Training in progress, step 20500, checkpoint

last-checkpoint/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c43564e930260fa3abbd24ef8bc8e2019043f8ab0ea3f357c88632def606095f
+oid sha256:4f41bb533aa86e301a1be3b65f96c1170612ae823deb447e90a35a957837ed0d
 size 242041896

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5157a79613ec248ac8f7d1c8901dc35a0467ddf892a8d4eace27612fd7b9473d
+oid sha256:5199d65e10b84399e00ecb642c52de2797b550f6e71e4e1f4b276272553ca552
 size 484163514

last-checkpoint/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:af796267785b8380aa702da74322c3a84f42d9cab004eb98c1af133c0218d64a
+oid sha256:d1cd63170bb513ed1093722cd85dcb9976e6bd105066dd8d5d629488d4e94f5b
 size 14244

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5b2103b3ba42c5a974b79f170c6d33390323651972016338595643ea512a6c2b
+oid sha256:53e446aca6ce45dd19f0dd3bb36af3837e8ac7287ccd924c8a3f9b7a954a486f
 size 1064

last-checkpoint/trainer_state.json

@@ -1,9 +1,9 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 4.987722529158994,
+  "epoch": 3.14610190300798,
   "eval_steps": 500,
-  "global_step": 32500,
+  "global_step": 20500,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -11,439 +11,285 @@
     {
       "epoch": 0.08,
       "learning_rate": 0.0009846531614487416,
-      "loss": 0.5762,
+      "loss": 0.3358,
       "step": 500
     },
     {
       "epoch": 0.15,
       "learning_rate": 0.0009693063228974831,
-      "loss": 0.6042,
+      "loss": 0.3723,
       "step": 1000
     },
     {
       "epoch": 0.23,
       "learning_rate": 0.0009539594843462247,
-      "loss": 0.6024,
+      "loss": 0.3901,
       "step": 1500
     },
     {
       "epoch": 0.31,
       "learning_rate": 0.0009386126457949663,
-      "loss": 0.6057,
+      "loss": 0.4022,
       "step": 2000
     },
     {
       "epoch": 0.38,
       "learning_rate": 0.0009232658072437078,
-      "loss": 0.6143,
+      "loss": 0.4109,
       "step": 2500
     },
     {
       "epoch": 0.46,
       "learning_rate": 0.0009079189686924494,
-      "loss": 0.6081,
+      "loss": 0.4132,
       "step": 3000
     },
     {
       "epoch": 0.54,
       "learning_rate": 0.000892572130141191,
-      "loss": 0.611,
+      "loss": 0.4197,
       "step": 3500
     },
     {
       "epoch": 0.61,
       "learning_rate": 0.0008772252915899325,
-      "loss": 0.6104,
+      "loss": 0.4227,
       "step": 4000
     },
     {
       "epoch": 0.69,
-      "learning_rate": 0.0008619091467157766,
-      "loss": 0.6154,
+      "learning_rate": 0.0008618784530386741,
+      "loss": 0.424,
       "step": 4500
     },
     {
       "epoch": 0.77,
-      "learning_rate": 0.0008465623081645181,
-      "loss": 0.6151,
+      "learning_rate": 0.0008465316144874156,
+      "loss": 0.4228,
       "step": 5000
     },
     {
       "epoch": 0.84,
-      "learning_rate": 0.0008312154696132597,
-      "loss": 0.611,
+      "learning_rate": 0.0008311847759361571,
+      "loss": 0.4246,
       "step": 5500
     },
     {
       "epoch": 0.92,
-      "learning_rate": 0.0008158686310620012,
-      "loss": 0.617,
+      "learning_rate": 0.0008158379373848988,
+      "loss": 0.4268,
       "step": 6000
     },
     {
       "epoch": 1.0,
-      "learning_rate": 0.0008005524861878454,
-      "loss": 0.6089,
+      "learning_rate": 0.0008005217925107428,
+      "loss": 0.4316,
       "step": 6500
     },
     {
       "epoch": 1.0,
