = 0.8554 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1288 | LCC = 0.9171 | SRCC = 0.9170 ] diff --git a/results_bvcc/w2v2_base/layer5/run1/config.gin b/results_bvcc/w2v2_base/layer5/run1/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer5/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 diff --git a/results_bvcc/w2v2_base/layer5/run1/model.pt b/results_bvcc/w2v2_base/layer5/run1/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..92816d9185741193398f4570146ac94454c24568 --- /dev/null +++ b/results_bvcc/w2v2_base/layer5/run1/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e271c97985a3dc4a4405fd43874dff59ac57cbaa96058513545bd9aeb049dda3 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer5/run1/model_best.pt b/results_bvcc/w2v2_base/layer5/run1/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..c7b98a85d54ed0f758bbf7bf4e2f6f98750d8da3 --- /dev/null +++ b/results_bvcc/w2v2_base/layer5/run1/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ca37cfba160d50901056ec0ddb90d01f25e9b4ff4a5cdd22bbcfb8572a12c22e +size 2749638 diff --git a/results_bvcc/w2v2_base/layer5/run1/train.log b/results_bvcc/w2v2_base/layer5/run1/train.log new file mode 100644 index 0000000000000000000000000000000000000000..0f5b03a37539021b5c3337c36ba56f669c76da40 --- /dev/null +++ b/results_bvcc/w2v2_base/layer5/run1/train.log @@ -0,0 +1,163 @@ +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=0.6072434588120534 +INFO:root: +[Valid][0][UTT][ MSE = 0.3426 | LCC = 0.7953 | SRCC = 0.7959 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2326 | LCC = 0.8528 | SRCC = 0.8521 ] +INFO:root:Average loss=0.2762049470956509 +INFO:root: +[Valid][1][UTT][ MSE = 0.2957 | LCC = 0.8291 | SRCC = 0.8294 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1788 | LCC = 0.8930 | SRCC = 0.8954 ] +INFO:root:Average loss=0.2402468598805941 +INFO:root: +[Valid][2][UTT][ MSE = 0.2810 | LCC = 0.8402 | SRCC = 0.8392 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1676 | LCC = 0.9056 | SRCC = 0.9019 ] +INFO:root:Average loss=0.2093752309297904 +INFO:root: +[Valid][3][UTT][ MSE = 0.2622 | LCC = 0.8472 | SRCC = 0.8464 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1469 | LCC = 0.9086 | SRCC = 0.9054 ] +INFO:root:Average loss=0.19134235038207129 +INFO:root: +[Valid][4][UTT][ MSE = 0.2669 | LCC = 0.8451 | SRCC = 0.8441 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1485 | LCC = 0.9104 | SRCC = 0.9059 ] +INFO:root:Average loss=0.17659575988849005 +INFO:root: +[Valid][5][UTT][ MSE = 0.2812 | LCC = 0.8447 | SRCC = 0.8441 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1525 | LCC = 0.9139 | SRCC = 0.9095 ] +INFO:root:Average loss=0.16496167723567057 +INFO:root: +[Valid][6][UTT][ MSE = 0.2535 | LCC = 0.8455 | SRCC = 0.8448 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1362 | LCC = 0.9118 | SRCC = 0.9095 ] +INFO:root:Average loss=0.15187348205691728 +INFO:root: +[Valid][7][UTT][ MSE = 0.2584 | LCC = 0.8451 | SRCC = 0.8448 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1398 | LCC = 0.9193 | SRCC = 0.9162 ] +INFO:root:Average loss=0.1430496895351471 +INFO:root: +[Valid][8][UTT][ MSE = 0.2631 | LCC = 0.8427 | SRCC = 0.8421 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1399 | LCC = 0.9133 | SRCC = 0.9083 ] +INFO:root:Average loss=0.13061018880361167 +INFO:root: +[Valid][9][UTT][ MSE = 0.2523 | LCC = 0.8431 | SRCC = 0.8425 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1256 | LCC = 0.9198 | SRCC = 0.9159 ] +INFO:root:Average loss=0.12383738255653626 +INFO:root: +[Valid][10][UTT][ MSE = 0.2512 | LCC = 0.8444 | SRCC = 0.8430 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1240 | LCC = 0.9191 | SRCC = 0.9148 ] +INFO:root:Average loss=0.11356295368228203 +INFO:root: +[Valid][11][UTT][ MSE = 0.2488 | LCC = 0.8453 | SRCC = 0.8436 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1273 | LCC = 0.9192 | SRCC = 0.9171 ] +INFO:root:Average loss=0.10458667977498128 +INFO:root: +[Valid][12][UTT][ MSE = 0.2645 | LCC = 0.8455 | SRCC = 0.8461 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1193 | LCC = 0.9259 | SRCC = 0.9230 ] +INFO:root:Average loss=0.10094026543008976 +INFO:root: +[Valid][13][UTT][ MSE = 0.2698 | LCC = 0.8431 | SRCC = 0.8429 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1263 | LCC = 0.9215 | SRCC = 0.9170 ] +INFO:root:Average loss=0.0926367540199023 +INFO:root: +[Valid][14][UTT][ MSE = 0.2548 | LCC = 0.8453 | SRCC = 0.8448 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1234 | LCC = 0.9230 | SRCC = 0.9179 ] +INFO:root:Average loss=0.08843183827896912 +INFO:root: +[Valid][15][UTT][ MSE = 0.2471 | LCC = 0.8429 | SRCC = 0.8447 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1316 | LCC = 0.9226 | SRCC = 0.9243 ] +INFO:root:Average loss=0.08706968139188412 +INFO:root: +[Valid][16][UTT][ MSE = 0.2534 | LCC = 0.8428 | SRCC = 0.8429 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1393 | LCC = 0.9194 | SRCC = 0.9158 ] +INFO:root:Average loss=0.08280902183972873 +INFO:root: +[Valid][17][UTT][ MSE = 0.2616 | LCC = 0.8395 | SRCC = 0.8407 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1328 | LCC = 0.9183 | SRCC = 0.9174 ] +INFO:root:Average loss=0.07873819596492328 +INFO:root: +[Valid][18][UTT][ MSE = 0.2470 | LCC = 0.8419 | SRCC = 0.8426 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1258 | LCC = 0.9207 | SRCC = 0.9213 ] +INFO:root:Average loss=0.07520532641464318 +INFO:root: +[Valid][19][UTT][ MSE = 0.2790 | LCC = 0.8402 | SRCC = 0.8411 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1268 | LCC = 0.9191 | SRCC = 0.9178 ] +INFO:root:Average loss=0.07073818362103058 +INFO:root: +[Valid][20][UTT][ MSE = 0.2463 | LCC = 0.8427 | SRCC = 0.8442 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1242 | LCC = 0.9280 | SRCC = 0.9298 ] +INFO:root:Average loss=0.06571514670474407 +INFO:root: +[Valid][21][UTT][ MSE = 0.2488 | LCC = 0.8432 | SRCC = 0.8428 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1160 | LCC = 0.9233 | SRCC = 0.9194 ] +INFO:root:Average loss=0.06407162809792237 +INFO:root: +[Valid][22][UTT][ MSE = 0.2760 | LCC = 0.8424 | SRCC = 0.8426 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1287 | LCC = 0.9194 | SRCC = 0.9170 ] +INFO:root:Average loss=0.059904623442353346 +INFO:root: +[Valid][23][UTT][ MSE = 0.2510 | LCC = 0.8409 | SRCC = 0.8409 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1227 | LCC = 0.9159 | SRCC = 0.9123 ] +INFO:root:Average loss=0.059470463496370196 +INFO:root: +[Valid][24][UTT][ MSE = 0.2494 | LCC = 0.8420 | SRCC = 0.8422 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1257 | LCC = 0.9185 | SRCC = 0.9148 ] +INFO:root:Average loss=0.057480395293961734 +INFO:root: +[Valid][25][UTT][ MSE = 0.2488 | LCC = 0.8415 | SRCC = 0.8416 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1326 | LCC = 0.9184 | SRCC = 0.9153 ] +INFO:root:Average loss=0.05278075271500991 +INFO:root: +[Valid][26][UTT][ MSE = 0.2679 | LCC = 0.8428 | SRCC = 0.8422 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1307 | LCC = 0.9205 | SRCC = 0.9150 ] +INFO:root:Average loss=0.05329860933125019 +INFO:root: +[Valid][27][UTT][ MSE = 0.2464 | LCC = 0.8420 | SRCC = 0.8418 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1250 | LCC = 0.9173 | SRCC = 0.9115 ] +INFO:root:Average loss=0.052810355925407164 +INFO:root: +[Valid][28][UTT][ MSE = 0.2717 | LCC = 0.8379 | SRCC = 0.8369 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1374 | LCC = 0.9131 | SRCC = 0.9044 ] +INFO:root:Average loss=0.0484368825952212 +INFO:root: +[Valid][29][UTT][ MSE = 0.2467 | LCC = 0.8439 | SRCC = 0.8435 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1210 | LCC = 0.9211 | SRCC = 0.9152 ] +INFO:root:Average loss=0.04494985735091644 +INFO:root: +[Valid][30][UTT][ MSE = 0.2502 | LCC = 0.8430 | SRCC = 0.8434 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1173 | LCC = 0.9244 | SRCC = 0.9215 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2646 | LCC = 0.8543 | SRCC = 0.8557 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1837 | LCC = 0.9090 | SRCC = 0.9077 ] diff --git a/results_bvcc/w2v2_base/layer5/run2/config.gin b/results_bvcc/w2v2_base/layer5/run2/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer5/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 diff --git a/results_bvcc/w2v2_base/layer5/run2/model.pt b/results_bvcc/w2v2_base/layer5/run2/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..586364f965629f0c005c801d551de626d06bb12e --- /dev/null +++ b/results_bvcc/w2v2_base/layer5/run2/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5cebe1d80d50fe2033efcdb0ac92c8cf71da9313094a9376094bd60ca269160d +size 2747979 diff --git a/results_bvcc/w2v2_base/layer5/run2/model_best.pt b/results_bvcc/w2v2_base/layer5/run2/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..945eaa54edbbfa2396a5b5d93b9fcee7e70e393c --- /dev/null +++ b/results_bvcc/w2v2_base/layer5/run2/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:551308909266fa52146478099de7a1291c84fc01febbd219f92e23d10cf40a21 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer5/run2/train.log b/results_bvcc/w2v2_base/layer5/run2/train.log new file mode 100644 index 0000000000000000000000000000000000000000..507711e2f680f7767db80f4d2b416b062b4a51d7 --- /dev/null +++ b/results_bvcc/w2v2_base/layer5/run2/train.log @@ -0,0 +1,163 @@ +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=0.5799742000989425 +INFO:root: +[Valid][0][UTT][ MSE = 0.4021 | LCC = 0.7800 | SRCC = 0.7871 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2986 | LCC = 0.8507 | SRCC = 0.8522 ] +INFO:root:Average loss=0.2917860427345985 +INFO:root: +[Valid][1][UTT][ MSE = 0.3038 | LCC = 0.8200 | SRCC = 0.8197 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1952 | LCC = 0.8877 | SRCC = 0.8869 ] +INFO:root:Average loss=0.2414740618222799 +INFO:root: +[Valid][2][UTT][ MSE = 0.2874 | LCC = 0.8370 | SRCC = 0.8348 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1766 | LCC = 0.9066 | SRCC = 0.9066 ] +INFO:root:Average loss=0.21003884535569411 +INFO:root: +[Valid][3][UTT][ MSE = 0.2881 | LCC = 0.8422 | SRCC = 0.8416 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1499 | LCC = 0.9161 | SRCC = 0.9148 ] +INFO:root:Average loss=0.1998395147996071 +INFO:root: +[Valid][4][UTT][ MSE = 0.2564 | LCC = 0.8454 | SRCC = 0.8445 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1406 | LCC = 0.9160 | SRCC = 0.9145 ] +INFO:root:Average loss=0.17790014468706572 +INFO:root: +[Valid][5][UTT][ MSE = 0.2572 | LCC = 0.8452 | SRCC = 0.8449 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1456 | LCC = 0.9166 | SRCC = 0.9173 ] +INFO:root:Average loss=0.1736166463830532 +INFO:root: +[Valid][6][UTT][ MSE = 0.2657 | LCC = 0.8464 | SRCC = 0.8446 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1361 | LCC = 0.9155 | SRCC = 0.9109 ] +INFO:root:Average loss=0.152551730378316 +INFO:root: +[Valid][7][UTT][ MSE = 0.2570 | LCC = 0.8419 | SRCC = 0.8410 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1333 | LCC = 0.9153 | SRCC = 0.9135 ] +INFO:root:Average loss=0.13934344836534598 +INFO:root: +[Valid][8][UTT][ MSE = 0.2502 | LCC = 0.8473 | SRCC = 0.8464 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1279 | LCC = 0.9223 | SRCC = 0.9219 ] +INFO:root:Average loss=0.13212403339835313 +INFO:root: +[Valid][9][UTT][ MSE = 0.2569 | LCC = 0.8458 | SRCC = 0.8447 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1203 | LCC = 0.9257 | SRCC = 0.9232 ] +INFO:root:Average loss=0.1260883825329634 +INFO:root: +[Valid][10][UTT][ MSE = 0.2504 | LCC = 0.8448 | SRCC = 0.8440 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1211 | LCC = 0.9181 | SRCC = 0.9161 ] +INFO:root:Average loss=0.1260175235951558 +INFO:root: +[Valid][11][UTT][ MSE = 0.2657 | LCC = 0.8447 | SRCC = 0.8443 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1677 | LCC = 0.9184 | SRCC = 0.9173 ] +INFO:root:Average loss=0.10781432353915313 +INFO:root: +[Valid][12][UTT][ MSE = 0.2420 | LCC = 0.8454 | SRCC = 0.8443 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1220 | LCC = 0.9219 | SRCC = 0.9181 ] +INFO:root:Average loss=0.0981959159939717 +INFO:root: +[Valid][13][UTT][ MSE = 0.2456 | LCC = 0.8430 | SRCC = 0.8428 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1269 | LCC = 0.9227 | SRCC = 0.9232 ] +INFO:root:Average loss=0.09300254247127435 +INFO:root: +[Valid][14][UTT][ MSE = 0.2473 | LCC = 0.8435 | SRCC = 0.8408 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1301 | LCC = 0.9215 | SRCC = 0.9161 ] +INFO:root:Average loss=0.08556896906632644 +INFO:root: +[Valid][15][UTT][ MSE = 0.2660 | LCC = 0.8387 | SRCC = 0.8392 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1579 | LCC = 0.9187 | SRCC = 0.9187 ] +INFO:root:Average loss=0.09022274508308141 +INFO:root: +[Valid][16][UTT][ MSE = 0.2535 | LCC = 0.8435 | SRCC = 0.8416 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1441 | LCC = 0.9231 | SRCC = 0.9224 ] +INFO:root:Average loss=0.07623327399293582 +INFO:root: +[Valid][17][UTT][ MSE = 0.2481 | LCC = 0.8438 | SRCC = 0.8412 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1125 | LCC = 0.9305 | SRCC = 0.9264 ] +INFO:root:Average loss=0.07697897514280601 +INFO:root: +[Valid][18][UTT][ MSE = 0.2581 | LCC = 0.8392 | SRCC = 0.8370 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1225 | LCC = 0.9269 | SRCC = 0.9259 ] +INFO:root:Average loss=0.0720921502663539 +INFO:root: +[Valid][19][UTT][ MSE = 0.2442 | LCC = 0.8418 | SRCC = 0.8403 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1100 | LCC = 0.9273 | SRCC = 0.9274 ] +INFO:root:Average loss=0.06823115619138265 +INFO:root: +[Valid][20][UTT][ MSE = 0.2523 | LCC = 0.8422 | SRCC = 0.8408 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1216 | LCC = 0.9224 | SRCC = 0.9239 ] +INFO:root:Average loss=0.06621394936855023 +INFO:root: +[Valid][21][UTT][ MSE = 0.2438 | LCC = 0.8440 | SRCC = 0.8426 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1204 | LCC = 0.9246 | SRCC = 0.9218 ] +INFO:root:Average loss=0.06224367672052139 +INFO:root: +[Valid][22][UTT][ MSE = 0.2464 | LCC = 0.8431 | SRCC = 0.8407 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1348 | LCC = 0.9195 | SRCC = 0.9164 ] +INFO:root:Average loss=0.062093708043297134 +INFO:root: +[Valid][23][UTT][ MSE = 0.2499 | LCC = 0.8422 | SRCC = 0.8403 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1379 | LCC = 0.9200 | SRCC = 0.9174 ] +INFO:root:Average loss=0.05720828148799065 +INFO:root: +[Valid][24][UTT][ MSE = 0.2480 | LCC = 0.8433 | SRCC = 0.8421 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1260 | LCC = 0.9215 | SRCC = 0.9179 ] +INFO:root:Average loss=0.05710547885451561 +INFO:root: +[Valid][25][UTT][ MSE = 0.2515 | LCC = 0.8445 | SRCC = 0.8414 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1201 | LCC = 0.9253 | SRCC = 0.9228 ] +INFO:root:Average loss=0.0530330117027729 +INFO:root: +[Valid][26][UTT][ MSE = 0.2542 | LCC = 0.8462 | SRCC = 0.8439 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1418 | LCC = 0.9282 | SRCC = 0.9270 ] +INFO:root:Average loss=0.04961686237500264 +INFO:root: +[Valid][27][UTT][ MSE = 0.2411 | LCC = 0.8495 | SRCC = 0.8478 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1177 | LCC = 0.9268 | SRCC = 0.9234 ] +INFO:root:Average loss=0.05038689915090799 +INFO:root: +[Valid][28][UTT][ MSE = 0.2454 | LCC = 0.8464 | SRCC = 0.8440 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1185 | LCC = 0.9269 | SRCC = 0.9225 ] +INFO:root:Average loss=0.05013405073147554 +INFO:root: +[Valid][29][UTT][ MSE = 0.2481 | LCC = 0.8469 | SRCC = 0.8444 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1207 | LCC = 0.9258 | SRCC = 0.9214 ] +INFO:root:Average loss=0.04676661907862394 +INFO:root: +[Valid][30][UTT][ MSE = 0.2542 | LCC = 0.8440 | SRCC = 0.8427 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1457 | LCC = 0.9241 | SRCC = 0.9227 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2572 | LCC = 0.8434 | SRCC = 0.8394 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1494 | LCC = 0.9152 | SRCC = 0.9175 ] diff --git a/results_bvcc/w2v2_base/layer5/run3/config.gin b/results_bvcc/w2v2_base/layer5/run3/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer5/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 diff --git a/results_bvcc/w2v2_base/layer5/run3/model.pt b/results_bvcc/w2v2_base/layer5/run3/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..d9a10973bc6b70a87733747d90489c851861ba59 --- /dev/null +++ b/results_bvcc/w2v2_base/layer5/run3/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:127ed34b90c484195f7b404589bc6afdc6e9efb03e4d8a82e086587e9df8de20 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer5/run3/model_best.pt b/results_bvcc/w2v2_base/layer5/run3/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..74be93ab850f3f8502b09735e8c74d459eda1327 --- /dev/null +++ b/results_bvcc/w2v2_base/layer5/run3/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d313c8f8fdd520fb873ff6f10cbdf3084053667198925969c4e9611fb049e1a5 +size 2749638 diff --git a/results_bvcc/w2v2_base/layer5/run3/train.log b/results_bvcc/w2v2_base/layer5/run3/train.log new file mode 100644 index 0000000000000000000000000000000000000000..c0c4d93bcaa44117f17ba1dc25a4504e622501fe --- /dev/null +++ b/results_bvcc/w2v2_base/layer5/run3/train.log @@ -0,0 +1,163 @@ +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=0.756594215066005 +INFO:root: +[Valid][0][UTT][ MSE = 0.4899 | LCC = 0.7613 | SRCC = 0.7633 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.3915 | LCC = 0.8287 | SRCC = 0.8321 ] +INFO:root:Average loss=0.3087254206721599 +INFO:root: +[Valid][1][UTT][ MSE = 0.3141 | LCC = 0.8219 | SRCC = 0.8261 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2175 | LCC = 0.8847 | SRCC = 0.8888 ] +INFO:root:Average loss=0.23653676742926621 +INFO:root: +[Valid][2][UTT][ MSE = 0.3360 | LCC = 0.8310 | SRCC = 0.8382 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.2503 | LCC = 0.8998 | SRCC = 0.9091 ] +INFO:root:Average loss=0.2065461447987801 +INFO:root: +[Valid][3][UTT][ MSE = 0.2713 | LCC = 0.8444 | SRCC = 0.8437 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1780 | LCC = 0.9098 | SRCC = 0.9104 ] +INFO:root:Average loss=0.1880288534821608 +INFO:root: +[Valid][4][UTT][ MSE = 0.2520 | LCC = 0.8453 | SRCC = 0.8440 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1527 | LCC = 0.9108 | SRCC = 0.9082 ] +INFO:root:Average loss=0.16750128815571466 +INFO:root: +[Valid][5][UTT][ MSE = 0.2509 | LCC = 0.8481 | SRCC = 0.8460 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1361 | LCC = 0.9224 | SRCC = 0.9216 ] +INFO:root:Average loss=0.15199235836282754 +INFO:root: +[Valid][6][UTT][ MSE = 0.2617 | LCC = 0.8465 | SRCC = 0.8447 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1404 | LCC = 0.9201 | SRCC = 0.9174 ] +INFO:root:Average loss=0.13471299237929857 +INFO:root: +[Valid][7][UTT][ MSE = 0.2720 | LCC = 0.8427 | SRCC = 0.8401 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1632 | LCC = 0.9223 | SRCC = 0.9202 ] +INFO:root:Average loss=0.13672850175927848 +INFO:root: +[Valid][8][UTT][ MSE = 0.2871 | LCC = 0.8396 | SRCC = 0.8369 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1896 | LCC = 0.9158 | SRCC = 0.9124 ] +INFO:root:Average loss=0.12199963399997124 +INFO:root: +[Valid][9][UTT][ MSE = 0.2604 | LCC = 0.8442 | SRCC = 0.8415 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1303 | LCC = 0.9234 | SRCC = 0.9182 ] +INFO:root:Average loss=0.11532726014653842 +INFO:root: +[Valid][10][UTT][ MSE = 0.2528 | LCC = 0.8432 | SRCC = 0.8409 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1243 | LCC = 0.9209 | SRCC = 0.9188 ] +INFO:root:Average loss=0.10079926281021191 +INFO:root: +[Valid][11][UTT][ MSE = 0.2579 | LCC = 0.8360 | SRCC = 0.8353 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1385 | LCC = 0.9207 | SRCC = 0.9205 ] +INFO:root:Average loss=0.09233648201020864 +INFO:root: +[Valid][12][UTT][ MSE = 0.2601 | LCC = 0.8408 | SRCC = 0.8399 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1480 | LCC = 0.9245 | SRCC = 0.9213 ] +INFO:root:Average loss=0.08630198488632838 +INFO:root: +[Valid][13][UTT][ MSE = 0.2588 | LCC = 0.8372 | SRCC = 0.8368 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1172 | LCC = 0.9255 | SRCC = 0.9236 ] +INFO:root:Average loss=0.08719470925056018 +INFO:root: +[Valid][14][UTT][ MSE = 0.2606 | LCC = 0.8409 | SRCC = 0.8390 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1527 | LCC = 0.9211 | SRCC = 0.9187 ] +INFO:root:Average loss=0.07698728187152973 +INFO:root: +[Valid][15][UTT][ MSE = 0.2545 | LCC = 0.8400 | SRCC = 0.8383 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1371 | LCC = 0.9235 | SRCC = 0.9204 ] +INFO:root:Average loss=0.07724875159179553 +INFO:root: +[Valid][16][UTT][ MSE = 0.2576 | LCC = 0.8381 | SRCC = 0.8370 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1337 | LCC = 0.9263 | SRCC = 0.9237 ] +INFO:root:Average loss=0.07191091890518482 +INFO:root: +[Valid][17][UTT][ MSE = 0.2564 | LCC = 0.8393 | SRCC = 0.8382 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1348 | LCC = 0.9270 | SRCC = 0.9241 ] +INFO:root:Average loss=0.06536132851854348 +INFO:root: +[Valid][18][UTT][ MSE = 0.2557 | LCC = 0.8394 | SRCC = 0.8377 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1313 | LCC = 0.9275 | SRCC = 0.9233 ] +INFO:root:Average loss=0.06396254715629113 +INFO:root: +[Valid][19][UTT][ MSE = 0.2590 | LCC = 0.8389 | SRCC = 0.8374 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1410 | LCC = 0.9215 | SRCC = 0.9153 ] +INFO:root:Average loss=0.06316086382437976 +INFO:root: +[Valid][20][UTT][ MSE = 0.2542 | LCC = 0.8354 | SRCC = 0.8347 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1167 | LCC = 0.9274 | SRCC = 0.9242 ] +INFO:root:Average loss=0.05734082392584055 +INFO:root: +[Valid][21][UTT][ MSE = 0.2703 | LCC = 0.8364 | SRCC = 0.8366 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1583 | LCC = 0.9229 | SRCC = 0.9221 ] +INFO:root:Average loss=0.05719747809836498 +INFO:root: +[Valid][22][UTT][ MSE = 0.2649 | LCC = 0.8363 | SRCC = 0.8369 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1389 | LCC = 0.9292 | SRCC = 0.9251 ] +INFO:root:Average loss=0.05455166435776613 +INFO:root: +[Valid][23][UTT][ MSE = 0.2707 | LCC = 0.8361 | SRCC = 0.8363 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1492 | LCC = 0.9298 | SRCC = 0.9281 ] +INFO:root:Average loss=0.049723502391806014 +INFO:root: +[Valid][24][UTT][ MSE = 0.2603 | LCC = 0.8321 | SRCC = 0.8321 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1282 | LCC = 0.9216 | SRCC = 0.9184 ] +INFO:root:Average loss=0.04996086647495245 +INFO:root: +[Valid][25][UTT][ MSE = 0.2722 | LCC = 0.8340 | SRCC = 0.8355 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1565 | LCC = 0.9240 | SRCC = 0.9220 ] +INFO:root:Average loss=0.050341144347420104 +INFO:root: +[Valid][26][UTT][ MSE = 0.2693 | LCC = 0.8361 | SRCC = 0.8353 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1546 | LCC = 0.9229 | SRCC = 0.9177 ] +INFO:root:Average loss=0.04670894126861523 +INFO:root: +[Valid][27][UTT][ MSE = 0.2743 | LCC = 0.8313 | SRCC = 0.8308 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1545 | LCC = 0.9236 | SRCC = 0.9168 ] +INFO:root:Average loss=0.04338333108581793 +INFO:root: +[Valid][28][UTT][ MSE = 0.2784 | LCC = 0.8340 | SRCC = 0.8347 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1626 | LCC = 0.9260 | SRCC = 0.9229 ] +INFO:root:Average loss=0.04315295275778343 +INFO:root: +[Valid][29][UTT][ MSE = 0.2560 | LCC = 0.8356 | SRCC = 0.8352 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1238 | LCC = 0.9264 | SRCC = 0.9236 ] +INFO:root:Average loss=0.043529249560565524 +INFO:root: +[Valid][30][UTT][ MSE = 0.2580 | LCC = 0.8327 | SRCC = 0.8322 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1274 | LCC = 0.9217 | SRCC = 0.9159 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2539 | LCC = 0.8500 | SRCC = 0.8522 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1524 | LCC = 0.9081 | SRCC = 0.9139 ] diff --git a/results_bvcc/w2v2_base/layer5/run4/config.gin b/results_bvcc/w2v2_base/layer5/run4/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer5/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 diff --git a/results_bvcc/w2v2_base/layer5/run4/model.pt b/results_bvcc/w2v2_base/layer5/run4/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..e4f08b46a43aeafba254d3ea3f0b3a98e93bf177 --- /dev/null +++ b/results_bvcc/w2v2_base/layer5/run4/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:71e77e166be02b23ffc14ea49660176b3b9004e6853e89cc9a75ebc7beb23b3c +size 2747979 diff --git a/results_bvcc/w2v2_base/layer5/run4/model_best.pt b/results_bvcc/w2v2_base/layer5/run4/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..0d7194376e12a724994762e05d94d3ca49389c74 --- /dev/null +++ b/results_bvcc/w2v2_base/layer5/run4/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e2ff254a30b7f7d73c3b4b29d05a92aa7075f85f5343065480faa6008268ce5d +size 2749702 diff --git a/results_bvcc/w2v2_base/layer5/run4/train.log b/results_bvcc/w2v2_base/layer5/run4/train.log new file mode 100644 index 0000000000000000000000000000000000000000..2e4fda9bea4cae1d08f308444ed053100ffb98ee --- /dev/null +++ b/results_bvcc/w2v2_base/layer5/run4/train.log @@ -0,0 +1,163 @@ +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=0.6242744884429834 +INFO:root: +[Valid][0][UTT][ MSE = 0.3371 | LCC = 0.7921 | SRCC = 0.7958 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2102 | LCC = 0.8584 | SRCC = 0.8626 ] +INFO:root:Average loss=0.27823792015894866 +INFO:root: +[Valid][1][UTT][ MSE = 0.3239 | LCC = 0.8300 | SRCC = 0.8286 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2010 | LCC = 0.8795 | SRCC = 0.8794 ] +INFO:root:Average loss=0.22932148571961966 +INFO:root: +[Valid][2][UTT][ MSE = 0.2731 | LCC = 0.8370 | SRCC = 0.8335 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1604 | LCC = 0.8970 | SRCC = 0.8928 ] +INFO:root:Average loss=0.20450924508846724 +INFO:root: +[Valid][3][UTT][ MSE = 0.2615 | LCC = 0.8444 | SRCC = 0.8432 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1445 | LCC = 0.9073 | SRCC = 0.9036 ] +INFO:root:Average loss=0.1840975851011582 +INFO:root: +[Valid][4][UTT][ MSE = 0.2627 | LCC = 0.8416 | SRCC = 0.8410 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1567 | LCC = 0.9113 | SRCC = 0.9113 ] +INFO:root:Average loss=0.1756601618268551 +INFO:root: +[Valid][5][UTT][ MSE = 0.2677 | LCC = 0.8472 | SRCC = 0.8462 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1479 | LCC = 0.9173 | SRCC = 0.9159 ] +INFO:root:Average loss=0.15283061363376105 +INFO:root: +[Valid][6][UTT][ MSE = 0.2383 | LCC = 0.8498 | SRCC = 0.8497 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1333 | LCC = 0.9133 | SRCC = 0.9107 ] +INFO:root:Average loss=0.14491501030249473 +INFO:root: +[Valid][7][UTT][ MSE = 0.2660 | LCC = 0.8490 | SRCC = 0.8475 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1359 | LCC = 0.9149 | SRCC = 0.9109 ] +INFO:root:Average loss=0.13402856981907135 +INFO:root: +[Valid][8][UTT][ MSE = 0.2420 | LCC = 0.8485 | SRCC = 0.8490 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1481 | LCC = 0.9119 | SRCC = 0.9126 ] +INFO:root:Average loss=0.1293724676928459 +INFO:root: +[Valid][9][UTT][ MSE = 0.2927 | LCC = 0.8446 | SRCC = 0.8446 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1501 | LCC = 0.9142 | SRCC = 0.9134 ] +INFO:root:Average loss=0.11750687123873295 +INFO:root: +[Valid][10][UTT][ MSE = 0.2516 | LCC = 0.8482 | SRCC = 0.8485 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1346 | LCC = 0.9104 | SRCC = 0.9085 ] +INFO:root:Average loss=0.11191435826894565 +INFO:root: +[Valid][11][UTT][ MSE = 0.2544 | LCC = 0.8481 | SRCC = 0.8469 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1351 | LCC = 0.9197 | SRCC = 0.9192 ] +INFO:root:Average loss=0.1067047730470315 +INFO:root: +[Valid][12][UTT][ MSE = 0.2680 | LCC = 0.8489 | SRCC = 0.8477 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1274 | LCC = 0.9218 | SRCC = 0.9196 ] +INFO:root:Average loss=0.10125979613990356 +INFO:root: +[Valid][13][UTT][ MSE = 0.2658 | LCC = 0.8486 | SRCC = 0.8486 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1361 | LCC = 0.9141 | SRCC = 0.9136 ] +INFO:root:Average loss=0.09854022093499318 +INFO:root: +[Valid][14][UTT][ MSE = 0.2394 | LCC = 0.8493 | SRCC = 0.8493 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1287 | LCC = 0.9182 | SRCC = 0.9166 ] +INFO:root:Average loss=0.0854317831973999 +INFO:root: +[Valid][15][UTT][ MSE = 0.2833 | LCC = 0.8469 | SRCC = 0.8461 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1437 | LCC = 0.9153 | SRCC = 0.9139 ] +INFO:root:Average loss=0.08403797256640899 +INFO:root: +[Valid][16][UTT][ MSE = 0.2469 | LCC = 0.8484 | SRCC = 0.8483 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1251 | LCC = 0.9154 | SRCC = 0.9152 ] +INFO:root:Average loss=0.0786969177902509 +INFO:root: +[Valid][17][UTT][ MSE = 0.2387 | LCC = 0.8486 | SRCC = 0.8481 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1224 | LCC = 0.9189 | SRCC = 0.9194 ] +INFO:root:Average loss=0.07396657416262688 +INFO:root: +[Valid][18][UTT][ MSE = 0.2610 | LCC = 0.8474 | SRCC = 0.8465 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1250 | LCC = 0.9209 | SRCC = 0.9193 ] +INFO:root:Average loss=0.06955351971854003 +INFO:root: +[Valid][19][UTT][ MSE = 0.2566 | LCC = 0.8437 | SRCC = 0.8426 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1355 | LCC = 0.9192 | SRCC = 0.9182 ] +INFO:root:Average loss=0.06807933093454593 +INFO:root: +[Valid][20][UTT][ MSE = 0.2633 | LCC = 0.8435 | SRCC = 0.8420 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1357 | LCC = 0.9114 | SRCC = 0.9093 ] +INFO:root:Average loss=0.07277613586913316 +INFO:root: +[Valid][21][UTT][ MSE = 0.2697 | LCC = 0.8468 | SRCC = 0.8460 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1391 | LCC = 0.9143 | SRCC = 0.9132 ] +INFO:root:Average loss=0.06137256954724972 +INFO:root: +[Valid][22][UTT][ MSE = 0.2422 | LCC = 0.8463 | SRCC = 0.8462 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1338 | LCC = 0.9129 | SRCC = 0.9138 ] +INFO:root:Average loss=0.05807210610080988 +INFO:root: +[Valid][23][UTT][ MSE = 0.2640 | LCC = 0.8481 | SRCC = 0.8477 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1351 | LCC = 0.9143 | SRCC = 0.9122 ] +INFO:root:Average loss=0.05795299844482006 +INFO:root: +[Valid][24][UTT][ MSE = 0.2456 | LCC = 0.8502 | SRCC = 0.8484 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1283 | LCC = 0.9193 | SRCC = 0.9180 ] +INFO:root:Average loss=0.05357959649215142 +INFO:root: +[Valid][25][UTT][ MSE = 0.2561 | LCC = 0.8449 | SRCC = 0.8442 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1229 | LCC = 0.9226 | SRCC = 0.9252 ] +INFO:root:Average loss=0.054397886380171165 +INFO:root: +[Valid][26][UTT][ MSE = 0.2411 | LCC = 0.8485 | SRCC = 0.8476 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1240 | LCC = 0.9192 | SRCC = 0.9190 ] +INFO:root:Average loss=0.050795077943266966 +INFO:root: +[Valid][27][UTT][ MSE = 0.2514 | LCC = 0.8481 | SRCC = 0.8476 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1238 | LCC = 0.9228 | SRCC = 0.9217 ] +INFO:root:Average loss=0.04788360111892988 +INFO:root: +[Valid][28][UTT][ MSE = 0.2766 | LCC = 0.8463 | SRCC = 0.8460 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1374 | LCC = 0.9157 | SRCC = 0.9149 ] +INFO:root:Average loss=0.04628427598911982 +INFO:root: +[Valid][29][UTT][ MSE = 0.2626 | LCC = 0.8476 | SRCC = 0.8470 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1277 | LCC = 0.9224 | SRCC = 0.9241 ] +INFO:root:Average loss=0.048210942353575655 +INFO:root: +[Valid][30][UTT][ MSE = 0.2435 | LCC = 0.8503 | SRCC = 0.8493 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1228 | LCC = 0.9227 | SRCC = 0.9226 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2657 | LCC = 0.8407 | SRCC = 0.8401 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1646 | LCC = 0.9060 | SRCC = 0.9107 ] diff --git a/results_bvcc/w2v2_base/layer6/run0/config.gin b/results_bvcc/w2v2_base/layer6/run0/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer6/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 diff --git a/results_bvcc/w2v2_base/layer6/run0/model.pt b/results_bvcc/w2v2_base/layer6/run0/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..3a45c02360f51bc2ec4e877ff028320c502ebd57 --- /dev/null +++ b/results_bvcc/w2v2_base/layer6/run0/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dda793eef2fb39487129f45e3e3de3834c505f72d243d12a38f1b414b0f47adf +size 2747979 diff --git a/results_bvcc/w2v2_base/layer6/run0/model_best.pt b/results_bvcc/w2v2_base/layer6/run0/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..d2c12373f9b1cd40ae5a65916d09e5dbef823983 --- /dev/null +++ b/results_bvcc/w2v2_base/layer6/run0/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cbd12dffd9d70c8ec7a356fc1ceba0a868fd82338fa0f25674acc9ff4619bca5 +size 2749638 diff --git a/results_bvcc/w2v2_base/layer6/run0/train.log b/results_bvcc/w2v2_base/layer6/run0/train.log new file mode 100644 index 0000000000000000000000000000000000000000..1ba16d570d17248a3e3fc31db01d5e48c01b6429 --- /dev/null +++ b/results_bvcc/w2v2_base/layer6/run0/train.log @@ -0,0 +1,163 @@ +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=0.6142755345656321 +INFO:root: +[Valid][0][UTT][ MSE = 0.3845 | LCC = 0.7761 | SRCC = 0.7756 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2882 | LCC = 0.8279 | SRCC = 0.8241 ] +INFO:root:Average loss=0.2968442866053337 +INFO:root: +[Valid][1][UTT][ MSE = 0.3273 | LCC = 0.8157 | SRCC = 0.8120 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2368 | LCC = 0.8727 | SRCC = 0.8654 ] +INFO:root:Average loss=0.2550998253700061 +INFO:root: +[Valid][2][UTT][ MSE = 0.3386 | LCC = 0.8242 | SRCC = 0.8267 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.2613 | LCC = 0.8846 | SRCC = 0.8860 ] +INFO:root:Average loss=0.21921340758219743 +INFO:root: +[Valid][3][UTT][ MSE = 0.2988 | LCC = 0.8326 | SRCC = 0.8325 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.2147 | LCC = 0.8957 | SRCC = 0.8914 ] +INFO:root:Average loss=0.19550346372983393 +INFO:root: +[Valid][4][UTT][ MSE = 0.2872 | LCC = 0.8376 | SRCC = 0.8384 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1924 | LCC = 0.9047 | SRCC = 0.9031 ] +INFO:root:Average loss=0.1779555648756333 +INFO:root: +[Valid][5][UTT][ MSE = 0.2809 | LCC = 0.8392 | SRCC = 0.8389 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1863 | LCC = 0.9111 | SRCC = 0.9076 ] +INFO:root:Average loss=0.17438662844972733 +INFO:root: +[Valid][6][UTT][ MSE = 0.2747 | LCC = 0.8380 | SRCC = 0.8369 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1597 | LCC = 0.9056 | SRCC = 0.8975 ] +INFO:root:Average loss=0.15484769279376054 +INFO:root: +[Valid][7][UTT][ MSE = 0.2865 | LCC = 0.8367 | SRCC = 0.8374 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1992 | LCC = 0.9058 | SRCC = 0.9014 ] +INFO:root:Average loss=0.14675012947275087 +INFO:root: +[Valid][8][UTT][ MSE = 0.3169 | LCC = 0.8347 | SRCC = 0.8377 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.2279 | LCC = 0.9096 | SRCC = 0.9072 ] +INFO:root:Average loss=0.1334612915913264 +INFO:root: +[Valid][9][UTT][ MSE = 0.3011 | LCC = 0.8344 | SRCC = 0.8335 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.2180 | LCC = 0.9028 | SRCC = 0.8941 ] +INFO:root:Average loss=0.12623011129788864 +INFO:root: +[Valid][10][UTT][ MSE = 0.2521 | LCC = 0.8378 | SRCC = 0.8363 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1364 | LCC = 0.9148 | SRCC = 0.9076 ] +INFO:root:Average loss=0.11651224251358937 +INFO:root: +[Valid][11][UTT][ MSE = 0.2811 | LCC = 0.8352 | SRCC = 0.8326 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1764 | LCC = 0.9190 | SRCC = 0.9123 ] +INFO:root:Average loss=0.10760256332846788 +INFO:root: +[Valid][12][UTT][ MSE = 0.3374 | LCC = 0.8342 | SRCC = 0.8358 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.2516 | LCC = 0.9126 | SRCC = 0.9121 ] +INFO:root:Average loss=0.10274191601918294 +INFO:root: +[Valid][13][UTT][ MSE = 0.3115 | LCC = 0.8314 | SRCC = 0.8319 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.2182 | LCC = 0.9128 | SRCC = 0.9098 ] +INFO:root:Average loss=0.09580978517157909 +INFO:root: +[Valid][14][UTT][ MSE = 0.2857 | LCC = 0.8367 | SRCC = 0.8365 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1889 | LCC = 0.9156 | SRCC = 0.9110 ] +INFO:root:Average loss=0.09109531830136593 +INFO:root: +[Valid][15][UTT][ MSE = 0.2657 | LCC = 0.8330 | SRCC = 0.8311 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1437 | LCC = 0.9150 | SRCC = 0.9050 ] +INFO:root:Average loss=0.0859944369070805 +INFO:root: +[Valid][16][UTT][ MSE = 0.2796 | LCC = 0.8353 | SRCC = 0.8338 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1746 | LCC = 0.9182 | SRCC = 0.9107 ] +INFO:root:Average loss=0.0802196588080663 +INFO:root: +[Valid][17][UTT][ MSE = 0.3109 | LCC = 0.8322 | SRCC = 0.8325 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.2135 | LCC = 0.9143 | SRCC = 0.9088 ] +INFO:root:Average loss=0.07699259456533653 +INFO:root: +[Valid][18][UTT][ MSE = 0.2779 | LCC = 0.8333 | SRCC = 0.8323 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1735 | LCC = 0.9132 | SRCC = 0.9070 ] +INFO:root:Average loss=0.07346126962548648 +INFO:root: +[Valid][19][UTT][ MSE = 0.2696 | LCC = 0.8341 | SRCC = 0.8317 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1503 | LCC = 0.9238 | SRCC = 0.9158 ] +INFO:root:Average loss=0.0706387522797554 +INFO:root: +[Valid][20][UTT][ MSE = 0.2678 | LCC = 0.8336 | SRCC = 0.8304 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1523 | LCC = 0.9203 | SRCC = 0.9125 ] +INFO:root:Average loss=0.06908275607304695 +INFO:root: +[Valid][21][UTT][ MSE = 0.2713 | LCC = 0.8317 | SRCC = 0.8311 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1524 | LCC = 0.9204 | SRCC = 0.9139 ] +INFO:root:Average loss=0.060633413875714325 +INFO:root: +[Valid][22][UTT][ MSE = 0.3036 | LCC = 0.8295 | SRCC = 0.8297 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1988 | LCC = 0.9172 | SRCC = 0.9107 ] +INFO:root:Average loss=0.062380208227878965 +INFO:root: +[Valid][23][UTT][ MSE = 0.2745 | LCC = 0.8334 | SRCC = 0.8329 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1625 | LCC = 0.9210 | SRCC = 0.9161 ] +INFO:root:Average loss=0.06279176817490505 +INFO:root: +[Valid][24][UTT][ MSE = 0.2638 | LCC = 0.8313 | SRCC = 0.8294 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1290 | LCC = 0.9250 | SRCC = 0.9138 ] +INFO:root:Average loss=0.05911798464755217 +INFO:root: +[Valid][25][UTT][ MSE = 0.2860 | LCC = 0.8269 | SRCC = 0.8246 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1702 | LCC = 0.9162 | SRCC = 0.9071 ] +INFO:root:Average loss=0.056809174995391794 +INFO:root: +[Valid][26][UTT][ MSE = 0.3114 | LCC = 0.8323 | SRCC = 0.8326 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.2032 | LCC = 0.9260 | SRCC = 0.9240 ] +INFO:root:Average loss=0.05394116430901564 +INFO:root: +[Valid][27][UTT][ MSE = 0.2764 | LCC = 0.8317 | SRCC = 0.8301 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1563 | LCC = 0.9245 | SRCC = 0.9145 ] +INFO:root:Average loss=0.05030747701246769 +INFO:root: +[Valid][28][UTT][ MSE = 0.2945 | LCC = 0.8297 | SRCC = 0.8295 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1786 | LCC = 0.9219 | SRCC = 0.9172 ] +INFO:root:Average loss=0.05010408019790283 +INFO:root: +[Valid][29][UTT][ MSE = 0.2722 | LCC = 0.8302 | SRCC = 0.8298 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1487 | LCC = 0.9223 | SRCC = 0.9150 ] +INFO:root:Average loss=0.04462467205639069 +INFO:root: +[Valid][30][UTT][ MSE = 0.2639 | LCC = 0.8306 | SRCC = 0.8283 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1291 | LCC = 0.9260 | SRCC = 0.9171 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2495 | LCC = 0.8474 | SRCC = 0.8474 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1211 | LCC = 0.9103 | SRCC = 0.9114 ] diff --git a/results_bvcc/w2v2_base/layer6/run1/config.gin b/results_bvcc/w2v2_base/layer6/run1/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer6/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 diff --git a/results_bvcc/w2v2_base/layer6/run1/model.pt b/results_bvcc/w2v2_base/layer6/run1/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..560ac7964e4053286489434afe9246a19c021021 --- /dev/null +++ b/results_bvcc/w2v2_base/layer6/run1/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cd31a2683f6e738ee74cedea98598b48d0bd5d422640432096530d0b7b071379 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer6/run1/model_best.pt b/results_bvcc/w2v2_base/layer6/run1/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..62f9fef01be645509a3a83d658b2c757fff798a7 --- /dev/null +++ b/results_bvcc/w2v2_base/layer6/run1/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4ce8c0082ac045fbfcc02407238bcdeb1fa5c0c2016f61d626e48c1b6a2fa40f +size 2749638 diff --git a/results_bvcc/w2v2_base/layer6/run1/train.log b/results_bvcc/w2v2_base/layer6/run1/train.log new file mode 100644 index 0000000000000000000000000000000000000000..22e39631eabdea2e498709413e25bc9cab7a3e64 --- /dev/null +++ b/results_bvcc/w2v2_base/layer6/run1/train.log @@ -0,0 +1,163 @@ +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=0.5580990799726584 +INFO:root: +[Valid][0][UTT][ MSE = 0.4009 | LCC = 0.7856 | SRCC = 0.7862 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.3008 | LCC = 0.8463 | SRCC = 0.8433 ] +INFO:root:Average loss=0.2924809897175202 +INFO:root: +[Valid][1][UTT][ MSE = 0.3548 | LCC = 0.8148 | SRCC = 0.8153 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2662 | LCC = 0.8802 | SRCC = 0.8781 ] +INFO:root:Average loss=0.247821477552255 +INFO:root: +[Valid][2][UTT][ MSE = 0.2875 | LCC = 0.8320 | SRCC = 0.8305 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1807 | LCC = 0.8990 | SRCC = 0.8894 ] +INFO:root:Average loss=0.22102429431218368 +INFO:root: +[Valid][3][UTT][ MSE = 0.3495 | LCC = 0.8319 | SRCC = 0.8338 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.2656 | LCC = 0.8987 | SRCC = 0.8954 ] +INFO:root:Average loss=0.20238577822844186 +INFO:root: +[Valid][4][UTT][ MSE = 0.3041 | LCC = 0.8350 | SRCC = 0.8351 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.2075 | LCC = 0.9075 | SRCC = 0.9044 ] +INFO:root:Average loss=0.17950892543945557 +INFO:root: +[Valid][5][UTT][ MSE = 0.3012 | LCC = 0.8347 | SRCC = 0.8341 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1931 | LCC = 0.9174 | SRCC = 0.9122 ] +INFO:root:Average loss=0.16318471099321658 +INFO:root: +[Valid][6][UTT][ MSE = 0.2858 | LCC = 0.8312 | SRCC = 0.8306 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1800 | LCC = 0.9131 | SRCC = 0.9086 ] +INFO:root:Average loss=0.15043960416164154 +INFO:root: +[Valid][7][UTT][ MSE = 0.2709 | LCC = 0.8355 | SRCC = 0.8340 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1584 | LCC = 0.9153 | SRCC = 0.9087 ] +INFO:root:Average loss=0.13779776380994382 +INFO:root: +[Valid][8][UTT][ MSE = 0.2796 | LCC = 0.8323 | SRCC = 0.8307 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1699 | LCC = 0.9157 | SRCC = 0.9101 ] +INFO:root:Average loss=0.12610350482356855 +INFO:root: +[Valid][9][UTT][ MSE = 0.3070 | LCC = 0.8293 | SRCC = 0.8284 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1984 | LCC = 0.9179 | SRCC = 0.9156 ] +INFO:root:Average loss=0.12092632427811623 +INFO:root: +[Valid][10][UTT][ MSE = 0.3131 | LCC = 0.8287 | SRCC = 0.8294 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.2114 | LCC = 0.9161 | SRCC = 0.9170 ] +INFO:root:Average loss=0.11831385155136769 +INFO:root: +[Valid][11][UTT][ MSE = 0.2664 | LCC = 0.8312 | SRCC = 0.8300 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1390 | LCC = 0.9174 | SRCC = 0.9097 ] +INFO:root:Average loss=0.10187734491550006 +INFO:root: +[Valid][12][UTT][ MSE = 0.2682 | LCC = 0.8273 | SRCC = 0.8259 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1401 | LCC = 0.9148 | SRCC = 0.9098 ] +INFO:root:Average loss=0.0973110672755119 +INFO:root: +[Valid][13][UTT][ MSE = 0.2766 | LCC = 0.8259 | SRCC = 0.8245 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1361 | LCC = 0.9227 | SRCC = 0.9163 ] +INFO:root:Average loss=0.0924737518414473 +INFO:root: +[Valid][14][UTT][ MSE = 0.2801 | LCC = 0.8253 | SRCC = 0.8238 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1435 | LCC = 0.9235 | SRCC = 0.9140 ] +INFO:root:Average loss=0.08497702492735325 +INFO:root: +[Valid][15][UTT][ MSE = 0.2617 | LCC = 0.8327 | SRCC = 0.8328 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1294 | LCC = 0.9265 | SRCC = 0.9228 ] +INFO:root:Average loss=0.08407594903539388 +INFO:root: +[Valid][16][UTT][ MSE = 0.3265 | LCC = 0.8278 | SRCC = 0.8287 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.2187 | LCC = 0.9192 | SRCC = 0.9163 ] +INFO:root:Average loss=0.0812506849089494 +INFO:root: +[Valid][17][UTT][ MSE = 0.2861 | LCC = 0.8273 | SRCC = 0.8270 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1740 | LCC = 0.9136 | SRCC = 0.9094 ] +INFO:root:Average loss=0.07659308574138543 +INFO:root: +[Valid][18][UTT][ MSE = 0.2745 | LCC = 0.8300 | SRCC = 0.8300 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1560 | LCC = 0.9195 | SRCC = 0.9164 ] +INFO:root:Average loss=0.07454659254887165 +INFO:root: +[Valid][19][UTT][ MSE = 0.3118 | LCC = 0.8252 | SRCC = 0.8262 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.2026 | LCC = 0.9143 | SRCC = 0.9126 ] +INFO:root:Average loss=0.06879815282538915 +INFO:root: +[Valid][20][UTT][ MSE = 0.2929 | LCC = 0.8264 | SRCC = 0.8263 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1780 | LCC = 0.9180 | SRCC = 0.9145 ] +INFO:root:Average loss=0.06742494758696128 +INFO:root: +[Valid][21][UTT][ MSE = 0.2918 | LCC = 0.8293 | SRCC = 0.8288 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1727 | LCC = 0.9225 | SRCC = 0.9182 ] +INFO:root:Average loss=0.06321026327518317 +INFO:root: +[Valid][22][UTT][ MSE = 0.3111 | LCC = 0.8278 | SRCC = 0.8283 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.2020 | LCC = 0.9191 | SRCC = 0.9138 ] +INFO:root:Average loss=0.0577335856281794 +INFO:root: +[Valid][23][UTT][ MSE = 0.2795 | LCC = 0.8294 | SRCC = 0.8283 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1640 | LCC = 0.9176 | SRCC = 0.9130 ] +INFO:root:Average loss=0.05590584391775804 +INFO:root: +[Valid][24][UTT][ MSE = 0.2822 | LCC = 0.8328 | SRCC = 0.8310 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1705 | LCC = 0.9210 | SRCC = 0.9151 ] +INFO:root:Average loss=0.05336940006758922 +INFO:root: +[Valid][25][UTT][ MSE = 0.2736 | LCC = 0.8324 | SRCC = 0.8310 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1541 | LCC = 0.9230 | SRCC = 0.9173 ] +INFO:root:Average loss=0.05054235305541601 +INFO:root: +[Valid][26][UTT][ MSE = 0.2641 | LCC = 0.8307 | SRCC = 0.8301 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1436 | LCC = 0.9134 | SRCC = 0.9085 ] +INFO:root:Average loss=0.053924188256645814 +INFO:root: +[Valid][27][UTT][ MSE = 0.3001 | LCC = 0.8274 | SRCC = 0.8274 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1877 | LCC = 0.9175 | SRCC = 0.9144 ] +INFO:root:Average loss=0.048522915643377185 +INFO:root: +[Valid][28][UTT][ MSE = 0.2785 | LCC = 0.8301 | SRCC = 0.8284 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1586 | LCC = 0.9204 | SRCC = 0.9117 ] +INFO:root:Average loss=0.04708896634670404 +INFO:root: +[Valid][29][UTT][ MSE = 0.2707 | LCC = 0.8295 | SRCC = 0.8284 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1403 | LCC = 0.9204 | SRCC = 0.9145 ] +INFO:root:Average loss=0.044791204592165276 +INFO:root: +[Valid][30][UTT][ MSE = 0.2986 | LCC = 0.8279 | SRCC = 0.8292 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1843 | LCC = 0.9172 | SRCC = 0.9153 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2531 | LCC = 0.8466 | SRCC = 0.8467 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1488 | LCC = 0.9040 | SRCC = 0.9053 ] diff --git a/results_bvcc/w2v2_base/layer6/run2/config.gin b/results_bvcc/w2v2_base/layer6/run2/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer6/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 diff --git a/results_bvcc/w2v2_base/layer6/run2/model.pt b/results_bvcc/w2v2_base/layer6/run2/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..e91a5e43242da288062a72cf3ea529a34eb88f9d --- /dev/null +++ b/results_bvcc/w2v2_base/layer6/run2/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:87de7d37a80e6796bfb29375fe5c063c923d9be8d7f32ac5e04dc0f8f028f1a8 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer6/run2/model_best.pt b/results_bvcc/w2v2_base/layer6/run2/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..c3c2d790662eecb7c7bff5f1bd185ee9c56192c3 --- /dev/null +++ b/results_bvcc/w2v2_base/layer6/run2/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d3ec547c6b398a3b90dacce780a5ff6f9d4e7c7249bfddf99b992e29e8b89045 +size 2749638 diff --git a/results_bvcc/w2v2_base/layer6/run2/train.log b/results_bvcc/w2v2_base/layer6/run2/train.log new file mode 100644 index 0000000000000000000000000000000000000000..4bb7c714b93ffa5f5e68cec42c3263beab6d5ef2 --- /dev/null +++ b/results_bvcc/w2v2_base/layer6/run2/train.log @@ -0,0 +1,163 @@ +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=0.6148088519007732 +INFO:root: +[Valid][0][UTT][ MSE = 0.4304 | LCC = 0.7457 | SRCC = 0.7628 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.3228 | LCC = 0.8167 | SRCC = 0.8270 ] +INFO:root:Average loss=0.3051064568452346 +INFO:root: +[Valid][1][UTT][ MSE = 0.3232 | LCC = 0.8130 | SRCC = 0.8169 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2292 | LCC = 0.8715 | SRCC = 0.8766 ] +INFO:root:Average loss=0.24819915493329367 +INFO:root: +[Valid][2][UTT][ MSE = 0.2963 | LCC = 0.8268 | SRCC = 0.8269 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1920 | LCC = 0.8827 | SRCC = 0.8824 ] +INFO:root:Average loss=0.21996593246093163 +INFO:root: +[Valid][3][UTT][ MSE = 0.2802 | LCC = 0.8375 | SRCC = 0.8374 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1743 | LCC = 0.9006 | SRCC = 0.9001 ] +INFO:root:Average loss=0.194535782512946 +INFO:root: +[Valid][4][UTT][ MSE = 0.2733 | LCC = 0.8351 | SRCC = 0.8349 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1455 | LCC = 0.9085 | SRCC = 0.9086 ] +INFO:root:Average loss=0.17612859358390173 +INFO:root: +[Valid][5][UTT][ MSE = 0.2927 | LCC = 0.8348 | SRCC = 0.8354 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.2079 | LCC = 0.8978 | SRCC = 0.8958 ] +INFO:root:Average loss=0.16859975065558386 +INFO:root: +[Valid][6][UTT][ MSE = 0.2742 | LCC = 0.8309 | SRCC = 0.8293 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1535 | LCC = 0.8992 | SRCC = 0.8958 ] +INFO:root:Average loss=0.1464075430845603 +INFO:root: +[Valid][7][UTT][ MSE = 0.2744 | LCC = 0.8329 | SRCC = 0.8319 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1513 | LCC = 0.9083 | SRCC = 0.9075 ] +INFO:root:Average loss=0.1382477230941638 +INFO:root: +[Valid][8][UTT][ MSE = 0.2706 | LCC = 0.8320 | SRCC = 0.8298 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1552 | LCC = 0.8999 | SRCC = 0.8935 ] +INFO:root:Average loss=0.12699348479509354 +INFO:root: +[Valid][9][UTT][ MSE = 0.2753 | LCC = 0.8345 | SRCC = 0.8340 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1701 | LCC = 0.9101 | SRCC = 0.9116 ] +INFO:root:Average loss=0.12009442148682399 +INFO:root: +[Valid][10][UTT][ MSE = 0.2652 | LCC = 0.8332 | SRCC = 0.8322 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1475 | LCC = 0.9094 | SRCC = 0.9078 ] +INFO:root:Average loss=0.1110131069062612 +INFO:root: +[Valid][11][UTT][ MSE = 0.2664 | LCC = 0.8342 | SRCC = 0.8336 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1419 | LCC = 0.9081 | SRCC = 0.9068 ] +INFO:root:Average loss=0.10226943210149422 +INFO:root: +[Valid][12][UTT][ MSE = 0.2645 | LCC = 0.8385 | SRCC = 0.8383 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1596 | LCC = 0.9130 | SRCC = 0.9130 ] +INFO:root:Average loss=0.09297347035354529 +INFO:root: +[Valid][13][UTT][ MSE = 0.2642 | LCC = 0.8332 | SRCC = 0.8319 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1524 | LCC = 0.9082 | SRCC = 0.9069 ] +INFO:root:Average loss=0.09251961269630836 +INFO:root: +[Valid][14][UTT][ MSE = 0.2669 | LCC = 0.8291 | SRCC = 0.8265 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1366 | LCC = 0.9133 | SRCC = 0.9081 ] +INFO:root:Average loss=0.08592875779439242 +INFO:root: +[Valid][15][UTT][ MSE = 0.2655 | LCC = 0.8340 | SRCC = 0.8337 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1505 | LCC = 0.9135 | SRCC = 0.9121 ] +INFO:root:Average loss=0.07729640746345887 +INFO:root: +[Valid][16][UTT][ MSE = 0.2576 | LCC = 0.8327 | SRCC = 0.8317 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1284 | LCC = 0.9138 | SRCC = 0.9113 ] +INFO:root:Average loss=0.07756937252214322 +INFO:root: +[Valid][17][UTT][ MSE = 0.2643 | LCC = 0.8350 | SRCC = 0.8346 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1508 | LCC = 0.9138 | SRCC = 0.9122 ] +INFO:root:Average loss=0.0758354917455178 +INFO:root: +[Valid][18][UTT][ MSE = 0.2639 | LCC = 0.8327 | SRCC = 0.8306 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1329 | LCC = 0.9146 | SRCC = 0.9092 ] +INFO:root:Average loss=0.06969426925747822 +INFO:root: +[Valid][19][UTT][ MSE = 0.2664 | LCC = 0.8311 | SRCC = 0.8303 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1408 | LCC = 0.9168 | SRCC = 0.9145 ] +INFO:root:Average