Training in progress, step 1300, checkpoint

checkpoint-1300/adapter_config.json

@@ -23,13 +23,13 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "v_proj",
-    "q_proj",
-    "gate_proj",
     "up_proj",
     "o_proj",
+    "v_proj",
+    "q_proj",
+    "down_proj",
     "k_proj",
-    "down_proj"
+    "gate_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

checkpoint-1300/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ed0b3ecbcbea7730073f18509d3fae79682af6ae8249e076b0342ff316d2f7f0
+oid sha256:9676e286f226904944b84a9822eea189e7b4af83a0642c8a417e6eb7df06d34b
 size 167832240

checkpoint-1300/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e441969ee559df349cd6c87df3fa5b926a5bd1f383e715665e55ed1078061c49
+oid sha256:a0568d0b174eaa2f3a427307f7a54cf1c63d42a9fd9b991a23d0315a476dff39
 size 335922386

checkpoint-1300/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:41a64e53f230dbc7bb191577f84942f863fa08696d47a151098d1a59eb5fc925
+oid sha256:bd6c3b79c39e6f73635cef961723ed64442bcc5b167613c9c44975a046d574bf
 size 14244

checkpoint-1300/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9f359ccf5a5e0648d6bcb1e4fb4c30fc000d4f24122a1ab52051f66607b5c401
+oid sha256:5a961339657365387feb4bb864ba7c8919a90799d14c2838ad249e04207dcd26
 size 1064

checkpoint-1300/trainer_state.json

@@ -1,7 +1,7 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 1.365546218487395,
+  "epoch": 1.36986301369863,
   "eval_steps": 50,
   "global_step": 1300,
   "is_hyper_param_search": false,
@@ -9,1126 +9,1126 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.01050420168067227,
-      "grad_norm": 2.5582146644592285,
-      "learning_rate": 4.98249299719888e-05,
-      "loss": 1.6787,
+      "epoch": 0.01053740779768177,
+      "grad_norm": 3.070249080657959,
+      "learning_rate": 4.982437653670531e-05,
+      "loss": 1.7879,
       "step": 10
     },
     {
-      "epoch": 0.02100840336134454,
-      "grad_norm": 0.9345070719718933,
-      "learning_rate": 4.96498599439776e-05,
-      "loss": 0.518,
+      "epoch": 0.02107481559536354,
+      "grad_norm": 1.702326774597168,
+      "learning_rate": 4.964875307341061e-05,
+      "loss": 0.5567,
       "step": 20
     },
     {
-      "epoch": 0.031512605042016806,
-      "grad_norm": 1.6358414888381958,
-      "learning_rate": 4.947478991596639e-05,
-      "loss": 0.4604,
+      "epoch": 0.03161222339304531,
+      "grad_norm": 1.1947294473648071,
+      "learning_rate": 4.947312961011591e-05,
+      "loss": 0.4493,
       "step": 30
     },
     {
-      "epoch": 0.04201680672268908,
-      "grad_norm": 0.7778844237327576,
-      "learning_rate": 4.9299719887955186e-05,
-      "loss": 0.3771,
+      "epoch": 0.04214963119072708,
+      "grad_norm": 0.9556658267974854,
+      "learning_rate": 4.929750614682122e-05,
+      "loss": 0.3728,
       "step": 40
     },
     {
-      "epoch": 0.052521008403361345,
-      "grad_norm": 0.7006077766418457,
-      "learning_rate": 4.912464985994398e-05,
-      "loss": 0.3842,
+      "epoch": 0.05268703898840885,
+      "grad_norm": 0.7952510714530945,
+      "learning_rate": 4.9121882683526524e-05,
+      "loss": 0.3535,
       "step": 50
     },
     {
-      "epoch": 0.052521008403361345,
-      "eval_loss": 0.42603224515914917,
-      "eval_runtime": 13.5673,
-      "eval_samples_per_second": 35.379,
-      "eval_steps_per_second": 2.211,
+      "epoch": 0.05268703898840885,
+      "eval_loss": 0.4311191439628601,
+      "eval_runtime": 13.6539,
+      "eval_samples_per_second": 35.155,
+      "eval_steps_per_second": 2.197,
       "step": 50
     },
     {
-      "epoch": 0.06302521008403361,
-      "grad_norm": 0.6415153741836548,
-      "learning_rate": 4.8949579831932775e-05,
-      "loss": 0.3399,
+      "epoch": 0.06322444678609063,
+      "grad_norm": 0.6962826251983643,
+      "learning_rate": 4.894625922023183e-05,
+      "loss": 0.3507,
       "step": 60
     },
     {
-      "epoch": 0.07352941176470588,
-      "grad_norm": 0.6030780076980591,
-      "learning_rate": 4.877450980392157e-05,
-      "loss": 0.3447,
+      "epoch": 0.0737618545837724,
+      "grad_norm": 0.6941961646080017,
+      "learning_rate": 4.877063575693713e-05,
+      "loss": 0.3585,
       "step": 70
     },
     {
-      "epoch": 0.08403361344537816,
-      "grad_norm": 0.688852071762085,
-      "learning_rate": 4.859943977591036e-05,
-      "loss": 0.3219,
+      "epoch": 0.08429926238145416,
+      "grad_norm": 0.6864392757415771,
+      "learning_rate": 4.8595012293642434e-05,
+      "loss": 0.3496,
       "step": 80
     },
     {
-      "epoch": 0.09453781512605042,
-      "grad_norm": 0.6371557712554932,
-      "learning_rate": 4.8424369747899164e-05,
-      "loss": 0.3379,
+      "epoch": 0.09483667017913593,
+      "grad_norm": 0.7322937846183777,
+      "learning_rate": 4.841938883034774e-05,
+      "loss": 0.3295,
       "step": 90
     },
     {
-      "epoch": 0.10504201680672269,
-      "grad_norm": 0.7739270329475403,
-      "learning_rate": 4.824929971988796e-05,
-      "loss": 0.3177,
+      "epoch": 0.1053740779768177,
+      "grad_norm": 0.6921488046646118,
+      "learning_rate": 4.824376536705304e-05,
+      "loss": 0.3357,
       "step": 100
     },
     {
-      "epoch": 0.10504201680672269,
-      "eval_loss": 0.3801896274089813,
-      "eval_runtime": 13.6107,
-      "eval_samples_per_second": 35.266,
-      "eval_steps_per_second": 2.204,
+      "epoch": 0.1053740779768177,
+      "eval_loss": 0.39120009541511536,
