End of training

all_results.json

@@ -1,14 +1,14 @@
 {
     "epoch": 10.0,
-    "eval_loss": 0.6743320226669312,
-    "eval_runtime": 128.2267,
+    "eval_loss": 0.6776940822601318,
+    "eval_runtime": 127.7608,
     "eval_samples": 532,
-    "eval_samples_per_second": 4.149,
-    "eval_steps_per_second": 0.523,
-    "eval_wer": 0.4303528958025326,
-    "train_loss": 0.6841171807712979,
-    "train_runtime": 19401.0857,
+    "eval_samples_per_second": 4.164,
+    "eval_steps_per_second": 0.524,
+    "eval_wer": 0.4203895352574404,
+    "train_loss": 0.6085318591859605,
+    "train_runtime": 19393.8759,
     "train_samples": 4617,
-    "train_samples_per_second": 2.38,
+    "train_samples_per_second": 2.381,
     "train_steps_per_second": 0.037
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 10.0,
-    "eval_loss": 0.6743320226669312,
-    "eval_runtime": 128.2267,
+    "eval_loss": 0.6776940822601318,
+    "eval_runtime": 127.7608,
     "eval_samples": 532,
-    "eval_samples_per_second": 4.149,
-    "eval_steps_per_second": 0.523,
-    "eval_wer": 0.4303528958025326
+    "eval_samples_per_second": 4.164,
+    "eval_steps_per_second": 0.524,
+    "eval_wer": 0.4203895352574404
 }

runs/Jul08_14-07-18_e43e87908b0c/events.out.tfevents.1657311726.e43e87908b0c.4172.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:457bb51610e36c4c8d553af5e35714bc681748f064ec0cd59d3f5fab4ed1af31
+size 358

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 10.0,
-    "train_loss": 0.6841171807712979,
-    "train_runtime": 19401.0857,
+    "train_loss": 0.6085318591859605,
+    "train_runtime": 19393.8759,
     "train_samples": 4617,
-    "train_samples_per_second": 2.38,
+    "train_samples_per_second": 2.381,
     "train_steps_per_second": 0.037
 }

trainer_state.json

@@ -1,5 +1,5 @@
 {
-  "best_metric": 0.6743320226669312,
+  "best_metric": 0.6776940822601318,
   "best_model_checkpoint": "ai-light-dance_singing3_ft_wav2vec2-large-xlsr-53-v1/checkpoint-720",
   "epoch": 9.996539792387543,
   "global_step": 720,
@@ -9,533 +9,533 @@
   "log_history": [
     {
       "epoch": 0.14,
-      "learning_rate": 5e-06,
-      "loss": 0.7293,
+      "learning_rate": 2.0000000000000003e-06,
+      "loss": 0.634,
       "step": 10
     },
     {
       "epoch": 0.28,
-      "learning_rate": 1e-05,
-      "loss": 0.7451,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 0.6419,
       "step": 20
     },
     {
       "epoch": 0.42,
-      "learning_rate": 1.5e-05,
-      "loss": 0.7259,
+      "learning_rate": 6e-06,
+      "loss": 0.6223,
       "step": 30
     },
     {
       "epoch": 0.55,
-      "learning_rate": 2e-05,
-      "loss": 0.7445,
+      "learning_rate": 8.000000000000001e-06,
+      "loss": 0.6284,
       "step": 40
     },
     {
       "epoch": 0.69,
-      "learning_rate": 2.5e-05,
-      "loss": 0.7222,
+      "learning_rate": 1e-05,
+      "loss": 0.6029,
       "step": 50
     },
     {
       "epoch": 0.83,
-      "learning_rate": 2.4626865671641793e-05,
-      "loss": 0.7336,
+      "learning_rate": 9.850746268656717e-06,
+      "loss": 0.6024,
       "step": 60
     },
     {
       "epoch": 0.97,
-      "learning_rate": 2.4253731343283584e-05,
-      "loss": 0.7295,
+      "learning_rate": 9.701492537313434e-06,
+      "loss": 0.6227,
       "step": 70
     },
     {
       "epoch": 1.0,
-      "eval_loss": 0.7118099927902222,
-      "eval_runtime": 130.35,
-      "eval_samples_per_second": 4.081,