loss=0.06590940406880318 +INFO:root: +[Valid][20][UTT][ MSE = 0.2663 | LCC = 0.8287 | SRCC = 0.8265 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1395 | LCC = 0.9084 | SRCC = 0.9016 ] +INFO:root:Average loss=0.06117814492720824 +INFO:root: +[Valid][21][UTT][ MSE = 0.2581 | LCC = 0.8352 | SRCC = 0.8346 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1347 | LCC = 0.9154 | SRCC = 0.9124 ] +INFO:root:Average loss=0.06007028246919314 +INFO:root: +[Valid][22][UTT][ MSE = 0.2656 | LCC = 0.8298 | SRCC = 0.8286 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1414 | LCC = 0.9111 | SRCC = 0.9072 ] +INFO:root:Average loss=0.05611395854980518 +INFO:root: +[Valid][23][UTT][ MSE = 0.2610 | LCC = 0.8329 | SRCC = 0.8325 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1337 | LCC = 0.9195 | SRCC = 0.9186 ] +INFO:root:Average loss=0.05435065886913202 +INFO:root: +[Valid][24][UTT][ MSE = 0.2644 | LCC = 0.8290 | SRCC = 0.8280 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1361 | LCC = 0.9137 | SRCC = 0.9110 ] +INFO:root:Average loss=0.05304712510834902 +INFO:root: +[Valid][25][UTT][ MSE = 0.2629 | LCC = 0.8325 | SRCC = 0.8322 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1409 | LCC = 0.9146 | SRCC = 0.9125 ] +INFO:root:Average loss=0.05281619383738591 +INFO:root: +[Valid][26][UTT][ MSE = 0.2646 | LCC = 0.8302 | SRCC = 0.8292 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1385 | LCC = 0.9154 | SRCC = 0.9104 ] +INFO:root:Average loss=0.048801469162870675 +INFO:root: +[Valid][27][UTT][ MSE = 0.2669 | LCC = 0.8314 | SRCC = 0.8302 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1441 | LCC = 0.9112 | SRCC = 0.9064 ] +INFO:root:Average loss=0.04623805401990047 +INFO:root: +[Valid][28][UTT][ MSE = 0.2671 | LCC = 0.8281 | SRCC = 0.8269 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1508 | LCC = 0.9055 | SRCC = 0.9015 ] +INFO:root:Average loss=0.04872281675059826 +INFO:root: +[Valid][29][UTT][ MSE = 0.2705 | LCC = 0.8294 | SRCC = 0.8284 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1507 | LCC = 0.9054 | SRCC = 0.9004 ] +INFO:root:Average loss=0.044115336635747016 +INFO:root: +[Valid][30][UTT][ MSE = 0.2648 | LCC = 0.8318 | SRCC = 0.8304 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1416 | LCC = 0.9143 | SRCC = 0.9115 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2389 | LCC = 0.8499 | SRCC = 0.8490 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1104 | LCC = 0.9187 | SRCC = 0.9191 ] diff --git a/results_bvcc/w2v2_base/layer6/run3/config.gin b/results_bvcc/w2v2_base/layer6/run3/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer6/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 diff --git a/results_bvcc/w2v2_base/layer6/run3/model.pt b/results_bvcc/w2v2_base/layer6/run3/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..d1bc048b17e47f5f6e17e087a1a2d50a3635a27b --- /dev/null +++ b/results_bvcc/w2v2_base/layer6/run3/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4097ec212063e87bc98945eb61b0ef37b86a8f5da85cc5240c25555cd9f01a37 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer6/run3/model_best.pt b/results_bvcc/w2v2_base/layer6/run3/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..e03c56f3a6b62899dd7e721978da9767bedf8d08 --- /dev/null +++ b/results_bvcc/w2v2_base/layer6/run3/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0b7335c210a2e94c4d268440f16c359657d43d4df70931762e05e911c506f606 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer6/run3/train.log b/results_bvcc/w2v2_base/layer6/run3/train.log new file mode 100644 index 0000000000000000000000000000000000000000..5224cef03a6a31f86031d773d997bcaa48805646 --- /dev/null +++ b/results_bvcc/w2v2_base/layer6/run3/train.log @@ -0,0 +1,163 @@ +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=0.6253060197983032 +INFO:root: +[Valid][0][UTT][ MSE = 0.3678 | LCC = 0.7725 | SRCC = 0.7779 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2585 | LCC = 0.8335 | SRCC = 0.8341 ] +INFO:root:Average loss=0.30844874871082795 +INFO:root: +[Valid][1][UTT][ MSE = 0.3309 | LCC = 0.8109 | SRCC = 0.8095 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2057 | LCC = 0.8759 | SRCC = 0.8723 ] +INFO:root:Average loss=0.2556785876170183 +INFO:root: +[Valid][2][UTT][ MSE = 0.3037 | LCC = 0.8285 | SRCC = 0.8283 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1921 | LCC = 0.9007 | SRCC = 0.8988 ] +INFO:root:Average loss=0.22205495547789794 +INFO:root: +[Valid][3][UTT][ MSE = 0.2887 | LCC = 0.8324 | SRCC = 0.8308 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1584 | LCC = 0.9048 | SRCC = 0.8999 ] +INFO:root:Average loss=0.20399859490302893 +INFO:root: +[Valid][4][UTT][ MSE = 0.2720 | LCC = 0.8397 | SRCC = 0.8378 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1487 | LCC = 0.9109 | SRCC = 0.9060 ] +INFO:root:Average loss=0.17854991239997056 +INFO:root: +[Valid][5][UTT][ MSE = 0.2740 | LCC = 0.8375 | SRCC = 0.8375 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1666 | LCC = 0.9128 | SRCC = 0.9081 ] +INFO:root:Average loss=0.17516542245180178 +INFO:root: +[Valid][6][UTT][ MSE = 0.2497 | LCC = 0.8409 | SRCC = 0.8406 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1307 | LCC = 0.9172 | SRCC = 0.9137 ] +INFO:root:Average loss=0.15547568026261452 +INFO:root: +[Valid][7][UTT][ MSE = 0.2614 | LCC = 0.8387 | SRCC = 0.8386 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1309 | LCC = 0.9226 | SRCC = 0.9197 ] +INFO:root:Average loss=0.1478640079880372 +INFO:root: +[Valid][8][UTT][ MSE = 0.2652 | LCC = 0.8431 | SRCC = 0.8417 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1378 | LCC = 0.9219 | SRCC = 0.9158 ] +INFO:root:Average loss=0.13735900389460418 +INFO:root: +[Valid][9][UTT][ MSE = 0.2812 | LCC = 0.8376 | SRCC = 0.8374 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1501 | LCC = 0.9192 | SRCC = 0.9141 ] +INFO:root:Average loss=0.12566122718346426 +INFO:root: +[Valid][10][UTT][ MSE = 0.2680 | LCC = 0.8407 | SRCC = 0.8394 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1556 | LCC = 0.9232 | SRCC = 0.9197 ] +INFO:root:Average loss=0.12023390218233451 +INFO:root: +[Valid][11][UTT][ MSE = 0.2555 | LCC = 0.8455 | SRCC = 0.8451 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1302 | LCC = 0.9308 | SRCC = 0.9285 ] +INFO:root:Average loss=0.10644523694347112 +INFO:root: +[Valid][12][UTT][ MSE = 0.2606 | LCC = 0.8389 | SRCC = 0.8391 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1274 | LCC = 0.9329 | SRCC = 0.9305 ] +INFO:root:Average loss=0.09991747265060742 +INFO:root: +[Valid][13][UTT][ MSE = 0.2779 | LCC = 0.8385 | SRCC = 0.8403 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1609 | LCC = 0.9272 | SRCC = 0.9279 ] +INFO:root:Average loss=0.09579188128312428 +INFO:root: +[Valid][14][UTT][ MSE = 0.2498 | LCC = 0.8425 | SRCC = 0.8419 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1286 | LCC = 0.9255 | SRCC = 0.9202 ] +INFO:root:Average loss=0.09460110102708523 +INFO:root: +[Valid][15][UTT][ MSE = 0.2580 | LCC = 0.8396 | SRCC = 0.8393 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1260 | LCC = 0.9260 | SRCC = 0.9216 ] +INFO:root:Average loss=0.08349252215180641 +INFO:root: +[Valid][16][UTT][ MSE = 0.2689 | LCC = 0.8356 | SRCC = 0.8367 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1385 | LCC = 0.9273 | SRCC = 0.9248 ] +INFO:root:Average loss=0.08226058130654004 +INFO:root: +[Valid][17][UTT][ MSE = 0.2605 | LCC = 0.8398 | SRCC = 0.8386 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1266 | LCC = 0.9323 | SRCC = 0.9284 ] +INFO:root:Average loss=0.0769474329188084 +INFO:root: +[Valid][18][UTT][ MSE = 0.2670 | LCC = 0.8364 | SRCC = 0.8348 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1396 | LCC = 0.9279 | SRCC = 0.9249 ] +INFO:root:Average loss=0.07554740052765761 +INFO:root: +[Valid][19][UTT][ MSE = 0.2577 | LCC = 0.8387 | SRCC = 0.8377 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1226 | LCC = 0.9298 | SRCC = 0.9242 ] +INFO:root:Average loss=0.06693541124845162 +INFO:root: +[Valid][20][UTT][ MSE = 0.2545 | LCC = 0.8406 | SRCC = 0.8403 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1239 | LCC = 0.9252 | SRCC = 0.9221 ] +INFO:root:Average loss=0.06550745895275703 +INFO:root: +[Valid][21][UTT][ MSE = 0.2716 | LCC = 0.8364 | SRCC = 0.8350 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1195 | LCC = 0.9309 | SRCC = 0.9257 ] +INFO:root:Average loss=0.06470328292403466 +INFO:root: +[Valid][22][UTT][ MSE = 0.2677 | LCC = 0.8356 | SRCC = 0.8340 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1326 | LCC = 0.9286 | SRCC = 0.9210 ] +INFO:root:Average loss=0.06058989298076202 +INFO:root: +[Valid][23][UTT][ MSE = 0.2646 | LCC = 0.8364 | SRCC = 0.8348 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1307 | LCC = 0.9252 | SRCC = 0.9204 ] +INFO:root:Average loss=0.05774165126375663 +INFO:root: +[Valid][24][UTT][ MSE = 0.2561 | LCC = 0.8392 | SRCC = 0.8377 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1206 | LCC = 0.9276 | SRCC = 0.9227 ] +INFO:root:Average loss=0.05572376013375246 +INFO:root: +[Valid][25][UTT][ MSE = 0.2763 | LCC = 0.8344 | SRCC = 0.8320 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1391 | LCC = 0.9216 | SRCC = 0.9126 ] +INFO:root:Average loss=0.06411140283139852 +INFO:root: +[Valid][26][UTT][ MSE = 0.2623 | LCC = 0.8385 | SRCC = 0.8377 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1335 | LCC = 0.9263 | SRCC = 0.9200 ] +INFO:root:Average loss=0.05482687046512579 +INFO:root: +[Valid][27][UTT][ MSE = 0.2735 | LCC = 0.8378 | SRCC = 0.8372 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1280 | LCC = 0.9222 | SRCC = 0.9161 ] +INFO:root:Average loss=0.04860908237214272 +INFO:root: +[Valid][28][UTT][ MSE = 0.2618 | LCC = 0.8397 | SRCC = 0.8388 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1416 | LCC = 0.9259 | SRCC = 0.9204 ] +INFO:root:Average loss=0.049004687975423455 +INFO:root: +[Valid][29][UTT][ MSE = 0.2611 | LCC = 0.8369 | SRCC = 0.8371 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1413 | LCC = 0.9218 | SRCC = 0.9185 ] +INFO:root:Average loss=0.04797896556556225 +INFO:root: +[Valid][30][UTT][ MSE = 0.2599 | LCC = 0.8404 | SRCC = 0.8402 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1289 | LCC = 0.9275 | SRCC = 0.9236 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2661 | LCC = 0.8415 | SRCC = 0.8412 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1570 | LCC = 0.9128 | SRCC = 0.9148 ] diff --git a/results_bvcc/w2v2_base/layer6/run4/config.gin b/results_bvcc/w2v2_base/layer6/run4/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer6/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 diff --git a/results_bvcc/w2v2_base/layer6/run4/model.pt b/results_bvcc/w2v2_base/layer6/run4/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..c0a1430f79ac6d988afe5b880a83da67953b1ae8 --- /dev/null +++ b/results_bvcc/w2v2_base/layer6/run4/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fc2ec5c70a430e135d37741acbc407900721ba1b176f34a2b0e7d06acbf7d7db +size 2747979 diff --git a/results_bvcc/w2v2_base/layer6/run4/model_best.pt b/results_bvcc/w2v2_base/layer6/run4/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..b429e34597b6ab85989cac2cb6d5cbe7fbd08929 --- /dev/null +++ b/results_bvcc/w2v2_base/layer6/run4/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ed771bd16a518830e0ec315c56013c5154df4862eadf725f7628e4b67d8107d7 +size 2749638 diff --git a/results_bvcc/w2v2_base/layer6/run4/train.log b/results_bvcc/w2v2_base/layer6/run4/train.log new file mode 100644 index 0000000000000000000000000000000000000000..eb3ea86ff8bfcf9119293f9ca9072147c958d88b --- /dev/null +++ b/results_bvcc/w2v2_base/layer6/run4/train.log @@ -0,0 +1,163 @@ +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=0.5963552355384215 +INFO:root: +[Valid][0][UTT][ MSE = 0.4000 | LCC = 0.7651 | SRCC = 0.7772 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2908 | LCC = 0.8409 | SRCC = 0.8496 ] +INFO:root:Average loss=0.28679868521598667 +INFO:root: +[Valid][1][UTT][ MSE = 0.3094 | LCC = 0.8142 | SRCC = 0.8165 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2004 | LCC = 0.8808 | SRCC = 0.8798 ] +INFO:root:Average loss=0.23730231210207328 +INFO:root: +[Valid][2][UTT][ MSE = 0.3887 | LCC = 0.8207 | SRCC = 0.8282 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.3107 | LCC = 0.8852 | SRCC = 0.8891 ] +INFO:root:Average loss=0.21367023388544717 +INFO:root: +[Valid][3][UTT][ MSE = 0.2684 | LCC = 0.8385 | SRCC = 0.8384 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1569 | LCC = 0.9084 | SRCC = 0.9042 ] +INFO:root:Average loss=0.19115894669905686 +INFO:root: +[Valid][4][UTT][ MSE = 0.2743 | LCC = 0.8380 | SRCC = 0.8386 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1561 | LCC = 0.9087 | SRCC = 0.9055 ] +INFO:root:Average loss=0.17733120603057054 +INFO:root: +[Valid][5][UTT][ MSE = 0.2615 | LCC = 0.8398 | SRCC = 0.8403 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1432 | LCC = 0.9124 | SRCC = 0.9107 ] +INFO:root:Average loss=0.1701529005017036 +INFO:root: +[Valid][6][UTT][ MSE = 0.2729 | LCC = 0.8387 | SRCC = 0.8380 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1581 | LCC = 0.9131 | SRCC = 0.9090 ] +INFO:root:Average loss=0.14848821486035982 +INFO:root: +[Valid][7][UTT][ MSE = 0.2728 | LCC = 0.8357 | SRCC = 0.8363 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1611 | LCC = 0.9141 | SRCC = 0.9102 ] +INFO:root:Average loss=0.13683981715868682 +INFO:root: +[Valid][8][UTT][ MSE = 0.2913 | LCC = 0.8368 | SRCC = 0.8373 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1826 | LCC = 0.9188 | SRCC = 0.9174 ] +INFO:root:Average loss=0.13097603848347297 +INFO:root: +[Valid][9][UTT][ MSE = 0.2769 | LCC = 0.8351 | SRCC = 0.8382 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1694 | LCC = 0.9125 | SRCC = 0.9127 ] +INFO:root:Average loss=0.11670426489450993 +INFO:root: +[Valid][10][UTT][ MSE = 0.2673 | LCC = 0.8346 | SRCC = 0.8363 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1514 | LCC = 0.9131 | SRCC = 0.9110 ] +INFO:root:Average loss=0.11439765741427739 +INFO:root: +[Valid][11][UTT][ MSE = 0.2874 | LCC = 0.8378 | SRCC = 0.8374 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1502 | LCC = 0.9161 | SRCC = 0.9099 ] +INFO:root:Average loss=0.10493602627554001 +INFO:root: +[Valid][12][UTT][ MSE = 0.2592 | LCC = 0.8355 | SRCC = 0.8353 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1355 | LCC = 0.9178 | SRCC = 0.9162 ] +INFO:root:Average loss=0.09973463100882676 +INFO:root: +[Valid][13][UTT][ MSE = 0.2782 | LCC = 0.8334 | SRCC = 0.8354 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1705 | LCC = 0.9152 | SRCC = 0.9151 ] +INFO:root:Average loss=0.09029431879902497 +INFO:root: +[Valid][14][UTT][ MSE = 0.2763 | LCC = 0.8313 | SRCC = 0.8312 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1500 | LCC = 0.9234 | SRCC = 0.9216 ] +INFO:root:Average loss=0.0845990947041756 +INFO:root: +[Valid][15][UTT][ MSE = 0.2675 | LCC = 0.8329 | SRCC = 0.8348 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1446 | LCC = 0.9186 | SRCC = 0.9157 ] +INFO:root:Average loss=0.0832715008694392 +INFO:root: +[Valid][16][UTT][ MSE = 0.2680 | LCC = 0.8328 | SRCC = 0.8352 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1415 | LCC = 0.9162 | SRCC = 0.9123 ] +INFO:root:Average loss=0.07774784893561633 +INFO:root: +[Valid][17][UTT][ MSE = 0.2694 | LCC = 0.8314 | SRCC = 0.8333 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1477 | LCC = 0.9202 | SRCC = 0.9194 ] +INFO:root:Average loss=0.07409244503539342 +INFO:root: +[Valid][18][UTT][ MSE = 0.2708 | LCC = 0.8329 | SRCC = 0.8345 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1488 | LCC = 0.9217 | SRCC = 0.9207 ] +INFO:root:Average loss=0.06987095151383144 +INFO:root: +[Valid][19][UTT][ MSE = 0.2727 | LCC = 0.8299 | SRCC = 0.8300 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1445 | LCC = 0.9218 | SRCC = 0.9151 ] +INFO:root:Average loss=0.06661270597042182 +INFO:root: +[Valid][20][UTT][ MSE = 0.2784 | LCC = 0.8289 | SRCC = 0.8295 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1380 | LCC = 0.9227 | SRCC = 0.9191 ] +INFO:root:Average loss=0.0642196020254722 +INFO:root: +[Valid][21][UTT][ MSE = 0.2676 | LCC = 0.8319 | SRCC = 0.8339 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1385 | LCC = 0.9187 | SRCC = 0.9164 ] +INFO:root:Average loss=0.060042137518907204 +INFO:root: +[Valid][22][UTT][ MSE = 0.2788 | LCC = 0.8273 | SRCC = 0.8271 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1491 | LCC = 0.9145 | SRCC = 0.9080 ] +INFO:root:Average loss=0.06390066163089031 +INFO:root: +[Valid][23][UTT][ MSE = 0.2968 | LCC = 0.8364 | SRCC = 0.8388 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1518 | LCC = 0.9230 | SRCC = 0.9199 ] +INFO:root:Average loss=0.06161712511227681 +INFO:root: +[Valid][24][UTT][ MSE = 0.2695 | LCC = 0.8358 | SRCC = 0.8375 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1474 | LCC = 0.9199 | SRCC = 0.9168 ] +INFO:root:Average loss=0.05450218509978209 +INFO:root: +[Valid][25][UTT][ MSE = 0.2840 | LCC = 0.8316 | SRCC = 0.8345 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1679 | LCC = 0.9199 | SRCC = 0.9193 ] +INFO:root:Average loss=0.05524119596259716 +INFO:root: +[Valid][26][UTT][ MSE = 0.2685 | LCC = 0.8323 | SRCC = 0.8335 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1424 | LCC = 0.9182 | SRCC = 0.9122 ] +INFO:root:Average loss=0.05104101731035954 +INFO:root: +[Valid][27][UTT][ MSE = 0.2722 | LCC = 0.8305 | SRCC = 0.8310 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1424 | LCC = 0.9227 | SRCC = 0.9203 ] +INFO:root:Average loss=0.04708080800871054 +INFO:root: +[Valid][28][UTT][ MSE = 0.2698 | LCC = 0.8322 | SRCC = 0.8341 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1427 | LCC = 0.9218 | SRCC = 0.9183 ] +INFO:root:Average loss=0.046625118129528485 +INFO:root: +[Valid][29][UTT][ MSE = 0.2679 | LCC = 0.8345 | SRCC = 0.8371 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1483 | LCC = 0.9206 | SRCC = 0.9170 ] +INFO:root:Average loss=0.043330371547012754 +INFO:root: +[Valid][30][UTT][ MSE = 0.2664 | LCC = 0.8323 | SRCC = 0.8341 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1404 | LCC = 0.9204 | SRCC = 0.9162 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2602 | LCC = 0.8496 | SRCC = 0.8494 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1587 | LCC = 0.9163 | SRCC = 0.9173 ] diff --git a/results_bvcc/w2v2_base/layer7/run0/config.gin b/results_bvcc/w2v2_base/layer7/run0/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer7/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 diff --git a/results_bvcc/w2v2_base/layer7/run0/model.pt b/results_bvcc/w2v2_base/layer7/run0/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..3b7374ae2397a4f97e8bf43d616adbfac465452d --- /dev/null +++ b/results_bvcc/w2v2_base/layer7/run0/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:919865caefb4fcf56fee9a2ca9c53e31030af2193ba7def746ede39e90a5537d +size 2747979 diff --git a/results_bvcc/w2v2_base/layer7/run0/model_best.pt b/results_bvcc/w2v2_base/layer7/run0/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..5d5aebd8a1cf8bdf69244f8ac03f7ba156f8d101 --- /dev/null +++ b/results_bvcc/w2v2_base/layer7/run0/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7105d4c8028c0fc880c17d31a4f366e5982c0d4f76fdafafd1c238c50a445eb8 +size 2749638 diff --git a/results_bvcc/w2v2_base/layer7/run0/train.log b/results_bvcc/w2v2_base/layer7/run0/train.log new file mode 100644 index 0000000000000000000000000000000000000000..825db2c12e543697449e8c028d18ed4aaa7ec8a7 --- /dev/null +++ b/results_bvcc/w2v2_base/layer7/run0/train.log @@ -0,0 +1,163 @@ +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=0.5978212465460484 +INFO:root: +[Valid][0][UTT][ MSE = 0.4081 | LCC = 0.7677 | SRCC = 0.7807 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2989 | LCC = 0.8367 | SRCC = 0.8461 ] +INFO:root:Average loss=0.303463050761284 +INFO:root: +[Valid][1][UTT][ MSE = 0.3102 | LCC = 0.8074 | SRCC = 0.8082 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2005 | LCC = 0.8698 | SRCC = 0.8666 ] +INFO:root:Average loss=0.2511971122943438 +INFO:root: +[Valid][2][UTT][ MSE = 0.3019 | LCC = 0.8215 | SRCC = 0.8246 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.2044 | LCC = 0.8849 | SRCC = 0.8834 ] +INFO:root:Average loss=0.22302705011306664 +INFO:root: +[Valid][3][UTT][ MSE = 0.2737 | LCC = 0.8289 | SRCC = 0.8286 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1659 | LCC = 0.8903 | SRCC = 0.8838 ] +INFO:root:Average loss=0.19639931733791643 +INFO:root: +[Valid][4][UTT][ MSE = 0.2821 | LCC = 0.8274 | SRCC = 0.8259 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1873 | LCC = 0.8895 | SRCC = 0.8816 ] +INFO:root:Average loss=0.18285693486149496 +INFO:root: +[Valid][5][UTT][ MSE = 0.2706 | LCC = 0.8325 | SRCC = 0.8316 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1710 | LCC = 0.8958 | SRCC = 0.8896 ] +INFO:root:Average loss=0.16687579777760383 +INFO:root: +[Valid][6][UTT][ MSE = 0.2605 | LCC = 0.8379 | SRCC = 0.8371 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1406 | LCC = 0.9099 | SRCC = 0.9049 ] +INFO:root:Average loss=0.1591080631583165 +INFO:root: +[Valid][7][UTT][ MSE = 0.2723 | LCC = 0.8314 | SRCC = 0.8333 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1740 | LCC = 0.8998 | SRCC = 0.8981 ] +INFO:root:Average loss=0.14707918570209771 +INFO:root: +[Valid][8][UTT][ MSE = 0.2701 | LCC = 0.8331 | SRCC = 0.8312 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1656 | LCC = 0.9003 | SRCC = 0.8915 ] +INFO:root:Average loss=0.12953216706713042 +INFO:root: +[Valid][9][UTT][ MSE = 0.2611 | LCC = 0.8335 | SRCC = 0.8330 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1479 | LCC = 0.9039 | SRCC = 0.9003 ] +INFO:root:Average loss=0.12520346847864297 +INFO:root: +[Valid][10][UTT][ MSE = 0.2760 | LCC = 0.8277 | SRCC = 0.8276 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1822 | LCC = 0.8950 | SRCC = 0.8896 ] +INFO:root:Average loss=0.11418244720269473 +INFO:root: +[Valid][11][UTT][ MSE = 0.2661 | LCC = 0.8282 | SRCC = 0.8272 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1533 | LCC = 0.9028 | SRCC = 0.8945 ] +INFO:root:Average loss=0.1116176066108239 +INFO:root: +[Valid][12][UTT][ MSE = 0.2625 | LCC = 0.8326 | SRCC = 0.8310 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1560 | LCC = 0.9044 | SRCC = 0.8955 ] +INFO:root:Average loss=0.10481898094981144 +INFO:root: +[Valid][13][UTT][ MSE = 0.2646 | LCC = 0.8283 | SRCC = 0.8260 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1496 | LCC = 0.9000 | SRCC = 0.8873 ] +INFO:root:Average loss=0.09592643981942764 +INFO:root: +[Valid][14][UTT][ MSE = 0.2789 | LCC = 0.8277 | SRCC = 0.8278 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1837 | LCC = 0.8997 | SRCC = 0.8931 ] +INFO:root:Average loss=0.08931358459477241 +INFO:root: +[Valid][15][UTT][ MSE = 0.2711 | LCC = 0.8261 | SRCC = 0.8244 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1616 | LCC = 0.8990 | SRCC = 0.8898 ] +INFO:root:Average loss=0.09198857213442142 +INFO:root: +[Valid][16][UTT][ MSE = 0.2632 | LCC = 0.8283 | SRCC = 0.8261 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1554 | LCC = 0.8995 | SRCC = 0.8913 ] +INFO:root:Average loss=0.08289125069784813 +INFO:root: +[Valid][17][UTT][ MSE = 0.2694 | LCC = 0.8291 | SRCC = 0.8277 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1363 | LCC = 0.9106 | SRCC = 0.9053 ] +INFO:root:Average loss=0.0767666227542437 +INFO:root: +[Valid][18][UTT][ MSE = 0.2700 | LCC = 0.8267 | SRCC = 0.8253 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1574 | LCC = 0.9030 | SRCC = 0.8985 ] +INFO:root:Average loss=0.07800056861761288 +INFO:root: +[Valid][19][UTT][ MSE = 0.2855 | LCC = 0.8236 | SRCC = 0.8220 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1829 | LCC = 0.9017 | SRCC = 0.8932 ] +INFO:root:Average loss=0.0812075703094403 +INFO:root: +[Valid][20][UTT][ MSE = 0.2829 | LCC = 0.8253 | SRCC = 0.8248 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1825 | LCC = 0.9024 | SRCC = 0.8979 ] +INFO:root:Average loss=0.06885565067522037 +INFO:root: +[Valid][21][UTT][ MSE = 0.2788 | LCC = 0.8251 | SRCC = 0.8242 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1714 | LCC = 0.9056 | SRCC = 0.8979 ] +INFO:root:Average loss=0.06562454778796588 +INFO:root: +[Valid][22][UTT][ MSE = 0.2909 | LCC = 0.8219 | SRCC = 0.8190 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1814 | LCC = 0.9087 | SRCC = 0.9039 ] +INFO:root:Average loss=0.06368741197272754 +INFO:root: +[Valid][23][UTT][ MSE = 0.2679 | LCC = 0.8272 | SRCC = 0.8243 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1463 | LCC = 0.9084 | SRCC = 0.8987 ] +INFO:root:Average loss=0.06033387894813831 +INFO:root: +[Valid][24][UTT][ MSE = 0.2732 | LCC = 0.8256 | SRCC = 0.8230 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1623 | LCC = 0.9060 | SRCC = 0.8968 ] +INFO:root:Average loss=0.0570622846388664 +INFO:root: +[Valid][25][UTT][ MSE = 0.2847 | LCC = 0.8228 | SRCC = 0.8211 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1770 | LCC = 0.9047 | SRCC = 0.8976 ] +INFO:root:Average loss=0.05616430088113516 +INFO:root: +[Valid][26][UTT][ MSE = 0.2700 | LCC = 0.8283 | SRCC = 0.8253 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1449 | LCC = 0.9088 | SRCC = 0.8965 ] +INFO:root:Average loss=0.05522516441459839 +INFO:root: +[Valid][27][UTT][ MSE = 0.2609 | LCC = 0.8313 | SRCC = 0.8287 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1329 | LCC = 0.9143 | SRCC = 0.9059 ] +INFO:root:Average loss=0.050518371331959196 +INFO:root: +[Valid][28][UTT][ MSE = 0.2721 | LCC = 0.8245 | SRCC = 0.8223 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1576 | LCC = 0.9030 | SRCC = 0.8931 ] +INFO:root:Average loss=0.05198379496160226 +INFO:root: +[Valid][29][UTT][ MSE = 0.2779 | LCC = 0.8272 | SRCC = 0.8245 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1686 | LCC = 0.9097 | SRCC = 0.9011 ] +INFO:root:Average loss=0.05053733881467428 +INFO:root: +[Valid][30][UTT][ MSE = 0.2662 | LCC = 0.8295 | SRCC = 0.8279 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1498 | LCC = 0.9064 | SRCC = 0.8970 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2734 | LCC = 0.8426 | SRCC = 0.8449 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1659 | LCC = 0.9168 | SRCC = 0.9197 ] diff --git a/results_bvcc/w2v2_base/layer7/run1/config.gin b/results_bvcc/w2v2_base/layer7/run1/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer7/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 diff --git a/results_bvcc/w2v2_base/layer7/run1/model.pt