-      "eval_bleu": 3.6822,
-      "eval_gen_len": 18.235,
-      "eval_loss": 0.506125271320343,
-      "eval_runtime": 27.1471,
-      "eval_samples_per_second": 36.836,
-      "eval_steps_per_second": 2.321,
+      "eval_bleu": 4.2677,
+      "eval_gen_len": 18.024,
+      "eval_loss": 0.45959702134132385,
+      "eval_runtime": 23.3818,
+      "eval_samples_per_second": 42.768,
+      "eval_steps_per_second": 2.694,
       "step": 6516
     },
     {
       "epoch": 1.07,
-      "learning_rate": 0.0007852056476365869,
-      "loss": 0.5657,
+      "learning_rate": 0.0007851749539594844,
+      "loss": 0.3897,
       "step": 7000
     },
     {
       "epoch": 1.15,
-      "learning_rate": 0.0007698588090853285,
-      "loss": 0.5623,
+      "learning_rate": 0.0007698281154082259,
+      "loss": 0.3901,
       "step": 7500
     },
     {
       "epoch": 1.23,
-      "learning_rate": 0.0007545119705340699,
-      "loss": 0.5669,
+      "learning_rate": 0.0007544812768569675,
+      "loss": 0.3965,
       "step": 8000
     },
     {
       "epoch": 1.3,
-      "learning_rate": 0.000739195825659914,
-      "loss": 0.5793,
+      "learning_rate": 0.0007391651319828116,
+      "loss": 0.394,
       "step": 8500
     },
     {
       "epoch": 1.38,
-      "learning_rate": 0.0007238489871086557,
-      "loss": 0.5743,
+      "learning_rate": 0.0007238182934315531,
+      "loss": 0.4003,
       "step": 9000
     },
     {
       "epoch": 1.46,
-      "learning_rate": 0.0007085328422344997,
-      "loss": 0.5722,
+      "learning_rate": 0.0007084714548802946,
+      "loss": 0.403,
       "step": 9500
     },
     {
       "epoch": 1.53,
-      "learning_rate": 0.0006931860036832413,
-      "loss": 0.5749,
+      "learning_rate": 0.0006931246163290362,
+      "loss": 0.4067,
       "step": 10000
     },
     {
       "epoch": 1.61,
-      "learning_rate": 0.0006778391651319828,
-      "loss": 0.5776,
+      "learning_rate": 0.0006778084714548804,
+      "loss": 0.4067,
       "step": 10500
     },
     {
       "epoch": 1.69,
-      "learning_rate": 0.0006624923265807244,
-      "loss": 0.5694,
+      "learning_rate": 0.0006624616329036218,
+      "loss": 0.4047,
       "step": 11000
     },
     {
       "epoch": 1.76,
-      "learning_rate": 0.0006471454880294659,
-      "loss": 0.5716,
+      "learning_rate": 0.0006471147943523635,
+      "loss": 0.4032,
       "step": 11500
     },
     {
       "epoch": 1.84,
-      "learning_rate": 0.0006317986494782076,
-      "loss": 0.5665,
+      "learning_rate": 0.000631767955801105,
+      "loss": 0.4077,
       "step": 12000
     },
     {
       "epoch": 1.92,
-      "learning_rate": 0.000616451810926949,
-      "loss": 0.5671,
+      "learning_rate": 0.0006164825046040515,
+      "loss": 0.4062,
       "step": 12500
     },
     {
       "epoch": 2.0,
-      "learning_rate": 0.0006011049723756906,
-      "loss": 0.5687,
+      "learning_rate": 0.0006011356660527931,
+      "loss": 0.4074,
       "step": 13000
     },
     {
       "epoch": 2.0,
-      "eval_bleu": 3.7119,
-      "eval_gen_len": 18.222,
-      "eval_loss": 0.495604932308197,
-      "eval_runtime": 23.8741,
-      "eval_samples_per_second": 41.886,
-      "eval_steps_per_second": 2.639,
+      "eval_bleu": 4.2659,
+      "eval_gen_len": 18.019,
+      "eval_loss": 0.45938295125961304,
+      "eval_runtime": 24.1279,