-      "eval_steps_per_second": 0.514,
-      "eval_wer": 0.47457736067365175,
+      "eval_loss": 0.7283244729042053,
+      "eval_runtime": 128.3245,
+      "eval_samples_per_second": 4.146,
+      "eval_steps_per_second": 0.522,
+      "eval_wer": 0.46589959503760364,
       "step": 72
     },
     {
       "epoch": 1.11,
-      "learning_rate": 2.3880597014925373e-05,
-      "loss": 0.767,
+      "learning_rate": 9.552238805970149e-06,
+      "loss": 0.6452,
       "step": 80
     },
     {
       "epoch": 1.25,
-      "learning_rate": 2.3507462686567168e-05,
-      "loss": 0.729,
+      "learning_rate": 9.402985074626867e-06,
+      "loss": 0.6194,
       "step": 90
     },
     {
       "epoch": 1.39,
-      "learning_rate": 2.3134328358208956e-05,
-      "loss": 0.7306,
+      "learning_rate": 9.253731343283582e-06,
+      "loss": 0.6242,
       "step": 100
     },
     {
       "epoch": 1.53,
-      "learning_rate": 2.2761194029850747e-05,
-      "loss": 0.7362,
+      "learning_rate": 9.104477611940299e-06,
+      "loss": 0.6169,
       "step": 110
     },
     {
       "epoch": 1.66,
-      "learning_rate": 2.238805970149254e-05,
-      "loss": 0.7128,
+      "learning_rate": 8.955223880597016e-06,
+      "loss": 0.6081,
       "step": 120
     },
     {
       "epoch": 1.8,
-      "learning_rate": 2.201492537313433e-05,
-      "loss": 0.7593,
+      "learning_rate": 8.805970149253732e-06,
+      "loss": 0.6511,
       "step": 130
     },
     {
       "epoch": 1.94,
-      "learning_rate": 2.164179104477612e-05,
-      "loss": 0.7062,
+      "learning_rate": 8.656716417910447e-06,
+      "loss": 0.6027,
       "step": 140
     },
     {
       "epoch": 2.0,
-      "eval_loss": 0.7108114361763,
-      "eval_runtime": 129.5999,
-      "eval_samples_per_second": 4.105,
+      "eval_loss": 0.7168184518814087,
+      "eval_runtime": 129.6874,
+      "eval_samples_per_second": 4.102,
       "eval_steps_per_second": 0.517,
-      "eval_wer": 0.4676994279102655,
+      "eval_wer": 0.4427588866748088,
       "step": 144
     },
     {
       "epoch": 2.08,
-      "learning_rate": 2.126865671641791e-05,
-      "loss": 0.7321,
+      "learning_rate": 8.507462686567165e-06,
+      "loss": 0.6228,
       "step": 150
     },
     {
       "epoch": 2.22,
-      "learning_rate": 2.0895522388059702e-05,
-      "loss": 0.7123,
+      "learning_rate": 8.35820895522388e-06,
+      "loss": 0.6092,
       "step": 160
     },
     {
       "epoch": 2.36,
-      "learning_rate": 2.0522388059701493e-05,
-      "loss": 0.7404,
+      "learning_rate": 8.208955223880599e-06,
+      "loss": 0.625,
       "step": 170
     },
     {
       "epoch": 2.5,
-      "learning_rate": 2.0149253731343285e-05,
-      "loss": 0.7048,
+      "learning_rate": 8.059701492537314e-06,
+      "loss": 0.5997,
       "step": 180
     },
     {
       "epoch": 2.64,
-      "learning_rate": 1.9776119402985073e-05,
-      "loss": 0.7002,
+      "learning_rate": 7.91044776119403e-06,
+      "loss": 0.5925,
       "step": 190
     },
     {
       "epoch": 2.78,
-      "learning_rate": 1.9402985074626868e-05,
-      "loss": 0.6914,
+      "learning_rate": 7.761194029850747e-06,
+      "loss": 0.5989,
       "step": 200
     },
     {
       "epoch": 2.91,
-      "learning_rate": 1.9029850746268656e-05,
-      "loss": 0.6932,
+      "learning_rate": 7.611940298507463e-06,
+      "loss": 0.5973,
       "step": 210
     },
     {