b/results_bvcc/w2v2_base/layer7/run1/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..a82e29faa1f6534cbfbe764ad59714912a12b75d --- /dev/null +++ b/results_bvcc/w2v2_base/layer7/run1/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cd9f0337337c5542ced8c4bdd84ed97623e9db1a9a84115bc407ed967805a60d +size 2747979 diff --git a/results_bvcc/w2v2_base/layer7/run1/model_best.pt b/results_bvcc/w2v2_base/layer7/run1/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..4439bf9afb286793c9d0d0cfdd1a08e4f023f04f --- /dev/null +++ b/results_bvcc/w2v2_base/layer7/run1/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:da37317e9f599561d21c18baaab3d1d9d6c226ed33e300202d9b652e040da983 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer7/run1/train.log b/results_bvcc/w2v2_base/layer7/run1/train.log new file mode 100644 index 0000000000000000000000000000000000000000..7ac24cd1d7a134b4660555c86b68cf0992c55580 --- /dev/null +++ b/results_bvcc/w2v2_base/layer7/run1/train.log @@ -0,0 +1,163 @@ +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=0.5258757166373424 +INFO:root: +[Valid][0][UTT][ MSE = 0.3487 | LCC = 0.7866 | SRCC = 0.7886 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2195 | LCC = 0.8677 | SRCC = 0.8700 ] +INFO:root:Average loss=0.28900523713001836 +INFO:root: +[Valid][1][UTT][ MSE = 0.3128 | LCC = 0.8172 | SRCC = 0.8192 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2046 | LCC = 0.8871 | SRCC = 0.8864 ] +INFO:root:Average loss=0.24587987172297943 +INFO:root: +[Valid][2][UTT][ MSE = 0.3325 | LCC = 0.8265 | SRCC = 0.8296 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.2255 | LCC = 0.8984 | SRCC = 0.8977 ] +INFO:root:Average loss=0.22315142341913322 +INFO:root: +[Valid][3][UTT][ MSE = 0.2789 | LCC = 0.8339 | SRCC = 0.8347 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1820 | LCC = 0.8979 | SRCC = 0.8957 ] +INFO:root:Average loss=0.19970538180608016 +INFO:root: +[Valid][4][UTT][ MSE = 0.3161 | LCC = 0.8251 | SRCC = 0.8245 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.2162 | LCC = 0.8996 | SRCC = 0.8952 ] +INFO:root:Average loss=0.1876436809125619 +INFO:root: +[Valid][5][UTT][ MSE = 0.3393 | LCC = 0.8302 | SRCC = 0.8324 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.2514 | LCC = 0.9005 | SRCC = 0.8996 ] +INFO:root:Average loss=0.17231222948966882 +INFO:root: +[Valid][6][UTT][ MSE = 0.2543 | LCC = 0.8387 | SRCC = 0.8394 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1429 | LCC = 0.9061 | SRCC = 0.9031 ] +INFO:root:Average loss=0.1657789025742274 +INFO:root: +[Valid][7][UTT][ MSE = 0.2899 | LCC = 0.8347 | SRCC = 0.8340 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1886 | LCC = 0.9056 | SRCC = 0.9023 ] +INFO:root:Average loss=0.1425107805392681 +INFO:root: +[Valid][8][UTT][ MSE = 0.2695 | LCC = 0.8368 | SRCC = 0.8365 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1769 | LCC = 0.9008 | SRCC = 0.8961 ] +INFO:root:Average loss=0.13217434430351624 +INFO:root: +[Valid][9][UTT][ MSE = 0.2669 | LCC = 0.8344 | SRCC = 0.8333 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1661 | LCC = 0.9023 | SRCC = 0.8953 ] +INFO:root:Average loss=0.12174825227031341 +INFO:root: +[Valid][10][UTT][ MSE = 0.2847 | LCC = 0.8305 | SRCC = 0.8295 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1864 | LCC = 0.8984 | SRCC = 0.8969 ] +INFO:root:Average loss=0.12135265824886468 +INFO:root: +[Valid][11][UTT][ MSE = 0.2815 | LCC = 0.8278 | SRCC = 0.8258 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1655 | LCC = 0.8944 | SRCC = 0.8852 ] +INFO:root:Average loss=0.11416030617860648 +INFO:root: +[Valid][12][UTT][ MSE = 0.3036 | LCC = 0.8283 | SRCC = 0.8255 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.2042 | LCC = 0.9002 | SRCC = 0.8987 ] +INFO:root:Average loss=0.10353490786674695 +INFO:root: +[Valid][13][UTT][ MSE = 0.2784 | LCC = 0.8308 | SRCC = 0.8285 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1789 | LCC = 0.8995 | SRCC = 0.8968 ] +INFO:root:Average loss=0.10067092694151096 +INFO:root: +[Valid][14][UTT][ MSE = 0.2919 | LCC = 0.8279 | SRCC = 0.8269 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1967 | LCC = 0.8961 | SRCC = 0.8966 ] +INFO:root:Average loss=0.08513917009799908 +INFO:root: +[Valid][15][UTT][ MSE = 0.2858 | LCC = 0.8249 | SRCC = 0.8231 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1926 | LCC = 0.8900 | SRCC = 0.8866 ] +INFO:root:Average loss=0.08916611606493974 +INFO:root: +[Valid][16][UTT][ MSE = 0.2870 | LCC = 0.8281 | SRCC = 0.8274 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1910 | LCC = 0.8962 | SRCC = 0.8959 ] +INFO:root:Average loss=0.08755979042213696 +INFO:root: +[Valid][17][UTT][ MSE = 0.2865 | LCC = 0.8305 | SRCC = 0.8286 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1875 | LCC = 0.9010 | SRCC = 0.8994 ] +INFO:root:Average loss=0.08293805734660381 +INFO:root: +[Valid][18][UTT][ MSE = 0.2623 | LCC = 0.8294 | SRCC = 0.8277 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1478 | LCC = 0.8974 | SRCC = 0.8930 ] +INFO:root:Average loss=0.07417367030985844 +INFO:root: +[Valid][19][UTT][ MSE = 0.2875 | LCC = 0.8276 | SRCC = 0.8259 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1927 | LCC = 0.8939 | SRCC = 0.8911 ] +INFO:root:Average loss=0.0729318384367686 +INFO:root: +[Valid][20][UTT][ MSE = 0.2649 | LCC = 0.8313 | SRCC = 0.8290 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1513 | LCC = 0.9031 | SRCC = 0.8982 ] +INFO:root:Average loss=0.06625771656250343 +INFO:root: +[Valid][21][UTT][ MSE = 0.2722 | LCC = 0.8263 | SRCC = 0.8241 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1648 | LCC = 0.8944 | SRCC = 0.8862 ] +INFO:root:Average loss=0.0643700314924503 +INFO:root: +[Valid][22][UTT][ MSE = 0.2669 | LCC = 0.8281 | SRCC = 0.8257 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1511 | LCC = 0.9026 | SRCC = 0.8963 ] +INFO:root:Average loss=0.06401710651623897 +INFO:root: +[Valid][23][UTT][ MSE = 0.2646 | LCC = 0.8324 | SRCC = 0.8305 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1464 | LCC = 0.9044 | SRCC = 0.8960 ] +INFO:root:Average loss=0.0599608699289652 +INFO:root: +[Valid][24][UTT][ MSE = 0.2770 | LCC = 0.8294 | SRCC = 0.8279 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1692 | LCC = 0.9040 | SRCC = 0.8995 ] +INFO:root:Average loss=0.05807262291319859 +INFO:root: +[Valid][25][UTT][ MSE = 0.2901 | LCC = 0.8283 | SRCC = 0.8262 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1899 | LCC = 0.9003 | SRCC = 0.8974 ] +INFO:root:Average loss=0.05630910716568812 +INFO:root: +[Valid][26][UTT][ MSE = 0.2840 | LCC = 0.8259 | SRCC = 0.8237 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1791 | LCC = 0.8986 | SRCC = 0.8902 ] +INFO:root:Average loss=0.053701904196387686 +INFO:root: +[Valid][27][UTT][ MSE = 0.2639 | LCC = 0.8309 | SRCC = 0.8287 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1488 | LCC = 0.8991 | SRCC = 0.8915 ] +INFO:root:Average loss=0.048629035791143395 +INFO:root: +[Valid][28][UTT][ MSE = 0.2716 | LCC = 0.8285 | SRCC = 0.8259 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1649 | LCC = 0.8986 | SRCC = 0.8910 ] +INFO:root:Average loss=0.04828035172361594 +INFO:root: +[Valid][29][UTT][ MSE = 0.2761 | LCC = 0.8261 | SRCC = 0.8234 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1536 | LCC = 0.8979 | SRCC = 0.8897 ] +INFO:root:Average loss=0.050154062489477486 +INFO:root: +[Valid][30][UTT][ MSE = 0.2821 | LCC = 0.8232 | SRCC = 0.8204 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1720 | LCC = 0.8939 | SRCC = 0.8873 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2672 | LCC = 0.8387 | SRCC = 0.8408 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1549 | LCC = 0.9089 | SRCC = 0.9121 ] diff --git a/results_bvcc/w2v2_base/layer7/run2/config.gin b/results_bvcc/w2v2_base/layer7/run2/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer7/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 diff --git a/results_bvcc/w2v2_base/layer7/run2/model.pt b/results_bvcc/w2v2_base/layer7/run2/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..8bb397dd4c727b04909d3e8da37ec5ff6b53a213 --- /dev/null +++ b/results_bvcc/w2v2_base/layer7/run2/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:65b32a30da9f2f4c1148ecb1d08096c5483d7fa80ddd56706782a2eec6f1b53a +size 2747979 diff --git a/results_bvcc/w2v2_base/layer7/run2/model_best.pt b/results_bvcc/w2v2_base/layer7/run2/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..51ef2aca2924be41be32fbb12a3a6ec17db7c9f9 --- /dev/null +++ b/results_bvcc/w2v2_base/layer7/run2/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3027fa3117ed4b177eea4746787cc248687fc362a17ba9b0b143adf852fdde5c +size 2749702 diff --git a/results_bvcc/w2v2_base/layer7/run2/train.log b/results_bvcc/w2v2_base/layer7/run2/train.log new file mode 100644 index 0000000000000000000000000000000000000000..5f62a4d8321f42469a28b8f67c9c2c3643019e61 --- /dev/null +++ b/results_bvcc/w2v2_base/layer7/run2/train.log @@ -0,0 +1,163 @@ +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=0.5639554445560162 +INFO:root: +[Valid][0][UTT][ MSE = 0.3716 | LCC = 0.7767 | SRCC = 0.7811 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2562 | LCC = 0.8423 | SRCC = 0.8443 ] +INFO:root:Average loss=0.29860492528249055 +INFO:root: +[Valid][1][UTT][ MSE = 0.3260 | LCC = 0.8182 | SRCC = 0.8184 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2086 | LCC = 0.8816 | SRCC = 0.8753 ] +INFO:root:Average loss=0.2475828780577733 +INFO:root: +[Valid][2][UTT][ MSE = 0.2972 | LCC = 0.8262 | SRCC = 0.8263 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1895 | LCC = 0.8945 | SRCC = 0.8889 ] +INFO:root:Average loss=0.22062990145805556 +INFO:root: +[Valid][3][UTT][ MSE = 0.2738 | LCC = 0.8321 | SRCC = 0.8313 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1780 | LCC = 0.8955 | SRCC = 0.8866 ] +INFO:root:Average loss=0.19229804103573164 +INFO:root: +[Valid][4][UTT][ MSE = 0.2696 | LCC = 0.8389 | SRCC = 0.8401 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1768 | LCC = 0.9007 | SRCC = 0.8973 ] +INFO:root:Average loss=0.17953554196999624 +INFO:root: +[Valid][5][UTT][ MSE = 0.2585 | LCC = 0.8443 | SRCC = 0.8439 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1518 | LCC = 0.9061 | SRCC = 0.8986 ] +INFO:root:Average loss=0.16468006028578833 +INFO:root: +[Valid][6][UTT][ MSE = 0.2460 | LCC = 0.8459 | SRCC = 0.8456 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1383 | LCC = 0.9113 | SRCC = 0.9068 ] +INFO:root:Average loss=0.1510708942436255 +INFO:root: +[Valid][7][UTT][ MSE = 0.2591 | LCC = 0.8449 | SRCC = 0.8439 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1588 | LCC = 0.9109 | SRCC = 0.9031 ] +INFO:root:Average loss=0.13623575321756876 +INFO:root: +[Valid][8][UTT][ MSE = 0.2824 | LCC = 0.8458 | SRCC = 0.8475 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1979 | LCC = 0.9150 | SRCC = 0.9100 ] +INFO:root:Average loss=0.1321171051225601 +INFO:root: +[Valid][9][UTT][ MSE = 0.2642 | LCC = 0.8434 | SRCC = 0.8426 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1655 | LCC = 0.9144 | SRCC = 0.9080 ] +INFO:root:Average loss=0.12167189566370769 +INFO:root: +[Valid][10][UTT][ MSE = 0.2638 | LCC = 0.8466 | SRCC = 0.8452 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1518 | LCC = 0.9140 | SRCC = 0.9071 ] +INFO:root:Average loss=0.11481097674904725 +INFO:root: +[Valid][11][UTT][ MSE = 0.2478 | LCC = 0.8437 | SRCC = 0.8427 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1465 | LCC = 0.9069 | SRCC = 0.8991 ] +INFO:root:Average loss=0.11209707668958566 +INFO:root: +[Valid][12][UTT][ MSE = 0.2479 | LCC = 0.8466 | SRCC = 0.8457 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1383 | LCC = 0.9118 | SRCC = 0.9012 ] +INFO:root:Average loss=0.09865253208539425 +INFO:root: +[Valid][13][UTT][ MSE = 0.2451 | LCC = 0.8463 | SRCC = 0.8464 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1529 | LCC = 0.9083 | SRCC = 0.9020 ] +INFO:root:Average loss=0.09169891963784511 +INFO:root: +[Valid][14][UTT][ MSE = 0.2434 | LCC = 0.8440 | SRCC = 0.8430 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1326 | LCC = 0.9074 | SRCC = 0.8994 ] +INFO:root:Average loss=0.0948208875189989 +INFO:root: +[Valid][15][UTT][ MSE = 0.2518 | LCC = 0.8403 | SRCC = 0.8380 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1507 | LCC = 0.9080 | SRCC = 0.9002 ] +INFO:root:Average loss=0.0871514977935033 +INFO:root: +[Valid][16][UTT][ MSE = 0.2548 | LCC = 0.8432 | SRCC = 0.8427 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1380 | LCC = 0.9092 | SRCC = 0.9043 ] +INFO:root:Average loss=0.07666673019337349 +INFO:root: +[Valid][17][UTT][ MSE = 0.2688 | LCC = 0.8390 | SRCC = 0.8377 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1455 | LCC = 0.9090 | SRCC = 0.9012 ] +INFO:root:Average loss=0.07521377145670928 +INFO:root: +[Valid][18][UTT][ MSE = 0.2655 | LCC = 0.8296 | SRCC = 0.8270 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1564 | LCC = 0.9030 | SRCC = 0.8934 ] +INFO:root:Average loss=0.0739270489758406 +INFO:root: +[Valid][19][UTT][ MSE = 0.2537 | LCC = 0.8403 | SRCC = 0.8379 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1453 | LCC = 0.9149 | SRCC = 0.9083 ] +INFO:root:Average loss=0.07125966045527886 +INFO:root: +[Valid][20][UTT][ MSE = 0.2809 | LCC = 0.8391 | SRCC = 0.8385 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1937 | LCC = 0.9072 | SRCC = 0.9028 ] +INFO:root:Average loss=0.06885050862836532 +INFO:root: +[Valid][21][UTT][ MSE = 0.2664 | LCC = 0.8341 | SRCC = 0.8315 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1614 | LCC = 0.9072 | SRCC = 0.8992 ] +INFO:root:Average loss=0.06359508017507884 +INFO:root: +[Valid][22][UTT][ MSE = 0.2599 | LCC = 0.8391 | SRCC = 0.8378 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1465 | LCC = 0.9105 | SRCC = 0.9031 ] +INFO:root:Average loss=0.06002769411469881 +INFO:root: +[Valid][23][UTT][ MSE = 0.2534 | LCC = 0.8374 | SRCC = 0.8357 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1414 | LCC = 0.9098 | SRCC = 0.9024 ] +INFO:root:Average loss=0.05749400127201508 +INFO:root: +[Valid][24][UTT][ MSE = 0.2654 | LCC = 0.8340 | SRCC = 0.8317 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1577 | LCC = 0.8998 | SRCC = 0.8876 ] +INFO:root:Average loss=0.05724178947126254 +INFO:root: +[Valid][25][UTT][ MSE = 0.2543 | LCC = 0.8394 | SRCC = 0.8375 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1605 | LCC = 0.9000 | SRCC = 0.8916 ] +INFO:root:Average loss=0.053545947497089706 +INFO:root: +[Valid][26][UTT][ MSE = 0.2601 | LCC = 0.8350 | SRCC = 0.8324 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1474 | LCC = 0.9096 | SRCC = 0.8975 ] +INFO:root:Average loss=0.05259323320709742 +INFO:root: +[Valid][27][UTT][ MSE = 0.2569 | LCC = 0.8364 | SRCC = 0.8350 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1482 | LCC = 0.9079 | SRCC = 0.9012 ] +INFO:root:Average loss=0.0504357045611892 +INFO:root: +[Valid][28][UTT][ MSE = 0.2608 | LCC = 0.8377 | SRCC = 0.8356 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1567 | LCC = 0.9072 | SRCC = 0.8995 ] +INFO:root:Average loss=0.047473481665245995 +INFO:root: +[Valid][29][UTT][ MSE = 0.2641 | LCC = 0.8365 | SRCC = 0.8348 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1487 | LCC = 0.9095 | SRCC = 0.8998 ] +INFO:root:Average loss=0.045812071301043034 +INFO:root: +[Valid][30][UTT][ MSE = 0.2563 | LCC = 0.8392 | SRCC = 0.8367 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1513 | LCC = 0.9036 | SRCC = 0.8960 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2630 | LCC = 0.8441 | SRCC = 0.8456 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1580 | LCC = 0.9180 | SRCC = 0.9200 ] diff --git a/results_bvcc/w2v2_base/layer7/run3/config.gin b/results_bvcc/w2v2_base/layer7/run3/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer7/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 diff --git a/results_bvcc/w2v2_base/layer7/run3/model.pt b/results_bvcc/w2v2_base/layer7/run3/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..5b8c0e034b6570f62768e5b29fe6b92d8275f50c --- /dev/null +++ b/results_bvcc/w2v2_base/layer7/run3/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7df359e7046896dad7e86753aa0818e628f43770c1b7376988b76a487deab2c1 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer7/run3/model_best.pt b/results_bvcc/w2v2_base/layer7/run3/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..9f3eaaddcd58cb6d0b0448d2cb5e4235f6d37398 --- /dev/null +++ b/results_bvcc/w2v2_base/layer7/run3/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cfe404507ea125f549ad694c6e713a71c14646d03a5a5b726818c6409805b7a1 +size 2749638 diff --git a/results_bvcc/w2v2_base/layer7/run3/train.log b/results_bvcc/w2v2_base/layer7/run3/train.log new file mode 100644 index 0000000000000000000000000000000000000000..0f3232efd4d80ce5d225f8fee6647b4211798a8f --- /dev/null +++ b/results_bvcc/w2v2_base/layer7/run3/train.log @@ -0,0 +1,163 @@ +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=0.6214169282943774 +INFO:root: +[Valid][0][UTT][ MSE = 0.4067 | LCC = 0.7786 | SRCC = 0.7874 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.3198 | LCC = 0.8378 | SRCC = 0.8447 ] +INFO:root:Average loss=0.2947407787044843 +INFO:root: +[Valid][1][UTT][ MSE = 0.3452 | LCC = 0.8167 | SRCC = 0.8180 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2591 | LCC = 0.8721 | SRCC = 0.8700 ] +INFO:root:Average loss=0.24651684268162802 +INFO:root: +[Valid][2][UTT][ MSE = 0.3800 | LCC = 0.8297 | SRCC = 0.8344 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.3151 | LCC = 0.8859 | SRCC = 0.8854 ] +INFO:root:Average loss=0.21856177120636672 +INFO:root: +[Valid][3][UTT][ MSE = 0.2995 | LCC = 0.8410 | SRCC = 0.8449 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.2251 | LCC = 0.8967 | SRCC = 0.8976 ] +INFO:root:Average loss=0.19965830674538246 +INFO:root: +[Valid][4][UTT][ MSE = 0.2760 | LCC = 0.8436 | SRCC = 0.8446 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1884 | LCC = 0.9055 | SRCC = 0.9035 ] +INFO:root:Average loss=0.17834219575310364 +INFO:root: +[Valid][5][UTT][ MSE = 0.2841 | LCC = 0.8471 | SRCC = 0.8473 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.2137 | LCC = 0.9029 | SRCC = 0.8984 ] +INFO:root:Average loss=0.16011094397459275 +INFO:root: +[Valid][6][UTT][ MSE = 0.3154 | LCC = 0.8418 | SRCC = 0.8412 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.2452 | LCC = 0.9037 | SRCC = 0.8967 ] +INFO:root:Average loss=0.15789123156513923 +INFO:root: +[Valid][7][UTT][ MSE = 0.2587 | LCC = 0.8455 | SRCC = 0.8452 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1657 | LCC = 0.9113 | SRCC = 0.9051 ] +INFO:root:Average loss=0.13562970083111373 +INFO:root: +[Valid][8][UTT][ MSE = 0.3716 | LCC = 0.8362 | SRCC = 0.8385 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.2987 | LCC = 0.9062 | SRCC = 0.9057 ] +INFO:root:Average loss=0.12922286012997994 +INFO:root: +[Valid][9][UTT][ MSE = 0.3077 | LCC = 0.8398 | SRCC = 0.8373 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.2338 | LCC = 0.9036 | SRCC = 0.8960 ] +INFO:root:Average loss=0.11750638160185936 +INFO:root: +[Valid][10][UTT][ MSE = 0.2985 | LCC = 0.8416 | SRCC = 0.8404 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.2159 | LCC = 0.9105 | SRCC = 0.9074 ] +INFO:root:Average loss=0.10597328889446381 +INFO:root: +[Valid][11][UTT][ MSE = 0.2922 | LCC = 0.8411 | SRCC = 0.8402 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.2188 | LCC = 0.9054 | SRCC = 0.9014 ] +INFO:root:Average loss=0.10290138538067158 +INFO:root: +[Valid][12][UTT][ MSE = 0.3094 | LCC = 0.8406 | SRCC = 0.8409 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.2357 | LCC = 0.9062 | SRCC = 0.9050 ] +INFO:root:Average loss=0.09875065460801125 +INFO:root: +[Valid][13][UTT][ MSE = 0.2638 | LCC = 0.8390 | SRCC = 0.8369 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1659 | LCC = 0.9111 | SRCC = 0.9040 ] +INFO:root:Average loss=0.09232557077820484 +INFO:root: +[Valid][14][UTT][ MSE = 0.2972 | LCC = 0.8405 | SRCC = 0.8376 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.2139 | LCC = 0.9133 | SRCC = 0.9070 ] +INFO:root:Average loss=0.08539257594981255 +INFO:root: +[Valid][15][UTT][ MSE = 0.2733 | LCC = 0.8383 | SRCC = 0.8362 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1715 | LCC = 0.9159 | SRCC = 0.9104 ] +INFO:root:Average loss=0.08127660223115714 +INFO:root: +[Valid][16][UTT][ MSE = 0.2556 | LCC = 0.8405 | SRCC = 0.8379 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1599 | LCC = 0.9088 | SRCC = 0.8997 ] +INFO:root:Average loss=0.0788978624324768 +INFO:root: +[Valid][17][UTT][ MSE = 0.2843 | LCC = 0.8390 | SRCC = 0.8361 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1995 | LCC = 0.9107 | SRCC = 0.9010 ] +INFO:root:Average loss=0.07432154246056691 +INFO:root: +[Valid][18][UTT][ MSE = 0.2635 | LCC = 0.8439 | SRCC = 0.8429 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1755 | LCC = 0.9140 | SRCC = 0.9099 ] +INFO:root:Average loss=0.07895530679096015 +INFO:root: +[Valid][19][UTT][ MSE = 0.3081 | LCC = 0.8389 | SRCC = 0.8377 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.2332 | LCC = 0.9086 | SRCC = 0.9028 ] +INFO:root:Average loss=0.0678338829236917 +INFO:root: +[Valid][20][UTT][ MSE = 0.2775 | LCC = 0.8408 | SRCC = 0.8381 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1922 | LCC = 0.9104 | SRCC = 0.9012 ] +INFO:root:Average loss=0.06680328355958828 +INFO:root: +[Valid][21][UTT][ MSE = 0.2996 | LCC = 0.8381 | SRCC = 0.8358 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.2313 | LCC = 0.9003 | SRCC = 0.8920 ] +INFO:root:Average loss=0.061287448287774354 +INFO:root: +[Valid][22][UTT][ MSE = 0.2792 | LCC = 0.8409 | SRCC = 0.8393 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1997 | LCC = 0.9061 | SRCC = 0.9004 ] +INFO:root:Average loss=0.060699537062110044 +INFO:root: +[Valid][23][UTT][ MSE = 0.3474 | LCC = 0.8402 | SRCC = 0.8400 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.2904 | LCC = 0.9041 | SRCC = 0.9027 ] +INFO:root:Average loss=0.057173243270088464 +INFO:root: +[Valid][24][UTT][ MSE = 0.2809 | LCC = 0.8432 | SRCC = 0.8410 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.2061 | LCC = 0.9068 | SRCC = 0.8997 ] +INFO:root:Average loss=0.05872795730829239 +INFO:root: +[Valid][25][UTT][ MSE = 0.3014 | LCC = 0.8447 | SRCC = 0.8431 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.2288 | LCC = 0.9096 | SRCC = 0.9034 ] +INFO:root:Average loss=0.05295657958739843 +INFO:root: +[Valid][26][UTT][ MSE = 0.2682 | LCC = 0.8440 | SRCC = 0.8422 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1813 | LCC = 0.9129 | SRCC = 0.9054 ] +INFO:root:Average loss=0.051070804277864784 +INFO:root: +[Valid][27][UTT][ MSE = 0.2901 | LCC = 0.8440 | SRCC = 0.8418 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.2127 | LCC = 0.9094 | SRCC = 0.9040 ] +INFO:root:Average loss=0.05149437628017786 +INFO:root: +[Valid][28][UTT][ MSE = 0.2847 | LCC = 0.8445 | SRCC = 0.8424 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.2073 | LCC = 0.9095 | SRCC = 0.9039 ] +INFO:root:Average loss=0.04850772047081055 +INFO:root: +[Valid][29][UTT][ MSE = 0.2931 | LCC = 0.8435 | SRCC = 0.8425 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.2239 | LCC = 0.9041 | SRCC = 0.8976 ] +INFO:root:Average loss=0.04669602857663845 +INFO:root: +[Valid][30][UTT][ MSE = 0.2547 | LCC = 0.8419 | SRCC = 0.8391 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1603 | LCC = 0.9100 | SRCC = 0.9035 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2641 | LCC = 0.8476 | SRCC = 0.8466 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1409 | LCC = 0.9016 | SRCC = 0.9030 ] diff --git a/results_bvcc/w2v2_base/layer7/run4/config.gin b/results_bvcc/w2v2_base/layer7/run4/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer7/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 diff --git a/results_bvcc/w2v2_base/layer7/run4/model.pt b/results_bvcc/w2v2_base/layer7/run4/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..bb9ba8e7d74e3bbe7dcbb44964bb5b701c981253 --- /dev/null +++ b/results_bvcc/w2v2_base/layer7/run4/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6705642e206e80d8f13a6f2a4eda2b9fdc4146a131ebeda65e9b5770f4c89e46 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer7/run4/model_best.pt b/results_bvcc/w2v2_base/layer7/run4/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..0609e5d87078b12ba9c2d8422a51588e53e70608 --- /dev/null +++ b/results_bvcc/w2v2_base/layer7/run4/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d274ec14b9d1cf5b31fef9b6b772bac74654c349a9c9be432a5395269c443c59 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer7/run4/train.log b/results_bvcc/w2v2_base/layer7/run4/train.log new file mode 100644 index 0000000000000000000000000000000000000000..b3ac18abf4f23e1ffaa650fe9242326078858b8a --- /dev/null +++ b/results_bvcc/w2v2_base/layer7/run4/train.log @@ -0,0 +1,163 @@ +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=0.5781776650975912 +INFO:root: +[Valid][0][UTT][ MSE = 0.3686 | LCC = 0.7838 | SRCC = 0.7865 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2509 | LCC = 0.8563 | SRCC = 0.8542 ] +INFO:root:Average loss=0.2922423413166633 +INFO:root: +[Valid][1][UTT][ MSE = 0.3146 | LCC = 0.8234 | SRCC = 0.8236 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1933 | LCC = 0.8881 | SRCC = 0.8849 ] +INFO:root:Average loss=0.2463863885555512 +INFO:root: +[Valid][2][UTT][ MSE = 0.2873 | LCC = 0.8347 | SRCC = 0.8333 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1707 | LCC = 0.9047 | SRCC = 0.8988 ] +INFO:root:Average loss=0.21543644273128265 +INFO:root: +[Valid][3][UTT][ MSE = 0.2808 | LCC = 0.8408 | SRCC = 0.8407 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1816 | LCC = 0.9019 | SRCC = 0.8966 ] +INFO:root:Average loss=0.18680457532023773 +INFO:root: +[Valid][4][UTT][ MSE = 0.2643 | LCC = 0.8422 | SRCC = 0.8413 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1692 | LCC = 0.9063 | SRCC = 0.9000 ] +INFO:root:Average loss=0.1874806518929127 +INFO:root: +[Valid][5][UTT][ MSE = 0.3031 | LCC = 0.8403 | SRCC = 0.8411 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.2185 | LCC = 0.9074 | SRCC = 0.9060 ] +INFO:root:Average loss=0.16520769387865678 +INFO:root: +[Valid][6][UTT][ MSE = 0.2552 | LCC = 0.8412 | SRCC = 0.8426 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1434 | LCC = 0.9067 | SRCC = 0.9037 ] +INFO:root:Average loss=0.15246930259924668 +INFO:root: +[Valid][7][UTT][ MSE = 0.2520 | LCC = 0.8469 | SRCC = 0.8468 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1592 | LCC = 0.9108 | SRCC = 0.9054 ] +INFO:root:Average loss=0.14004423001255745 +INFO:root: +[Valid][8][UTT][ MSE = 0.2600 | LCC = 0.8397 | SRCC = 0.8392 