+      "eval_samples_per_second": 41.446,
+      "eval_steps_per_second": 2.611,
       "step": 13032
     },
     {
       "epoch": 2.07,
       "learning_rate": 0.0005857888275015347,
-      "loss": 0.5307,
+      "loss": 0.3667,
       "step": 13500
     },
     {
       "epoch": 2.15,
       "learning_rate": 0.0005704419889502763,
-      "loss": 0.5325,
+      "loss": 0.3674,
       "step": 14000
     },
     {
       "epoch": 2.23,
       "learning_rate": 0.0005550951503990178,
-      "loss": 0.5353,
+      "loss": 0.3749,
       "step": 14500
     },
     {
       "epoch": 2.3,
       "learning_rate": 0.0005397483118477594,
-      "loss": 0.5289,
+      "loss": 0.3707,
       "step": 15000
     },
     {
       "epoch": 2.38,
       "learning_rate": 0.0005244014732965009,
-      "loss": 0.5321,
+      "loss": 0.3676,
       "step": 15500
     },
     {
       "epoch": 2.46,
-      "learning_rate": 0.000509085328422345,
-      "loss": 0.5271,
+      "learning_rate": 0.0005090546347452425,
+      "loss": 0.3761,
       "step": 16000
     },
     {
       "epoch": 2.53,
       "learning_rate": 0.0004937384898710866,
-      "loss": 0.5295,
+      "loss": 0.3795,
       "step": 16500
     },
     {
       "epoch": 2.61,
       "learning_rate": 0.00047839165131982814,
-      "loss": 0.524,
+      "loss": 0.377,
       "step": 17000
     },
     {
       "epoch": 2.69,
       "learning_rate": 0.0004630448127685697,
-      "loss": 0.526,
+      "loss": 0.3804,
       "step": 17500
     },
     {
       "epoch": 2.76,
-      "learning_rate": 0.00044772866789441376,
-      "loss": 0.5323,
+      "learning_rate": 0.00044769797421731125,
+      "loss": 0.3824,
       "step": 18000
     },
     {
       "epoch": 2.84,
       "learning_rate": 0.00043238182934315537,
-      "loss": 0.5236,
+      "loss": 0.3724,
       "step": 18500
     },
     {
       "epoch": 2.92,
       "learning_rate": 0.00041703499079189687,
-      "loss": 0.5266,
+      "loss": 0.3795,
       "step": 19000
     },
     {
       "epoch": 2.99,
       "learning_rate": 0.0004016881522406384,
-      "loss": 0.525,
+      "loss": 0.3833,
       "step": 19500
     },
     {
       "epoch": 3.0,
-      "eval_bleu": 3.8722,
-      "eval_gen_len": 18.219,
-      "eval_loss": 0.4868793785572052,
-      "eval_runtime": 23.2067,
-      "eval_samples_per_second": 43.091,
-      "eval_steps_per_second": 2.715,
+      "eval_bleu": 4.3571,
+      "eval_gen_len": 18.015,
+      "eval_loss": 0.4461449086666107,
+      "eval_runtime": 23.8381,
+      "eval_samples_per_second": 41.95,
+      "eval_steps_per_second": 2.643,
       "step": 19548
     },
     {
       "epoch": 3.07,
-      "learning_rate": 0.00038637200736648254,
-      "loss": 0.4934,
+      "learning_rate": 0.00038634131368938,
+      "loss": 0.3423,
       "step": 20000
     },
     {
       "epoch": 3.15,
-      "learning_rate": 0.0003710251688152241,
-      "loss": 0.4869,
+      "learning_rate": 0.0003710558624923266,
+      "loss": 0.3375,
       "step": 20500
-    },
-    {
-      "epoch": 3.22,
-      "learning_rate": 0.0003556783302639656,
-      "loss": 0.4917,
-      "step": 21000
-    },
-    {
-      "epoch": 3.3,
-      "learning_rate": 0.00034033149171270715,
-      "loss": 0.4841,
-      "step": 21500
-    },
-    {
-      "epoch": 3.38,
-      "learning_rate": 0.00032501534683855126,
-      "loss": 0.4906,
-      "step": 22000
-    },