+      "eval_runtime": 13.7031,
+      "eval_samples_per_second": 35.029,
+      "eval_steps_per_second": 2.189,
       "step": 100
     },
     {
-      "epoch": 0.11554621848739496,
-      "grad_norm": 0.649507462978363,
-      "learning_rate": 4.807422969187675e-05,
-      "loss": 0.3415,
+      "epoch": 0.11591148577449947,
+      "grad_norm": 0.6553240418434143,
+      "learning_rate": 4.8068141903758344e-05,
+      "loss": 0.3105,
       "step": 110
     },
     {
-      "epoch": 0.12605042016806722,
-      "grad_norm": 0.594717264175415,
-      "learning_rate": 4.7899159663865554e-05,
-      "loss": 0.3325,
+      "epoch": 0.12644889357218125,
+      "grad_norm": 0.5637819170951843,
+      "learning_rate": 4.789251844046364e-05,
+      "loss": 0.3164,
       "step": 120
     },
     {
-      "epoch": 0.13655462184873948,
-      "grad_norm": 0.627918541431427,
-      "learning_rate": 4.772408963585435e-05,
-      "loss": 0.3222,
+      "epoch": 0.136986301369863,
+      "grad_norm": 0.6341928839683533,
+      "learning_rate": 4.7716894977168955e-05,
+      "loss": 0.304,
       "step": 130
     },
     {
-      "epoch": 0.14705882352941177,
-      "grad_norm": 0.5384674668312073,
-      "learning_rate": 4.7549019607843135e-05,
-      "loss": 0.3426,
+      "epoch": 0.1475237091675448,
+      "grad_norm": 0.5917785167694092,
+      "learning_rate": 4.754127151387426e-05,
+      "loss": 0.3234,
       "step": 140
     },
     {
-      "epoch": 0.15756302521008403,
-      "grad_norm": 0.5673420429229736,
-      "learning_rate": 4.7373949579831936e-05,
-      "loss": 0.3061,
+      "epoch": 0.15806111696522657,
+      "grad_norm": 0.5884453654289246,
+      "learning_rate": 4.736564805057956e-05,
+      "loss": 0.317,
       "step": 150
     },
     {
-      "epoch": 0.15756302521008403,
-      "eval_loss": 0.3653399348258972,
-      "eval_runtime": 13.5947,
-      "eval_samples_per_second": 35.308,
-      "eval_steps_per_second": 2.207,
+      "epoch": 0.15806111696522657,
+      "eval_loss": 0.37688738107681274,
+      "eval_runtime": 13.6535,
+      "eval_samples_per_second": 35.156,
+      "eval_steps_per_second": 2.197,
       "step": 150
     },
     {
-      "epoch": 0.16806722689075632,
-      "grad_norm": 0.6111018657684326,
-      "learning_rate": 4.719887955182073e-05,
-      "loss": 0.3271,
+      "epoch": 0.16859852476290832,
+      "grad_norm": 0.5819964408874512,
+      "learning_rate": 4.7190024587284866e-05,
+      "loss": 0.2992,
       "step": 160
     },
     {
-      "epoch": 0.17857142857142858,
-      "grad_norm": 0.7422594428062439,
-      "learning_rate": 4.7023809523809525e-05,
-      "loss": 0.315,
+      "epoch": 0.1791359325605901,
+      "grad_norm": 0.689468264579773,
+      "learning_rate": 4.7014401123990165e-05,
+      "loss": 0.3168,
       "step": 170
     },
     {
-      "epoch": 0.18907563025210083,
-      "grad_norm": 0.7226534485816956,
-      "learning_rate": 4.684873949579832e-05,
-      "loss": 0.3031,
+      "epoch": 0.18967334035827185,
+      "grad_norm": 0.6950872540473938,
+      "learning_rate": 4.683877766069547e-05,
+      "loss": 0.3041,
       "step": 180
     },
     {
-      "epoch": 0.19957983193277312,
-      "grad_norm": 0.6302976012229919,
-      "learning_rate": 4.667366946778712e-05,
-      "loss": 0.3161,
+      "epoch": 0.20021074815595363,
+      "grad_norm": 0.8322122097015381,
+      "learning_rate": 4.6663154197400776e-05,
+      "loss": 0.3028,
       "step": 190
     },
     {
-      "epoch": 0.21008403361344538,
-      "grad_norm": 0.6225076913833618,
-      "learning_rate": 4.6498599439775914e-05,
-      "loss": 0.3038,
+      "epoch": 0.2107481559536354,
+      "grad_norm": 0.5850774645805359,
+      "learning_rate": 4.6487530734106075e-05,
+      "loss": 0.2992,
       "step": 200
     },
     {
-      "epoch": 0.21008403361344538,
-      "eval_loss": 0.35061606764793396,
-      "eval_runtime": 13.5616,
-      "eval_samples_per_second": 35.394,
-      "eval_steps_per_second": 2.212,
+      "epoch": 0.2107481559536354,
+      "eval_loss": 0.36230018734931946,
+      "eval_runtime": 13.6165,
+      "eval_samples_per_second": 35.251,
+      "eval_steps_per_second": 2.203,
       "step": 200
     },
     {
-      "epoch": 0.22058823529411764,
-      "grad_norm": 0.6001319885253906,
-      "learning_rate": 4.632352941176471e-05,
-      "loss": 0.3129,
+      "epoch": 0.22128556375131717,
+      "grad_norm": 0.6668715476989746,
+      "learning_rate": 4.631190727081138e-05,
+      "loss": 0.2924,
       "step": 210
     },
     {
-      "epoch": 0.23109243697478993,
-      "grad_norm": 0.5385990142822266,
-      "learning_rate": 4.61484593837535e-05,
-      "loss": 0.2991,
+      "epoch": 0.23182297154899895,
+      "grad_norm": 0.4749641418457031,
+      "learning_rate": 4.6136283807516686e-05,
+      "loss": 0.3017,
       "step": 220
     },
     {
-      "epoch": 0.2415966386554622,
-      "grad_norm": 0.4513624906539917,
-      "learning_rate": 4.59733893557423e-05,
-      "loss": 0.2896,
+      "epoch": 0.24236037934668073,
+      "grad_norm": 0.6381515860557556,
+      "learning_rate": 4.596066034422199e-05,
+      "loss": 0.2887,
       "step": 230
     },
     {
-      "epoch": 0.25210084033613445,
-      "grad_norm": 0.6142160892486572,
-      "learning_rate": 4.579831932773109e-05,
-      "loss": 0.3059,
+      "epoch": 0.2528977871443625,
+      "grad_norm": 0.49952977895736694,
