add updated model files, new architecture

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7e16a00f3b7f5e5a385a1ce482a06886aaa158f7810c2d751cc6c06faba66b9d
+oid sha256:e6a46a29427f2b7642e604221ccc3a99c7344d84b80ecb503770cc76eb6db8c8
 size 308321413

rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7a8af7acfeb1e1d6a4eb4294cdf97bc63445757ad0d0aa78a19b5d357906885c
+size 14575

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:52ee137ddf3e7c84e65c0bbe4aa2de396127a8cd6c5276dbbba60ed19c1bcc8f
+size 627

trainer_state.json

@@ -1,377 +1,279 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 24.983073798239676,
-  "global_step": 3075,
+  "epoch": 16.995936461026968,
+  "global_step": 3834,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.3861538956236535,
-      "eval_bleu": 58.02299639521676,
-      "eval_loss": 0.3069784641265869,
-      "eval_perfect_accuracy": 0.0773067331670823,
-      "eval_runtime": 1188.6804,
-      "eval_samples_per_second": 3.036,
-      "eval_simple_accuracy": 0.6000514485996334,
-      "eval_steps_per_second": 0.051,
-      "eval_token_accuracy": 0.4887938518923438,
-      "step": 123
+      "eval_bleu": 69.52679870560905,
+      "eval_loss": 0.16526192426681519,
+      "eval_perfect_accuracy": 0.10141313383208644,
+      "eval_runtime": 3340.3895,
+      "eval_samples_per_second": 1.08,
+      "eval_steps_per_second": 0.045,
+      "eval_subword_accuracy": 0.5594287933042841,
+      "eval_token_accuracy": 0.6964062297769533,
+      "eval_word_accuracy": 0.4744973429025366,
+      "step": 225
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.4548114329807414,
-      "eval_bleu": 65.34000395485323,
-      "eval_loss": 0.21667715907096863,
-      "eval_perfect_accuracy": 0.10446106954835134,
-      "eval_runtime": 1040.2638,
-      "eval_samples_per_second": 3.469,
-      "eval_simple_accuracy": 0.6507732373082983,
-      "eval_steps_per_second": 0.059,
-      "eval_token_accuracy": 0.5523261421727808,
-      "step": 246
+      "eval_bleu": 74.31204559162617,
+      "eval_loss": 0.11631891876459122,
+      "eval_perfect_accuracy": 0.14990302022720975,
+      "eval_runtime": 3269.4249,
+      "eval_samples_per_second": 1.104,
+      "eval_steps_per_second": 0.046,
+      "eval_subword_accuracy": 0.6142826689405244,
+      "eval_token_accuracy": 0.7389895544659987,
+      "eval_word_accuracy": 0.532967032967033,
+      "step": 451
     },
     {
-      "epoch": 3.0,
-      "eval_accuracy": 0.49269371274754853,
-      "eval_bleu": 68.74113419920226,
-      "eval_loss": 0.18049955368041992,
-      "eval_perfect_accuracy": 0.11831532280410086,
-      "eval_runtime": 1030.6108,
-      "eval_samples_per_second": 3.502,
-      "eval_simple_accuracy": 0.667243478818176,
-      "eval_steps_per_second": 0.059,
-      "eval_token_accuracy": 0.5896942895597,
-      "step": 369
-    },
-    {
-      "epoch": 4.0,
-      "eval_accuracy": 0.5115833811985449,
-      "eval_bleu": 70.81806278651923,
-      "eval_loss": 0.1621144413948059,
-      "eval_perfect_accuracy": 0.13022998060404545,
-      "eval_runtime": 1015.4894,
-      "eval_samples_per_second": 3.554,
-      "eval_simple_accuracy": 0.6820154445082647,
-      "eval_steps_per_second": 0.06,
-      "eval_token_accuracy": 0.6063884102367733,
-      "step": 492
-    },
-    {
-      "epoch": 4.06,
-      "learning_rate": 4.186991869918699e-05,
-      "loss": 0.6023,
+      "epoch": 2.22,
+      "learning_rate": 4.555555555555556e-05,
+      "loss": 0.6045,
       "step": 500
     },