-    {
-      "epoch": 3.45,
-      "learning_rate": 0.0003096685082872928,
-      "loss": 0.4884,
-      "step": 22500
-    },
-    {
-      "epoch": 3.53,
-      "learning_rate": 0.0002943216697360344,
-      "loss": 0.486,
-      "step": 23000
-    },
-    {
-      "epoch": 3.61,
-      "learning_rate": 0.00027897483118477593,
-      "loss": 0.4932,
-      "step": 23500
-    },
-    {
-      "epoch": 3.68,
-      "learning_rate": 0.00026365868631062,
-      "loss": 0.4951,
-      "step": 24000
-    },
-    {
-      "epoch": 3.76,
-      "learning_rate": 0.0002483118477593616,
-      "loss": 0.4896,
-      "step": 24500
-    },
-    {
-      "epoch": 3.84,
-      "learning_rate": 0.00023296500920810313,
-      "loss": 0.4832,
-      "step": 25000
-    },
-    {
-      "epoch": 3.91,
-      "learning_rate": 0.00021761817065684468,
-      "loss": 0.4874,
-      "step": 25500
-    },
-    {
-      "epoch": 3.99,
-      "learning_rate": 0.00020230202578268877,
-      "loss": 0.488,
-      "step": 26000
-    },
-    {
-      "epoch": 4.0,
-      "eval_bleu": 3.7886,
-      "eval_gen_len": 18.23,
-      "eval_loss": 0.4795362949371338,
-      "eval_runtime": 23.2422,
-      "eval_samples_per_second": 43.025,
-      "eval_steps_per_second": 2.711,
-      "step": 26064
-    },
-    {
-      "epoch": 4.07,
-      "learning_rate": 0.00018695518723143033,
-      "loss": 0.4583,
-      "step": 26500
-    },
-    {
-      "epoch": 4.14,
-      "learning_rate": 0.00017160834868017188,
-      "loss": 0.4495,
-      "step": 27000
-    },
-    {
-      "epoch": 4.22,
-      "learning_rate": 0.00015626151012891346,
-      "loss": 0.454,
-      "step": 27500
-    },
-    {
-      "epoch": 4.3,
-      "learning_rate": 0.00014094536525475752,
-      "loss": 0.4547,
-      "step": 28000
-    },
-    {
-      "epoch": 4.37,
-      "learning_rate": 0.00012559852670349908,
-      "loss": 0.4588,
-      "step": 28500
-    },
-    {
-      "epoch": 4.45,
-      "learning_rate": 0.00011025168815224063,
-      "loss": 0.453,
-      "step": 29000
-    },
-    {
-      "epoch": 4.53,
-      "learning_rate": 9.49048496009822e-05,
-      "loss": 0.4549,
-      "step": 29500
-    },
-    {
-      "epoch": 4.6,
-      "learning_rate": 7.955801104972376e-05,
-      "loss": 0.4533,
-      "step": 30000
-    },
-    {
-      "epoch": 4.68,
-      "learning_rate": 6.424186617556783e-05,
-      "loss": 0.4495,
-      "step": 30500
-    },
-    {
-      "epoch": 4.76,
-      "learning_rate": 4.8895027624309394e-05,
-      "loss": 0.4555,
-      "step": 31000
-    },
-    {
-      "epoch": 4.83,
-      "learning_rate": 3.3548189073050956e-05,
-      "loss": 0.4563,
-      "step": 31500
-    },
-    {
-      "epoch": 4.91,
-      "learning_rate": 1.820135052179251e-05,
-      "loss": 0.4563,
-      "step": 32000
-    },
-    {
-      "epoch": 4.99,
-      "learning_rate": 2.885205647636587e-06,
-      "loss": 0.4527,
-      "step": 32500
     }
   ],
   "logging_steps": 500,
   "max_steps": 32580,
   "num_train_epochs": 5,
   "save_steps": 500,
-  "total_flos": 1.7233985138589696e+16,
+  "total_flos": 1.0876549321457664e+16,
   "trial_name": null,
   "trial_params": null
 }

last-checkpoint/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bbbc9f6fd02ff4264ade394c37847b822a605f5a9ef106ad63ba6639ed6bf7c2
+oid sha256:0d9e5fd144c9e568216db4505e71aa6bc722764749fdf639a5e84c071e512165
 size 4792