+      "learning_rate": 4.57850368809273e-05,
+      "loss": 0.2833,
       "step": 240
     },
     {
-      "epoch": 0.26260504201680673,
-      "grad_norm": 0.6714802384376526,
-      "learning_rate": 4.562324929971989e-05,
-      "loss": 0.2897,
+      "epoch": 0.26343519494204426,
+      "grad_norm": 0.699518084526062,
+      "learning_rate": 4.5609413417632596e-05,
+      "loss": 0.2859,
       "step": 250
     },
     {
-      "epoch": 0.26260504201680673,
-      "eval_loss": 0.3456435203552246,
-      "eval_runtime": 13.5552,
-      "eval_samples_per_second": 35.411,
-      "eval_steps_per_second": 2.213,
+      "epoch": 0.26343519494204426,
+      "eval_loss": 0.3563433885574341,
+      "eval_runtime": 13.6275,
+      "eval_samples_per_second": 35.223,
+      "eval_steps_per_second": 2.201,
       "step": 250
     },
     {
-      "epoch": 0.27310924369747897,
-      "grad_norm": 0.6518235206604004,
-      "learning_rate": 4.5448179271708687e-05,
-      "loss": 0.312,
+      "epoch": 0.273972602739726,
+      "grad_norm": 0.5912085175514221,
+      "learning_rate": 4.54337899543379e-05,
+      "loss": 0.302,
       "step": 260
     },
     {
-      "epoch": 0.28361344537815125,
-      "grad_norm": 0.6250632405281067,
-      "learning_rate": 4.527310924369748e-05,
-      "loss": 0.2959,
+      "epoch": 0.2845100105374078,
+      "grad_norm": 0.6353363990783691,
+      "learning_rate": 4.525816649104321e-05,
+      "loss": 0.3112,
       "step": 270
     },
     {
-      "epoch": 0.29411764705882354,
-      "grad_norm": 0.5683826804161072,
-      "learning_rate": 4.5098039215686275e-05,
-      "loss": 0.3027,
+      "epoch": 0.2950474183350896,
+      "grad_norm": 0.5483567118644714,
+      "learning_rate": 4.5082543027748506e-05,
+      "loss": 0.2808,
       "step": 280
     },
     {
-      "epoch": 0.30462184873949577,
-      "grad_norm": 0.560312807559967,
-      "learning_rate": 4.4922969187675076e-05,
-      "loss": 0.3002,
+      "epoch": 0.3055848261327713,
+      "grad_norm": 0.5003193616867065,
+      "learning_rate": 4.490691956445381e-05,
+      "loss": 0.3185,
       "step": 290
     },
     {
-      "epoch": 0.31512605042016806,
-      "grad_norm": 0.66291743516922,
-      "learning_rate": 4.474789915966387e-05,
-      "loss": 0.2925,
+      "epoch": 0.31612223393045313,
+      "grad_norm": 0.4919240176677704,
+      "learning_rate": 4.473129610115912e-05,
+      "loss": 0.2883,
       "step": 300
     },
     {
-      "epoch": 0.31512605042016806,
-      "eval_loss": 0.3431606888771057,
-      "eval_runtime": 13.5629,
-      "eval_samples_per_second": 35.391,
-      "eval_steps_per_second": 2.212,
+      "epoch": 0.31612223393045313,
+      "eval_loss": 0.35315924882888794,
+      "eval_runtime": 13.6133,
+      "eval_samples_per_second": 35.26,
+      "eval_steps_per_second": 2.204,
       "step": 300
     },
     {
-      "epoch": 0.32563025210084034,
-      "grad_norm": 0.6478439569473267,
-      "learning_rate": 4.4572829131652665e-05,
-      "loss": 0.2893,
+      "epoch": 0.3266596417281349,
+      "grad_norm": 0.5145038366317749,
+      "learning_rate": 4.455567263786442e-05,
+      "loss": 0.3084,
       "step": 310
     },
     {
-      "epoch": 0.33613445378151263,
-      "grad_norm": 0.5832348465919495,
-      "learning_rate": 4.439775910364146e-05,
-      "loss": 0.2842,
+      "epoch": 0.33719704952581664,
+      "grad_norm": 0.6343855261802673,
+      "learning_rate": 4.438004917456973e-05,
+      "loss": 0.295,
       "step": 320
     },
     {
-      "epoch": 0.34663865546218486,
-      "grad_norm": 0.525932252407074,
-      "learning_rate": 4.422268907563025e-05,
-      "loss": 0.2837,
+      "epoch": 0.34773445732349845,
+      "grad_norm": 0.5336400270462036,
+      "learning_rate": 4.420442571127503e-05,
+      "loss": 0.2882,
       "step": 330
     },
     {
-      "epoch": 0.35714285714285715,
-      "grad_norm": 0.5487508177757263,
-      "learning_rate": 4.404761904761905e-05,
-      "loss": 0.2706,
+      "epoch": 0.3582718651211802,
+      "grad_norm": 0.49547308683395386,
+      "learning_rate": 4.4028802247980333e-05,
+      "loss": 0.3084,
       "step": 340
     },
     {
-      "epoch": 0.36764705882352944,
-      "grad_norm": 0.5392388701438904,
-      "learning_rate": 4.387254901960784e-05,
-      "loss": 0.2835,
+      "epoch": 0.36880927291886195,
+      "grad_norm": 0.5037292242050171,
+      "learning_rate": 4.385317878468563e-05,
+      "loss": 0.3084,
       "step": 350
     },
     {
-      "epoch": 0.36764705882352944,
-      "eval_loss": 0.33528536558151245,
-      "eval_runtime": 13.5508,
-      "eval_samples_per_second": 35.422,
-      "eval_steps_per_second": 2.214,
+      "epoch": 0.36880927291886195,
+      "eval_loss": 0.34608179330825806,
+      "eval_runtime": 13.6537,
+      "eval_samples_per_second": 35.155,
+      "eval_steps_per_second": 2.197,
       "step": 350
     },
     {
-      "epoch": 0.37815126050420167,
-      "grad_norm": 0.6706260442733765,
-      "learning_rate": 4.369747899159664e-05,
-      "loss": 0.2844,
+      "epoch": 0.3793466807165437,
+      "grad_norm": 0.5956543684005737,
+      "learning_rate": 4.367755532139094e-05,
+      "loss": 0.2763,
       "step": 360
     },
     {
-      "epoch": 0.38865546218487396,
-      "grad_norm": 0.6042625904083252,
-      "learning_rate": 4.352240896358544e-05,
-      "loss": 0.2758,
+      "epoch": 0.3898840885142255,
+      "grad_norm": 0.6263634562492371,
