Upload 10 files

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6f5204db756e34f799719358f88672c8b4c6026f91eebda145d29f6b5a859364
+oid sha256:bb6f65f29ee2dd243f8694ffbcb7bbaa69f03ab6b54e8e0304daed4f8c389605
 size 377851056

optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8e5c4216be93855fbc013153efec4d3f85293e3472110511180806826aaf4fe2
+oid sha256:eefd1ccbd216487e6fef630abe43cad9988fd32f3971fdd0e671c57ca15ee35b
 size 722217338

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5c63dd48407ceec67dee01431e0a481bb387803d9ef4f6a5541ecfead1b611e8
+oid sha256:c4ba484b726cd999ba2ea5517937ac5134e6dd5ba461411ddcfcc27176ea6f26
 size 377899102

rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b3913ae9493cc4baad846c1b707d069986928d1245d876521c5d6546dda5464f
+oid sha256:d2f26a93007e1df9fe2d108b521dc1e9f1a2714dcc89ec6017185eb07f7a2e36
 size 14244

scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ea297f7ab56af8b17d03b96627bb16862b6bf5fc7088f8cd1f0a0f4093d4464c
+oid sha256:82aca86b1ee054cd7ebba1ff6f00f6a548e6c9f0be9128f918ee36a8e37a51bd
 size 1064

trainer_state.json

@@ -1,298 +1,398 @@
 {
-  "best_metric": 0.05392587557435036,
-  "best_model_checkpoint": "./wav2vec2-base-demo/checkpoint-7000",
-  "epoch": 14.553014553014552,
+  "best_metric": 0.04381340742111206,
+  "best_model_checkpoint": "./wav2vec2-base-demo/checkpoint-7500",
+  "epoch": 20.0,
   "eval_steps": 500,
-  "global_step": 7000,
+  "global_step": 9620,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 1.0395010395010396,
-      "grad_norm": 2.4276885986328125,
-      "learning_rate": 4.99e-05,
-      "loss": 0.0935,
+      "grad_norm": 2.0576889514923096,
+      "learning_rate": 5e-05,
+      "loss": 0.0875,
       "step": 500
     },
     {
       "epoch": 1.0395010395010396,
-      "eval_cer": 0.03704485933935964,
-      "eval_loss": 0.07600321620702744,
-      "eval_mer": 0.10765687967760507,
-      "eval_runtime": 34.3532,
-      "eval_samples_per_second": 44.741,
-      "eval_steps_per_second": 5.618,
-      "eval_wer": 0.1080737444373808,
-      "eval_wil": 0.19400502379268658,
-      "eval_wip": 0.8059949762073134,
+      "eval_cer": 0.035779866801874644,
+      "eval_loss": 0.06287091970443726,
+      "eval_mer": 0.10693836308521387,
+      "eval_runtime": 38.7834,
+      "eval_samples_per_second": 39.63,
+      "eval_steps_per_second": 4.976,
+      "eval_wer": 0.1073481347909388,
+      "eval_wil": 0.19249465427434787,
+      "eval_wip": 0.8075053457256521,
       "step": 500
     },
     {
       "epoch": 2.079002079002079,
-      "grad_norm": 1.2353624105453491,
-      "learning_rate": 4.8671291355389544e-05,
-      "loss": 0.0914,
+      "grad_norm": 2.8653175830841064,
+      "learning_rate": 4.7258771929824566e-05,
+      "loss": 0.0756,
       "step": 1000
     },
     {
       "epoch": 2.079002079002079,
-      "eval_cer": 0.037467263664779926,
-      "eval_loss": 0.08513512462377548,
-      "eval_mer": 0.1085740644640489,
-      "eval_runtime": 34.2139,
-      "eval_samples_per_second": 44.923,
-      "eval_steps_per_second": 5.641,
-      "eval_wer": 0.10882505923828238,
-      "eval_wil": 0.19587239193616368,
-      "eval_wip": 0.8041276080638363,
+      "eval_cer": 0.03587579137775098,
+      "eval_loss": 0.07413601130247116,
+      "eval_mer": 0.10741198965245755,
+      "eval_runtime": 38.5405,