     {
-      "epoch": 5.0,
-      "eval_accuracy": 0.528468272733082,
-      "eval_bleu": 72.98982499035274,
-      "eval_loss": 0.15042348206043243,
-      "eval_perfect_accuracy": 0.13577168190634525,
-      "eval_runtime": 1033.5998,
-      "eval_samples_per_second": 3.492,
-      "eval_simple_accuracy": 0.6923445587577481,
-      "eval_steps_per_second": 0.059,
-      "eval_token_accuracy": 0.6224039874752381,
-      "step": 615
-    },
-    {
-      "epoch": 6.0,
-      "eval_accuracy": 0.5492471073070218,
-      "eval_bleu": 74.56394107515801,
-      "eval_loss": 0.14072531461715698,
-      "eval_perfect_accuracy": 0.14657799944582986,
-      "eval_runtime": 1023.0499,
-      "eval_samples_per_second": 3.528,
-      "eval_simple_accuracy": 0.7019495958155016,
-      "eval_steps_per_second": 0.06,
-      "eval_token_accuracy": 0.6440957362498019,
-      "step": 738
-    },
-    {
-      "epoch": 7.0,
-      "eval_accuracy": 0.564568482688391,
-      "eval_bleu": 75.61656636338704,
-      "eval_loss": 0.13451269268989563,
-      "eval_perfect_accuracy": 0.14934885009697976,
-      "eval_runtime": 1024.191,
-      "eval_samples_per_second": 3.524,
-      "eval_simple_accuracy": 0.7093622708757638,
-      "eval_steps_per_second": 0.06,
-      "eval_token_accuracy": 0.6577774949083504,
-      "step": 861
+      "epoch": 3.0,
+      "eval_bleu": 76.85411902074411,
+      "eval_loss": 0.09909386187791824,
+      "eval_perfect_accuracy": 0.18952618453865336,
+      "eval_runtime": 3381.7443,
+      "eval_samples_per_second": 1.067,
+      "eval_steps_per_second": 0.045,
+      "eval_subword_accuracy": 0.6512524901453821,
+      "eval_token_accuracy": 0.7695926757936676,
+      "eval_word_accuracy": 0.5709718317623064,
+      "step": 676
     },
     {
-      "epoch": 7.99,
-      "eval_accuracy": 0.5669532973180321,
-      "eval_bleu": 76.49084761884065,
-      "eval_loss": 0.13158030807971954,
-      "eval_perfect_accuracy": 0.14768633970628983,
-      "eval_runtime": 1031.2577,
-      "eval_samples_per_second": 3.5,
-      "eval_simple_accuracy": 0.7093948790077678,
-      "eval_steps_per_second": 0.059,
-      "eval_token_accuracy": 0.661253716075824,
-      "step": 984
+      "epoch": 4.0,
+      "eval_bleu": 79.14575222404268,
+      "eval_loss": 0.09123498946428299,
+      "eval_perfect_accuracy": 0.21169298974785258,
+      "eval_runtime": 3480.4117,
+      "eval_samples_per_second": 1.037,
+      "eval_steps_per_second": 0.043,
+      "eval_subword_accuracy": 0.6773669070746654,
+      "eval_token_accuracy": 0.7950973443948806,
+      "eval_word_accuracy": 0.5997376762487704,
+      "step": 902
     },
     {
-      "epoch": 8.12,
-      "learning_rate": 3.373983739837399e-05,
-      "loss": 0.1587,
+      "epoch": 4.43,
+      "learning_rate": 4.111111111111111e-05,
+      "loss": 0.1137,
       "step": 1000
     },
     {
-      "epoch": 8.99,
-      "eval_accuracy": 0.5799008390541571,
-      "eval_bleu": 77.12937097778037,
-      "eval_loss": 0.12549354135990143,
-      "eval_perfect_accuracy": 0.16098642283180936,
-      "eval_runtime": 1023.3584,
-      "eval_samples_per_second": 3.527,
-      "eval_simple_accuracy": 0.7213640986524281,
-      "eval_steps_per_second": 0.06,
-      "eval_token_accuracy": 0.6679379608441394,
-      "step": 1107
-    },
-    {
-      "epoch": 9.99,
-      "eval_accuracy": 0.5873817235028894,
-      "eval_bleu": 77.65888247465986,
-      "eval_loss": 0.12297692894935608,
-      "eval_perfect_accuracy": 0.16181767802715433,