+      "learning_rate": 4.3501931858096244e-05,
+      "loss": 0.3025,
       "step": 370
     },
     {
-      "epoch": 0.39915966386554624,
-      "grad_norm": 0.534008264541626,
-      "learning_rate": 4.334733893557423e-05,
-      "loss": 0.2918,
+      "epoch": 0.40042149631190727,
+      "grad_norm": 0.4832920730113983,
+      "learning_rate": 4.332630839480154e-05,
+      "loss": 0.2792,
       "step": 380
     },
     {
-      "epoch": 0.4096638655462185,
-      "grad_norm": 0.48162588477134705,
-      "learning_rate": 4.317226890756303e-05,
-      "loss": 0.273,
+      "epoch": 0.410958904109589,
+      "grad_norm": 0.4969714879989624,
+      "learning_rate": 4.3150684931506855e-05,
+      "loss": 0.2704,
       "step": 390
     },
     {
-      "epoch": 0.42016806722689076,
-      "grad_norm": 0.5669644474983215,
-      "learning_rate": 4.2997198879551826e-05,
-      "loss": 0.285,
+      "epoch": 0.4214963119072708,
+      "grad_norm": 0.568900465965271,
+      "learning_rate": 4.297506146821216e-05,
+      "loss": 0.2734,
       "step": 400
     },
     {
-      "epoch": 0.42016806722689076,
-      "eval_loss": 0.3348632752895355,
-      "eval_runtime": 13.5507,
-      "eval_samples_per_second": 35.423,
-      "eval_steps_per_second": 2.214,
+      "epoch": 0.4214963119072708,
+      "eval_loss": 0.34151414036750793,
+      "eval_runtime": 13.664,
+      "eval_samples_per_second": 35.129,
+      "eval_steps_per_second": 2.196,
       "step": 400
     },
     {
-      "epoch": 0.43067226890756305,
-      "grad_norm": 0.6257824897766113,
-      "learning_rate": 4.2822128851540614e-05,
-      "loss": 0.299,
+      "epoch": 0.4320337197049526,
+      "grad_norm": 0.5023282170295715,
+      "learning_rate": 4.279943800491746e-05,
+      "loss": 0.2845,
       "step": 410
     },
     {
-      "epoch": 0.4411764705882353,
-      "grad_norm": 0.5430576205253601,
-      "learning_rate": 4.2647058823529415e-05,
-      "loss": 0.2868,
+      "epoch": 0.44257112750263433,
+      "grad_norm": 0.531538188457489,
+      "learning_rate": 4.2623814541622765e-05,
+      "loss": 0.276,
       "step": 420
     },
     {
-      "epoch": 0.45168067226890757,
-      "grad_norm": 0.5633955597877502,
-      "learning_rate": 4.247198879551821e-05,
-      "loss": 0.2589,
+      "epoch": 0.45310853530031614,
+      "grad_norm": 0.6686979532241821,
+      "learning_rate": 4.2448191078328064e-05,
+      "loss": 0.283,
       "step": 430
     },
     {
-      "epoch": 0.46218487394957986,
-      "grad_norm": 0.5294789671897888,
-      "learning_rate": 4.2296918767507e-05,
-      "loss": 0.2777,
+      "epoch": 0.4636459430979979,
+      "grad_norm": 0.5194190144538879,
+      "learning_rate": 4.227256761503337e-05,
+      "loss": 0.2636,
       "step": 440
     },
     {
-      "epoch": 0.4726890756302521,
-      "grad_norm": 0.5480856895446777,
-      "learning_rate": 4.21218487394958e-05,
-      "loss": 0.2704,
+      "epoch": 0.47418335089567965,
+      "grad_norm": 0.6105541586875916,
+      "learning_rate": 4.2096944151738675e-05,
+      "loss": 0.2713,
       "step": 450
     },
     {
-      "epoch": 0.4726890756302521,
-      "eval_loss": 0.329515278339386,
-      "eval_runtime": 13.5423,
-      "eval_samples_per_second": 35.445,
-      "eval_steps_per_second": 2.215,
+      "epoch": 0.47418335089567965,
+      "eval_loss": 0.33879777789115906,
+      "eval_runtime": 13.6666,
+      "eval_samples_per_second": 35.122,
+      "eval_steps_per_second": 2.195,
       "step": 450
     },
     {
-      "epoch": 0.4831932773109244,
-      "grad_norm": 0.5051332116127014,
-      "learning_rate": 4.19467787114846e-05,
-      "loss": 0.2438,
+      "epoch": 0.48472075869336145,
+      "grad_norm": 0.4929138123989105,
+      "learning_rate": 4.1921320688443974e-05,
+      "loss": 0.261,
       "step": 460
     },
     {
-      "epoch": 0.49369747899159666,
-      "grad_norm": 0.6251511573791504,
-      "learning_rate": 4.177170868347339e-05,
-      "loss": 0.2748,
+      "epoch": 0.4952581664910432,
+      "grad_norm": 0.584095299243927,
+      "learning_rate": 4.174569722514928e-05,
+      "loss": 0.2714,
       "step": 470
     },
     {
-      "epoch": 0.5042016806722689,
-      "grad_norm": 0.4729413092136383,
-      "learning_rate": 4.159663865546219e-05,
-      "loss": 0.2689,
+      "epoch": 0.505795574288725,
+      "grad_norm": 0.5386167764663696,
+      "learning_rate": 4.1570073761854585e-05,
+      "loss": 0.2855,
       "step": 480
     },
     {
-      "epoch": 0.5147058823529411,
-      "grad_norm": 0.5220003724098206,
-      "learning_rate": 4.142156862745099e-05,
-      "loss": 0.2899,
+      "epoch": 0.5163329820864068,
+      "grad_norm": 0.5819774270057678,
+      "learning_rate": 4.139445029855989e-05,
+      "loss": 0.2911,
       "step": 490
     },
     {
-      "epoch": 0.5252100840336135,
-      "grad_norm": 0.54283207654953,
-      "learning_rate": 4.1246498599439776e-05,
-      "loss": 0.272,
+      "epoch": 0.5268703898840885,
+      "grad_norm": 0.5875944495201111,
+      "learning_rate": 4.12188268352652e-05,
+      "loss": 0.2619,
       "step": 500
     },
     {
-      "epoch": 0.5252100840336135,
-      "eval_loss": 0.32714489102363586,
-      "eval_runtime": 13.5497,
-      "eval_samples_per_second": 35.425,
-      "eval_steps_per_second": 2.214,
+      "epoch": 0.5268703898840885,
+      "eval_loss": 0.3364439010620117,
+      "eval_runtime": 13.6731,
+      "eval_samples_per_second": 35.105,