+      "eval_samples_per_second": 39.88,
+      "eval_steps_per_second": 5.008,
+      "eval_wer": 0.10762988842556069,
+      "eval_wil": 0.1943762121634257,
+      "eval_wip": 0.8056237878365743,
       "step": 1000
     },
     {
       "epoch": 3.1185031185031185,
-      "grad_norm": 0.4545057415962219,
-      "learning_rate": 4.733724653148346e-05,
-      "loss": 0.1007,
+      "grad_norm": 2.5419676303863525,
+      "learning_rate": 4.451754385964912e-05,
+      "loss": 0.0846,
       "step": 1500
     },
     {
       "epoch": 3.1185031185031185,
-      "eval_cer": 0.03667877559066205,
-      "eval_loss": 0.08273813873529434,
-      "eval_mer": 0.10847477196628565,
-      "eval_runtime": 34.1503,
-      "eval_samples_per_second": 45.007,
-      "eval_steps_per_second": 5.651,
-      "eval_wer": 0.10859388545338959,
-      "eval_wil": 0.1960486347087158,
-      "eval_wip": 0.8039513652912842,
+      "eval_cer": 0.03573875626935621,
+      "eval_loss": 0.06114184856414795,
+      "eval_mer": 0.10740366250983036,
+      "eval_runtime": 38.4569,
+      "eval_samples_per_second": 39.967,
+      "eval_steps_per_second": 5.019,
+      "eval_wer": 0.10774258987940945,
+      "eval_wil": 0.19442229772038444,
+      "eval_wip": 0.8055777022796156,
       "step": 1500
     },
     {
       "epoch": 4.158004158004158,
-      "grad_norm": 0.7860220074653625,
-      "learning_rate": 4.600586979722519e-05,
-      "loss": 0.0883,
+      "grad_norm": 0.5870616436004639,
+      "learning_rate": 4.178179824561404e-05,
+      "loss": 0.0702,
       "step": 2000
     },
     {
       "epoch": 4.158004158004158,
-      "eval_cer": 0.03734054236715384,
-      "eval_loss": 0.08130443096160889,
-      "eval_mer": 0.10836359444348377,
-      "eval_runtime": 34.1136,
-      "eval_samples_per_second": 45.055,
-      "eval_steps_per_second": 5.658,
-      "eval_wer": 0.10865167889961279,
-      "eval_wil": 0.19535745081720401,
-      "eval_wip": 0.804642549182796,
+      "eval_cer": 0.03601282648614575,
+      "eval_loss": 0.06569497287273407,
+      "eval_mer": 0.1072850653356514,
+      "eval_runtime": 38.397,
+      "eval_samples_per_second": 40.029,
+      "eval_steps_per_second": 5.026,
+      "eval_wer": 0.10779894060633383,
+      "eval_wil": 0.19276907013348765,
+      "eval_wip": 0.8072309298665123,
       "step": 2000
     },
     {
       "epoch": 5.197505197505198,
-      "grad_norm": 5.241793632507324,
-      "learning_rate": 4.46718249733191e-05,
-      "loss": 0.0988,
+      "grad_norm": 0.7855449318885803,
+      "learning_rate": 3.90405701754386e-05,
+      "loss": 0.0783,
       "step": 2500
     },
     {
       "epoch": 5.197505197505198,
-      "eval_cer": 0.03660837486975867,
-      "eval_loss": 0.07141255587339401,
-      "eval_mer": 0.10702128887093983,
-      "eval_runtime": 34.0472,
-      "eval_samples_per_second": 45.143,
-      "eval_steps_per_second": 5.669,
-      "eval_wer": 0.10720684274403283,
-      "eval_wil": 0.1936232791597472,
-      "eval_wip": 0.8063767208402528,
+      "eval_cer": 0.03561542467180091,
+      "eval_loss": 0.059767913073301315,
+      "eval_mer": 0.10633408071748879,
+      "eval_runtime": 38.4232,
+      "eval_samples_per_second": 40.002,
+      "eval_steps_per_second": 5.023,
+      "eval_wer": 0.10689732897554378,
+      "eval_wil": 0.19087190993998937,
+      "eval_wip": 0.8091280900600106,