-      "eval_runtime": 1023.9043,
-      "eval_samples_per_second": 3.525,
-      "eval_simple_accuracy": 0.7230901124023623,
-      "eval_steps_per_second": 0.06,
-      "eval_token_accuracy": 0.6762557947545564,
-      "step": 1230
+      "epoch": 5.0,
+      "eval_bleu": 80.22128752790911,
+      "eval_loss": 0.08361164480447769,
+      "eval_perfect_accuracy": 0.23330562482682182,
+      "eval_runtime": 1727.4258,
+      "eval_samples_per_second": 2.089,
+      "eval_steps_per_second": 0.087,
+      "eval_subword_accuracy": 0.693643785404865,
+      "eval_token_accuracy": 0.8034821726091302,
+      "eval_word_accuracy": 0.6188809276376962,
+      "step": 1127
     },
     {
-      "epoch": 10.99,
-      "eval_accuracy": 0.59429296131896,
-      "eval_bleu": 78.01343118892193,
-      "eval_loss": 0.12066468596458435,
-      "eval_perfect_accuracy": 0.171515655306179,
-      "eval_runtime": 1021.8584,
-      "eval_samples_per_second": 3.532,
-      "eval_simple_accuracy": 0.7276157260621433,
-      "eval_steps_per_second": 0.06,
-      "eval_token_accuracy": 0.6818008877615727,
+      "epoch": 6.0,
+      "eval_bleu": 80.47331007551738,
+      "eval_loss": 0.07888099551200867,
+      "eval_perfect_accuracy": 0.25270157938487114,
+      "eval_runtime": 1706.4952,
+      "eval_samples_per_second": 2.115,
+      "eval_steps_per_second": 0.088,
+      "eval_subword_accuracy": 0.6985128247973933,
+      "eval_token_accuracy": 0.8036176790040939,
+      "eval_word_accuracy": 0.6238234160841085,
       "step": 1353
     },
     {
-      "epoch": 12.0,
-      "eval_accuracy": 0.6051917739086551,
-      "eval_bleu": 78.9493434619066,
-      "eval_loss": 0.11863834410905838,
-      "eval_perfect_accuracy": 0.17594901634801885,
-      "eval_runtime": 1010.8004,
-      "eval_samples_per_second": 3.57,
-      "eval_simple_accuracy": 0.7328412818571789,
-      "eval_steps_per_second": 0.06,
-      "eval_token_accuracy": 0.6936664143325764,
-      "step": 1477
-    },
-    {
-      "epoch": 12.19,
-      "learning_rate": 2.5609756097560977e-05,
-      "loss": 0.1271,
+      "epoch": 6.65,
+      "learning_rate": 3.6666666666666666e-05,
+      "loss": 0.0886,
       "step": 1500
     },
     {
-      "epoch": 13.0,
-      "eval_accuracy": 0.6046047631112237,
-      "eval_bleu": 79.73687831704848,
-      "eval_loss": 0.11646737158298492,
-      "eval_perfect_accuracy": 0.17650318647824884,
-      "eval_runtime": 1016.1798,
-      "eval_samples_per_second": 3.552,
-      "eval_simple_accuracy": 0.733183430453509,
-      "eval_steps_per_second": 0.06,
-      "eval_token_accuracy": 0.6896060298961236,
-      "step": 1600
-    },
-    {
-      "epoch": 14.0,
-      "eval_accuracy": 0.6090549311381985,
-      "eval_bleu": 79.73048335421603,
-      "eval_loss": 0.11454274505376816,
-      "eval_perfect_accuracy": 0.17955112219451372,
-      "eval_runtime": 1021.0729,
-      "eval_samples_per_second": 3.535,
-      "eval_simple_accuracy": 0.7366154820326104,
-      "eval_steps_per_second": 0.06,
-      "eval_token_accuracy": 0.6950767769510844,
-      "step": 1723
+      "epoch": 7.0,
+      "eval_bleu": 81.35496460228585,
+      "eval_loss": 0.07820562273263931,
+      "eval_perfect_accuracy": 0.257689110556941,
+      "eval_runtime": 1696.8822,
+      "eval_samples_per_second": 2.127,
+      "eval_steps_per_second": 0.089,
+      "eval_subword_accuracy": 0.7077631303620004,
+      "eval_token_accuracy": 0.8164888051893702,
+      "eval_word_accuracy": 0.6347371318469585,