+      "eval_steps_per_second": 2.194,
       "step": 500
     },
     {
-      "epoch": 0.5357142857142857,
-      "grad_norm": 0.5851682424545288,
-      "learning_rate": 4.107142857142857e-05,
-      "loss": 0.2691,
+      "epoch": 0.5374077976817703,
+      "grad_norm": 0.6705812811851501,
+      "learning_rate": 4.1043203371970496e-05,
+      "loss": 0.2792,
       "step": 510
     },
     {
-      "epoch": 0.5462184873949579,
-      "grad_norm": 0.6026607751846313,
-      "learning_rate": 4.089635854341737e-05,
-      "loss": 0.2716,
+      "epoch": 0.547945205479452,
+      "grad_norm": 0.5975854396820068,
+      "learning_rate": 4.08675799086758e-05,
+      "loss": 0.2611,
       "step": 520
     },
     {
-      "epoch": 0.5567226890756303,
-      "grad_norm": 0.522422730922699,
-      "learning_rate": 4.0721288515406165e-05,
-      "loss": 0.2774,
+      "epoch": 0.5584826132771338,
+      "grad_norm": 0.5905255675315857,
+      "learning_rate": 4.069195644538111e-05,
+      "loss": 0.2688,
       "step": 530
     },
     {
-      "epoch": 0.5672268907563025,
-      "grad_norm": 0.516901433467865,
-      "learning_rate": 4.054621848739496e-05,
-      "loss": 0.2726,
+      "epoch": 0.5690200210748156,
+      "grad_norm": 0.6222755312919617,
+      "learning_rate": 4.0516332982086406e-05,
+      "loss": 0.2968,
       "step": 540
     },
     {
-      "epoch": 0.5777310924369747,
-      "grad_norm": 0.667030394077301,
-      "learning_rate": 4.0371148459383754e-05,
-      "loss": 0.2622,
+      "epoch": 0.5795574288724974,
+      "grad_norm": 0.5604814291000366,
+      "learning_rate": 4.034070951879171e-05,
+      "loss": 0.2623,
       "step": 550
     },
     {
-      "epoch": 0.5777310924369747,
-      "eval_loss": 0.3262839615345001,
-      "eval_runtime": 13.5448,
-      "eval_samples_per_second": 35.438,
-      "eval_steps_per_second": 2.215,
+      "epoch": 0.5795574288724974,
+      "eval_loss": 0.3329848647117615,
+      "eval_runtime": 13.6573,
+      "eval_samples_per_second": 35.146,
+      "eval_steps_per_second": 2.197,
       "step": 550
     },
     {
-      "epoch": 0.5882352941176471,
-      "grad_norm": 0.542658269405365,
-      "learning_rate": 4.0196078431372555e-05,
-      "loss": 0.2572,
+      "epoch": 0.5900948366701791,
+      "grad_norm": 0.552557647228241,
+      "learning_rate": 4.016508605549702e-05,
+      "loss": 0.2579,
       "step": 560
     },
     {
-      "epoch": 0.5987394957983193,
-      "grad_norm": 0.5408573746681213,
-      "learning_rate": 4.002100840336135e-05,
-      "loss": 0.2636,
+      "epoch": 0.6006322444678609,
+      "grad_norm": 0.7151882648468018,
+      "learning_rate": 3.998946259220232e-05,
+      "loss": 0.263,
       "step": 570
     },
     {
-      "epoch": 0.6092436974789915,
-      "grad_norm": 0.5691037774085999,
-      "learning_rate": 3.984593837535014e-05,
-      "loss": 0.268,
+      "epoch": 0.6111696522655427,
+      "grad_norm": 0.7065399289131165,
+      "learning_rate": 3.981383912890763e-05,
+      "loss": 0.2626,
       "step": 580
     },
     {
-      "epoch": 0.6197478991596639,
-      "grad_norm": 0.5530794858932495,
-      "learning_rate": 3.967086834733894e-05,
-      "loss": 0.2583,
+      "epoch": 0.6217070600632244,
+      "grad_norm": 0.5590985417366028,
+      "learning_rate": 3.963821566561293e-05,
+      "loss": 0.2614,
       "step": 590
     },
     {
-      "epoch": 0.6302521008403361,
-      "grad_norm": 0.546229362487793,
-      "learning_rate": 3.949579831932773e-05,
-      "loss": 0.2622,
+      "epoch": 0.6322444678609063,
+      "grad_norm": 0.6018164157867432,
+      "learning_rate": 3.946259220231823e-05,
+      "loss": 0.2547,
       "step": 600
     },
     {
-      "epoch": 0.6302521008403361,
-      "eval_loss": 0.3219989836215973,
-      "eval_runtime": 13.5524,
-      "eval_samples_per_second": 35.418,
-      "eval_steps_per_second": 2.214,
+      "epoch": 0.6322444678609063,
+      "eval_loss": 0.3322373926639557,
+      "eval_runtime": 13.6386,
+      "eval_samples_per_second": 35.194,
+      "eval_steps_per_second": 2.2,
       "step": 600
     },
     {
-      "epoch": 0.6407563025210085,
-      "grad_norm": 0.5098925232887268,
-      "learning_rate": 3.9320728291316526e-05,
-      "loss": 0.2553,
+      "epoch": 0.642781875658588,
+      "grad_norm": 0.6659051179885864,
+      "learning_rate": 3.928696873902353e-05,
+      "loss": 0.2768,
       "step": 610
     },
     {
-      "epoch": 0.6512605042016807,
-      "grad_norm": 0.5201871991157532,
-      "learning_rate": 3.914565826330533e-05,
-      "loss": 0.2584,
+      "epoch": 0.6533192834562698,
+      "grad_norm": 0.6509780883789062,
+      "learning_rate": 3.911134527572884e-05,
+      "loss": 0.272,
       "step": 620
     },
     {
-      "epoch": 0.6617647058823529,
-      "grad_norm": 0.47408100962638855,
-      "learning_rate": 3.897058823529412e-05,
-      "loss": 0.2686,
+      "epoch": 0.6638566912539515,
+      "grad_norm": 0.5752527713775635,
+      "learning_rate": 3.893572181243414e-05,
+      "loss": 0.2625,
       "step": 630
     },
     {
-      "epoch": 0.6722689075630253,
-      "grad_norm": 0.5591098666191101,
-      "learning_rate": 3.8795518207282915e-05,
-      "loss": 0.2772,
+      "epoch": 0.6743940990516333,
+      "grad_norm": 0.4737485349178314,
+      "learning_rate": 3.876009834913945e-05,
+      "loss": 0.2668,
       "step": 640
     },
     {
-      "epoch": 0.6827731092436975,