       "step": 2500
     },
     {
       "epoch": 6.237006237006237,
-      "grad_norm": 1.3025474548339844,
-      "learning_rate": 4.334044823906083e-05,
-      "loss": 0.0962,
+      "grad_norm": 1.664757251739502,
+      "learning_rate": 3.630482456140351e-05,
+      "loss": 0.0743,
       "step": 3000
     },
     {
       "epoch": 6.237006237006237,
-      "eval_cer": 0.035636844921291995,
-      "eval_loss": 0.06766052544116974,
-      "eval_mer": 0.1056455798396863,
-      "eval_runtime": 34.1329,
-      "eval_samples_per_second": 45.03,
-      "eval_steps_per_second": 5.654,
-      "eval_wer": 0.10587759348089927,
-      "eval_wil": 0.1915670539174873,
-      "eval_wip": 0.8084329460825127,
+      "eval_cer": 0.03521802285745608,
+      "eval_loss": 0.058559250086545944,
+      "eval_mer": 0.10588895120583286,
+      "eval_runtime": 38.4379,
+      "eval_samples_per_second": 39.987,
+      "eval_steps_per_second": 5.021,
+      "eval_wer": 0.10639017243322439,
+      "eval_wil": 0.1914301768679808,
+      "eval_wip": 0.8085698231320192,
       "step": 3000
     },
     {
       "epoch": 7.276507276507276,
-      "grad_norm": 1.1820029020309448,
-      "learning_rate": 4.200907150480256e-05,
-      "loss": 0.0951,
+      "grad_norm": 1.1673611402511597,
+      "learning_rate": 3.356907894736842e-05,
+      "loss": 0.0706,
       "step": 3500
     },
     {
       "epoch": 7.276507276507276,
-      "eval_cer": 0.03551012362366591,
-      "eval_loss": 0.06142089143395424,
-      "eval_mer": 0.10593122547888299,
-      "eval_runtime": 34.0865,
-      "eval_samples_per_second": 45.091,
-      "eval_steps_per_second": 5.662,
-      "eval_wer": 0.10610876726579206,
-      "eval_wil": 0.19198195094690818,
-      "eval_wip": 0.8080180490530918,
+      "eval_cer": 0.033943596349384715,
+      "eval_loss": 0.0515948086977005,
+      "eval_mer": 0.10461261868644306,
+      "eval_runtime": 38.4144,
+      "eval_samples_per_second": 40.011,
+      "eval_steps_per_second": 5.024,
+      "eval_wer": 0.10492505353319058,
+      "eval_wil": 0.18992491039188297,
+      "eval_wip": 0.810075089608117,
       "step": 3500
     },
     {
       "epoch": 8.316008316008316,
-      "grad_norm": 3.633211851119995,
-      "learning_rate": 4.0677694770544295e-05,
-      "loss": 0.1161,
+      "grad_norm": 1.4960205554962158,
+      "learning_rate": 3.0833333333333335e-05,
+      "loss": 0.0861,
       "step": 4000
     },
     {
       "epoch": 8.316008316008316,
-      "eval_cer": 0.03551012362366591,
-      "eval_loss": 0.06067756190896034,
-      "eval_mer": 0.10610691424946658,
-      "eval_runtime": 34.0596,
-      "eval_samples_per_second": 45.127,
-      "eval_steps_per_second": 5.667,
-      "eval_wer": 0.10633994105068485,
-      "eval_wil": 0.19244783032392676,
-      "eval_wip": 0.8075521696760732,
+      "eval_cer": 0.03432729465289007,
+      "eval_loss": 0.05460890755057335,
+      "eval_mer": 0.1054821478774248,
+      "eval_runtime": 38.5397,
+      "eval_samples_per_second": 39.881,
+      "eval_steps_per_second": 5.008,
+      "eval_wer": 0.10571396371013186,
+      "eval_wil": 0.19116659580295647,
+      "eval_wip": 0.8088334041970435,
       "step": 4000
     },
     {
       "epoch": 9.355509355509355,
-      "grad_norm": 1.8992377519607544,
-      "learning_rate": 3.934364994663821e-05,