       "epoch": 3.0,
-      "eval_loss": 0.7351313829421997,
-      "eval_runtime": 127.3643,
-      "eval_samples_per_second": 4.177,
-      "eval_steps_per_second": 0.526,
-      "eval_wer": 0.48511923892781383,
+      "eval_loss": 0.706081211566925,
+      "eval_runtime": 130.3715,
+      "eval_samples_per_second": 4.081,
+      "eval_steps_per_second": 0.514,
+      "eval_wer": 0.4312528122388635,
       "step": 216
     },
     {
       "epoch": 3.06,
-      "learning_rate": 1.865671641791045e-05,
-      "loss": 0.7105,
+      "learning_rate": 7.46268656716418e-06,
+      "loss": 0.6123,
       "step": 220
     },
     {
       "epoch": 3.19,
-      "learning_rate": 1.828358208955224e-05,
-      "loss": 0.702,
+      "learning_rate": 7.313432835820896e-06,
+      "loss": 0.6008,
       "step": 230
     },
     {
       "epoch": 3.33,
-      "learning_rate": 1.791044776119403e-05,
-      "loss": 0.7372,
+      "learning_rate": 7.164179104477612e-06,
+      "loss": 0.6303,
       "step": 240
     },
     {
       "epoch": 3.47,
-      "learning_rate": 1.7537313432835823e-05,
-      "loss": 0.6776,
+      "learning_rate": 7.014925373134329e-06,
+      "loss": 0.5887,
       "step": 250
     },
     {
       "epoch": 3.61,
-      "learning_rate": 1.716417910447761e-05,
-      "loss": 0.6749,
+      "learning_rate": 6.865671641791045e-06,
+      "loss": 0.5873,
       "step": 260
     },
     {
       "epoch": 3.75,
-      "learning_rate": 1.6791044776119406e-05,
-      "loss": 0.6834,
+      "learning_rate": 6.7164179104477625e-06,
+      "loss": 0.589,
       "step": 270
     },
     {
       "epoch": 3.89,
-      "learning_rate": 1.6417910447761194e-05,
-      "loss": 0.6674,
+      "learning_rate": 6.567164179104478e-06,
+      "loss": 0.5773,
       "step": 280
     },
     {
       "epoch": 4.0,
-      "eval_loss": 0.7145407795906067,
-      "eval_runtime": 127.1129,
-      "eval_samples_per_second": 4.185,
-      "eval_steps_per_second": 0.527,
-      "eval_wer": 0.4479012663109854,
+      "eval_loss": 0.7067840695381165,
+      "eval_runtime": 128.8032,
+      "eval_samples_per_second": 4.13,
+      "eval_steps_per_second": 0.52,
+      "eval_wer": 0.4336311628205952,
       "step": 288
     },
     {
       "epoch": 4.03,
-      "learning_rate": 1.6044776119402986e-05,
-      "loss": 0.702,
+      "learning_rate": 6.417910447761194e-06,
+      "loss": 0.6105,
       "step": 290
     },
     {
       "epoch": 4.17,
-      "learning_rate": 1.5671641791044777e-05,
-      "loss": 0.6806,
+      "learning_rate": 6.2686567164179116e-06,
+      "loss": 0.5964,
       "step": 300
     },
     {
       "epoch": 4.3,
-      "learning_rate": 1.529850746268657e-05,
-      "loss": 0.683,
+      "learning_rate": 6.119402985074627e-06,
+      "loss": 0.5959,
       "step": 310
     },
     {
       "epoch": 4.44,
-      "learning_rate": 1.4925373134328357e-05,
-      "loss": 0.6647,
+      "learning_rate": 5.970149253731343e-06,
+      "loss": 0.5843,
       "step": 320
     },
     {
       "epoch": 4.58,
-      "learning_rate": 1.455223880597015e-05,
-      "loss": 0.7013,
+      "learning_rate": 5.820895522388061e-06,
+      "loss": 0.6068,
       "step": 330
     },
     {
       "epoch": 4.72,
-      "learning_rate": 1.417910447761194e-05,
-      "loss": 0.6824,
+      "learning_rate": 5.671641791044776e-06,
+      "loss": 0.6045,
       "step": 340
     },
     {
       "epoch": 4.86,
-      "learning_rate": 1.3805970149253733e-05,