-      "grad_norm": 0.5344163179397583,
-      "learning_rate": 3.862044817927171e-05,
-      "loss": 0.263,
+      "epoch": 0.684931506849315,
+      "grad_norm": 0.548322319984436,
+      "learning_rate": 3.8584474885844754e-05,
+      "loss": 0.2597,
       "step": 650
     },
     {
-      "epoch": 0.6827731092436975,
-      "eval_loss": 0.31990015506744385,
-      "eval_runtime": 13.5501,
-      "eval_samples_per_second": 35.424,
-      "eval_steps_per_second": 2.214,
+      "epoch": 0.684931506849315,
+      "eval_loss": 0.3307412266731262,
+      "eval_runtime": 13.6407,
+      "eval_samples_per_second": 35.189,
+      "eval_steps_per_second": 2.199,
       "step": 650
     },
     {
-      "epoch": 0.6932773109243697,
-      "grad_norm": 0.6538853049278259,
-      "learning_rate": 3.844537815126051e-05,
-      "loss": 0.2605,
+      "epoch": 0.6954689146469969,
+      "grad_norm": 0.5635099411010742,
+      "learning_rate": 3.840885142255005e-05,
+      "loss": 0.2453,
       "step": 660
     },
     {
-      "epoch": 0.7037815126050421,
-      "grad_norm": 0.43679994344711304,
-      "learning_rate": 3.82703081232493e-05,
-      "loss": 0.2486,
+      "epoch": 0.7060063224446786,
+      "grad_norm": 0.6012313365936279,
+      "learning_rate": 3.823322795925536e-05,
+      "loss": 0.2695,
       "step": 670
     },
     {
-      "epoch": 0.7142857142857143,
-      "grad_norm": 0.5602915287017822,
-      "learning_rate": 3.809523809523809e-05,
-      "loss": 0.2359,
+      "epoch": 0.7165437302423604,
+      "grad_norm": 0.6025511026382446,
+      "learning_rate": 3.8057604495960664e-05,
+      "loss": 0.2599,
       "step": 680
     },
     {
-      "epoch": 0.7247899159663865,
-      "grad_norm": 0.5354353189468384,
-      "learning_rate": 3.792016806722689e-05,
-      "loss": 0.2541,
+      "epoch": 0.7270811380400422,
+      "grad_norm": 0.5996225476264954,
+      "learning_rate": 3.788198103266596e-05,
+      "loss": 0.2624,
       "step": 690
     },
     {
-      "epoch": 0.7352941176470589,
-      "grad_norm": 0.5954485535621643,
-      "learning_rate": 3.774509803921569e-05,
-      "loss": 0.2649,
+      "epoch": 0.7376185458377239,
+      "grad_norm": 0.5269689559936523,
+      "learning_rate": 3.770635756937127e-05,
+      "loss": 0.2498,
       "step": 700
     },
     {
-      "epoch": 0.7352941176470589,
-      "eval_loss": 0.3204093873500824,
-      "eval_runtime": 13.5523,
-      "eval_samples_per_second": 35.418,
-      "eval_steps_per_second": 2.214,
+      "epoch": 0.7376185458377239,
+      "eval_loss": 0.3312172591686249,
+      "eval_runtime": 13.6169,
+      "eval_samples_per_second": 35.25,
+      "eval_steps_per_second": 2.203,
       "step": 700
     },
     {
-      "epoch": 0.7457983193277311,
-      "grad_norm": 0.6093761324882507,
-      "learning_rate": 3.757002801120448e-05,
-      "loss": 0.2456,
+      "epoch": 0.7481559536354057,
+      "grad_norm": 0.6115732192993164,
+      "learning_rate": 3.7530734106076575e-05,
+      "loss": 0.2566,
       "step": 710
     },
     {
-      "epoch": 0.7563025210084033,
-      "grad_norm": 0.4796586334705353,
-      "learning_rate": 3.739495798319328e-05,
-      "loss": 0.253,
+      "epoch": 0.7586933614330874,
+      "grad_norm": 0.4686366319656372,
+      "learning_rate": 3.7355110642781874e-05,
+      "loss": 0.2325,
       "step": 720
     },
     {
-      "epoch": 0.7668067226890757,
-      "grad_norm": 0.5846813917160034,
-      "learning_rate": 3.721988795518208e-05,
-      "loss": 0.2442,
+      "epoch": 0.7692307692307693,
+      "grad_norm": 0.5317673087120056,
+      "learning_rate": 3.717948717948718e-05,
+      "loss": 0.2429,
       "step": 730
     },
     {
-      "epoch": 0.7773109243697479,
-      "grad_norm": 0.4811939597129822,
-      "learning_rate": 3.704481792717087e-05,
-      "loss": 0.2522,
+      "epoch": 0.779768177028451,
+      "grad_norm": 0.5183236598968506,
+      "learning_rate": 3.7003863716192485e-05,
+      "loss": 0.2562,
       "step": 740
     },
     {
-      "epoch": 0.7878151260504201,
-      "grad_norm": 0.5718042850494385,
-      "learning_rate": 3.6869747899159665e-05,
-      "loss": 0.2562,
+      "epoch": 0.7903055848261328,
+      "grad_norm": 0.5319241881370544,
+      "learning_rate": 3.682824025289779e-05,
+      "loss": 0.2356,
       "step": 750
     },
     {
-      "epoch": 0.7878151260504201,
-      "eval_loss": 0.32099905610084534,
-      "eval_runtime": 13.5408,
-      "eval_samples_per_second": 35.448,
-      "eval_steps_per_second": 2.216,
+      "epoch": 0.7903055848261328,
+      "eval_loss": 0.3285733163356781,
+      "eval_runtime": 13.6213,
+      "eval_samples_per_second": 35.239,
+      "eval_steps_per_second": 2.202,
       "step": 750
     },
     {
-      "epoch": 0.7983193277310925,
-      "grad_norm": 0.5630698204040527,
-      "learning_rate": 3.669467787114846e-05,
-      "loss": 0.2476,
+      "epoch": 0.8008429926238145,
+      "grad_norm": 0.5814469456672668,
+      "learning_rate": 3.6652616789603096e-05,
+      "loss": 0.2428,
       "step": 760
     },
     {
-      "epoch": 0.8088235294117647,
-      "grad_norm": 0.6513442397117615,
-      "learning_rate": 3.6519607843137254e-05,
-      "loss": 0.2544,
+      "epoch": 0.8113804004214963,
+      "grad_norm": 0.5138055086135864,
+      "learning_rate": 3.6476993326308395e-05,
+      "loss": 0.2639,
       "step": 770
     },
     {
-      "epoch": 0.819327731092437,
-      "grad_norm": 0.6139647960662842,