-      "loss": 0.1202,
+      "grad_norm": 1.2034363746643066,
+      "learning_rate": 2.80921052631579e-05,
+      "loss": 0.0815,
       "step": 4500
     },
     {
       "epoch": 9.355509355509355,
-      "eval_cer": 0.036453493283771225,
-      "eval_loss": 0.07183138281106949,
-      "eval_mer": 0.10711813567143516,
-      "eval_runtime": 34.0394,
-      "eval_samples_per_second": 45.154,
-      "eval_steps_per_second": 5.67,
-      "eval_wer": 0.10732242963647923,
-      "eval_wil": 0.19300207421517157,
-      "eval_wip": 0.8069979257848284,
+      "eval_cer": 0.033710636665113605,
+      "eval_loss": 0.04747992008924484,
+      "eval_mer": 0.10431998202348183,
+      "eval_runtime": 38.4001,
+      "eval_samples_per_second": 40.026,
+      "eval_steps_per_second": 5.026,
+      "eval_wer": 0.10464329989856869,
+      "eval_wil": 0.18925902394749305,
+      "eval_wip": 0.810740976052507,
       "step": 4500
     },
     {
       "epoch": 10.395010395010395,
-      "grad_norm": 1.5443813800811768,
-      "learning_rate": 3.800960512273213e-05,
-      "loss": 0.1431,
+      "grad_norm": 1.3969591856002808,
+      "learning_rate": 2.535087719298246e-05,
+      "loss": 0.0945,
       "step": 5000
     },
     {
       "epoch": 10.395010395010395,
-      "eval_cer": 0.036031088958350936,
-      "eval_loss": 0.0653519257903099,
-      "eval_mer": 0.10687726984492996,
-      "eval_runtime": 34.4227,
-      "eval_samples_per_second": 44.651,
-      "eval_steps_per_second": 5.607,
-      "eval_wer": 0.10714904929780963,
-      "eval_wil": 0.19299958643561255,
-      "eval_wip": 0.8070004135643875,
+      "eval_cer": 0.03349138049168197,
+      "eval_loss": 0.04973128065466881,
+      "eval_mer": 0.10454392081880554,
+      "eval_runtime": 38.8567,
+      "eval_samples_per_second": 39.556,
+      "eval_steps_per_second": 4.967,
+      "eval_wer": 0.10475600135241744,
+      "eval_wil": 0.19024756262985354,
+      "eval_wip": 0.8097524373701465,
       "step": 5000
     },
     {
       "epoch": 11.434511434511435,
-      "grad_norm": 0.9764054417610168,
-      "learning_rate": 3.667556029882604e-05,
-      "loss": 0.1356,
+      "grad_norm": 1.1010193824768066,
+      "learning_rate": 2.260964912280702e-05,
+      "loss": 0.0887,
       "step": 5500
     },
     {
       "epoch": 11.434511434511435,
-      "eval_cer": 0.03546788319112388,
-      "eval_loss": 0.07734616100788116,
-      "eval_mer": 0.10516307479543621,
-      "eval_runtime": 34.1476,
-      "eval_samples_per_second": 45.01,
-      "eval_steps_per_second": 5.652,
-      "eval_wer": 0.10547303935733687,
-      "eval_wil": 0.18957490448529868,
-      "eval_wip": 0.8104250955147013,
+      "eval_cer": 0.03318990325321347,
+      "eval_loss": 0.047050829976797104,
+      "eval_mer": 0.10347152005392653,
+      "eval_runtime": 38.4763,
+      "eval_samples_per_second": 39.947,
+      "eval_steps_per_second": 5.016,
+      "eval_wer": 0.10379803899470304,
+      "eval_wil": 0.1878148971119521,
+      "eval_wip": 0.8121851028880479,
       "step": 5500
     },
     {
       "epoch": 12.474012474012474,
-      "grad_norm": 1.104174017906189,
-      "learning_rate": 3.5341515474919964e-05,
-      "loss": 0.1561,
+      "grad_norm": 0.8793329000473022,
+      "learning_rate": 1.986842105263158e-05,
+      "loss": 0.1057,
       "step": 6000
     },