+      "step": 1579
     },
     {
-      "epoch": 15.0,
-      "eval_accuracy": 0.6126954921803128,
-      "eval_bleu": 79.95232854185171,
-      "eval_loss": 0.11355158686637878,
-      "eval_perfect_accuracy": 0.18149071765031866,
-      "eval_runtime": 1019.1512,
-      "eval_samples_per_second": 3.541,
-      "eval_simple_accuracy": 0.738064270532627,
-      "eval_steps_per_second": 0.06,
-      "eval_token_accuracy": 0.6964121435142594,
-      "step": 1846
+      "epoch": 8.0,
+      "eval_bleu": 81.4155970043395,
+      "eval_loss": 0.07889819890260696,
+      "eval_perfect_accuracy": 0.2454973676918814,
+      "eval_runtime": 1666.9192,
+      "eval_samples_per_second": 2.165,
+      "eval_steps_per_second": 0.091,
+      "eval_subword_accuracy": 0.6984033613445378,
+      "eval_token_accuracy": 0.8176890756302521,
+      "eval_word_accuracy": 0.6222492535662943,
+      "step": 1804
     },
     {
-      "epoch": 16.0,
-      "eval_accuracy": 0.6142770873724692,
-      "eval_bleu": 80.17800049091872,
-      "eval_loss": 0.11310549825429916,
-      "eval_perfect_accuracy": 0.17871986699916875,
-      "eval_runtime": 1000.0367,
-      "eval_samples_per_second": 3.609,
-      "eval_simple_accuracy": 0.7383911260846073,
-      "eval_steps_per_second": 0.061,
-      "eval_token_accuracy": 0.6993293710072148,
-      "step": 1969
+      "epoch": 8.87,
+      "learning_rate": 3.222222222222223e-05,
+      "loss": 0.0768,
+      "step": 2000
     },
     {
-      "epoch": 16.25,
-      "learning_rate": 1.747967479674797e-05,
-      "loss": 0.1109,
-      "step": 2000
+      "epoch": 9.0,
+      "eval_bleu": 81.70774187369526,
+      "eval_loss": 0.07279801368713379,
+      "eval_perfect_accuracy": 0.27847049044056527,
+      "eval_runtime": 1589.4085,
+      "eval_samples_per_second": 2.271,
+      "eval_steps_per_second": 0.095,
+      "eval_subword_accuracy": 0.7123310810810811,
+      "eval_token_accuracy": 0.8106841216216216,
+      "eval_word_accuracy": 0.6412536119137586,
+      "step": 2030
     },
     {
-      "epoch": 17.0,
-      "eval_accuracy": 0.6165277159040286,
-      "eval_bleu": 80.11868616750657,
-      "eval_loss": 0.11200028657913208,
-      "eval_perfect_accuracy": 0.17899695206428373,
-      "eval_runtime": 994.6715,
-      "eval_samples_per_second": 3.628,
-      "eval_simple_accuracy": 0.7413924775663463,
-      "eval_steps_per_second": 0.061,
-      "eval_token_accuracy": 0.6986571628587793,
-      "step": 2092
+      "epoch": 10.0,
+      "eval_bleu": 82.60753055516308,
+      "eval_loss": 0.07139766961336136,
+      "eval_perfect_accuracy": 0.29897478525907456,
+      "eval_runtime": 1630.9624,
+      "eval_samples_per_second": 2.213,
+      "eval_steps_per_second": 0.093,
+      "eval_subword_accuracy": 0.7325649377766642,
+      "eval_token_accuracy": 0.8289902280130294,
+      "eval_word_accuracy": 0.6648324437352116,
+      "step": 2255
     },
     {
-      "epoch": 18.0,
-      "eval_accuracy": 0.622873435767464,
-      "eval_bleu": 80.72391323382952,
-      "eval_loss": 0.1101086214184761,
-      "eval_perfect_accuracy": 0.18703241895261846,
-      "eval_runtime": 1022.8027,
-      "eval_samples_per_second": 3.529,
-      "eval_simple_accuracy": 0.7436717883731981,
-      "eval_steps_per_second": 0.06,
-      "eval_token_accuracy": 0.7055916363060352,
-      "step": 2215
+      "epoch": 11.0,
+      "eval_bleu": 83.62166331521699,
+      "eval_loss": 0.06930351257324219,
+      "eval_perfect_accuracy": 0.30950401773344416,