-      "loss": 0.6746,
+      "learning_rate": 5.522388059701493e-06,
+      "loss": 0.5906,
       "step": 350
     },
     {
       "epoch": 5.0,
-      "learning_rate": 1.3432835820895523e-05,
-      "loss": 0.6784,
+      "learning_rate": 5.37313432835821e-06,
+      "loss": 0.5965,
       "step": 360
     },
     {
       "epoch": 5.0,
-      "eval_loss": 0.7020928859710693,
-      "eval_runtime": 128.5118,
-      "eval_samples_per_second": 4.14,
-      "eval_steps_per_second": 0.521,
-      "eval_wer": 0.4596001799832873,
+      "eval_loss": 0.7037749886512756,
+      "eval_runtime": 129.081,
+      "eval_samples_per_second": 4.121,
+      "eval_steps_per_second": 0.519,
+      "eval_wer": 0.4339525615478563,
       "step": 360
     },
     {
       "epoch": 5.14,
-      "learning_rate": 1.3059701492537313e-05,
-      "loss": 0.6613,
+      "learning_rate": 5.2238805970149255e-06,
+      "loss": 0.5885,
       "step": 370
     },
     {
       "epoch": 5.28,
-      "learning_rate": 1.2686567164179105e-05,
-      "loss": 0.6596,
+      "learning_rate": 5.074626865671642e-06,
+      "loss": 0.5907,
       "step": 380
     },
     {
       "epoch": 5.42,
-      "learning_rate": 1.2313432835820896e-05,
-      "loss": 0.6584,
+      "learning_rate": 4.925373134328359e-06,
+      "loss": 0.5862,
       "step": 390
     },
     {
       "epoch": 5.55,
-      "learning_rate": 1.1940298507462686e-05,
-      "loss": 0.6721,
+      "learning_rate": 4.7761194029850745e-06,
+      "loss": 0.6002,
       "step": 400
     },
     {
       "epoch": 5.69,
-      "learning_rate": 1.1567164179104478e-05,
-      "loss": 0.679,
+      "learning_rate": 4.626865671641791e-06,
+      "loss": 0.6035,
       "step": 410
     },
     {
       "epoch": 5.83,
-      "learning_rate": 1.119402985074627e-05,
-      "loss": 0.7022,
+      "learning_rate": 4.477611940298508e-06,
+      "loss": 0.6163,
       "step": 420
     },
     {
       "epoch": 5.97,
-      "learning_rate": 1.082089552238806e-05,
-      "loss": 0.6493,
+      "learning_rate": 4.3283582089552236e-06,
+      "loss": 0.5854,
       "step": 430
     },
     {
       "epoch": 6.0,
-      "eval_loss": 0.7246174216270447,
-      "eval_runtime": 129.6572,
-      "eval_samples_per_second": 4.103,
-      "eval_steps_per_second": 0.517,
-      "eval_wer": 0.4651282380921772,
+      "eval_loss": 0.705437958240509,
+      "eval_runtime": 128.9432,
+      "eval_samples_per_second": 4.126,
+      "eval_steps_per_second": 0.52,
+      "eval_wer": 0.43086713376615027,
       "step": 432
     },
     {
       "epoch": 6.11,
-      "learning_rate": 1.0447761194029851e-05,
-      "loss": 0.7152,
+      "learning_rate": 4.17910447761194e-06,
+      "loss": 0.6481,
       "step": 440
     },
     {
       "epoch": 6.25,
-      "learning_rate": 1.0074626865671643e-05,
-      "loss": 0.6432,
+      "learning_rate": 4.029850746268657e-06,
+      "loss": 0.5806,
       "step": 450
     },
     {
       "epoch": 6.39,
-      "learning_rate": 9.701492537313434e-06,
-      "loss": 0.6463,
+      "learning_rate": 3.8805970149253735e-06,
+      "loss": 0.5838,
       "step": 460
     },
     {
       "epoch": 6.53,
-      "learning_rate": 9.328358208955226e-06,
-      "loss": 0.6777,
+      "learning_rate": 3.73134328358209e-06,
+      "loss": 0.6078,
       "step": 470
     },
     {
       "epoch": 6.66,
-      "learning_rate": 8.955223880597016e-06,
-      "loss": 0.676,