     {
       "epoch": 12.474012474012474,
-      "eval_cer": 0.035862127228182814,
-      "eval_loss": 0.06988665461540222,
-      "eval_mer": 0.10506015773415463,
-      "eval_runtime": 34.0048,
-      "eval_samples_per_second": 45.2,
-      "eval_steps_per_second": 5.676,
-      "eval_wer": 0.10547303935733687,
-      "eval_wil": 0.1890727356961288,
-      "eval_wip": 0.8109272643038712,
+      "eval_cer": 0.03343656644832406,
+      "eval_loss": 0.044334810227155685,
+      "eval_mer": 0.1031848564848621,
+      "eval_runtime": 38.3629,
+      "eval_samples_per_second": 40.065,
+      "eval_steps_per_second": 5.031,
+      "eval_wer": 0.10351628536008115,
+      "eval_wil": 0.1874799681437187,
+      "eval_wip": 0.8125200318562813,
       "step": 6000
     },
     {
       "epoch": 13.513513513513514,
-      "grad_norm": 4.416908264160156,
-      "learning_rate": 3.400747065101387e-05,
-      "loss": 0.134,
+      "grad_norm": 2.145364761352539,
+      "learning_rate": 1.712719298245614e-05,
+      "loss": 0.0852,
       "step": 6500
     },
     {
       "epoch": 13.513513513513514,
-      "eval_cer": 0.0349750781448002,
-      "eval_loss": 0.0634101927280426,
-      "eval_mer": 0.10508709193678625,
-      "eval_runtime": 34.1003,
-      "eval_samples_per_second": 45.073,
-      "eval_steps_per_second": 5.66,
-      "eval_wer": 0.10529965901866728,
-      "eval_wil": 0.19050757507724403,
-      "eval_wip": 0.809492424922756,
+      "eval_cer": 0.03316249623153452,
+      "eval_loss": 0.05074913427233696,
+      "eval_mer": 0.10275280898876404,
+      "eval_runtime": 38.4426,
+      "eval_samples_per_second": 39.982,
+      "eval_steps_per_second": 5.02,
+      "eval_wer": 0.10306547954468613,
+      "eval_wil": 0.18655687555087064,
+      "eval_wip": 0.8134431244491294,
       "step": 6500
     },
     {
       "epoch": 14.553014553014552,
-      "grad_norm": 0.7748957276344299,
-      "learning_rate": 3.267342582710779e-05,
-      "loss": 0.1304,
+      "grad_norm": 1.7652612924575806,
+      "learning_rate": 1.4385964912280702e-05,
+      "loss": 0.0839,
       "step": 7000
     },
     {
       "epoch": 14.553014553014552,
-      "eval_cer": 0.03490467742389682,
-      "eval_loss": 0.05392587557435036,
-      "eval_mer": 0.10498097544102386,
-      "eval_runtime": 34.1635,
-      "eval_samples_per_second": 44.99,
-      "eval_steps_per_second": 5.649,
-      "eval_wer": 0.10524186557244408,
-      "eval_wil": 0.18975677659084123,
-      "eval_wip": 0.8102432234091588,
+      "eval_cer": 0.03292953654726341,
+      "eval_loss": 0.04428742080926895,
+      "eval_mer": 0.1023710529272952,
+      "eval_runtime": 38.4249,
+      "eval_samples_per_second": 40.0,
+      "eval_steps_per_second": 5.023,
+      "eval_wer": 0.10267102445621548,
+      "eval_wil": 0.18581708506190875,
+      "eval_wip": 0.8141829149380913,
       "step": 7000
+    },
+    {
+      "epoch": 15.592515592515593,
+      "grad_norm": 0.4182071089744568,
+      "learning_rate": 1.1655701754385966e-05,
+      "loss": 0.114,
+      "step": 7500
+    },
+    {
+      "epoch": 15.592515592515593,
+      "eval_cer": 0.03295694356894236,
+      "eval_loss": 0.04381340742111206,
+      "eval_mer": 0.10235380034829504,
+      "eval_runtime": 38.4514,
+      "eval_samples_per_second": 39.973,
+      "eval_steps_per_second": 5.019,