+      "eval_runtime": 1612.5857,
+      "eval_samples_per_second": 2.238,
+      "eval_steps_per_second": 0.094,
+      "eval_subword_accuracy": 0.739740238184157,
+      "eval_token_accuracy": 0.8358438109465123,
+      "eval_word_accuracy": 0.6716549295774648,
+      "step": 2481
     },
     {
-      "epoch": 19.0,
-      "eval_accuracy": 0.6222941661905669,
-      "eval_bleu": 80.76914319903946,
-      "eval_loss": 0.11085215210914612,
-      "eval_perfect_accuracy": 0.18647824882238848,
-      "eval_runtime": 1021.7653,
-      "eval_samples_per_second": 3.532,
-      "eval_simple_accuracy": 0.7428426331492414,
-      "eval_steps_per_second": 0.06,
-      "eval_token_accuracy": 0.7042150701453691,
-      "step": 2338
+      "epoch": 11.08,
+      "learning_rate": 2.777777777777778e-05,
+      "loss": 0.0681,
+      "step": 2500
     },
     {
-      "epoch": 19.99,
-      "eval_accuracy": 0.6245798719005644,
-      "eval_bleu": 80.78924704054619,
-      "eval_loss": 0.11003228276968002,
-      "eval_perfect_accuracy": 0.19229703518980326,
-      "eval_runtime": 1016.0765,
-      "eval_samples_per_second": 3.552,
-      "eval_simple_accuracy": 0.7465914135328809,
-      "eval_steps_per_second": 0.06,
-      "eval_token_accuracy": 0.7053395903354683,
-      "step": 2461
+      "epoch": 12.0,
+      "eval_bleu": 84.02807089662639,
+      "eval_loss": 0.06798376142978668,
+      "eval_perfect_accuracy": 0.3081185924078692,
+      "eval_runtime": 1680.6877,
+      "eval_samples_per_second": 2.147,
+      "eval_steps_per_second": 0.09,
+      "eval_subword_accuracy": 0.745770423991727,
+      "eval_token_accuracy": 0.8428541882109617,
+      "eval_word_accuracy": 0.6793930762489044,
+      "step": 2707
     },
     {
-      "epoch": 20.31,
-      "learning_rate": 9.34959349593496e-06,
-      "loss": 0.1009,
-      "step": 2500
+      "epoch": 13.0,
+      "eval_bleu": 84.42354785113844,
+      "eval_loss": 0.06762586534023285,
+      "eval_perfect_accuracy": 0.31698531449154893,
+      "eval_runtime": 1751.9665,
+      "eval_samples_per_second": 2.06,
+      "eval_steps_per_second": 0.086,
+      "eval_subword_accuracy": 0.7447953313190679,
+      "eval_token_accuracy": 0.8407350689127105,
+      "eval_word_accuracy": 0.67777227994942,
+      "step": 2932
     },
     {
-      "epoch": 20.99,
-      "eval_accuracy": 0.6269267364414843,
-      "eval_bleu": 81.12519074317272,
-      "eval_loss": 0.10916706919670105,
-      "eval_perfect_accuracy": 0.1886949293433084,
-      "eval_runtime": 1038.5849,
-      "eval_samples_per_second": 3.475,
-      "eval_simple_accuracy": 0.7457342213764668,
-      "eval_steps_per_second": 0.059,
-      "eval_token_accuracy": 0.708499841420869,
-      "step": 2584
+      "epoch": 13.3,
+      "learning_rate": 2.3333333333333336e-05,
+      "loss": 0.0619,
+      "step": 3000
     },
     {
-      "epoch": 21.99,
-      "eval_accuracy": 0.6295486935866983,
-      "eval_bleu": 81.28558003169574,
-      "eval_loss": 0.10952065885066986,
-      "eval_perfect_accuracy": 0.1886949293433084,
-      "eval_runtime": 1037.9043,
-      "eval_samples_per_second": 3.477,
-      "eval_simple_accuracy": 0.7463499604117181,
-      "eval_steps_per_second": 0.059,
-      "eval_token_accuracy": 0.7121773555027712,
-      "step": 2707
+      "epoch": 14.0,
+      "eval_bleu": 84.59233008431931,
+      "eval_loss": 0.06648800522089005,
+      "eval_perfect_accuracy": 0.32696037683568857,
+      "eval_runtime": 1738.746,
+      "eval_samples_per_second": 2.076,