-      "learning_rate": 3.634453781512605e-05,
-      "loss": 0.258,
+      "epoch": 0.821917808219178,
+      "grad_norm": 0.5066888332366943,
+      "learning_rate": 3.63013698630137e-05,
+      "loss": 0.2373,
       "step": 780
     },
     {
-      "epoch": 0.8298319327731093,
-      "grad_norm": 0.5916554927825928,
-      "learning_rate": 3.616946778711485e-05,
-      "loss": 0.2415,
+      "epoch": 0.8324552160168599,
+      "grad_norm": 0.513317883014679,
+      "learning_rate": 3.6125746399719e-05,
+      "loss": 0.2462,
       "step": 790
     },
     {
-      "epoch": 0.8403361344537815,
-      "grad_norm": 0.5163634419441223,
-      "learning_rate": 3.5994397759103643e-05,
-      "loss": 0.252,
+      "epoch": 0.8429926238145417,
+      "grad_norm": 0.5878031253814697,
+      "learning_rate": 3.5950122936424305e-05,
+      "loss": 0.2397,
       "step": 800
     },
     {
-      "epoch": 0.8403361344537815,
-      "eval_loss": 0.3215568959712982,
-      "eval_runtime": 13.5293,
-      "eval_samples_per_second": 35.479,
-      "eval_steps_per_second": 2.217,
+      "epoch": 0.8429926238145417,
+      "eval_loss": 0.3288480043411255,
+      "eval_runtime": 13.6316,
+      "eval_samples_per_second": 35.212,
+      "eval_steps_per_second": 2.201,
       "step": 800
     },
     {
-      "epoch": 0.8508403361344538,
-      "grad_norm": 0.5768859386444092,
-      "learning_rate": 3.581932773109244e-05,
-      "loss": 0.2421,
+      "epoch": 0.8535300316122234,
+      "grad_norm": 0.6269211769104004,
+      "learning_rate": 3.577449947312961e-05,
+      "loss": 0.2418,
       "step": 810
     },
     {
-      "epoch": 0.8613445378151261,
-      "grad_norm": 0.6197952032089233,
-      "learning_rate": 3.564425770308123e-05,
-      "loss": 0.2672,
+      "epoch": 0.8640674394099052,
+      "grad_norm": 0.5279048085212708,
+      "learning_rate": 3.5598876009834916e-05,
+      "loss": 0.246,
       "step": 820
     },
     {
-      "epoch": 0.8718487394957983,
-      "grad_norm": 0.5396980047225952,
-      "learning_rate": 3.546918767507003e-05,
-      "loss": 0.2393,
+      "epoch": 0.8746048472075869,
+      "grad_norm": 0.5837969183921814,
+      "learning_rate": 3.542325254654022e-05,
+      "loss": 0.256,
       "step": 830
     },
     {
-      "epoch": 0.8823529411764706,
-      "grad_norm": 0.5783377885818481,
-      "learning_rate": 3.529411764705883e-05,
-      "loss": 0.2374,
+      "epoch": 0.8851422550052687,
+      "grad_norm": 0.5384166240692139,
+      "learning_rate": 3.524762908324553e-05,
+      "loss": 0.2412,
       "step": 840
     },
     {
-      "epoch": 0.8928571428571429,
-      "grad_norm": 0.5808666944503784,
-      "learning_rate": 3.511904761904762e-05,
-      "loss": 0.2405,
+      "epoch": 0.8956796628029505,
+      "grad_norm": 0.5514136552810669,
+      "learning_rate": 3.507200561995083e-05,
+      "loss": 0.251,
       "step": 850
     },
     {
-      "epoch": 0.8928571428571429,
-      "eval_loss": 0.3207303583621979,
-      "eval_runtime": 13.5299,
-      "eval_samples_per_second": 35.477,
-      "eval_steps_per_second": 2.217,
+      "epoch": 0.8956796628029505,
+      "eval_loss": 0.3275190591812134,
+      "eval_runtime": 13.654,
+      "eval_samples_per_second": 35.155,
+      "eval_steps_per_second": 2.197,
       "step": 850
     },
     {
-      "epoch": 0.9033613445378151,
-      "grad_norm": 0.4931146204471588,
-      "learning_rate": 3.4943977591036416e-05,
-      "loss": 0.2382,
+      "epoch": 0.9062170706006323,
+      "grad_norm": 0.5456790328025818,
+      "learning_rate": 3.489638215665613e-05,
+      "loss": 0.2337,
       "step": 860
     },
     {
-      "epoch": 0.9138655462184874,
-      "grad_norm": 0.6456460952758789,
-      "learning_rate": 3.476890756302521e-05,
-      "loss": 0.2416,
+      "epoch": 0.916754478398314,
+      "grad_norm": 0.5794851779937744,
+      "learning_rate": 3.472075869336143e-05,
+      "loss": 0.2462,
       "step": 870
     },
     {
-      "epoch": 0.9243697478991597,
-      "grad_norm": 0.5459381937980652,
-      "learning_rate": 3.4593837535014004e-05,
-      "loss": 0.2611,
+      "epoch": 0.9272918861959958,
+      "grad_norm": 0.5296761393547058,
+      "learning_rate": 3.454513523006674e-05,
+      "loss": 0.2357,
       "step": 880
     },
     {
-      "epoch": 0.9348739495798319,
-      "grad_norm": 0.5317162275314331,
-      "learning_rate": 3.4418767507002805e-05,
-      "loss": 0.2383,
+      "epoch": 0.9378292939936775,
+      "grad_norm": 0.5668390989303589,
+      "learning_rate": 3.436951176677204e-05,
+      "loss": 0.2439,
       "step": 890
     },
     {
-      "epoch": 0.9453781512605042,
-      "grad_norm": 0.5790566205978394,
-      "learning_rate": 3.42436974789916e-05,
-      "loss": 0.2455,
+      "epoch": 0.9483667017913593,
+      "grad_norm": 0.5026710629463196,
+      "learning_rate": 3.419388830347735e-05,
+      "loss": 0.2301,
       "step": 900
     },
     {
-      "epoch": 0.9453781512605042,
-      "eval_loss": 0.31993839144706726,
-      "eval_runtime": 13.5323,
-      "eval_samples_per_second": 35.471,
-      "eval_steps_per_second": 2.217,
+      "epoch": 0.9483667017913593,
+      "eval_loss": 0.32623717188835144,
+      "eval_runtime": 13.6644,
+      "eval_samples_per_second": 35.128,
+      "eval_steps_per_second": 2.195,
       "step": 900
     },
     {
-      "epoch": 0.9558823529411765,
-      "grad_norm": 0.5805277228355408,