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1472 | LCC = 0.9133 | SRCC = 0.9080 ] +INFO:root:Average loss=0.12891729615437678 +INFO:root: +[Valid][9][UTT][ MSE = 0.2571 | LCC = 0.8410 | SRCC = 0.8403 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1595 | LCC = 0.9097 | SRCC = 0.9055 ] +INFO:root:Average loss=0.11967659206726612 +INFO:root: +[Valid][10][UTT][ MSE = 0.2616 | LCC = 0.8418 | SRCC = 0.8426 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1262 | LCC = 0.9215 | SRCC = 0.9167 ] +INFO:root:Average loss=0.10985743636504197 +INFO:root: +[Valid][11][UTT][ MSE = 0.2534 | LCC = 0.8393 | SRCC = 0.8395 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1475 | LCC = 0.9131 | SRCC = 0.9084 ] +INFO:root:Average loss=0.10445524083498196 +INFO:root: +[Valid][12][UTT][ MSE = 0.2566 | LCC = 0.8404 | SRCC = 0.8391 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1558 | LCC = 0.9085 | SRCC = 0.9021 ] +INFO:root:Average loss=0.09762112070352603 +INFO:root: +[Valid][13][UTT][ MSE = 0.2481 | LCC = 0.8423 | SRCC = 0.8411 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1270 | LCC = 0.9214 | SRCC = 0.9163 ] +INFO:root:Average loss=0.09220594473374195 +INFO:root: +[Valid][14][UTT][ MSE = 0.2601 | LCC = 0.8399 | SRCC = 0.8378 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1581 | LCC = 0.9144 | SRCC = 0.9103 ] +INFO:root:Average loss=0.08448165454543553 +INFO:root: +[Valid][15][UTT][ MSE = 0.2650 | LCC = 0.8365 | SRCC = 0.8363 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1443 | LCC = 0.9062 | SRCC = 0.9024 ] +INFO:root:Average loss=0.08019758254671708 +INFO:root: +[Valid][16][UTT][ MSE = 0.2661 | LCC = 0.8352 | SRCC = 0.8346 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1515 | LCC = 0.9122 | SRCC = 0.9072 ] +INFO:root:Average loss=0.08243693492542474 +INFO:root: +[Valid][17][UTT][ MSE = 0.2631 | LCC = 0.8346 | SRCC = 0.8356 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1356 | LCC = 0.9084 | SRCC = 0.9026 ] +INFO:root:Average loss=0.07301904466480781 +INFO:root: +[Valid][18][UTT][ MSE = 0.2674 | LCC = 0.8337 | SRCC = 0.8323 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1480 | LCC = 0.9105 | SRCC = 0.9041 ] +INFO:root:Average loss=0.07148055928066754 +INFO:root: +[Valid][19][UTT][ MSE = 0.2657 | LCC = 0.8370 | SRCC = 0.8362 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1448 | LCC = 0.9146 | SRCC = 0.9092 ] +INFO:root:Average loss=0.06632295269996692 +INFO:root: +[Valid][20][UTT][ MSE = 0.2509 | LCC = 0.8384 | SRCC = 0.8379 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1351 | LCC = 0.9163 | SRCC = 0.9123 ] +INFO:root:Average loss=0.06568409411762005 +INFO:root: +[Valid][21][UTT][ MSE = 0.2591 | LCC = 0.8327 | SRCC = 0.8326 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1331 | LCC = 0.9085 | SRCC = 0.9029 ] +INFO:root:Average loss=0.06465235739373244 +INFO:root: +[Valid][22][UTT][ MSE = 0.2636 | LCC = 0.8335 | SRCC = 0.8315 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1489 | LCC = 0.9073 | SRCC = 0.8998 ] +INFO:root:Average loss=0.05863045538083101 +INFO:root: +[Valid][23][UTT][ MSE = 0.2713 | LCC = 0.8323 | SRCC = 0.8309 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1575 | LCC = 0.9081 | SRCC = 0.8998 ] +INFO:root:Average loss=0.05474566176342659 +INFO:root: +[Valid][24][UTT][ MSE = 0.2651 | LCC = 0.8334 | SRCC = 0.8324 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1418 | LCC = 0.9060 | SRCC = 0.8998 ] +INFO:root:Average loss=0.056819174820796035 +INFO:root: +[Valid][25][UTT][ MSE = 0.2757 | LCC = 0.8328 | SRCC = 0.8323 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1443 | LCC = 0.9072 | SRCC = 0.8990 ] +INFO:root:Average loss=0.0519372932612896 +INFO:root: +[Valid][26][UTT][ MSE = 0.2652 | LCC = 0.8373 | SRCC = 0.8358 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1626 | LCC = 0.9127 | SRCC = 0.9070 ] +INFO:root:Average loss=0.054447044021426104 +INFO:root: +[Valid][27][UTT][ MSE = 0.2627 | LCC = 0.8377 | SRCC = 0.8366 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1414 | LCC = 0.9124 | SRCC = 0.9058 ] +INFO:root:Average loss=0.04741895824479751 +INFO:root: +[Valid][28][UTT][ MSE = 0.2610 | LCC = 0.8348 | SRCC = 0.8340 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1466 | LCC = 0.9064 | SRCC = 0.9011 ] +INFO:root:Average loss=0.04748764696220557 +INFO:root: +[Valid][29][UTT][ MSE = 0.2597 | LCC = 0.8363 | SRCC = 0.8347 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1494 | LCC = 0.9099 | SRCC = 0.9045 ] +INFO:root:Average loss=0.045857588187433206 +INFO:root: +[Valid][30][UTT][ MSE = 0.2652 | LCC = 0.8343 | SRCC = 0.8333 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1518 | LCC = 0.9036 | SRCC = 0.8985 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2445 | LCC = 0.8465 | SRCC = 0.8462 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1268 | LCC = 0.9067 | SRCC = 0.9085 ] diff --git a/results_bvcc/w2v2_base/layer8/run0/config.gin b/results_bvcc/w2v2_base/layer8/run0/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer8/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 diff --git a/results_bvcc/w2v2_base/layer8/run0/model.pt b/results_bvcc/w2v2_base/layer8/run0/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..62041dd73511dd08381a0a118f115706ae12f9d8 --- /dev/null +++ b/results_bvcc/w2v2_base/layer8/run0/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:fc0c7cee8d308d6ba23ed035390f6b1aca895eee9f2171edbe829b08256f19a1 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer8/run0/model_best.pt b/results_bvcc/w2v2_base/layer8/run0/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..ed8eff1741ad647e9b40a8d0a3dd048ccfb13861 --- /dev/null +++ b/results_bvcc/w2v2_base/layer8/run0/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:acb35d2b9857bba42b86f9ac2b7eecfadf9d87dd2e911678eb381753dda8c187 +size 2749638 diff --git a/results_bvcc/w2v2_base/layer8/run0/train.log b/results_bvcc/w2v2_base/layer8/run0/train.log new file mode 100644 index 0000000000000000000000000000000000000000..303e83945f1b1335d9ffe5551d08d43a87e0b17c --- /dev/null +++ b/results_bvcc/w2v2_base/layer8/run0/train.log @@ -0,0 +1,163 @@ +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=0.5085832958038037 +INFO:root: +[Valid][0][UTT][ MSE = 0.3479 | LCC = 0.7970 | SRCC = 0.7979 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2279 | LCC = 0.8516 | SRCC = 0.8497 ] +INFO:root:Average loss=0.2764177970015086 +INFO:root: +[Valid][1][UTT][ MSE = 0.2905 | LCC = 0.8361 | SRCC = 0.8348 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1705 | LCC = 0.8996 | SRCC = 0.8986 ] +INFO:root:Average loss=0.23178240198355454 +INFO:root: +[Valid][2][UTT][ MSE = 0.2719 | LCC = 0.8424 | SRCC = 0.8401 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1685 | LCC = 0.9054 | SRCC = 0.9006 ] +INFO:root:Average loss=0.21265333336897385 +INFO:root: +[Valid][3][UTT][ MSE = 0.2512 | LCC = 0.8545 | SRCC = 0.8536 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1562 | LCC = 0.9139 | SRCC = 0.9085 ] +INFO:root:Average loss=0.18886187682167077 +INFO:root: +[Valid][4][UTT][ MSE = 0.2383 | LCC = 0.8503 | SRCC = 0.8500 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1331 | LCC = 0.9111 | SRCC = 0.9098 ] +INFO:root:Average loss=0.18046243335956183 +INFO:root: +[Valid][5][UTT][ MSE = 0.2645 | LCC = 0.8551 | SRCC = 0.8549 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1785 | LCC = 0.9139 | SRCC = 0.9120 ] +INFO:root:Average loss=0.16494930850771758 +INFO:root: +[Valid][6][UTT][ MSE = 0.2443 | LCC = 0.8552 | SRCC = 0.8551 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1475 | LCC = 0.9209 | SRCC = 0.9205 ] +INFO:root:Average loss=0.15529667500119942 +INFO:root: +[Valid][7][UTT][ MSE = 0.2328 | LCC = 0.8547 | SRCC = 0.8545 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1363 | LCC = 0.9187 | SRCC = 0.9153 ] +INFO:root:Average loss=0.14334505815536547 +INFO:root: +[Valid][8][UTT][ MSE = 0.2353 | LCC = 0.8551 | SRCC = 0.8517 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1389 | LCC = 0.9146 | SRCC = 0.9060 ] +INFO:root:Average loss=0.1398928832167234 +INFO:root: +[Valid][9][UTT][ MSE = 0.2524 | LCC = 0.8504 | SRCC = 0.8470 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1595 | LCC = 0.9169 | SRCC = 0.9073 ] +INFO:root:Average loss=0.1248478644933456 +INFO:root: +[Valid][10][UTT][ MSE = 0.2394 | LCC = 0.8507 | SRCC = 0.8484 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1369 | LCC = 0.9211 | SRCC = 0.9169 ] +INFO:root:Average loss=0.11977466749839294 +INFO:root: +[Valid][11][UTT][ MSE = 0.2293 | LCC = 0.8525 | SRCC = 0.8523 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1185 | LCC = 0.9223 | SRCC = 0.9190 ] +INFO:root:Average loss=0.11583082616711274 +INFO:root: +[Valid][12][UTT][ MSE = 0.2577 | LCC = 0.8451 | SRCC = 0.8464 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1616 | LCC = 0.9186 | SRCC = 0.9206 ] +INFO:root:Average loss=0.10719523855890983 +INFO:root: +[Valid][13][UTT][ MSE = 0.2430 | LCC = 0.8490 | SRCC = 0.8471 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1371 | LCC = 0.9230 | SRCC = 0.9176 ] +INFO:root:Average loss=0.10327025044422883 +INFO:root: +[Valid][14][UTT][ MSE = 0.2376 | LCC = 0.8490 | SRCC = 0.8467 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1219 | LCC = 0.9271 | SRCC = 0.9226 ] +INFO:root:Average loss=0.09925453121272418 +INFO:root: +[Valid][15][UTT][ MSE = 0.2550 | LCC = 0.8468 | SRCC = 0.8457 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1593 | LCC = 0.9203 | SRCC = 0.9197 ] +INFO:root:Average loss=0.09990211547567295 +INFO:root: +[Valid][16][UTT][ MSE = 0.2380 | LCC = 0.8522 | SRCC = 0.8493 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1286 | LCC = 0.9249 | SRCC = 0.9193 ] +INFO:root:Average loss=0.08926289308911715 +INFO:root: +[Valid][17][UTT][ MSE = 0.2391 | LCC = 0.8497 | SRCC = 0.8472 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1320 | LCC = 0.9202 | SRCC = 0.9152 ] +INFO:root:Average loss=0.09007783348743732 +INFO:root: +[Valid][18][UTT][ MSE = 0.2437 | LCC = 0.8522 | SRCC = 0.8488 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1283 | LCC = 0.9257 | SRCC = 0.9177 ] +INFO:root:Average loss=0.07774474887320629 +INFO:root: +[Valid][19][UTT][ MSE = 0.2396 | LCC = 0.8496 | SRCC = 0.8480 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1304 | LCC = 0.9252 | SRCC = 0.9222 ] +INFO:root:Average loss=0.07737403329557334 +INFO:root: +[Valid][20][UTT][ MSE = 0.2519 | LCC = 0.8459 | SRCC = 0.8428 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1466 | LCC = 0.9212 | SRCC = 0.9178 ] +INFO:root:Average loss=0.07374340343551758 +INFO:root: +[Valid][21][UTT][ MSE = 0.2393 | LCC = 0.8455 | SRCC = 0.8437 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1211 | LCC = 0.9218 | SRCC = 0.9181 ] +INFO:root:Average loss=0.06902880852039044 +INFO:root: +[Valid][22][UTT][ MSE = 0.2483 | LCC = 0.8444 | SRCC = 0.8424 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1553 | LCC = 0.9134 | SRCC = 0.9078 ] +INFO:root:Average loss=0.06574013733710998 +INFO:root: +[Valid][23][UTT][ MSE = 0.2560 | LCC = 0.8464 | SRCC = 0.8469 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1589 | LCC = 0.9224 | SRCC = 0.9263 ] +INFO:root:Average loss=0.06524491744736831 +INFO:root: +[Valid][24][UTT][ MSE = 0.2515 | LCC = 0.8456 | SRCC = 0.8428 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1460 | LCC = 0.9218 | SRCC = 0.9186 ] +INFO:root:Average loss=0.06061217537484108 +INFO:root: +[Valid][25][UTT][ MSE = 0.2574 | LCC = 0.8444 | SRCC = 0.8440 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1604 | LCC = 0.9190 | SRCC = 0.9181 ] +INFO:root:Average loss=0.058810039781607114 +INFO:root: +[Valid][26][UTT][ MSE = 0.2533 | LCC = 0.8468 | SRCC = 0.8462 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1526 | LCC = 0.9237 | SRCC = 0.9250 ] +INFO:root:Average loss=0.06629020835344608 +INFO:root: +[Valid][27][UTT][ MSE = 0.2451 | LCC = 0.8492 | SRCC = 0.8459 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1261 | LCC = 0.9268 | SRCC = 0.9201 ] +INFO:root:Average loss=0.05829571249584357 +INFO:root: +[Valid][28][UTT][ MSE = 0.2401 | LCC = 0.8483 | SRCC = 0.8447 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1158 | LCC = 0.9273 | SRCC = 0.9210 ] +INFO:root:Average loss=0.05691393116154732 +INFO:root: +[Valid][29][UTT][ MSE = 0.2532 | LCC = 0.8455 | SRCC = 0.8443 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1549 | LCC = 0.9205 | SRCC = 0.9192 ] +INFO:root:Average loss=0.05289071702804321 +INFO:root: +[Valid][30][UTT][ MSE = 0.2438 | LCC = 0.8482 | SRCC = 0.8461 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1339 | LCC = 0.9241 | SRCC = 0.9223 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2427 | LCC = 0.8596 | SRCC = 0.8611 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1370 | LCC = 0.9147 | SRCC = 0.9168 ] diff --git a/results_bvcc/w2v2_base/layer8/run1/config.gin b/results_bvcc/w2v2_base/layer8/run1/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer8/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 diff --git a/results_bvcc/w2v2_base/layer8/run1/model.pt b/results_bvcc/w2v2_base/layer8/run1/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..fdce846bef86da0deddba5dd059ecee9927affcd --- /dev/null +++ b/results_bvcc/w2v2_base/layer8/run1/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:14ad655f388a60eaf9e333f4fff899215f2fb59d4e6a17b673dd8f0878f6d872 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer8/run1/model_best.pt b/results_bvcc/w2v2_base/layer8/run1/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..e8e17914c7d4b61dac8f23e4e02541efb7682562 --- /dev/null +++ b/results_bvcc/w2v2_base/layer8/run1/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:85beb439fde05abc6c1ca9a0d249cba47605f1bfae9c6c0679bdc4f2684dd9be +size 2749702 diff --git a/results_bvcc/w2v2_base/layer8/run1/train.log b/results_bvcc/w2v2_base/layer8/run1/train.log new file mode 100644 index 0000000000000000000000000000000000000000..75d731a6bba319d419d217bc58066ee27497878a --- /dev/null +++ b/results_bvcc/w2v2_base/layer8/run1/train.log @@ -0,0 +1,163 @@ +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=0.683724256280141 +INFO:root: +[Valid][0][UTT][ MSE = 0.3987 | LCC = 0.7714 | SRCC = 0.7879 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.3119 | LCC = 0.8266 | SRCC = 0.8429 ] +INFO:root:Average loss=0.30227775203111845 +INFO:root: +[Valid][1][UTT][ MSE = 0.3174 | LCC = 0.8377 | SRCC = 0.8360 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2258 | LCC = 0.8927 | SRCC = 0.8890 ] +INFO:root:Average loss=0.2539854947573099 +INFO:root: +[Valid][2][UTT][ MSE = 0.2946 | LCC = 0.8501 | SRCC = 0.8461 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.2055 | LCC = 0.9061 | SRCC = 0.8966 ] +INFO:root:Average loss=0.22066620603585854 +INFO:root: +[Valid][3][UTT][ MSE = 0.2694 | LCC = 0.8589 | SRCC = 0.8565 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1946 | LCC = 0.9106 | SRCC = 0.9052 ] +INFO:root:Average loss=0.19890172015397978 +INFO:root: +[Valid][4][UTT][ MSE = 0.2378 | LCC = 0.8654 | SRCC = 0.8626 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1479 | LCC = 0.9193 | SRCC = 0.9138 ] +INFO:root:Average loss=0.18546318979217455 +INFO:root: +[Valid][5][UTT][ MSE = 0.2445 | LCC = 0.8636 | SRCC = 0.8606 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1628 | LCC = 0.9177 | SRCC = 0.9101 ] +INFO:root:Average loss=0.16843986845551392 +INFO:root: +[Valid][6][UTT][ MSE = 0.2625 | LCC = 0.8629 | SRCC = 0.8624 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1943 | LCC = 0.9188 | SRCC = 0.9149 ] +INFO:root:Average loss=0.154103428316422 +INFO:root: +[Valid][7][UTT][ MSE = 0.2480 | LCC = 0.8655 | SRCC = 0.8639 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1713 | LCC = 0.9236 | SRCC = 0.9215 ] +INFO:root:Average loss=0.14508407887740013 +INFO:root: +[Valid][8][UTT][ MSE = 0.2587 | LCC = 0.8644 | SRCC = 0.8619 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1859 | LCC = 0.9255 | SRCC = 0.9199 ] +INFO:root:Average loss=0.1386538359981317 +INFO:root: +[Valid][9][UTT][ MSE = 0.3141 | LCC = 0.8601 | SRCC = 0.8597 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.2581 | LCC = 0.9200 | SRCC = 0.9177 ] +INFO:root:Average loss=0.1337322104626741 +INFO:root: +[Valid][10][UTT][ MSE = 0.2400 | LCC = 0.8657 | SRCC = 0.8647 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1628 | LCC = 0.9255 | SRCC = 0.9228 ] +INFO:root:Average loss=0.11942044865244474 +INFO:root: +[Valid][11][UTT][ MSE = 0.2620 | LCC = 0.8614 | SRCC = 0.8620 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1889 | LCC = 0.9236 | SRCC = 0.9231 ] +INFO:root:Average loss=0.11379621894313739 +INFO:root: +[Valid][12][UTT][ MSE = 0.2596 | LCC = 0.8596 | SRCC = 0.8591 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1896 | LCC = 0.9173 | SRCC = 0.9168 ] +INFO:root:Average loss=0.10941385544645481 +INFO:root: +[Valid][13][UTT][ MSE = 0.2326 | LCC = 0.8617 | SRCC = 0.8606 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1435 | LCC = 0.9230 | SRCC = 0.9214 ] +INFO:root:Average loss=0.10438355755729553 +INFO:root: +[Valid][14][UTT][ MSE = 0.2815 | LCC = 0.8617 | SRCC = 0.8614 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.2182 | LCC = 0.9226 | SRCC = 0.9218 ] +INFO:root:Average loss=0.09861305050360851 +INFO:root: +[Valid][15][UTT][ MSE = 0.2431 | LCC = 0.8592 | SRCC = 0.8575 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1643 | LCC = 0.9220 | SRCC = 0.9185 ] +INFO:root:Average loss=0.089461686853797 +INFO:root: +[Valid][16][UTT][ MSE = 0.2486 | LCC = 0.8627 | SRCC = 0.8606 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1662 | LCC = 0.9312 | SRCC = 0.9268 ] +INFO:root:Average loss=0.08774355096885791 +INFO:root: +[Valid][17][UTT][ MSE = 0.2416 | LCC = 0.8628 | SRCC = 0.8600 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1540 | LCC = 0.9285 | SRCC = 0.9251 ] +INFO:root:Average loss=0.08036915413462199 +INFO:root: +[Valid][18][UTT][ MSE = 0.2487 | LCC = 0.8577 | SRCC = 0.8563 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1659 | LCC = 0.9241 | SRCC = 0.9235 ] +INFO:root:Average loss=0.07784261879248497 +INFO:root: +[Valid][19][UTT][ MSE = 0.2308 | LCC = 0.8594 | SRCC = 0.8583 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1140 | LCC = 0.9271 | SRCC = 0.9228 ] +INFO:root:Average loss=0.08034861617936538 +INFO:root: +[Valid][20][UTT][ MSE = 0.2537 | LCC = 0.8563 | SRCC = 0.8560 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1683 | LCC = 0.9239 | SRCC = 0.9251 ] +INFO:root:Average loss=0.07009926505195789 +INFO:root: +[Valid][21][UTT][ MSE = 0.2334 | LCC = 0.8570 | SRCC = 0.8552 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1410 | LCC = 0.9254 | SRCC = 0.9238 ] +INFO:root:Average loss=0.07273554071210898 +INFO:root: +[Valid][22][UTT][ MSE = 0.2538 | LCC = 0.8561 | SRCC = 0.8543 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1649 | LCC = 0.9267 | SRCC = 0.9248 ] +INFO:root:Average loss=0.061960883056506134 +INFO:root: +[Valid][23][UTT][ MSE = 0.2259 | LCC = 0.8594 | SRCC = 0.8566 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1281 | LCC = 0.9257 | SRCC = 0.9242 ] +INFO:root:Average loss=0.06391249950497578 +INFO:root: +[Valid][24][UTT][ MSE = 0.2436 | LCC = 0.8580 | SRCC = 0.8554 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1552 | LCC = 0.9257 | SRCC = 0.9229 ] +INFO:root:Average loss=0.05825810234707136 +INFO:root: +[Valid][25][UTT][ MSE = 0.2418 | LCC = 0.8595 | SRCC = 0.8579 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1573 | LCC = 0.9264 | SRCC = 0.9254 ] +INFO:root:Average loss=0.05849276258586309 +INFO:root: +[Valid][26][UTT][ MSE = 0.2565 | LCC = 0.8575 | SRCC = 0.8571 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1793 | LCC = 0.9223 | SRCC = 0.9207 ] +INFO:root:Average loss=0.055277945091709114 +INFO:root: +[Valid][27][UTT][ MSE = 0.2377 | LCC = 0.8622 | SRCC = 0.8605 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1522 | LCC = 0.9272 | SRCC = 0.9290 ] +INFO:root:Average loss=0.054125680516545586 +INFO:root: +[Valid][28][UTT][ MSE = 0.2236 | LCC = 0.8606 | SRCC = 0.8600 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1267 | LCC = 0.9288 | SRCC = 0.9293 ] +INFO:root:Average loss=0.05384236497756762 +INFO:root: +[Valid][29][UTT][ MSE = 0.2731 | LCC = 0.8557 | SRCC = 0.8557 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1985 | LCC = 0.9216 | SRCC = 0.9206 ] +INFO:root:Average loss=0.052630212205724836 +INFO:root: +[Valid][30][UTT][ MSE = 0.2433 | LCC = 0.8568 | SRCC = 0.8556 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1560 | LCC = 0.9249 | SRCC = 0.9244 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2396 | LCC = 0.8624 | SRCC = 0.8617 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1233 | LCC = 0.9166 | SRCC = 0.9179 ] diff --git a/results_bvcc/w2v2_base/layer8/run2/config.gin b/results_bvcc/w2v2_base/layer8/run2/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer8/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 diff --git a/results_bvcc/w2v2_base/layer8/run2/model.pt b/results_bvcc/w2v2_base/layer8/run2/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..330260152543f617a8d5287f947d09e72ce84d18 --- /dev/null +++ b/results_bvcc/w2v2_base/layer8/run2/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:cc626141970258e8339a8b1e545e12f38f93830c6cadb7a539dbbf79d6187d7a +size 2747979 diff --git a/results_bvcc/w2v2_base/layer8/run2/model_best.pt b/results_bvcc/w2v2_base/layer8/run2/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..53c70a295966cddd64835848afd896a79c927f51 --- /dev/null +++ b/results_bvcc/w2v2_base/layer8/run2/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2c71fec74d9a3de43dbc6efdc70e0c1a3066cc3ed3d91da721ad2404fade99e9 +size 2749638 diff --git a/results_bvcc/w2v2_base/layer8/run2/train.log b/results_bvcc/w2v2_base/layer8/run2/train.log new file mode 100644 index 0000000000000000000000000000000000000000..c7e23f37092f610fb60cb0cc5bc4a6661b8da143 --- /dev/null +++ b/results_bvcc/w2v2_base/layer8/run2/train.log @@ -0,0 +1,163 @@ +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=0.5680788084864616 +INFO:root: +[Valid][0][UTT][ MSE = 0.3323 | LCC = 0.7984 | SRCC = 0.8051 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2255 | LCC = 0.8540 | SRCC = 0.8609 ] +INFO:root:Average loss=0.27302612841893464 +INFO:root: +[Valid][1][UTT][ MSE = 0.2848 | LCC = 0.8370 | SRCC = 0.8371 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1685 | LCC = 0.8959 | SRCC = 0.8938 ] +INFO:root:Average loss=0.23560917434784082 +INFO:root: +[Valid][2][UTT][ MSE = 0.2799 | LCC = 0.8463 | SRCC = 0.8450 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1609 | LCC = 0.9107 | SRCC = 0.9086 ] +INFO:root:Average loss=0.21046095532484543 +INFO:root: +[Valid][3][UTT][ MSE = 0.2485 | LCC = 0.8519 | SRCC = 0.8496 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1318 | LCC = 0.9169 | SRCC = 0.9135 ] +INFO:root:Average loss=0.1967449783323667 +INFO:root: +[Valid][4][UTT][ MSE = 0.2400 | LCC = 0.8568 | SRCC = 0.8562 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1277 | LCC = 0.9229 | SRCC = 0.9238 ] +INFO:root:Average loss=0.17864594952418253 +INFO:root: +[Valid][5][UTT][ MSE = 0.2385 | LCC = 0.8580 | SRCC = 0.8561 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1231 | LCC = 0.9230 | SRCC = 0.9196 ] +INFO:root:Average loss=0.1717398352921009 +INFO:root: +[Valid][6][UTT][ MSE = 0.2339 | LCC = 0.8619 | SRCC = 0.8593 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1210 | LCC = 0.9251 | SRCC = 0.9203 ] +INFO:root:Average loss=0.1480703452267708 +INFO:root: +[Valid][7][UTT][ MSE = 0.2431 | LCC = 0.8554 | SRCC = 0.8544 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1204 | LCC = 0.9206 | SRCC = 0.9168 ] +INFO:root:Average loss=0.13980287714646414 +INFO:root: +[Valid][8][UTT][ MSE = 0.2241 | LCC = 0.8586 | SRCC = 0.8567 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1232 | LCC = 0.9231 | SRCC = 0.9199 ] +INFO:root:Average loss=0.1378768358666163 +INFO:root: +[Valid][9][UTT][ MSE = 0.2490 | LCC = 0.8583 | SRCC = 0.8565 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1224 | LCC = 0.9255 | SRCC = 0.9209 ] +INFO:root:Average loss=0.1296186972505007 +INFO:root: +[Valid][10][UTT][ MSE = 0.2497 | LCC = 0.8572 | SRCC = 0.8552 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1247 | LCC = 0.9231 | SRCC = 0.9185 ] +INFO:root:Average loss=0.12862621371944746 +INFO:root: +[Valid][11][UTT][ MSE = 0.2854 | LCC = 0.8564 | SRCC = 0.8549 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1499 | LCC = 0.9199 | SRCC = 0.9159 ] +INFO:root:Average loss=0.11450745776677743 +INFO:root: +[Valid][12][UTT][ MSE = 0.2351 | LCC = 0.8551 | SRCC = 0.8532 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1291 | LCC = 0.9260 | SRCC = 0.9238 ] +INFO:root:Average loss=0.10432536054689151 +INFO:root: +[Valid][13][UTT][ MSE = 0.2282 | LCC = 0.8578 | SRCC = 0.8555 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1171 | LCC = 0.9227 | SRCC = 0.9178 ] +INFO:root:Average loss=0.1005809577420736 +INFO:root: +[Valid][14][UTT][ MSE = 0.2329 | LCC = 0.8550 | SRCC = 0.8534 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1228 | LCC = 0.9204 | SRCC = 0.9155 ] +INFO:root:Average loss=0.10624509791915233 +INFO:root: +[Valid][15][UTT][ MSE = 0.2245 | LCC = 0.8590 | SRCC = 0.8567 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1127 | LCC = 0.9243 | SRCC = 0.9197 ] +INFO:root:Average loss=0.08895588365311806 +INFO:root: +[Valid][16][UTT][ MSE = 0.2411 | LCC = 0.8508 | SRCC = 0.8484 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1281 | LCC = 0.9151 | SRCC = 0.9094 ] +INFO:root:Average loss=0.08507792384196551 +INFO:root: +[Valid][17][UTT][ MSE = 0.2493 | LCC = 0.8559 | SRCC = 0.8544 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1244 | LCC = 0.9227 | SRCC = 0.9149 ] +INFO:root:Average loss=0.08081028801508439 +INFO:root: +[Valid][18][UTT][ MSE = 0.2331 | LCC = 0.8519 | SRCC = 0.8508 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1256 | LCC = 0.9179 | SRCC = 0.9137 ] +INFO:root:Average loss=0.08020964436806165 +INFO:root: +[Valid][19][UTT][ MSE = 0.2555 | LCC = 0.8503 | SRCC = 0.8479 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1253 | LCC = 0.9242 | SRCC = 0.9198 ] +INFO:root:Average loss=0.08820658043409005 +INFO:root: +[Valid][20][UTT][ MSE = 0.2368 | LCC = 0.8516 | SRCC = 0.8502 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1252 | LCC = 0.9242 | SRCC = 0.9211 ] +INFO:root:Average loss=0.07124130012324223 +INFO:root: +[Valid][21][UTT][ MSE = 0.2536 | LCC = 0.8512 | SRCC = 0.8502 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1276 | LCC = 0.9237 | SRCC = 0.9223 ] +INFO:root:Average loss=0.06594321241554542 +INFO:root: +[Valid][22][UTT][ MSE = 0.2365 | LCC = 0.8500 | SRCC = 0.8485 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1123 | LCC = 0.9251 | SRCC = 0.9210 ] +INFO:root:Average loss=0.07075058793028195 +INFO:root: +[Valid][23][UTT][ MSE = 0.2318 | LCC = 0.8513 | SRCC = 0.8503 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1120 | LCC = 0.9247 | SRCC = 0.9224 ] +INFO:root:Average loss=0.064077103510499 +INFO:root: +[Valid][24][UTT][ MSE = 0.2326 | LCC = 0.8542 | SRCC = 0.8526 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1137 | LCC = 0.9296 | SRCC = 0.9259 ] +INFO:root:Average loss=0.05531275129088989 +INFO:root: +[Valid][25][UTT][ MSE = 0.2385 | LCC = 0.8516 | SRCC = 0.8504 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1139 | LCC = 0.9263 | SRCC = 0.9234 ] +INFO:root:Average loss=0.05669853392128761 +INFO:root: +[Valid][26][UTT][ MSE = 0.2418 | LCC = 0.8482 | SRCC = 0.8472 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1270 | LCC = 0.9241 | SRCC = 0.9217 ] +INFO:root:Average loss=0.05418940542791134 +INFO:root: +[Valid][27][UTT][ MSE = 0.2382 | LCC = 0.8493 | SRCC = 0.8473 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1320 | LCC = 0.9211 | SRCC = 0.9158 ] +INFO:root:Average loss=0.05830511818520534 +INFO:root: +[Valid][28][UTT][ MSE = 0.2537 | LCC = 0.8457 | SRCC = 0.8437 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1144 | LCC = 0.9234 | SRCC = 0.9202 ] +INFO:root:Average loss=0.049411399146685235 +INFO:root: +[Valid][29][UTT][ MSE = 0.2510 | LCC = 0.8485 | SRCC = 0.8459 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1233 | LCC = 0.9215 | SRCC = 0.9110 ] +INFO:root:Average loss=0.04859461344014376 +INFO:root: +[Valid][30][UTT][ MSE = 0.2333 | LCC = 0.8495 | SRCC = 0.8491 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1192 | LCC = 0.9228 | SRCC = 0.9188 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2448 | LCC = 0.8627 | SRCC = 0.8633 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1590 | LCC = 0.9200 | SRCC = 0.9222 ] diff --git a/results_bvcc/w2v2_base/layer8/run3/config.gin b/results_bvcc/w2v2_base/layer8/run3/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer8/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 diff --git a/results_bvcc/w2v2_base/layer8/run3/model.pt b/results_bvcc/w2v2_base/layer8/run3/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..7ff8e7bfdb4856b0e436330486942d307f4a035a --- /dev/null +++ b/results_bvcc/w2v2_base/layer8/run3/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e9b1adbdd1edb5a326377c2b46462552fd66fe1dc4184cdeccfe3b244ee62fb4 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer8/run3/model_best.pt b/results_bvcc/w2v2_base/layer8/run3/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..e03b42f21f4e1705820c95ac74626bd40dc2fe1d --- /dev/null +++ b/results_bvcc/w2v2_base/layer8/run3/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:326658f39868287f56a72a5106bc47334ccd21154ef770e47958ecf0cb9eb9a9 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer8/run3/train.log b/results_bvcc/w2v2_base/layer8/run3/train.log new file mode 100644 index 0000000000000000000000000000000000000000..86781e6d94e58105b7ace205c57b4212f0e34e4e --- /dev/null +++ b/results_bvcc/w2v2_base/layer8/run3/train.log @@ -0,0 +1,163 @@ +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=0.5400142952417716 +INFO:root: +[Valid][0][UTT][ MSE = 0.3479 | LCC = 0.7968 | SRCC = 0.7987 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2287 | LCC = 0.8607 | SRCC = 0.8611 ] +INFO:root:Average loss=0.2873571331684406 +INFO:root: +[Valid][1][UTT][ MSE = 0.3014 | LCC = 0.8283 | SRCC = 0.8313 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2060 | LCC = 0.8891 | SRCC = 0.8930 ] +INFO:root:Average loss=0.24200314493515554 +INFO:root: +[Valid][2][UTT][ MSE = 0.2861 | LCC = 0.8422 | SRCC = 0.8413 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1704 | LCC = 0.9044 | SRCC = 0.9005 ] +INFO:root:Average loss=0.21858992007298347 +INFO:root: +[Valid][3][UTT][ MSE = 0.2798 | LCC = 0.8461 | SRCC = 0.8476 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1998 | LCC = 0.9050 | SRCC = 0.9080 ] +INFO:root:Average loss=0.19658920818414444 +INFO:root: +[Valid][4][UTT][ MSE = 0.2480 | LCC = 0.8512 | SRCC = 0.8498 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1476 | LCC = 0.9103 | SRCC = 0.9093 ] +INFO:root:Average loss=0.18550774435966444 +INFO:root: +[Valid][5][UTT][ MSE = 0.2613 | LCC = 0.8544 | SRCC = 0.8530 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1440 | LCC = 0.9183 | SRCC = 0.9159 ] +INFO:root:Average loss=0.1777514731272673 +INFO:root: +[Valid][6][UTT][ MSE = 0.2480 | LCC = 0.8539 | SRCC = 0.8520 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1300 | LCC = 0.9175 | SRCC = 0.9111 ] +INFO:root:Average loss=0.16125448181843147 +INFO:root: +[Valid][7][UTT][ MSE = 0.2394 | LCC = 0.8550 | SRCC = 0.8548 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1489 | LCC = 0.9144 | SRCC = 0.9124 ] +INFO:root:Average loss=0.14927348781090516 +INFO:root: +[Valid][8][UTT][ MSE = 0.2614 | LCC = 0.8536 | SRCC = 0.8513 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1515 | LCC = 0.9176 | SRCC = 0.9130 ] +INFO:root:Average loss=0.14059587319691977 +INFO:root: +[Valid][9][UTT][ MSE = 0.2614 | LCC = 0.8547 | SRCC = 0.8538 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1361 | LCC = 0.9198 | SRCC = 0.9171 ] +INFO:root:Average loss=0.1336527508802903 +INFO:root: +[Valid][10][UTT][ MSE = 0.2261 | LCC = 0.8593 | SRCC = 0.8584 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1199 | LCC = 0.9234 | SRCC = 0.9249 ] +INFO:root:Average loss=0.12281636836437079 +INFO:root: +[Valid][11][UTT][ MSE = 0.2200 | LCC = 0.8608 | SRCC = 0.8588 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1198 | LCC = 0.9232 | SRCC = 0.9211 ] +INFO:root:Average loss=0.11573554957524324 +INFO:root: +[Valid][12][UTT][ MSE = 0.2315 | LCC = 0.8581 | SRCC = 0.8560 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1330 | LCC = 0.9222 | SRCC = 0.9179 ] +INFO:root:Average loss=0.10926913107052827 +INFO:root: +[Valid][13][UTT][ MSE = 0.2349 | LCC = 0.8572 | SRCC = 0.8548 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1304 | LCC = 0.9251 | SRCC = 0.9219 ] +INFO:root:Average loss=0.10396642581774639 +INFO:root: +[Valid][14][UTT][ MSE = 0.2294 | LCC = 0.8586 | SRCC = 0.8583 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1250 | LCC = 0.9270 | SRCC = 0.9246 ] +INFO:root:Average loss=0.09849048190965103 +INFO:root: +[Valid][15][UTT][ MSE = 0.2242 | LCC = 0.8596 | SRCC = 0.8593 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1136 | LCC = 0.9272 | SRCC = 0.9248 ] +INFO:root:Average loss=0.08961267778888726 +INFO:root: +[Valid][16][UTT][ MSE = 0.2373 | LCC = 0.8537 | SRCC = 0.8526 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1161 | LCC = 0.9279 | SRCC = 0.9236 ] +INFO:root:Average loss=0.08647735603153706 +INFO:root: +[Valid][17][UTT][ MSE = 0.2249 | LCC = 0.8580 | SRCC = 0.8558 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1104 | LCC = 0.9286 | SRCC = 0.9238 ] +INFO:root:Average loss=0.08365402108010574 +INFO:root: +[Valid][18][UTT][ MSE = 0.2261 | LCC = 0.8606 | SRCC = 0.8590 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1195 | LCC = 0.9308 | SRCC = 0.9270 ] +INFO:root:Average loss=0.0823587863586652 +INFO:root: +[Valid][19][UTT][ MSE = 0.2314 | LCC = 0.8581 | SRCC = 0.8581 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1350 | LCC = 0.9289 | SRCC = 0.9304 ] +INFO:root:Average loss=0.08102096206484696 +INFO:root: +[Valid][20][UTT][ MSE = 0.2313 | LCC = 0.8599 | SRCC = 0.8596 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1433 | LCC = 0.9256 | SRCC = 0.9268 ] +INFO:root:Average loss=0.07122003764678271 +INFO:root: +[Valid][21][UTT][ MSE = 0.2253 | LCC = 0.8547 | SRCC = 0.8536 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1146 | LCC = 0.9275 | SRCC = 0.9249 ] +INFO:root:Average loss=0.07570496129875 +INFO:root: +[Valid][22][UTT][ MSE = 0.2326 | LCC = 0.8568 | SRCC = 0.8562 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1356 | LCC = 0.9275 | SRCC = 0.9271 ] +INFO:root:Average loss=0.06793811611640148 +INFO:root: +[Valid][23][UTT][ MSE = 0.2383 | LCC = 0.8574 | SRCC = 0.8541 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1367 | LCC = 0.9277 | SRCC = 0.9218 ] +INFO:root:Average loss=0.06977727526846604 +INFO:root: +[Valid][24][UTT][ MSE = 0.2273 | LCC = 0.8588 | SRCC = 0.8564 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1226 | LCC = 0.9279 | SRCC = 0.9248 ] +INFO:root:Average loss=0.06550530864833257 +INFO:root: +[Valid][25][UTT][ MSE = 0.2238 | LCC = 0.8588 | SRCC = 0.8572 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1142 | LCC = 0.9291 | SRCC = 0.9259 ] +INFO:root:Average loss=0.06214213519333265 +INFO:root: +[Valid][26][UTT][ MSE = 0.2236 | LCC = 0.8608 | SRCC = 0.8587 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1119 | LCC = 0.9318 | SRCC = 0.9283 ] +INFO:root:Average loss=0.05913168559662807 +INFO:root: +[Valid][27][UTT][ MSE = 0.2287 | LCC = 0.8550 | SRCC = 0.8532 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1162 | LCC = 0.9252 | SRCC = 0.9205 ] +INFO:root:Average loss=0.05950829152686474 +INFO:root: +[Valid][28][UTT][ MSE = 0.2296 | LCC = 0.8576 | SRCC = 0.8554 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1200 | LCC = 0.9291 | SRCC = 0.9265 ] +INFO:root:Average loss=0.05610970923533806 +INFO:root: +[Valid][29][UTT][ MSE = 0.2342 | LCC = 0.8575 | SRCC = 0.8558 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1173 | LCC = 0.9293 | SRCC = 0.9241 ] +INFO:root:Average loss=0.05461709989378086 +INFO:root: +[Valid][30][UTT][ MSE = 0.2312 | LCC = 0.8574 | SRCC = 0.8553 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1246 | LCC = 0.9302 | SRCC = 0.9259 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2541 | LCC = 0.8418 | SRCC = 0.8425 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1314 | LCC = 0.9168 | SRCC = 0.9196 ] diff --git a/results_bvcc/w2v2_base/layer8/run4/config.gin b/results_bvcc/w2v2_base/layer8/run4/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer8/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 diff --git a/results_bvcc/w2v2_base/layer8/run4/model.pt b/results_bvcc/w2v2_base/layer8/run4/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..f6a6c6bbfa64a86e5ed6a37543973a08132d06b0 --- /dev/null +++ b/results_bvcc/w2v2_base/layer8/run4/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e015ad1d2495facd7b7fb690eb96aa734779859c50f296d667ef2c11ae712309 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer8/run4/model_best.pt b/results_bvcc/w2v2_base/layer8/run4/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..aa2c9ba0bb2972d60f5f98aeeed2f2b392a08002 --- /dev/null +++ b/results_bvcc/w2v2_base/layer8/run4/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:afd8492b307370fe47aaa03755821a88894c5b8032b899b25926809638a731ea +size 2749702 diff --git a/results_bvcc/w2v2_base/layer8/run4/train.log b/results_bvcc/w2v2_base/layer8/run4/train.log new file mode 100644 index 0000000000000000000000000000000000000000..9e3d34203d88ef2887c883c069a451e486744459 --- /dev/null +++ b/results_bvcc/w2v2_base/layer8/run4/train.log @@ -0,0 +1,163 @@ +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=0.5615891115023539 +INFO:root: +[Valid][0][UTT][ MSE = 0.3295 | LCC = 0.8037 | SRCC = 0.8057 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2288 | LCC = 0.8591 | SRCC = 0.8591 ] +INFO:root:Average loss=0.2699241066972415 +INFO:root: +[Valid][1][UTT][ MSE = 0.2945 | LCC = 0.8392 | SRCC = 0.8381 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1786 | LCC = 0.9066 | SRCC = 0.9047 ] +INFO:root:Average loss=0.23674970444960472 +INFO:root: +[Valid][2][UTT][ MSE = 0.2738 | LCC = 0.8496 | SRCC = 0.8489 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1778 | LCC = 0.9135 | SRCC = 0.9115 ] +INFO:root:Average loss=0.20859502676205757 +INFO:root: +[Valid][3][UTT][ MSE = 0.2549 | LCC = 0.8548 | SRCC = 0.8542 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1637 | LCC = 0.9190 | SRCC = 0.9175 ] +INFO:root:Average loss=0.19741387512439337 +INFO:root: +[Valid][4][UTT][ MSE = 0.3098 | LCC = 0.8548 | SRCC = 0.8560 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.2344 | LCC = 0.9174 | SRCC = 0.9155 ] +INFO:root:Average loss=0.17613845863021338 +INFO:root: +[Valid][5][UTT][ MSE = 0.2475 | LCC = 0.8586 | SRCC = 0.8576 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1549 | LCC = 0.9232 | SRCC = 0.9173 ] +INFO:root:Average loss=0.17385753062673104 +INFO:root: +[Valid][6][UTT][ MSE = 0.2380 | LCC = 0.8601 | SRCC = 0.8583 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1355 | LCC = 0.9236 | SRCC = 0.9181 ] +INFO:root:Average loss=0.17289526464465338 +INFO:root: +[Valid][7][UTT][ MSE = 0.2293 | LCC = 0.8621 | SRCC = 0.8616 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1392 | LCC = 0.9229 | SRCC = 0.9173 ] +INFO:root:Average loss=0.14882277181515327 +INFO:root: +[Valid][8][UTT][ MSE = 0.2283 | LCC = 0.8625 | SRCC = 0.8612 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1256 | LCC = 0.9276 | SRCC = 0.9225 ] +INFO:root:Average loss=0.14228027485884154 +INFO:root: +[Valid][9][UTT][ MSE = 0.2421 | LCC = 0.8587 | SRCC = 0.8571 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1449 | LCC = 0.9266 | SRCC = 0.9212 ] +INFO:root:Average loss=0.1329594285060198 +INFO:root: +[Valid][10][UTT][ MSE = 0.2335 | LCC = 0.8595 | SRCC = 0.8577 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1269 | LCC = 0.9276 | SRCC = 0.9213 ] +INFO:root:Average loss=0.1273114541784311 +INFO:root: +[Valid][11][UTT][ MSE = 0.2318 | LCC = 0.8598 | SRCC = 0.8588 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1162 | LCC = 0.9334 | SRCC = 0.9271 ] +INFO:root:Average loss=0.11716745402186345 +INFO:root: +[Valid][12][UTT][ MSE = 0.2356 | LCC = 0.8577 | SRCC = 0.8565 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1355 | LCC = 0.9268 | SRCC = 0.9229 ] +INFO:root:Average loss=0.11691695222487816 +INFO:root: +[Valid][13][UTT][ MSE = 0.2474 | LCC = 0.8559 | SRCC = 0.8534 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1446 | LCC = 0.9261 | SRCC = 0.9185 ] +INFO:root:Average loss=0.10546949811470814 +INFO:root: +[Valid][14][UTT][ MSE = 0.2421 | LCC = 0.8558 | SRCC = 0.8557 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1360 | LCC = 0.9313 | SRCC = 0.9267 ] +INFO:root:Average loss=0.10120275750374183 +INFO:root: +[Valid][15][UTT][ MSE = 0.2476 | LCC = 0.8566 | SRCC = 0.8565 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1561 | LCC = 0.9250 | SRCC = 0.9217 ] +INFO:root:Average loss=0.0924875758206233 +INFO:root: +[Valid][16][UTT][ MSE = 0.2342 | LCC = 0.8565 | SRCC = 0.8553 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1334 | LCC = 0.9279 | SRCC = 0.9178 ] +INFO:root:Average loss=0.08661637550745255 +INFO:root: +[Valid][17][UTT][ MSE = 0.2424 | LCC = 0.8538 | SRCC = 0.8509 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1443 | LCC = 0.9215 | SRCC = 0.9100 ] +INFO:root:Average loss=0.08385322510431974 +INFO:root: +[Valid][18][UTT][ MSE = 0.2581 | LCC = 0.8493 | SRCC = 0.8473 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1619 | LCC = 0.9232 | SRCC = 0.9133 ] +INFO:root:Average loss=0.08161849084381874 +INFO:root: +[Valid][19][UTT][ MSE = 0.2454 | LCC = 0.8505 | SRCC = 0.8486 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1419 | LCC = 0.9237 | SRCC = 0.9179 ] +INFO:root:Average loss=0.073384275373358 +INFO:root: +[Valid][20][UTT][ MSE = 0.2408 | LCC = 0.8507 | SRCC = 0.8493 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1330 | LCC = 0.9275 | SRCC = 0.9181 ] +INFO:root:Average loss=0.07226721560343718 +INFO:root: +[Valid][21][UTT][ MSE = 0.2669 | LCC = 0.8509 | SRCC = 0.8496 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1732 | LCC = 0.9253 | SRCC = 0.9182 ] +INFO:root:Average loss=0.07067048277419347 +INFO:root: +[Valid][22][UTT][ MSE = 0.2352 | LCC = 0.8537 | SRCC = 0.8523 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1236 | LCC = 0.9285 | SRCC = 0.9197 ] +INFO:root:Average loss=0.06722476806205052 +INFO:root: +[Valid][23][UTT][ MSE = 0.2596 | LCC = 0.8499 | SRCC = 0.8481 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1620 | LCC = 0.9272 | SRCC = 0.9217 ] +INFO:root:Average loss=0.06924948416268215 +INFO:root: +[Valid][24][UTT][ MSE = 0.2422 | LCC = 0.8488 | SRCC = 0.8471 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1286 | LCC = 0.9287 | SRCC = 0.9189 ] +INFO:root:Average loss=0.06186086631929263 +INFO:root: +[Valid][25][UTT][ MSE = 0.2588 | LCC = 0.8478 | SRCC = 0.8449 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1606 | LCC = 0.9234 | SRCC = 0.9175 ] +INFO:root:Average loss=0.05809476661185423 +INFO:root: +[Valid][26][UTT][ MSE = 0.2418 | LCC = 0.8519 | SRCC = 0.8502 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1322 | LCC = 0.9299 | SRCC = 0.9240 ] +INFO:root:Average loss=0.0555392015629854 +INFO:root: +[Valid][27][UTT][ MSE = 0.2496 | LCC = 0.8444 | SRCC = 0.8427 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1335 | LCC = 0.9218 | SRCC = 0.9122 ] +INFO:root:Average loss=0.0578689274784082 +INFO:root: +[Valid][28][UTT][ MSE = 0.2686 | LCC = 0.8490 | SRCC = 0.8476 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1758 | LCC = 0.9225 | SRCC = 0.9160 ] +INFO:root:Average loss=0.05634371559016216 +INFO:root: +[Valid][29][UTT][ MSE = 0.2706 | LCC = 0.8496 | SRCC = 0.8489 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1764 | LCC = 0.9249 | SRCC = 0.9214 ] +INFO:root:Average loss=0.05324583132870686 +INFO:root: +[Valid][30][UTT][ MSE = 0.2655 | LCC = 0.8493 | SRCC = 0.8471 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1705 | LCC = 0.9238 | SRCC = 0.9178 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2393 | LCC = 0.8546 | SRCC = 0.8582 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1386 | LCC = 0.9148 | SRCC = 0.9187 ] diff --git a/results_bvcc/w2v2_base/layer9/run0/config.gin b/results_bvcc/w2v2_base/layer9/run0/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer9/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 diff --git a/results_bvcc/w2v2_base/layer9/run0/model.pt b/results_bvcc/w2v2_base/layer9/run0/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..b9013215ad59fe08f1785769b0e62d6cb07c7568 --- /dev/null +++ b/results_bvcc/w2v2_base/layer9/run0/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0cd78cd8bb568dcdf5e2c84f5bef454511a84fe62ed10c629c188f1b231bd045 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer9/run0/model_best.pt b/results_bvcc/w2v2_base/layer9/run0/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..2332d86b571c2dcc48b8798f69858e2a6ea43644 --- /dev/null +++ b/results_bvcc/w2v2_base/layer9/run0/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a41aff4e7b7cb123db4a28e414fd9b1f9125002f074bf52322853fb00b84b1ca +size 2749702 diff --git a/results_bvcc/w2v2_base/layer9/run0/train.log b/results_bvcc/w2v2_base/layer9/run0/train.log new file mode 100644 index 0000000000000000000000000000000000000000..273e3aff5dc3a540b56043d6a33bdd72dfb9c8e5 --- /dev/null +++ b/results_bvcc/w2v2_base/layer9/run0/train.log @@ -0,0 +1,163 @@ +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=0.6092293864259353 +INFO:root: +[Valid][0][UTT][ MSE = 0.3464 | LCC = 0.7880 | SRCC = 0.7947 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2361 | LCC = 0.8394 | SRCC = 0.8446 ] +INFO:root:Average loss=0.2693075795586293 +INFO:root: +[Valid][1][UTT][ MSE = 0.2831 | LCC = 0.8348 | SRCC = 0.8388 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1803 | LCC = 0.8862 | SRCC = 0.8870 ] +INFO:root:Average loss=0.23559129467377296 +INFO:root: +[Valid][2][UTT][ MSE = 0.2848 | LCC = 0.8500 | SRCC = 0.8516 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1703 | LCC = 0.9063 | SRCC = 0.9060 ] +INFO:root:Average loss=0.21324823261835638 +INFO:root: +[Valid][3][UTT][ MSE = 0.2455 | LCC = 0.8554 | SRCC = 0.8562 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1445 | LCC = 0.9092 | SRCC = 0.9075 ] +INFO:root:Average loss=0.1977092592188945 +INFO:root: +[Valid][4][UTT][ MSE = 0.2556 | LCC = 0.8554 | SRCC = 0.8565 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1744 | LCC = 0.9098 | SRCC = 0.9106 ] +INFO:root:Average loss=0.19152175291226461 +INFO:root: +[Valid][5][UTT][ MSE = 0.2590 | LCC = 0.8583 | SRCC = 0.8561 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1525 | LCC = 0.9153 | SRCC = 0.9091 ] +INFO:root:Average loss=0.1781023736947622 +INFO:root: +[Valid][6][UTT][ MSE = 0.2461 | LCC = 0.8549 | SRCC = 0.8536 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1426 | LCC = 0.9149 | SRCC = 0.9123 ] +INFO:root:Average loss=0.1645132340490818 +INFO:root: +[Valid][7][UTT][ MSE = 0.2495 | LCC = 0.8602 | SRCC = 0.8571 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1402 | LCC = 0.9169 | SRCC = 0.9105 ] +INFO:root:Average loss=0.15929321514872405 +INFO:root: +[Valid][8][UTT][ MSE = 0.2431 | LCC = 0.8582 | SRCC = 0.8551 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1481 | LCC = 0.9128 | SRCC = 0.9049 ] +INFO:root:Average loss=0.15305849699637827 +INFO:root: +[Valid][9][UTT][ MSE = 0.2375 | LCC = 0.8585 | SRCC = 0.8570 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1390 | LCC = 0.9197 | SRCC = 0.9149 ] +INFO:root:Average loss=0.13854364831096086 +INFO:root: +[Valid][10][UTT][ MSE = 0.2578 | LCC = 0.8586 | SRCC = 0.8569 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1481 | LCC = 0.9226 | SRCC = 0.9164 ] +INFO:root:Average loss=0.12910349695728376 +INFO:root: +[Valid][11][UTT][ MSE = 0.2616 | LCC = 0.8560 | SRCC = 0.8536 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1693 | LCC = 0.9216 | SRCC = 0.9195 ] +INFO:root:Average loss=0.13119826121972158 +INFO:root: +[Valid][12][UTT][ MSE = 0.2269 | LCC = 0.8612 | SRCC = 0.8592 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1313 | LCC = 0.9215 | SRCC = 0.9190 ] +INFO:root:Average loss=0.11794060287185204 +INFO:root: +[Valid][13][UTT][ MSE = 0.2735 | LCC = 0.8556 | SRCC = 0.8543 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1925 | LCC = 0.9170 | SRCC = 0.9141 ] +INFO:root:Average loss=0.11566729423327324 +INFO:root: +[Valid][14][UTT][ MSE = 0.2311 | LCC = 0.8617 | SRCC = 0.8593 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1311 | LCC = 0.9275 | SRCC = 0.9237 ] +INFO:root:Average loss=0.10748401241233715 +INFO:root: +[Valid][15][UTT][ MSE = 0.2520 | LCC = 0.8579 | SRCC = 0.8535 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1556 | LCC = 0.9224 | SRCC = 0.9149 ] +INFO:root:Average loss=0.10183319267936242 +INFO:root: +[Valid][16][UTT][ MSE = 0.2356 | LCC = 0.8543 | SRCC = 0.8509 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1232 | LCC = 0.9246 | SRCC = 0.9209 ] +INFO:root:Average loss=0.09677598348412758 +INFO:root: +[Valid][17][UTT][ MSE = 0.2331 | LCC = 0.8577 | SRCC = 0.8557 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1219 | LCC = 0.9234 | SRCC = 0.9190 ] +INFO:root:Average loss=0.09118171088779584 +INFO:root: +[Valid][18][UTT][ MSE = 0.2314 | LCC = 0.8567 | SRCC = 0.8541 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1229 | LCC = 0.9261 | SRCC = 0.9232 ] +INFO:root:Average loss=0.08547160225227857 +INFO:root: +[Valid][19][UTT][ MSE = 0.2272 | LCC = 0.8592 | SRCC = 0.8562 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1266 | LCC = 0.9282 | SRCC = 0.9242 ] +INFO:root:Average loss=0.08951480046678813 +INFO:root: +[Valid][20][UTT][ MSE = 0.2380 | LCC = 0.8585 | SRCC = 0.8545 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1357 | LCC = 0.9293 | SRCC = 0.9238 ] +INFO:root:Average loss=0.08036335610235348 +INFO:root: +[Valid][21][UTT][ MSE = 0.2615 | LCC = 0.8557 | SRCC = 0.8529 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1703 | LCC = 0.9245 | SRCC = 0.9200 ] +INFO:root:Average loss=0.07637957540842202 +INFO:root: +[Valid][22][UTT][ MSE = 0.2312 | LCC = 0.8591 | SRCC = 0.8553 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1237 | LCC = 0.9312 | SRCC = 0.9252 ] +INFO:root:Average loss=0.07832169633072156 +INFO:root: +[Valid][23][UTT][ MSE = 0.2280 | LCC = 0.8582 | SRCC = 0.8546 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1168 | LCC = 0.9272 | SRCC = 0.9215 ] +INFO:root:Average loss=0.07261175996600053 +INFO:root: +[Valid][24][UTT][ MSE = 0.2317 | LCC = 0.8623 | SRCC = 0.8593 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1326 | LCC = 0.9317 | SRCC = 0.9291 ] +INFO:root:Average loss=0.07138573063107637 +INFO:root: +[Valid][25][UTT][ MSE = 0.2464 | LCC = 0.8587 | SRCC = 0.8569 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1477 | LCC = 0.9309 | SRCC = 0.9283 ] +INFO:root:Average loss=0.06573987542054592 +INFO:root: +[Valid][26][UTT][ MSE = 0.2375 | LCC = 0.8565 | SRCC = 0.8533 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1159 | LCC = 0.9308 | SRCC = 0.9235 ] +INFO:root:Average loss=0.06677169476946195 +INFO:root: +[Valid][27][UTT][ MSE = 0.2389 | LCC = 0.8571 | SRCC = 0.8539 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1383 | LCC = 0.9299 | SRCC = 0.9269 ] +INFO:root:Average loss=0.0662077929920111 +INFO:root: +[Valid][28][UTT][ MSE = 0.2257 | LCC = 0.8573 | SRCC = 0.8556 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1048 | LCC = 0.9312 | SRCC = 0.9275 ] +INFO:root:Average loss=0.061594242039017186 +INFO:root: +[Valid][29][UTT][ MSE = 0.2309 | LCC = 0.8588 | SRCC = 0.8568 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1264 | LCC = 0.9311 | SRCC = 0.9298 ] +INFO:root:Average loss=0.05597598611926421 +INFO:root: +[Valid][30][UTT][ MSE = 0.2214 | LCC = 0.8628 | SRCC = 0.8611 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1118 | LCC = 0.9353 | SRCC = 0.9345 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2419 | LCC = 0.8533 | SRCC = 0.8518 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1345 | LCC = 0.9207 | SRCC = 0.9211 ] diff --git a/results_bvcc/w2v2_base/layer9/run1/config.gin b/results_bvcc/w2v2_base/layer9/run1/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer9/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 diff --git a/results_bvcc/w2v2_base/layer9/run1/model.pt b/results_bvcc/w2v2_base/layer9/run1/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..790dc66341cae576c2c423204612e73aa03de704 --- /dev/null +++ b/results_bvcc/w2v2_base/layer9/run1/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:89f88df7fd3fd978d6614d438660ae1e98afdeb27896e2a5b81570be57557ff6 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer9/run1/model_best.pt b/results_bvcc/w2v2_base/layer9/run1/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..6080a8c8f720eac1aa92af0e4fd3a442ef3f0472 --- /dev/null +++ b/results_bvcc/w2v2_base/layer9/run1/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:b4cb0be3c5a861db9587d889f4c2bf7b501c38bdb18e49cd5b2ed065592cf2fd +size 2749702 diff --git a/results_bvcc/w2v2_base/layer9/run1/train.log b/results_bvcc/w2v2_base/layer9/run1/train.log new file mode 100644 index 0000000000000000000000000000000000000000..6f1a0daedba50e5b93d7480335cfbb2128914dbf --- /dev/null +++ b/results_bvcc/w2v2_base/layer9/run1/train.log @@ -0,0 +1,163 @@ +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=0.4906021694724376 +INFO:root: +[Valid][0][UTT][ MSE = 0.3310 | LCC = 0.8081 | SRCC = 0.8111 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2115 | LCC = 0.8622 | SRCC = 0.8650 ] +INFO:root:Average loss=0.2987025968539409 +INFO:root: +[Valid][1][UTT][ MSE = 0.3370 | LCC = 0.8306 | SRCC = 0.8315 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2132 | LCC = 0.8794 | SRCC = 0.8828 ] +INFO:root:Average loss=0.23938435125045288 +INFO:root: +[Valid][2][UTT][ MSE = 0.3048 | LCC = 0.8427 | SRCC = 0.8411 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.2206 | LCC = 0.8986 | SRCC = 0.8976 ] +INFO:root:Average loss=0.22200090151566726 +INFO:root: +[Valid][3][UTT][ MSE = 0.2597 | LCC = 0.8555 | SRCC = 0.8527 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1592 | LCC = 0.9111 | SRCC = 0.9084 ] +INFO:root:Average loss=0.2058709802535864 +INFO:root: +[Valid][4][UTT][ MSE = 0.2961 | LCC = 0.8517 | SRCC = 0.8506 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.2085 | LCC = 0.9120 | SRCC = 0.9143 ] +INFO:root:Average loss=0.1890944758286843 +INFO:root: +[Valid][5][UTT][ MSE = 0.2412 | LCC = 0.8590 | SRCC = 0.8552 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1421 | LCC = 0.9136 | SRCC = 0.9115 ] +INFO:root:Average loss=0.18050124830542466 +INFO:root: +[Valid][6][UTT][ MSE = 0.2407 | LCC = 0.8604 | SRCC = 0.8579 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1498 | LCC = 0.9099 | SRCC = 0.9079 ] +INFO:root:Average loss=0.16650586508405515 +INFO:root: +[Valid][7][UTT][ MSE = 0.2355 | LCC = 0.8610 | SRCC = 0.8584 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1593 | LCC = 0.9103 | SRCC = 0.9095 ] +INFO:root:Average loss=0.16550859388632652 +INFO:root: +[Valid][8][UTT][ MSE = 0.2596 | LCC = 0.8554 | SRCC = 0.8571 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1831 | LCC = 0.9108 | SRCC = 0.9139 ] +INFO:root:Average loss=0.15212558439144722 +INFO:root: +[Valid][9][UTT][ MSE = 0.2458 | LCC = 0.8544 | SRCC = 0.8521 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1561 | LCC = 0.9094 | SRCC = 0.9090 ] +INFO:root:Average loss=0.14105830064568764 +INFO:root: +[Valid][10][UTT][ MSE = 0.2568 | LCC = 0.8546 | SRCC = 0.8518 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1626 | LCC = 0.9145 | SRCC = 0.9109 ] +INFO:root:Average loss=0.13410663843536988 +INFO:root: +[Valid][11][UTT][ MSE = 0.2451 | LCC = 0.8604 | SRCC = 0.8579 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1303 | LCC = 0.9186 | SRCC = 0.9151 ] +INFO:root:Average loss=0.133896415718855 +INFO:root: +[Valid][12][UTT][ MSE = 0.2387 | LCC = 0.8599 | SRCC = 0.8571 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1277 | LCC = 0.9200 | SRCC = 0.9193 ] +INFO:root:Average loss=0.12418619696146403 +INFO:root: +[Valid][13][UTT][ MSE = 0.2315 | LCC = 0.8574 | SRCC = 0.8549 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1218 | LCC = 0.9207 | SRCC = 0.9189 ] +INFO:root:Average loss=0.11751198367430614 +INFO:root: +[Valid][14][UTT][ MSE = 0.2396 | LCC = 0.8561 | SRCC = 0.8535 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1334 | LCC = 0.9244 | SRCC = 0.9248 ] +INFO:root:Average loss=0.11563433553927983 +INFO:root: +[Valid][15][UTT][ MSE = 0.2406 | LCC = 0.8558 | SRCC = 0.8541 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1394 | LCC = 0.9222 | SRCC = 0.9224 ] +INFO:root:Average loss=0.10864840829983735 +INFO:root: +[Valid][16][UTT][ MSE = 0.2424 | LCC = 0.8583 | SRCC = 0.8568 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1468 | LCC = 0.9210 | SRCC = 0.9221 ] +INFO:root:Average loss=0.09832426738471557 +INFO:root: +[Valid][17][UTT][ MSE = 0.2399 | LCC = 0.8576 | SRCC = 0.8546 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1311 | LCC = 0.9179 | SRCC = 0.9124 ] +INFO:root:Average loss=0.09449597013493378 +INFO:root: +[Valid][18][UTT][ MSE = 0.2276 | LCC = 0.8571 | SRCC = 0.8553 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1217 | LCC = 0.9204 | SRCC = 0.9182 ] +INFO:root:Average loss=0.09518912897851223 +INFO:root: +[Valid][19][UTT][ MSE = 0.2446 | LCC = 0.8541 | SRCC = 0.8503 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1343 | LCC = 0.9221 | SRCC = 0.9202 ] +INFO:root:Average loss=0.10134275486836067 +INFO:root: +[Valid][20][UTT][ MSE = 0.2293 | LCC = 0.8548 | SRCC = 0.8524 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1243 | LCC = 0.9193 | SRCC = 0.9207 ] +INFO:root:Average loss=0.08166889884533027 +INFO:root: +[Valid][21][UTT][ MSE = 0.2337 | LCC = 0.8540 | SRCC = 0.8512 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1271 | LCC = 0.9199 | SRCC = 0.9192 ] +INFO:root:Average loss=0.07851536500339325 +INFO:root: +[Valid][22][UTT][ MSE = 0.2379 | LCC = 0.8557 | SRCC = 0.8528 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1420 | LCC = 0.9203 | SRCC = 0.9215 ] +INFO:root:Average loss=0.07765501603866234 +INFO:root: +[Valid][23][UTT][ MSE = 0.2358 | LCC = 0.8542 | SRCC = 0.8507 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1218 | LCC = 0.9226 | SRCC = 0.9188 ] +INFO:root:Average loss=0.07735667095925564 +INFO:root: +[Valid][24][UTT][ MSE = 0.2353 | LCC = 0.8526 | SRCC = 0.8497 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1254 | LCC = 0.9209 | SRCC = 0.9200 ] +INFO:root:Average loss=0.07626928814137593 +INFO:root: +[Valid][25][UTT][ MSE = 0.2319 | LCC = 0.8532 | SRCC = 0.8509 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1212 | LCC = 0.9233 | SRCC = 0.9239 ] +INFO:root:Average loss=0.06676949383929753 +INFO:root: +[Valid][26][UTT][ MSE = 0.2398 | LCC = 0.8524 | SRCC = 0.8500 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1225 | LCC = 0.9217 | SRCC = 0.9208 ] +INFO:root:Average loss=0.06724720094830562 +INFO:root: +[Valid][27][UTT][ MSE = 0.2377 | LCC = 0.8517 | SRCC = 0.8505 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1299 | LCC = 0.9219 | SRCC = 0.9240 ] +INFO:root:Average loss=0.06576158865713157 +INFO:root: +[Valid][28][UTT][ MSE = 0.2312 | LCC = 0.8549 | SRCC = 0.8522 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1140 | LCC = 0.9272 | SRCC = 0.9239 ] +INFO:root:Average loss=0.06337594092847446 +INFO:root: +[Valid][29][UTT][ MSE = 0.2374 | LCC = 0.8538 | SRCC = 0.8504 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1221 | LCC = 0.9256 | SRCC = 0.9242 ] +INFO:root:Average loss=0.06110485671804501 +INFO:root: +[Valid][30][UTT][ MSE = 0.2479 | LCC = 0.8502 | SRCC = 0.8480 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1482 | LCC = 0.9216 | SRCC = 0.9221 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2406 | LCC = 0.8517 | SRCC = 0.8525 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1204 | LCC = 0.9148 | SRCC = 0.9154 ] diff --git a/results_bvcc/w2v2_base/layer9/run2/config.gin b/results_bvcc/w2v2_base/layer9/run2/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer9/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 diff --git a/results_bvcc/w2v2_base/layer9/run2/model.pt b/results_bvcc/w2v2_base/layer9/run2/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..57f761240d1958500d84581db992add46ae89c9f --- /dev/null +++ b/results_bvcc/w2v2_base/layer9/run2/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:eceb51c8799309f7e97c35e9a92275626ea18a8a2d0d7bab164cf87eb7e03a1e +size 2747979 diff --git a/results_bvcc/w2v2_base/layer9/run2/model_best.pt b/results_bvcc/w2v2_base/layer9/run2/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..f46fa02f5c69b35671801d4c0c4dd5e368af8595 --- /dev/null +++ b/results_bvcc/w2v2_base/layer9/run2/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2b562f1f6e398b29e1e85b63e30a65d592e204c61c28eb1c644d048c71ae7f40 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer9/run2/train.log b/results_bvcc/w2v2_base/layer9/run2/train.log new file mode 100644 index 0000000000000000000000000000000000000000..0cee9ef5fa4c4ccd4f3f2f0c1a71400efe9fc916 --- /dev/null +++ b/results_bvcc/w2v2_base/layer9/run2/train.log @@ -0,0 +1,163 @@ +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=0.6458497463892667 +INFO:root: +[Valid][0][UTT][ MSE = 0.4431 | LCC = 0.7654 | SRCC = 0.7860 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.3394 | LCC = 0.8386 | SRCC = 0.8505 ] +INFO:root:Average loss=0.29617693026860553 +INFO:root: +[Valid][1][UTT][ MSE = 0.3128 | LCC = 0.8304 | SRCC = 0.8345 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2308 | LCC = 0.8856 | SRCC = 0.8894 ] +INFO:root:Average loss=0.24034068294060537 +INFO:root: +[Valid][2][UTT][ MSE = 0.2719 | LCC = 0.8438 | SRCC = 0.8446 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1924 | LCC = 0.8979 | SRCC = 0.9006 ] +INFO:root:Average loss=0.2180850278490629 +INFO:root: +[Valid][3][UTT][ MSE = 0.2610 | LCC = 0.8499 | SRCC = 0.8476 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1572 | LCC = 0.9140 | SRCC = 0.9128 ] +INFO:root:Average loss=0.2068684404859176 +INFO:root: +[Valid][4][UTT][ MSE = 0.2922 | LCC = 0.8517 | SRCC = 0.8524 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.2158 | LCC = 0.9090 | SRCC = 0.9145 ] +INFO:root:Average loss=0.18888308298893464 +INFO:root: +[Valid][5][UTT][ MSE = 0.2701 | LCC = 0.8569 | SRCC = 0.8560 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1916 | LCC = 0.9113 | SRCC = 0.9147 ] +INFO:root:Average loss=0.17567143455529824 +INFO:root: +[Valid][6][UTT][ MSE = 0.2936 | LCC = 0.8589 | SRCC = 0.8576 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.2203 | LCC = 0.9150 | SRCC = 0.9154 ] +INFO:root:Average loss=0.16759646913179985 +INFO:root: +[Valid][7][UTT][ MSE = 0.2390 | LCC = 0.8610 | SRCC = 0.8590 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1484 | LCC = 0.9183 | SRCC = 0.9165 ] +INFO:root:Average loss=0.1622830180403514 +INFO:root: +[Valid][8][UTT][ MSE = 0.2627 | LCC = 0.8585 | SRCC = 0.8587 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1923 | LCC = 0.9147 | SRCC = 0.9163 ] +INFO:root:Average loss=0.16626749607997063 +INFO:root: +[Valid][9][UTT][ MSE = 0.2275 | LCC = 0.8582 | SRCC = 0.8576 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1404 | LCC = 0.9139 | SRCC = 0.9154 ] +INFO:root:Average loss=0.14980575767083046 +INFO:root: +[Valid][10][UTT][ MSE = 0.2323 | LCC = 0.8609 | SRCC = 0.8586 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1448 | LCC = 0.9186 | SRCC = 0.9170 ] +INFO:root:Average loss=0.13818957274540877 +INFO:root: +[Valid][11][UTT][ MSE = 0.2858 | LCC = 0.8555 | SRCC = 0.8558 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.2207 | LCC = 0.9112 | SRCC = 0.9118 ] +INFO:root:Average loss=0.13411271256896165 +INFO:root: +[Valid][12][UTT][ MSE = 0.2449 | LCC = 0.8632 | SRCC = 0.8603 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1710 | LCC = 0.9186 | SRCC = 0.9150 ] +INFO:root:Average loss=0.12913740560030326 +INFO:root: +[Valid][13][UTT][ MSE = 0.2349 | LCC = 0.8645 | SRCC = 0.8625 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1479 | LCC = 0.9217 | SRCC = 0.9164 ] +INFO:root:Average loss=0.12060663677178897 +INFO:root: +[Valid][14][UTT][ MSE = 0.2473 | LCC = 0.8623 | SRCC = 0.8611 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1771 | LCC = 0.9159 | SRCC = 0.9145 ] +INFO:root:Average loss=0.10985142250473683 +INFO:root: +[Valid][15][UTT][ MSE = 0.2278 | LCC = 0.8590 | SRCC = 0.8567 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1367 | LCC = 0.9174 | SRCC = 0.9129 ] +INFO:root:Average loss=0.10796022988282718 +INFO:root: +[Valid][16][UTT][ MSE = 0.2423 | LCC = 0.8612 | SRCC = 0.8596 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1530 | LCC = 0.9152 | SRCC = 0.9112 ] +INFO:root:Average loss=0.10152303241193295 +INFO:root: +[Valid][17][UTT][ MSE = 0.2577 | LCC = 0.8595 | SRCC = 0.8572 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1833 | LCC = 0.9181 | SRCC = 0.9143 ] +INFO:root:Average loss=0.10260800157602017 +INFO:root: +[Valid][18][UTT][ MSE = 0.2614 | LCC = 0.8575 | SRCC = 0.8565 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1884 | LCC = 0.9155 | SRCC = 0.9136 ] +INFO:root:Average loss=0.09271644724485202 +INFO:root: +[Valid][19][UTT][ MSE = 0.2429 | LCC = 0.8601 | SRCC = 0.8602 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1662 | LCC = 0.9186 | SRCC = 0.9192 ] +INFO:root:Average loss=0.09358113588621983 +INFO:root: +[Valid][20][UTT][ MSE = 0.2496 | LCC = 0.8614 | SRCC = 0.8608 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1727 | LCC = 0.9195 | SRCC = 0.9183 ] +INFO:root:Average loss=0.08826010511853756 +INFO:root: +[Valid][21][UTT][ MSE = 0.2776 | LCC = 0.8597 | SRCC = 0.8600 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.2085 | LCC = 0.9177 | SRCC = 0.9171 ] +INFO:root:Average loss=0.08861871531758553 +INFO:root: +[Valid][22][UTT][ MSE = 0.2372 | LCC = 0.8584 | SRCC = 0.8574 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1499 | LCC = 0.9202 | SRCC = 0.9175 ] +INFO:root:Average loss=0.08886236310578309 +INFO:root: +[Valid][23][UTT][ MSE = 0.2398 | LCC = 0.8580 | SRCC = 0.8570 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1588 | LCC = 0.9152 | SRCC = 0.9131 ] +INFO:root:Average loss=0.07984707805399711 +INFO:root: +[Valid][24][UTT][ MSE = 0.2856 | LCC = 0.8580 | SRCC = 0.8583 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.2238 | LCC = 0.9150 | SRCC = 0.9154 ] +INFO:root:Average loss=0.07138788489959179 +INFO:root: +[Valid][25][UTT][ MSE = 0.2291 | LCC = 0.8577 | SRCC = 0.8572 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1326 | LCC = 0.9235 | SRCC = 0.9187 ] +INFO:root:Average loss=0.06980475936180507 +INFO:root: +[Valid][26][UTT][ MSE = 0.2494 | LCC = 0.8555 | SRCC = 0.8545 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1616 | LCC = 0.9244 | SRCC = 0.9208 ] +INFO:root:Average loss=0.0740565094523705 +INFO:root: +[Valid][27][UTT][ MSE = 0.2391 | LCC = 0.8606 | SRCC = 0.8591 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1527 | LCC = 0.9257 | SRCC = 0.9228 ] +INFO:root:Average loss=0.06538710862589188 +INFO:root: +[Valid][28][UTT][ MSE = 0.2369 | LCC = 0.8570 | SRCC = 0.8576 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1538 | LCC = 0.9204 | SRCC = 0.9192 ] +INFO:root:Average loss=0.06615091325380863 +INFO:root: +[Valid][29][UTT][ MSE = 0.3064 | LCC = 0.8539 | SRCC = 0.8551 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.2375 | LCC = 0.9198 | SRCC = 0.9210 ] +INFO:root:Average loss=0.05938345776536526 +INFO:root: +[Valid][30][UTT][ MSE = 0.2609 | LCC = 0.8580 | SRCC = 0.8569 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1898 | LCC = 0.9189 | SRCC = 0.9140 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2427 | LCC = 0.8573 | SRCC = 0.8578 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1293 | LCC = 0.9182 | SRCC = 0.9196 ] diff --git a/results_bvcc/w2v2_base/layer9/run3/config.gin b/results_bvcc/w2v2_base/layer9/run3/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer9/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 diff --git a/results_bvcc/w2v2_base/layer9/run3/model.pt b/results_bvcc/w2v2_base/layer9/run3/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..149259fb7cfdd3d286ac1981fde7e198aadbec4d --- /dev/null +++ b/results_bvcc/w2v2_base/layer9/run3/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2ad59c6ba21ba250305cbe0668ed9876ed383619baee55285dc3fa4cf3775d9e +size 2747979 diff --git a/results_bvcc/w2v2_base/layer9/run3/model_best.pt b/results_bvcc/w2v2_base/layer9/run3/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..3c8bcf3dfe1aace411a149e3772f9702ec266c1c --- /dev/null +++ b/results_bvcc/w2v2_base/layer9/run3/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:2db433363d55ad7a0c687e57bd0fc492cbc2ebabcb715047462dc77e947aabd6 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer9/run3/train.log b/results_bvcc/w2v2_base/layer9/run3/train.log new file mode 100644 index 0000000000000000000000000000000000000000..12f0ebc81342963e124c0a6e6e6df1db14b6f80e --- /dev/null +++ b/results_bvcc/w2v2_base/layer9/run3/train.log @@ -0,0 +1,163 @@ +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=0.5315741673111916 +INFO:root: +[Valid][0][UTT][ MSE = 0.3157 | LCC = 0.8037 | SRCC = 0.8142 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2156 | LCC = 0.8624 | SRCC = 0.8714 ] +INFO:root:Average loss=0.28470108963740176 +INFO:root: +[Valid][1][UTT][ MSE = 0.2737 | LCC = 0.8357 | SRCC = 0.8361 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.1758 | LCC = 0.8881 | SRCC = 0.8886 ] +INFO:root:Average loss=0.24258504177515322 +INFO:root: +[Valid][2][UTT][ MSE = 0.2700 | LCC = 0.8473 | SRCC = 0.8465 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1631 | LCC = 0.8990 | SRCC = 0.8966 ] +INFO:root:Average loss=0.23840844879547754 +INFO:root: +[Valid][3][UTT][ MSE = 0.2672 | LCC = 0.8536 | SRCC = 0.8532 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1838 | LCC = 0.9057 | SRCC = 0.9056 ] +INFO:root:Average loss=0.21266147398795837 +INFO:root: +[Valid][4][UTT][ MSE = 0.3164 | LCC = 0.8518 | SRCC = 0.8548 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.2514 | LCC = 0.9019 | SRCC = 0.9045 ] +INFO:root:Average loss=0.19881071933569053 +INFO:root: +[Valid][5][UTT][ MSE = 0.2530 | LCC = 0.8564 | SRCC = 0.8542 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1658 | LCC = 0.9089 | SRCC = 0.9044 ] +INFO:root:Average loss=0.17918571380850595 +INFO:root: +[Valid][6][UTT][ MSE = 0.2375 | LCC = 0.8604 | SRCC = 0.8598 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1462 | LCC = 0.9162 | SRCC = 0.9157 ] +INFO:root:Average loss=0.166574482829907 +INFO:root: +[Valid][7][UTT][ MSE = 0.2280 | LCC = 0.8591 | SRCC = 0.8583 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1323 | LCC = 0.9171 | SRCC = 0.9163 ] +INFO:root:Average loss=0.1618993341540679 +INFO:root: +[Valid][8][UTT][ MSE = 0.2512 | LCC = 0.8627 | SRCC = 0.8606 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1728 | LCC = 0.9175 | SRCC = 0.9142 ] +INFO:root:Average loss=0.15824708715081215 +INFO:root: +[Valid][9][UTT][ MSE = 0.2412 | LCC = 0.8623 | SRCC = 0.8608 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1455 | LCC = 0.9254 | SRCC = 0.9196 ] +INFO:root:Average loss=0.1486571814196232 +INFO:root: +[Valid][10][UTT][ MSE = 0.2359 | LCC = 0.8620 | SRCC = 0.8595 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1270 | LCC = 0.9237 | SRCC = 0.9213 ] +INFO:root:Average loss=0.15305039840630996 +INFO:root: +[Valid][11][UTT][ MSE = 0.2294 | LCC = 0.8656 | SRCC = 0.8630 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1290 | LCC = 0.9250 | SRCC = 0.9200 ] +INFO:root:Average loss=0.1315722850461801 +INFO:root: +[Valid][12][UTT][ MSE = 0.2467 | LCC = 0.8604 | SRCC = 0.8610 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1576 | LCC = 0.9230 | SRCC = 0.9216 ] +INFO:root:Average loss=0.14162453779807457 +INFO:root: +[Valid][13][UTT][ MSE = 0.2350 | LCC = 0.8635 | SRCC = 0.8608 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1300 | LCC = 0.9222 | SRCC = 0.9170 ] +INFO:root:Average loss=0.1164270434050988 +INFO:root: +[Valid][14][UTT][ MSE = 0.2266 | LCC = 0.8606 | SRCC = 0.8583 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1256 | LCC = 0.9250 | SRCC = 0.9182 ] +INFO:root:Average loss=0.11853593568771313 +INFO:root: +[Valid][15][UTT][ MSE = 0.2292 | LCC = 0.8614 | SRCC = 0.8593 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1321 | LCC = 0.9245 | SRCC = 0.9188 ] +INFO:root:Average loss=0.10972892626737937 +INFO:root: +[Valid][16][UTT][ MSE = 0.2234 | LCC = 0.8614 | SRCC = 0.8600 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1110 | LCC = 0.9238 | SRCC = 0.9184 ] +INFO:root:Average loss=0.10610254400242598 +INFO:root: +[Valid][17][UTT][ MSE = 0.2216 | LCC = 0.8650 | SRCC = 0.8640 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1257 | LCC = 0.9225 | SRCC = 0.9178 ] +INFO:root:Average loss=0.10958677601928894 +INFO:root: +[Valid][18][UTT][ MSE = 0.2282 | LCC = 0.8598 | SRCC = 0.8585 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1262 | LCC = 0.9239 | SRCC = 0.9201 ] +INFO:root:Average loss=0.10009780096320006 +INFO:root: +[Valid][19][UTT][ MSE = 0.2315 | LCC = 0.8591 | SRCC = 0.8557 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1188 | LCC = 0.9254 | SRCC = 0.9193 ] +INFO:root:Average loss=0.09311500526009461 +INFO:root: +[Valid][20][UTT][ MSE = 0.2348 | LCC = 0.8581 | SRCC = 0.8555 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1366 | LCC = 0.9233 | SRCC = 0.9169 ] +INFO:root:Average loss=0.08736066114252959 +INFO:root: +[Valid][21][UTT][ MSE = 0.2399 | LCC = 0.8581 | SRCC = 0.8560 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1409 | LCC = 0.9250 | SRCC = 0.9192 ] +INFO:root:Average loss=0.08658311802607316 +INFO:root: +[Valid][22][UTT][ MSE = 0.2314 | LCC = 0.8605 | SRCC = 0.8582 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1265 | LCC = 0.9278 | SRCC = 0.9211 ] +INFO:root:Average loss=0.08099264412736282 +INFO:root: +[Valid][23][UTT][ MSE = 0.2357 | LCC = 0.8577 | SRCC = 0.8562 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1189 | LCC = 0.9277 | SRCC = 0.9206 ] +INFO:root:Average loss=0.08360019617546828 +INFO:root: +[Valid][24][UTT][ MSE = 0.2354 | LCC = 0.8604 | SRCC = 0.8589 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1283 | LCC = 0.9268 | SRCC = 0.9189 ] +INFO:root:Average loss=0.0733444409397168 +INFO:root: +[Valid][25][UTT][ MSE = 0.2219 | LCC = 0.8578 | SRCC = 0.8574 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1099 | LCC = 0.9279 | SRCC = 0.9238 ] +INFO:root:Average loss=0.08087875990149303 +INFO:root: +[Valid][26][UTT][ MSE = 0.2349 | LCC = 0.8588 | SRCC = 0.8570 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1335 | LCC = 0.9277 | SRCC = 0.9221 ] +INFO:root:Average loss=0.07024355423756135 +INFO:root: +[Valid][27][UTT][ MSE = 0.2316 | LCC = 0.8547 | SRCC = 0.8532 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1244 | LCC = 0.9222 | SRCC = 0.9175 ] +INFO:root:Average loss=0.07358622092467088 +INFO:root: +[Valid][28][UTT][ MSE = 0.2233 | LCC = 0.8575 | SRCC = 0.8558 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1125 | LCC = 0.9276 | SRCC = 0.9222 ] +INFO:root:Average loss=0.07223565322466385 +INFO:root: +[Valid][29][UTT][ MSE = 0.2287 | LCC = 0.8589 | SRCC = 0.8590 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1170 | LCC = 0.9300 | SRCC = 0.9243 ] +INFO:root:Average loss=0.0634167074966125 +INFO:root: +[Valid][30][UTT][ MSE = 0.2293 | LCC = 0.8604 | SRCC = 0.8578 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1309 | LCC = 0.9275 | SRCC = 0.9207 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2484 | LCC = 0.8527 | SRCC = 0.8500 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1436 | LCC = 0.9177 | SRCC = 0.9193 ] diff --git a/results_bvcc/w2v2_base/layer9/run4/config.gin b/results_bvcc/w2v2_base/layer9/run4/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base/layer9/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 diff --git a/results_bvcc/w2v2_base/layer9/run4/model.pt b/results_bvcc/w2v2_base/layer9/run4/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..30fe3fae65e09eaabfac0ed46305f799eda498f6 --- /dev/null +++ b/results_bvcc/w2v2_base/layer9/run4/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:1c14bd61885e96e7171d6a2bdeff356fc0516813060960476f35ae8ded72d447 +size 2747979 diff --git a/results_bvcc/w2v2_base/layer9/run4/model_best.pt b/results_bvcc/w2v2_base/layer9/run4/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..e9f687cd784a6067eae240f55108658d662bc38c --- /dev/null +++ b/results_bvcc/w2v2_base/layer9/run4/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4ebbc65243a72d6bc06d563e82db4e2cfb62e1d4930f37c4f5347710746ab4a0 +size 2749702 diff --git a/results_bvcc/w2v2_base/layer9/run4/train.log b/results_bvcc/w2v2_base/layer9/run4/train.log new file mode 100644 index 0000000000000000000000000000000000000000..cc03a862c5bf892f650a17b56a8780c49cb30083 --- /dev/null +++ b/results_bvcc/w2v2_base/layer9/run4/train.log @@ -0,0 +1,163 @@ +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=0.5980511823525796 +INFO:root: +[Valid][0][UTT][ MSE = 0.3506 | LCC = 0.7904 | SRCC = 0.8036 ] +INFO:root: +[Valid][0][SYS][ MSE = 0.2495 | LCC = 0.8533 | SRCC = 0.8614 ] +INFO:root:Average loss=0.2669436196104074 +INFO:root: +[Valid][1][UTT][ MSE = 0.2927 | LCC = 0.8390 | SRCC = 0.8429 ] +INFO:root: +[Valid][1][SYS][ MSE = 0.2025 | LCC = 0.8970 | SRCC = 0.8997 ] +INFO:root:Average loss=0.22954150174672788 +INFO:root: +[Valid][2][UTT][ MSE = 0.2597 | LCC = 0.8529 | SRCC = 0.8546 ] +INFO:root: +[Valid][2][SYS][ MSE = 0.1572 | LCC = 0.9157 | SRCC = 0.9166 ] +INFO:root:Average loss=0.21201087639499933 +INFO:root: +[Valid][3][UTT][ MSE = 0.2655 | LCC = 0.8592 | SRCC = 0.8584 ] +INFO:root: +[Valid][3][SYS][ MSE = 0.1590 | LCC = 0.9199 | SRCC = 0.9169 ] +INFO:root:Average loss=0.19595668312066641 +INFO:root: +[Valid][4][UTT][ MSE = 0.2439 | LCC = 0.8630 | SRCC = 0.8623 ] +INFO:root: +[Valid][4][SYS][ MSE = 0.1448 | LCC = 0.9191 | SRCC = 0.9165 ] +INFO:root:Average loss=0.18703980428668168 +INFO:root: +[Valid][5][UTT][ MSE = 0.2436 | LCC = 0.8655 | SRCC = 0.8655 ] +INFO:root: +[Valid][5][SYS][ MSE = 0.1478 | LCC = 0.9250 | SRCC = 0.9229 ] +INFO:root:Average loss=0.1817215696359292 +INFO:root: +[Valid][6][UTT][ MSE = 0.2433 | LCC = 0.8646 | SRCC = 0.8636 ] +INFO:root: +[Valid][6][SYS][ MSE = 0.1370 | LCC = 0.9265 | SRCC = 0.9252 ] +INFO:root:Average loss=0.16478320535941002 +INFO:root: +[Valid][7][UTT][ MSE = 0.2445 | LCC = 0.8646 | SRCC = 0.8636 ] +INFO:root: +[Valid][7][SYS][ MSE = 0.1394 | LCC = 0.9231 | SRCC = 0.9178 ] +INFO:root:Average loss=0.17197712377095833 +INFO:root: +[Valid][8][UTT][ MSE = 0.2399 | LCC = 0.8653 | SRCC = 0.8640 ] +INFO:root: +[Valid][8][SYS][ MSE = 0.1285 | LCC = 0.9235 | SRCC = 0.9217 ] +INFO:root:Average loss=0.1428469917139946 +INFO:root: +[Valid][9][UTT][ MSE = 0.2293 | LCC = 0.8651 | SRCC = 0.8643 ] +INFO:root: +[Valid][9][SYS][ MSE = 0.1248 | LCC = 0.9295 | SRCC = 0.9265 ] +INFO:root:Average loss=0.13779205026534888 +INFO:root: +[Valid][10][UTT][ MSE = 0.2227 | LCC = 0.8662 | SRCC = 0.8656 ] +INFO:root: +[Valid][10][SYS][ MSE = 0.1228 | LCC = 0.9287 | SRCC = 0.9259 ] +INFO:root:Average loss=0.13392920123460966 +INFO:root: +[Valid][11][UTT][ MSE = 0.2370 | LCC = 0.8609 | SRCC = 0.8596 ] +INFO:root: +[Valid][11][SYS][ MSE = 0.1237 | LCC = 0.9228 | SRCC = 0.9154 ] +INFO:root:Average loss=0.12532686470792845 +INFO:root: +[Valid][12][UTT][ MSE = 0.2343 | LCC = 0.8641 | SRCC = 0.8641 ] +INFO:root: +[Valid][12][SYS][ MSE = 0.1294 | LCC = 0.9252 | SRCC = 0.9206 ] +INFO:root:Average loss=0.12905744415445206 +INFO:root: +[Valid][13][UTT][ MSE = 0.2286 | LCC = 0.8641 | SRCC = 0.8633 ] +INFO:root: +[Valid][13][SYS][ MSE = 0.1374 | LCC = 0.9266 | SRCC = 0.9241 ] +INFO:root:Average loss=0.11065948382019997 +INFO:root: +[Valid][14][UTT][ MSE = 0.2194 | LCC = 0.8650 | SRCC = 0.8649 ] +INFO:root: +[Valid][14][SYS][ MSE = 0.1073 | LCC = 0.9330 | SRCC = 0.9280 ] +INFO:root:Average loss=0.10826678048723783 +INFO:root: +[Valid][15][UTT][ MSE = 0.2221 | LCC = 0.8646 | SRCC = 0.8651 ] +INFO:root: +[Valid][15][SYS][ MSE = 0.1148 | LCC = 0.9313 | SRCC = 0.9251 ] +INFO:root:Average loss=0.11520418104452965 +INFO:root: +[Valid][16][UTT][ MSE = 0.2297 | LCC = 0.8636 | SRCC = 0.8631 ] +INFO:root: +[Valid][16][SYS][ MSE = 0.1416 | LCC = 0.9231 | SRCC = 0.9185 ] +INFO:root:Average loss=0.09981599383247204 +INFO:root: +[Valid][17][UTT][ MSE = 0.2330 | LCC = 0.8603 | SRCC = 0.8602 ] +INFO:root: +[Valid][17][SYS][ MSE = 0.1244 | LCC = 0.9214 | SRCC = 0.9174 ] +INFO:root:Average loss=0.09710770310499729 +INFO:root: +[Valid][18][UTT][ MSE = 0.2410 | LCC = 0.8603 | SRCC = 0.8604 ] +INFO:root: +[Valid][18][SYS][ MSE = 0.1311 | LCC = 0.9229 | SRCC = 0.9182 ] +INFO:root:Average loss=0.08577284293297009 +INFO:root: +[Valid][19][UTT][ MSE = 0.2408 | LCC = 0.8601 | SRCC = 0.8605 ] +INFO:root: +[Valid][19][SYS][ MSE = 0.1247 | LCC = 0.9215 | SRCC = 0.9196 ] +INFO:root:Average loss=0.08453243187604806 +INFO:root: +[Valid][20][UTT][ MSE = 0.2465 | LCC = 0.8610 | SRCC = 0.8602 ] +INFO:root: +[Valid][20][SYS][ MSE = 0.1346 | LCC = 0.9212 | SRCC = 0.9159 ] +INFO:root:Average loss=0.09004265380402406 +INFO:root: +[Valid][21][UTT][ MSE = 0.2214 | LCC = 0.8612 | SRCC = 0.8610 ] +INFO:root: +[Valid][21][SYS][ MSE = 0.1272 | LCC = 0.9232 | SRCC = 0.9209 ] +INFO:root:Average loss=0.07454365706787659 +INFO:root: +[Valid][22][UTT][ MSE = 0.2370 | LCC = 0.8591 | SRCC = 0.8570 ] +INFO:root: +[Valid][22][SYS][ MSE = 0.1222 | LCC = 0.9291 | SRCC = 0.9223 ] +INFO:root:Average loss=0.07639849959657742 +INFO:root: +[Valid][23][UTT][ MSE = 0.2309 | LCC = 0.8587 | SRCC = 0.8588 ] +INFO:root: +[Valid][23][SYS][ MSE = 0.1195 | LCC = 0.9247 | SRCC = 0.9218 ] +INFO:root:Average loss=0.06665781006599084 +INFO:root: +[Valid][24][UTT][ MSE = 0.2496 | LCC = 0.8564 | SRCC = 0.8556 ] +INFO:root: +[Valid][24][SYS][ MSE = 0.1235 | LCC = 0.9270 | SRCC = 0.9221 ] +INFO:root:Average loss=0.07035466116399337 +INFO:root: +[Valid][25][UTT][ MSE = 0.2320 | LCC = 0.8614 | SRCC = 0.8619 ] +INFO:root: +[Valid][25][SYS][ MSE = 0.1149 | LCC = 0.9281 | SRCC = 0.9281 ] +INFO:root:Average loss=0.06847770096590886 +INFO:root: +[Valid][26][UTT][ MSE = 0.2498 | LCC = 0.8580 | SRCC = 0.8584 ] +INFO:root: +[Valid][26][SYS][ MSE = 0.1290 | LCC = 0.9230 | SRCC = 0.9214 ] +INFO:root:Average loss=0.06642000276881915 +INFO:root: +[Valid][27][UTT][ MSE = 0.2297 | LCC = 0.8596 | SRCC = 0.8590 ] +INFO:root: +[Valid][27][SYS][ MSE = 0.1348 | LCC = 0.9245 | SRCC = 0.9197 ] +INFO:root:Average loss=0.06052997765632776 +INFO:root: +[Valid][28][UTT][ MSE = 0.2333 | LCC = 0.8595 | SRCC = 0.8581 ] +INFO:root: +[Valid][28][SYS][ MSE = 0.1191 | LCC = 0.9261 | SRCC = 0.9207 ] +INFO:root:Average loss=0.060885779989453465 +INFO:root: +[Valid][29][UTT][ MSE = 0.2278 | LCC = 0.8585 | SRCC = 0.8579 ] +INFO:root: +[Valid][29][SYS][ MSE = 0.1166 | LCC = 0.9246 | SRCC = 0.9206 ] +INFO:root:Average loss=0.06202450996408096 +INFO:root: +[Valid][30][UTT][ MSE = 0.2369 | LCC = 0.8604 | SRCC = 0.8593 ] +INFO:root: +[Valid][30][SYS][ MSE = 0.1235 | LCC = 0.9271 | SRCC = 0.9227 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2405 | LCC = 0.8559 | SRCC = 0.8565 ] +INFO:root: +[Test][31][SYS][ MSE = 0.1231 | LCC = 0.9244 | SRCC = 0.9261 ] diff --git a/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run0/config.gin b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run0/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/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 diff --git a/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run0/model.pt b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run0/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..b2c12ab800e85bb2a667d9df87f0de57f33bcaa1 --- /dev/null +++ b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run0/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c6ceb7cef901e56b997ce7816ee22c09e9d8642b6d6269f4a557728a971cefff +size 2747979 diff --git a/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run0/model_best.pt b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run0/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..46f4d1f3ffcada54e6ec669658628c8978991384 --- /dev/null +++ b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run0/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:156c2feb093a56d5f10c66b8333b6dd358e2f698e2c4809e33a777be1779b26f +size 2749638 diff --git a/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run0/train.log b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run0/train.log new file mode 100644 index 0000000000000000000000000000000000000000..49aab9c4a7254b581313a0e588abda120e64fa3a --- /dev/null +++ b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/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=0.50291807949543 +INFO:root: +[Valid][0][UTT][ MSE = 0.3648 | LCC = 0.7637 | SRCC = 0.7663 ] +INFO:root:Average loss=0.29433613136792797 +INFO:root: +[Valid][1][UTT][ MSE = 0.3153 | LCC = 0.8090 | SRCC = 0.8112 ] +INFO:root:Average loss=0.25963135808706284 +INFO:root: +[Valid][2][UTT][ MSE = 0.2783 | LCC = 0.8225 | SRCC = 0.8217 ] +INFO:root:Average loss=0.23644148271817428 +INFO:root: +[Valid][3][UTT][ MSE = 0.2831 | LCC = 0.8308 | SRCC = 0.8300 ] +INFO:root:Average loss=0.21857103208700815 +INFO:root: +[Valid][4][UTT][ MSE = 0.2671 | LCC = 0.8334 | SRCC = 0.8315 ] +INFO:root:Average loss=0.20735690131401405 +INFO:root: +[Valid][5][UTT][ MSE = 0.2611 | LCC = 0.8349 | SRCC = 0.8331 ] +INFO:root:Average loss=0.1999413917462031 +INFO:root: +[Valid][6][UTT][ MSE = 0.2679 | LCC = 0.8429 | SRCC = 0.8404 ] +INFO:root:Average loss=0.19058068497822836 +INFO:root: +[Valid][7][UTT][ MSE = 0.2647 | LCC = 0.8390 | SRCC = 0.8396 ] +INFO:root:Average loss=0.18695699776976538 +INFO:root: +[Valid][8][UTT][ MSE = 0.2465 | LCC = 0.8402 | SRCC = 0.8399 ] +INFO:root:Average loss=0.18114551300039658 +INFO:root: +[Valid][9][UTT][ MSE = 0.2901 | LCC = 0.8391 | SRCC = 0.8394 ] +INFO:root:Average loss=0.17618295273337609 +INFO:root: +[Valid][10][UTT][ MSE = 0.2474 | LCC = 0.8455 | SRCC = 0.8437 ] +INFO:root:Average loss=0.15937039437584388 +INFO:root: +[Valid][11][UTT][ MSE = 0.2473 | LCC = 0.8485 | SRCC = 0.8484 ] +INFO:root:Average loss=0.1480565069195552 +INFO:root: +[Valid][12][UTT][ MSE = 0.2397 | LCC = 0.8508 | SRCC = 0.8491 ] +INFO:root:Average loss=0.14147661502162615 +INFO:root: +[Valid][13][UTT][ MSE = 0.2333 | LCC = 0.8510 | SRCC = 0.8510 ] +INFO:root:Average loss=0.13626917470724154 +INFO:root: +[Valid][14][UTT][ MSE = 0.2619 | LCC = 0.8469 | SRCC = 0.8472 ] +INFO:root:Average loss=0.13868246982112908 +INFO:root: +[Valid][15][UTT][ MSE = 0.2437 | LCC = 0.8534 | SRCC = 0.8534 ] +INFO:root:Average loss=0.1277079949967372 +INFO:root: +[Valid][16][UTT][ MSE = 0.2327 | LCC = 0.8526 | SRCC = 0.8510 ] +INFO:root:Average loss=0.11859527020118175 +INFO:root: +[Valid][17][UTT][ MSE = 0.2395 | LCC = 0.8528 | SRCC = 0.8513 ] +INFO:root:Average loss=0.1199362212075637 +INFO:root: +[Valid][18][UTT][ MSE = 0.2339 | LCC = 0.8536 | SRCC = 0.8520 ] +INFO:root:Average loss=0.11384502817422916 +INFO:root: +[Valid][19][UTT][ MSE = 0.2366 | LCC = 0.8503 | SRCC = 0.8486 ] +INFO:root:Average loss=0.11150806492719895 +INFO:root: +[Valid][20][UTT][ MSE = 0.2416 | LCC = 0.8491 | SRCC = 0.8476 ] +INFO:root:Average loss=0.11710820027077809 +INFO:root: +[Valid][21][UTT][ MSE = 0.2375 | LCC = 0.8493 | SRCC = 0.8487 ] +INFO:root:Average loss=0.09969277507983722 +INFO:root: +[Valid][22][UTT][ MSE = 0.2291 | LCC = 0.8539 | SRCC = 0.8541 ] +INFO:root:Average loss=0.09246688240613693 +INFO:root: +[Valid][23][UTT][ MSE = 0.2354 | LCC = 0.8541 | SRCC = 0.8519 ] +INFO:root:Average loss=0.09564988401073676 +INFO:root: +[Valid][24][UTT][ MSE = 0.2468 | LCC = 0.8493 | SRCC = 0.8485 ] +INFO:root:Average loss=0.09205347863145363 +INFO:root: +[Valid][25][UTT][ MSE = 0.2343 | LCC = 0.8523 | SRCC = 0.8513 ] +INFO:root:Average loss=0.09188771744569142 +INFO:root: +[Valid][26][UTT][ MSE = 0.2292 | LCC = 0.8530 | SRCC = 0.8507 ] +INFO:root:Average loss=0.08994028363854457 +INFO:root: +[Valid][27][UTT][ MSE = 0.2391 | LCC = 0.8518 | SRCC = 0.8522 ] +INFO:root:Average loss=0.08169009612920956 +INFO:root: +[Valid][28][UTT][ MSE = 0.2274 | LCC = 0.8540 | SRCC = 0.8534 ] +INFO:root:Average loss=0.0821145710368187 +INFO:root: +[Valid][29][UTT][ MSE = 0.2300 | LCC = 0.8551 | SRCC = 0.8534 ] +INFO:root:Average loss=0.08127826141814391 +INFO:root: +[Valid][30][UTT][ MSE = 0.2317 | LCC = 0.8523 | SRCC = 0.8512 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2367 | LCC = 0.8569 | SRCC = 0.8557 ] diff --git a/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run1/config.gin b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run1/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/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 diff --git a/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run1/model.pt b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run1/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..da1d278327aa91897f4e9b05fe77d5c774499b25 --- /dev/null +++ b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run1/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:51fa2e7214a13662edefd87c81e1e13f5f303767729bbfb68254ed69a031e43d +size 2747979 diff --git a/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run1/model_best.pt b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run1/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..cfb371f454dfa6216c4513824fbb92118f28cfba --- /dev/null +++ b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run1/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c5acb0f47e66a459777c7d32314474d7d7d8fcd26e9d15f76a9e3dbea0a673c8 +size 2749702 diff --git a/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run1/train.log b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run1/train.log new file mode 100644 index 0000000000000000000000000000000000000000..06ff2c87437ed62ff7aaff96b5629b3d87093ec5 --- /dev/null +++ b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/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=0.5855056796318445 +INFO:root: +[Valid][0][UTT][ MSE = 0.3978 | LCC = 0.7501 | SRCC = 0.7472 ] +INFO:root:Average loss=0.3052517265463487 +INFO:root: +[Valid][1][UTT][ MSE = 0.3064 | LCC = 0.7988 | SRCC = 0.8004 ] +INFO:root:Average loss=0.25929620766486877 +INFO:root: +[Valid][2][UTT][ MSE = 0.2865 | LCC = 0.8185 | SRCC = 0.8203 ] +INFO:root:Average loss=0.24464651178091001 +INFO:root: +[Valid][3][UTT][ MSE = 0.2874 | LCC = 0.8278 | SRCC = 0.8259 ] +INFO:root:Average loss=0.23131250943510961 +INFO:root: +[Valid][4][UTT][ MSE = 0.2881 | LCC = 0.8228 | SRCC = 0.8250 ] +INFO:root:Average loss=0.21194701785078415 +INFO:root: +[Valid][5][UTT][ MSE = 0.2773 | LCC = 0.8356 | SRCC = 0.8332 ] +INFO:root:Average loss=0.2097265955347281 +INFO:root: +[Valid][6][UTT][ MSE = 0.2949 | LCC = 0.8303 | SRCC = 0.8329 ] +INFO:root:Average loss=0.1927090436220169 +INFO:root: +[Valid][7][UTT][ MSE = 0.2632 | LCC = 0.8354 | SRCC = 0.8358 ] +INFO:root:Average loss=0.18096944871239173 +INFO:root: +[Valid][8][UTT][ MSE = 0.2530 | LCC = 0.8410 | SRCC = 0.8425 ] +INFO:root:Average loss=0.16814967540976328 +INFO:root: +[Valid][9][UTT][ MSE = 0.2561 | LCC = 0.8443 | SRCC = 0.8453 ] +INFO:root:Average loss=0.15789712487887114 +INFO:root: +[Valid][10][UTT][ MSE = 0.2409 | LCC = 0.8446 | SRCC = 0.8454 ] +INFO:root:Average loss=0.16342571950875795 +INFO:root: +[Valid][11][UTT][ MSE = 0.2449 | LCC = 0.8452 | SRCC = 0.8465 ] +INFO:root:Average loss=0.14636064750643876 +INFO:root: +[Valid][12][UTT][ MSE = 0.2314 | LCC = 0.8519 | SRCC = 0.8513 ] +INFO:root:Average loss=0.1424512203114155 +INFO:root: +[Valid][13][UTT][ MSE = 0.2388 | LCC = 0.8519 | SRCC = 0.8503 ] +INFO:root:Average loss=0.13201338138717872 +INFO:root: +[Valid][14][UTT][ MSE = 0.2377 | LCC = 0.8475 | SRCC = 0.8487 ] +INFO:root:Average loss=0.12394143478610577 +INFO:root: +[Valid][15][UTT][ MSE = 0.2473 | LCC = 0.8503 | SRCC = 0.8497 ] +INFO:root:Average loss=0.1235882163238831 +INFO:root: +[Valid][16][UTT][ MSE = 0.2420 | LCC = 0.8508 | SRCC = 0.8501 ] +INFO:root:Average loss=0.120010966864916 +INFO:root: +[Valid][17][UTT][ MSE = 0.2569 | LCC = 0.8464 | SRCC = 0.8474 ] +INFO:root:Average loss=0.11158457866463906 +INFO:root: +[Valid][18][UTT][ MSE = 0.2585 | LCC = 0.8518 | SRCC = 0.8521 ] +INFO:root:Average loss=0.11445855530790794 +INFO:root: +[Valid][19][UTT][ MSE = 0.2447 | LCC = 0.8494 | SRCC = 0.8493 ] +INFO:root:Average loss=0.1003785939552845 +INFO:root: +[Valid][20][UTT][ MSE = 0.2315 | LCC = 0.8523 | SRCC = 0.8519 ] +INFO:root:Average loss=0.10885728198366287 +INFO:root: +[Valid][21][UTT][ MSE = 0.2444 | LCC = 0.8478 | SRCC = 0.8494 ] +INFO:root:Average loss=0.09760843952878928 +INFO:root: +[Valid][22][UTT][ MSE = 0.2289 | LCC = 0.8545 | SRCC = 0.8545 ] +INFO:root:Average loss=0.09018352169256943 +INFO:root: +[Valid][23][UTT][ MSE = 0.2397 | LCC = 0.8516 | SRCC = 0.8509 ] +INFO:root:Average loss=0.08714846731760563 +INFO:root: +[Valid][24][UTT][ MSE = 0.2351 | LCC = 0.8501 | SRCC = 0.8497 ] +INFO:root:Average loss=0.0990918553314912 +INFO:root: +[Valid][25][UTT][ MSE = 0.2288 | LCC = 0.8536 | SRCC = 0.8524 ] +INFO:root:Average loss=0.08018444894024959 +INFO:root: +[Valid][26][UTT][ MSE = 0.2457 | LCC = 0.8529 | SRCC = 0.8519 ] +INFO:root:Average loss=0.08357787834337124 +INFO:root: +[Valid][27][UTT][ MSE = 0.2271 | LCC = 0.8544 | SRCC = 0.8537 ] +INFO:root:Average loss=0.07903660924579853 +INFO:root: +[Valid][28][UTT][ MSE = 0.2302 | LCC = 0.8560 | SRCC = 0.8530 ] +INFO:root:Average loss=0.07842154547763176 +INFO:root: +[Valid][29][UTT][ MSE = 0.2261 | LCC = 0.8546 | SRCC = 0.8540 ] +INFO:root:Average loss=0.0805959433126144 +INFO:root: +[Valid][30][UTT][ MSE = 0.2381 | LCC = 0.8528 | SRCC = 0.8536 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2611 | LCC = 0.8495 | SRCC = 0.8480 ] diff --git a/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run2/config.gin b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run2/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/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 diff --git a/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run2/model.pt b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run2/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..a663bf2e57fa52d648a44e7b7edb1f82b803156a --- /dev/null +++ b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run2/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3d0a3326fdab0fb6a625a52187ea407ee76f3521b0198cf888808d4734907008 +size 2747979 diff --git a/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run2/model_best.pt b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run2/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..e852a46c51c1c0471a7f246f10b38d98997a13c9 --- /dev/null +++ b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run2/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9e6ed07950306da79c174541e5b8daee4553219008098c2e4a8e52d81589b8d9 +size 2749638 diff --git a/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run2/train.log b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run2/train.log new file mode 100644 index 0000000000000000000000000000000000000000..61d7b2539c16a314d6f52d76ed0ce894cb3c9322 --- /dev/null +++ b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/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=0.5084759711455076 +INFO:root: +[Valid][0][UTT][ MSE = 0.3636 | LCC = 0.7557 | SRCC = 0.7621 ] +INFO:root:Average loss=0.30706647439644885 +INFO:root: +[Valid][1][UTT][ MSE = 0.3151 | LCC = 0.8007 | SRCC = 0.8030 ] +INFO:root:Average loss=0.2623538002371788 +INFO:root: +[Valid][2][UTT][ MSE = 0.3110 | LCC = 0.8220 | SRCC = 0.8213 ] +INFO:root:Average loss=0.2326728193423687 +INFO:root: +[Valid][3][UTT][ MSE = 0.2862 | LCC = 0.8305 | SRCC = 0.8274 ] +INFO:root:Average loss=0.22395901114512712 +INFO:root: +[Valid][4][UTT][ MSE = 0.2674 | LCC = 0.8304 | SRCC = 0.8302 ] +INFO:root:Average loss=0.20366994043191275 +INFO:root: +[Valid][5][UTT][ MSE = 0.2649 | LCC = 0.8383 | SRCC = 0.8377 ] +INFO:root:Average loss=0.1987848320068457 +INFO:root: +[Valid][6][UTT][ MSE = 0.2599 | LCC = 0.8394 | SRCC = 0.8388 ] +INFO:root:Average loss=0.18917677799860635 +INFO:root: +[Valid][7][UTT][ MSE = 0.2574 | LCC = 0.8371 | SRCC = 0.8392 ] +INFO:root:Average loss=0.18556514898171791 +INFO:root: +[Valid][8][UTT][ MSE = 0.2752 | LCC = 0.8450 | SRCC = 0.8449 ] +INFO:root:Average loss=0.1656024647064698 +INFO:root: +[Valid][9][UTT][ MSE = 0.2943 | LCC = 0.8426 | SRCC = 0.8427 ] +INFO:root:Average loss=0.16914869520144585 +INFO:root: +[Valid][10][UTT][ MSE = 0.2652 | LCC = 0.8474 | SRCC = 0.8485 ] +INFO:root:Average loss=0.1515972159611873 +INFO:root: +[Valid][11][UTT][ MSE = 0.2930 | LCC = 0.8430 | SRCC = 0.8433 ] +INFO:root:Average loss=0.14992079262932143 +INFO:root: +[Valid][12][UTT][ MSE = 0.2575 | LCC = 0.8471 | SRCC = 0.8489 ] +INFO:root:Average loss=0.1377835555527455 +INFO:root: +[Valid][13][UTT][ MSE = 0.2614 | LCC = 0.8454 | SRCC = 0.8459 ] +INFO:root:Average loss=0.13687396679933256 +INFO:root: +[Valid][14][UTT][ MSE = 0.2829 | LCC = 0.8425 | SRCC = 0.8444 ] +INFO:root:Average loss=0.12881493549316358 +INFO:root: +[Valid][15][UTT][ MSE = 0.2334 | LCC = 0.8490 | SRCC = 0.8508 ] +INFO:root:Average loss=0.12805823838481536 +INFO:root: +[Valid][16][UTT][ MSE = 0.2421 | LCC = 0.8479 | SRCC = 0.8489 ] +INFO:root:Average loss=0.11622702788848144 +INFO:root: +[Valid][17][UTT][ MSE = 0.2366 | LCC = 0.8495 | SRCC = 0.8493 ] +INFO:root:Average loss=0.11665664269373967 +INFO:root: +[Valid][18][UTT][ MSE = 0.2457 | LCC = 0.8517 | SRCC = 0.8528 ] +INFO:root:Average loss=0.1165073067905047 +INFO:root: +[Valid][19][UTT][ MSE = 0.2614 | LCC = 0.8506 | SRCC = 0.8523 ] +INFO:root:Average loss=0.10905655549886899 +INFO:root: +[Valid][20][UTT][ MSE = 0.2522 | LCC = 0.8492 | SRCC = 0.8492 ] +INFO:root:Average loss=0.10352906365043078 +INFO:root: +[Valid][21][UTT][ MSE = 0.2579 | LCC = 0.8501 | SRCC = 0.8494 ] +INFO:root:Average loss=0.09883613865344952 +INFO:root: +[Valid][22][UTT][ MSE = 0.2362 | LCC = 0.8513 | SRCC = 0.8503 ] +INFO:root:Average loss=0.09639730672232616 +INFO:root: +[Valid][23][UTT][ MSE = 0.2405 | LCC = 0.8516 | SRCC = 0.8516 ] +INFO:root:Average loss=0.09600095560726447 +INFO:root: +[Valid][24][UTT][ MSE = 0.2386 | LCC = 0.8493 | SRCC = 0.8493 ] +INFO:root:Average loss=0.09854865852647866 +INFO:root: +[Valid][25][UTT][ MSE = 0.2369 | LCC = 0.8525 | SRCC = 0.8520 ] +INFO:root:Average loss=0.09675059754114884 +INFO:root: +[Valid][26][UTT][ MSE = 0.2339 | LCC = 0.8524 | SRCC = 0.8516 ] +INFO:root:Average loss=0.08580169072135901 +INFO:root: +[Valid][27][UTT][ MSE = 0.2608 | LCC = 0.8526 | SRCC = 0.8525 ] +INFO:root:Average loss=0.08808129266477548 +INFO:root: +[Valid][28][UTT][ MSE = 0.2364 | LCC = 0.8499 | SRCC = 0.8494 ] +INFO:root:Average loss=0.07828553632283822 +INFO:root: +[Valid][29][UTT][ MSE = 0.2935 | LCC = 0.8474 | SRCC = 0.8467 ] +INFO:root:Average loss=0.07865113774553323 +INFO:root: +[Valid][30][UTT][ MSE = 0.2698 | LCC = 0.8502 | SRCC = 0.8491 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2426 | LCC = 0.8544 | SRCC = 0.8516 ] diff --git a/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run3/config.gin b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run3/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/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 diff --git a/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run3/model.pt b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run3/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..9ac5cfcd056ef7ff2eb00783e6a03ae85a2bb436 --- /dev/null +++ b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run3/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:8b5ce02354d70b122d45fb115cb3ab32fddd4ce964572e52ae365c3392978bf1 +size 2747979 diff --git a/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run3/model_best.pt b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run3/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..3f1b6ccbddfe40906ec5cde33fadce505df7d0c7 --- /dev/null +++ b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run3/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7345171922319d99ff59e87001743fd338cdc7357a8ec65e44fc78ed79450087 +size 2749638 diff --git a/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run3/train.log b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run3/train.log new file mode 100644 index 0000000000000000000000000000000000000000..bb23f126456293f251b211242149cd6eb0e46f58 --- /dev/null +++ b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/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=0.5462396721809338 +INFO:root: +[Valid][0][UTT][ MSE = 0.3721 | LCC = 0.7573 | SRCC = 0.7588 ] +INFO:root:Average loss=0.30311783097493344 +INFO:root: +[Valid][1][UTT][ MSE = 0.3117 | LCC = 0.8087 | SRCC = 0.8120 ] +INFO:root:Average loss=0.2613933258331739 +INFO:root: +[Valid][2][UTT][ MSE = 0.2977 | LCC = 0.8249 | SRCC = 0.8244 ] +INFO:root:Average loss=0.24301403206892502 +INFO:root: +[Valid][3][UTT][ MSE = 0.2749 | LCC = 0.8298 | SRCC = 0.8318 ] +INFO:root:Average loss=0.22846918858778784 +INFO:root: +[Valid][4][UTT][ MSE = 0.2659 | LCC = 0.8360 | SRCC = 0.8346 ] +INFO:root:Average loss=0.21675822597283584 +INFO:root: +[Valid][5][UTT][ MSE = 0.2769 | LCC = 0.8388 | SRCC = 0.8385 ] +INFO:root:Average loss=0.20353675595460793 +INFO:root: +[Valid][6][UTT][ MSE = 0.2532 | LCC = 0.8425 | SRCC = 0.8426 ] +INFO:root:Average loss=0.19074134901165962 +INFO:root: +[Valid][7][UTT][ MSE = 0.2795 | LCC = 0.8368 | SRCC = 0.8400 ] +INFO:root:Average loss=0.17998591533456093 +INFO:root: +[Valid][8][UTT][ MSE = 0.2501 | LCC = 0.8460 | SRCC = 0.8443 ] +INFO:root:Average loss=0.1763109790209012 +INFO:root: +[Valid][9][UTT][ MSE = 0.2568 | LCC = 0.8438 | SRCC = 0.8447 ] +INFO:root:Average loss=0.16469519508954805 +INFO:root: +[Valid][10][UTT][ MSE = 0.2435 | LCC = 0.8459 | SRCC = 0.8457 ] +INFO:root:Average loss=0.1605397509649778 +INFO:root: +[Valid][11][UTT][ MSE = 0.2494 | LCC = 0.8482 | SRCC = 0.8470 ] +INFO:root:Average loss=0.15335003649577117 +INFO:root: +[Valid][12][UTT][ MSE = 0.2423 | LCC = 0.8472 | SRCC = 0.8472 ] +INFO:root:Average loss=0.13942107777947035 +INFO:root: +[Valid][13][UTT][ MSE = 0.2630 | LCC = 0.8444 | SRCC = 0.8445 ] +INFO:root:Average loss=0.1374869487988643 +INFO:root: +[Valid][14][UTT][ MSE = 0.2374 | LCC = 0.8507 | SRCC = 0.8482 ] +INFO:root:Average loss=0.1394726654084829 +INFO:root: +[Valid][15][UTT][ MSE = 0.2345 | LCC = 0.8538 | SRCC = 0.8528 ] +INFO:root:Average loss=0.13249869300768927 +INFO:root: +[Valid][16][UTT][ MSE = 0.2290 | LCC = 0.8562 | SRCC = 0.8550 ] +INFO:root:Average loss=0.12491989804384036 +INFO:root: +[Valid][17][UTT][ MSE = 0.2395 | LCC = 0.8495 | SRCC = 0.8503 ] +INFO:root:Average loss=0.11946285248566897 +INFO:root: +[Valid][18][UTT][ MSE = 0.2651 | LCC = 0.8452 | SRCC = 0.8443 ] +INFO:root:Average loss=0.11300687386821477 +INFO:root: +[Valid][19][UTT][ MSE = 0.2403 | LCC = 0.8548 | SRCC = 0.8524 ] +INFO:root:Average loss=0.10669383397087073 +INFO:root: +[Valid][20][UTT][ MSE = 0.2285 | LCC = 0.8543 | SRCC = 0.8531 ] +INFO:root:Average loss=0.10929267490521455 +INFO:root: +[Valid][21][UTT][ MSE = 0.2453 | LCC = 0.8541 | SRCC = 0.8528 ] +INFO:root:Average loss=0.10815160463635738 +INFO:root: +[Valid][22][UTT][ MSE = 0.2351 | LCC = 0.8555 | SRCC = 0.8549 ] +INFO:root:Average loss=0.09860006844003995 +INFO:root: +[Valid][23][UTT][ MSE = 0.2252 | LCC = 0.8541 | SRCC = 0.8542 ] +INFO:root:Average loss=0.09188081567677167 +INFO:root: +[Valid][24][UTT][ MSE = 0.2395 | LCC = 0.8519 | SRCC = 0.8494 ] +INFO:root:Average loss=0.08857262000823632 +INFO:root: +[Valid][25][UTT][ MSE = 0.2565 | LCC = 0.8458 | SRCC = 0.8477 ] +INFO:root:Average loss=0.09461229977508386 +INFO:root: +[Valid][26][UTT][ MSE = 0.2383 | LCC = 0.8502 | SRCC = 0.8487 ] +INFO:root:Average loss=0.08317802364054398 +INFO:root: +[Valid][27][UTT][ MSE = 0.2265 | LCC = 0.8533 | SRCC = 0.8528 ] +INFO:root:Average loss=0.0819217039224429 +INFO:root: +[Valid][28][UTT][ MSE = 0.2298 | LCC = 0.8521 | SRCC = 0.8514 ] +INFO:root:Average loss=0.08475901258106415 +INFO:root: +[Valid][29][UTT][ MSE = 0.2453 | LCC = 0.8506 | SRCC = 0.8499 ] +INFO:root:Average loss=0.07875150399139294 +INFO:root: +[Valid][30][UTT][ MSE = 0.2363 | LCC = 0.8512 | SRCC = 0.8505 ] +INFO:root: +[Test][31][UTT][ MSE = 0.2250 | LCC = 0.8602 | SRCC = 0.8596 ] diff --git a/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run4/config.gin b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run4/config.gin new file mode 100644 index 0000000000000000000000000000000000000000..d36a6964ab640c490fd7442e2d46a6fbe5fc84aa --- /dev/null +++ b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/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 diff --git a/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run4/model.pt b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run4/model.pt new file mode 100644 index 0000000000000000000000000000000000000000..016d1d00ce5b33523cc32155f4bb68275dcf4ed2 --- /dev/null +++ b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run4/model.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0572cd32e25005d4bd606a3fdaec59fbc864ae5b7f8465248074176661808766 +size 2747979 diff --git a/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run4/model_best.pt b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run4/model_best.pt new file mode 100644 index 0000000000000000000000000000000000000000..b7d7dea621139f99882fa851295033161c54d246 --- /dev/null +++ b/results_bvcc/w2v2_base_bvcc_finetuned_10epoch/layer0/run4/model_best.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ff349146033e40e55ba1f1b1780e6caa47c9c28c2c5762ccc0824d2c1bcecbad +size 2749702