+      "eval_steps_per_second": 0.087,
+      "eval_subword_accuracy": 0.7536297454036935,
+      "eval_token_accuracy": 0.8476946489532349,
+      "eval_word_accuracy": 0.6883173734610123,
+      "step": 3158
     },
     {
-      "epoch": 22.99,
-      "eval_accuracy": 0.6287138141188421,
-      "eval_bleu": 80.93511268093461,
-      "eval_loss": 0.1084115132689476,
-      "eval_perfect_accuracy": 0.19479080077583819,
-      "eval_runtime": 1032.8176,
-      "eval_samples_per_second": 3.494,
-      "eval_simple_accuracy": 0.7478732859319451,
-      "eval_steps_per_second": 0.059,
-      "eval_token_accuracy": 0.7086084306754697,
-      "step": 2830
+      "epoch": 15.0,
+      "eval_bleu": 84.49240456076029,
+      "eval_loss": 0.06621185690164566,
+      "eval_perfect_accuracy": 0.32917705735660846,
+      "eval_runtime": 1720.559,
+      "eval_samples_per_second": 2.098,
+      "eval_steps_per_second": 0.088,
+      "eval_subword_accuracy": 0.7496996063807748,
+      "eval_token_accuracy": 0.8439196188108556,
+      "eval_word_accuracy": 0.6839415256100242,
+      "step": 3383
     },
     {
-      "epoch": 24.0,
-      "eval_accuracy": 0.6307677671011321,
-      "eval_bleu": 81.11966959311552,
-      "eval_loss": 0.10846742242574692,
-      "eval_perfect_accuracy": 0.19201995012468828,
-      "eval_runtime": 1015.3974,
-      "eval_samples_per_second": 3.554,
-      "eval_simple_accuracy": 0.7478197443947611,
-      "eval_steps_per_second": 0.06,
-      "eval_token_accuracy": 0.7126502394317065,
-      "step": 2954
+      "epoch": 15.52,
+      "learning_rate": 1.888888888888889e-05,
+      "loss": 0.0574,
+      "step": 3500
     },
     {
-      "epoch": 24.37,
-      "learning_rate": 1.2195121951219514e-06,
-      "loss": 0.0953,
-      "step": 3000
+      "epoch": 16.0,
+      "eval_bleu": 85.02924908908608,
+      "eval_loss": 0.06474177539348602,
+      "eval_perfect_accuracy": 0.343585480742588,
+      "eval_runtime": 1678.6132,
+      "eval_samples_per_second": 2.15,
+      "eval_steps_per_second": 0.09,
+      "eval_subword_accuracy": 0.7591370348981107,
+      "eval_token_accuracy": 0.8501773781041169,
+      "eval_word_accuracy": 0.6949701058636388,
+      "step": 3609
     },
     {
-      "epoch": 24.98,
-      "eval_accuracy": 0.6323030591218893,
-      "eval_bleu": 81.29208359004284,
-      "eval_loss": 0.10844007134437561,
-      "eval_perfect_accuracy": 0.19257412025491827,
-      "eval_runtime": 1029.0079,
-      "eval_samples_per_second": 3.507,
-      "eval_simple_accuracy": 0.7477571722935489,
-      "eval_steps_per_second": 0.059,
-      "eval_token_accuracy": 0.7139641781581867,
-      "step": 3075
+      "epoch": 17.0,
+      "eval_bleu": 84.79034637013908,
+      "eval_loss": 0.06463250517845154,
+      "eval_perfect_accuracy": 0.34968135217511775,
+      "eval_runtime": 1711.9697,
+      "eval_samples_per_second": 2.108,
+      "eval_steps_per_second": 0.088,
+      "eval_subword_accuracy": 0.762099607356892,
+      "eval_token_accuracy": 0.8493490390576566,
+      "eval_word_accuracy": 0.7003675462175655,
+      "step": 3834
     }
   ],
-  "max_steps": 3075,
+  "max_steps": 5625,
   "num_train_epochs": 25,
-  "total_flos": 9686690034941952.0,
+  "total_flos": 6098403774431232.0,
   "trial_name": null,
   "trial_params": null
 }

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9fb8a43a695feecbdd75975e3545d24ccd9aed989a14888afb3e50c07eee1c23
+oid sha256:4ade5cc7246ed0457ed40a82cf7baea9db562870c60bef51dc078fc52843a21b
 size 4219