+      "eval_wer": 0.10267102445621548,
+      "eval_wil": 0.18560348362381918,
+      "eval_wip": 0.8143965163761808,
+      "step": 7500
+    },
+    {
+      "epoch": 16.632016632016633,
+      "grad_norm": 0.8875317573547363,
+      "learning_rate": 8.914473684210526e-06,
+      "loss": 0.1054,
+      "step": 8000
+    },
+    {
+      "epoch": 16.632016632016633,
+      "eval_cer": 0.03250472771123962,
+      "eval_loss": 0.04421629384160042,
+      "eval_mer": 0.10242860355295705,
+      "eval_runtime": 38.4795,
+      "eval_samples_per_second": 39.943,
+      "eval_steps_per_second": 5.016,
+      "eval_wer": 0.10267102445621548,
+      "eval_wil": 0.18642151998260104,
+      "eval_wip": 0.813578480017399,
+      "step": 8000
+    },
+    {
+      "epoch": 17.671517671517673,
+      "grad_norm": 1.46517813205719,
+      "learning_rate": 6.1787280701754385e-06,
+      "loss": 0.108,
+      "step": 8500
+    },
+    {
+      "epoch": 17.671517671517673,
+      "eval_cer": 0.03273768739551073,
+      "eval_loss": 0.044399794191122055,
+      "eval_mer": 0.1021680521231184,
+      "eval_runtime": 38.4253,
+      "eval_samples_per_second": 40.0,
+      "eval_steps_per_second": 5.023,
+      "eval_wer": 0.10250197227544236,
+      "eval_wil": 0.18531456248554645,
+      "eval_wip": 0.8146854375144535,
+      "step": 8500
+    },
+    {
+      "epoch": 18.71101871101871,
+      "grad_norm": 0.9103938341140747,
+      "learning_rate": 3.4429824561403515e-06,
+      "loss": 0.1081,
+      "step": 9000
+    },
+    {
+      "epoch": 18.71101871101871,
+      "eval_cer": 0.03260065228711596,
+      "eval_loss": 0.044667795300483704,
+      "eval_mer": 0.10185861081475658,
+      "eval_runtime": 38.5859,
+      "eval_samples_per_second": 39.833,
+      "eval_steps_per_second": 5.002,
+      "eval_wer": 0.10222021864082047,
+      "eval_wil": 0.18466398681705298,
+      "eval_wip": 0.815336013182947,
+      "step": 9000
+    },
+    {
+      "epoch": 19.75051975051975,
+      "grad_norm": 1.2316771745681763,
+      "learning_rate": 7.017543859649123e-07,
+      "loss": 0.1122,
+      "step": 9500
+    },
+    {
+      "epoch": 19.75051975051975,
+      "eval_cer": 0.03240880313536328,
+      "eval_loss": 0.04449043795466423,
+      "eval_mer": 0.10180347210517444,
+      "eval_runtime": 38.5699,
+      "eval_samples_per_second": 39.85,
+      "eval_steps_per_second": 5.004,
+      "eval_wer": 0.10210751718697171,
+      "eval_wil": 0.1850184634725608,
+      "eval_wip": 0.8149815365274392,
+      "step": 9500
     }
   ],
   "logging_steps": 500,
-  "max_steps": 19240,
+  "max_steps": 9620,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 40,
+  "num_train_epochs": 20,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -301,12 +401,12 @@
         "should_evaluate": false,
         "should_log": false,
         "should_save": true,
-        "should_training_stop": false
+        "should_training_stop": true
       },
       "attributes": {}
     }
   },
-  "total_flos": 4.4824764428392996e+18,
+  "total_flos": 6.154380357141717e+18,
   "train_batch_size": 16,
   "trial_name": null,
   "trial_params": null

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:775696d645371d4025f4adc46331697961085adc3d0812695e8af16d5dc92a47
+oid sha256:38d136ab42c7a1566f953cd7f5347049927e6445ad163737d0a0ffc5e5be98b2
 size 5176