+      "learning_rate": 3.582089552238806e-06,
+      "loss": 0.6097,
       "step": 480
     },
     {
       "epoch": 6.8,
-      "learning_rate": 8.582089552238805e-06,
-      "loss": 0.6484,
+      "learning_rate": 3.447761194029851e-06,
+      "loss": 0.5909,
       "step": 490
     },
     {
       "epoch": 6.94,
-      "learning_rate": 8.208955223880597e-06,
-      "loss": 0.6338,
+      "learning_rate": 3.2985074626865677e-06,
+      "loss": 0.5832,
       "step": 500
     },
     {
       "epoch": 7.0,
-      "eval_loss": 0.7020286917686462,
-      "eval_runtime": 128.8637,
-      "eval_samples_per_second": 4.128,
-      "eval_steps_per_second": 0.52,
-      "eval_wer": 0.44552291572925373,
+      "eval_loss": 0.699042022228241,
+      "eval_runtime": 127.7248,
+      "eval_samples_per_second": 4.165,
+      "eval_steps_per_second": 0.525,
+      "eval_wer": 0.42527479591180817,
       "step": 504
     },
     {
       "epoch": 7.08,
-      "learning_rate": 7.835820895522389e-06,
-      "loss": 0.6875,
+      "learning_rate": 3.1492537313432834e-06,
+      "loss": 0.6283,
       "step": 510
     },
     {
       "epoch": 7.22,
-      "learning_rate": 7.4626865671641785e-06,
-      "loss": 0.6397,
+      "learning_rate": 3e-06,
+      "loss": 0.5905,
       "step": 520
     },
     {
       "epoch": 7.36,
-      "learning_rate": 7.08955223880597e-06,
-      "loss": 0.6695,
+      "learning_rate": 2.8507462686567167e-06,
+      "loss": 0.6143,
       "step": 530
     },
     {
       "epoch": 7.5,
-      "learning_rate": 6.716417910447762e-06,
-      "loss": 0.6582,
+      "learning_rate": 2.7014925373134333e-06,
+      "loss": 0.6072,
       "step": 540
     },
     {
       "epoch": 7.64,
-      "learning_rate": 6.343283582089552e-06,
-      "loss": 0.6335,
+      "learning_rate": 2.5522388059701496e-06,
+      "loss": 0.5894,
       "step": 550
     },
     {
       "epoch": 7.78,
-      "learning_rate": 5.970149253731343e-06,
-      "loss": 0.6542,
+      "learning_rate": 2.402985074626866e-06,
+      "loss": 0.604,
       "step": 560
     },
     {
       "epoch": 7.91,
-      "learning_rate": 5.597014925373135e-06,
-      "loss": 0.6464,
+      "learning_rate": 2.2537313432835824e-06,
+      "loss": 0.5968,
       "step": 570
     },
     {
       "epoch": 8.0,
-      "eval_loss": 0.6980531811714172,
-      "eval_runtime": 129.6171,
-      "eval_samples_per_second": 4.104,
-      "eval_steps_per_second": 0.517,
-      "eval_wer": 0.439737738638555,
+      "eval_loss": 0.698808491230011,
+      "eval_runtime": 129.2631,
+      "eval_samples_per_second": 4.116,
+      "eval_steps_per_second": 0.518,
+      "eval_wer": 0.42315356431188533,
       "step": 576
     },
     {
       "epoch": 8.06,
-      "learning_rate": 5.2238805970149255e-06,
-      "loss": 0.6731,
+      "learning_rate": 2.1044776119402986e-06,
+      "loss": 0.6225,
       "step": 580
     },
     {
       "epoch": 8.19,
-      "learning_rate": 4.850746268656717e-06,
-      "loss": 0.632,
+      "learning_rate": 1.955223880597015e-06,
+      "loss": 0.5931,
       "step": 590
     },
     {
       "epoch": 8.33,
-      "learning_rate": 4.477611940298508e-06,
-      "loss": 0.644,
+      "learning_rate": 1.8059701492537314e-06,
+      "loss": 0.6011,
       "step": 600
     },
     {
       "epoch": 8.47,
-      "learning_rate": 4.1044776119402985e-06,
-      "loss": 0.6391,
+      "learning_rate": 1.6567164179104479e-06,
+      "loss": 0.5992,
       "step": 610
     },
     {
       "epoch": 8.61,
-      "learning_rate": 3.7313432835820893e-06,
-      "loss": 0.6551,
+      "learning_rate": 1.5074626865671643e-06,
+      "loss": 0.6182,
       "step": 620
     },
     {
       "epoch": 8.75,
-      "learning_rate": 3.358208955223881e-06,
-      "loss": 0.6572,
+      "learning_rate": 1.3582089552238805e-06,
+      "loss": 0.6231,
       "step": 630
     },
     {
       "epoch": 8.89,
-      "learning_rate": 2.9850746268656716e-06,
-      "loss": 0.6227,
+      "learning_rate": 1.2089552238805971e-06,
+      "loss": 0.5895,
       "step": 640
     },
     {
       "epoch": 9.0,
-      "eval_loss": 0.6833487153053284,
-      "eval_runtime": 129.1282,
-      "eval_samples_per_second": 4.12,
-      "eval_steps_per_second": 0.519,
-      "eval_wer": 0.43639519187504017,
+      "eval_loss": 0.6866586208343506,
+      "eval_runtime": 126.5198,
+      "eval_samples_per_second": 4.205,
+      "eval_steps_per_second": 0.53,
+      "eval_wer": 0.4221893681301022,
       "step": 648
     },
     {
       "epoch": 9.03,
-      "learning_rate": 2.6119402985074627e-06,
-      "loss": 0.6722,
+      "learning_rate": 1.0597014925373136e-06,
+      "loss": 0.6387,
       "step": 650
     },
     {
       "epoch": 9.17,
-      "learning_rate": 2.238805970149254e-06,
-      "loss": 0.6375,
+      "learning_rate": 9.104477611940299e-07,
+      "loss": 0.6074,
       "step": 660
     },
     {
       "epoch": 9.3,
-      "learning_rate": 1.8656716417910446e-06,
-      "loss": 0.6752,
+      "learning_rate": 7.611940298507463e-07,
+      "loss": 0.6425,
       "step": 670
     },
     {
       "epoch": 9.44,
-      "learning_rate": 1.4925373134328358e-06,
-      "loss": 0.6428,
+      "learning_rate": 6.119402985074627e-07,
+      "loss": 0.6131,
       "step": 680
     },
     {
       "epoch": 9.58,
-      "learning_rate": 1.119402985074627e-06,
-      "loss": 0.6649,
+      "learning_rate": 4.6268656716417914e-07,
+      "loss": 0.6328,
       "step": 690
     },
     {
       "epoch": 9.72,
-      "learning_rate": 7.462686567164179e-07,
-      "loss": 0.6376,
+      "learning_rate": 3.134328358208955e-07,
+      "loss": 0.61,
       "step": 700
     },
     {
       "epoch": 9.86,
-      "learning_rate": 3.7313432835820895e-07,
-      "loss": 0.6634,
+      "learning_rate": 1.6417910447761195e-07,
+      "loss": 0.6417,
       "step": 710
     },
     {
       "epoch": 10.0,
-      "learning_rate": 0.0,
-      "loss": 0.6627,
+      "learning_rate": 1.492537313432836e-08,
+      "loss": 0.6378,
       "step": 720
     },
     {
       "epoch": 10.0,
-      "eval_loss": 0.6743320226669312,
-      "eval_runtime": 130.183,
-      "eval_samples_per_second": 4.087,
-      "eval_steps_per_second": 0.515,
-      "eval_wer": 0.4303528958025326,
+      "eval_loss": 0.6776940822601318,
+      "eval_runtime": 127.2273,
+      "eval_samples_per_second": 4.181,
+      "eval_steps_per_second": 0.527,
+      "eval_wer": 0.4203895352574404,
       "step": 720
     },
     {
       "epoch": 10.0,
       "step": 720,
       "total_flos": 2.7197958559499194e+19,
-      "train_loss": 0.6841171807712979,
-      "train_runtime": 19401.0857,
-      "train_samples_per_second": 2.38,
+      "train_loss": 0.6085318591859605,
+      "train_runtime": 19393.8759,
+      "train_samples_per_second": 2.381,
       "train_steps_per_second": 0.037
     }
   ],