-      "learning_rate": 3.4068627450980394e-05,
-      "loss": 0.2393,
+      "epoch": 0.958904109589041,
+      "grad_norm": 0.5280706286430359,
+      "learning_rate": 3.4018264840182654e-05,
+      "loss": 0.238,
       "step": 910
     },
     {
-      "epoch": 0.9663865546218487,
-      "grad_norm": 0.6085871458053589,
-      "learning_rate": 3.389355742296919e-05,
-      "loss": 0.2505,
+      "epoch": 0.9694415173867229,
+      "grad_norm": 0.5578323006629944,
+      "learning_rate": 3.384264137688795e-05,
+      "loss": 0.2484,
       "step": 920
     },
     {
-      "epoch": 0.976890756302521,
-      "grad_norm": 0.6251375079154968,
-      "learning_rate": 3.371848739495799e-05,
-      "loss": 0.2378,
+      "epoch": 0.9799789251844047,
+      "grad_norm": 0.5756859183311462,
+      "learning_rate": 3.366701791359326e-05,
+      "loss": 0.2539,
       "step": 930
     },
     {
-      "epoch": 0.9873949579831933,
-      "grad_norm": 0.6169071197509766,
-      "learning_rate": 3.3543417366946776e-05,
-      "loss": 0.239,
+      "epoch": 0.9905163329820864,
+      "grad_norm": 0.6118686199188232,
+      "learning_rate": 3.3491394450298564e-05,
+      "loss": 0.2401,
       "step": 940
     },
     {
-      "epoch": 0.9978991596638656,
-      "grad_norm": 0.5361204147338867,
-      "learning_rate": 3.336834733893557e-05,
-      "loss": 0.2484,
+      "epoch": 1.0010537407797682,
+      "grad_norm": 0.43512389063835144,
+      "learning_rate": 3.331577098700386e-05,
+      "loss": 0.2334,
       "step": 950
     },
     {
-      "epoch": 0.9978991596638656,
-      "eval_loss": 0.31845951080322266,
-      "eval_runtime": 13.5483,
-      "eval_samples_per_second": 35.429,
-      "eval_steps_per_second": 2.214,
+      "epoch": 1.0010537407797682,
+      "eval_loss": 0.32700982689857483,
+      "eval_runtime": 13.6561,
+      "eval_samples_per_second": 35.149,
+      "eval_steps_per_second": 2.197,
       "step": 950
     },
     {
-      "epoch": 1.0084033613445378,
-      "grad_norm": 0.47690704464912415,
-      "learning_rate": 3.319327731092437e-05,
-      "loss": 0.2006,
+      "epoch": 1.01159114857745,
+      "grad_norm": 0.5436535477638245,
+      "learning_rate": 3.314014752370917e-05,
+      "loss": 0.1933,
       "step": 960
     },
     {
-      "epoch": 1.01890756302521,
-      "grad_norm": 0.7460448741912842,
-      "learning_rate": 3.3018207282913166e-05,
-      "loss": 0.1983,
+      "epoch": 1.0221285563751317,
+      "grad_norm": 0.5048521161079407,
+      "learning_rate": 3.2964524060414474e-05,
+      "loss": 0.2088,
       "step": 970
     },
     {
-      "epoch": 1.0294117647058822,
-      "grad_norm": 0.5729458332061768,
-      "learning_rate": 3.284313725490196e-05,
-      "loss": 0.1989,
+      "epoch": 1.0326659641728135,
+      "grad_norm": 0.6231564283370972,
+      "learning_rate": 3.278890059711977e-05,
+      "loss": 0.2146,
       "step": 980
     },
     {
-      "epoch": 1.0399159663865547,
-      "grad_norm": 0.5524929761886597,
-      "learning_rate": 3.266806722689076e-05,
-      "loss": 0.2039,
+      "epoch": 1.0432033719704952,
+      "grad_norm": 0.6846581101417542,
+      "learning_rate": 3.261327713382508e-05,
+      "loss": 0.2149,
       "step": 990
     },
     {
-      "epoch": 1.050420168067227,
-      "grad_norm": 0.6426274180412292,
-      "learning_rate": 3.2492997198879555e-05,
-      "loss": 0.2184,
+      "epoch": 1.053740779768177,
+      "grad_norm": 0.5640810132026672,
+      "learning_rate": 3.2437653670530384e-05,
+      "loss": 0.2078,
       "step": 1000
     },
     {
-      "epoch": 1.050420168067227,
-      "eval_loss": 0.3305407762527466,
-      "eval_runtime": 13.5592,
-      "eval_samples_per_second": 35.4,
-      "eval_steps_per_second": 2.213,
+      "epoch": 1.053740779768177,
+      "eval_loss": 0.3363134264945984,
+      "eval_runtime": 13.6484,
+      "eval_samples_per_second": 35.169,
+      "eval_steps_per_second": 2.198,
       "step": 1000
     },
     {
-      "epoch": 1.0609243697478992,
-      "grad_norm": 0.4944634437561035,
-      "learning_rate": 3.231792717086835e-05,
-      "loss": 0.2042,
+      "epoch": 1.064278187565859,
+      "grad_norm": 0.6279656887054443,
+      "learning_rate": 3.226203020723569e-05,
+      "loss": 0.2004,
       "step": 1010
     },
     {
-      "epoch": 1.0714285714285714,
-      "grad_norm": 0.5576530694961548,
-      "learning_rate": 3.2142857142857144e-05,
-      "loss": 0.2029,
+      "epoch": 1.0748155953635405,
+      "grad_norm": 0.5957475304603577,
+      "learning_rate": 3.2086406743940996e-05,
+      "loss": 0.1952,
       "step": 1020
     },
     {
-      "epoch": 1.0819327731092436,
-      "grad_norm": 0.674849271774292,
-      "learning_rate": 3.196778711484594e-05,
-      "loss": 0.2026,
+      "epoch": 1.0853530031612224,
+      "grad_norm": 0.6246820092201233,
+      "learning_rate": 3.1910783280646294e-05,
+      "loss": 0.2071,
       "step": 1030
     },
     {
-      "epoch": 1.092436974789916,
-      "grad_norm": 0.5424471497535706,
-      "learning_rate": 3.179271708683473e-05,
-      "loss": 0.2027,
+      "epoch": 1.095890410958904,
+      "grad_norm": 0.5296807885169983,
+      "learning_rate": 3.17351598173516e-05,
+      "loss": 0.2077,
       "step": 1040
     },
     {
-      "epoch": 1.1029411764705883,
-      "grad_norm": 0.6491550207138062,
-      "learning_rate": 3.161764705882353e-05,
-      "loss": 0.2029,
+      "epoch": 1.106427818756586,
+      "grad_norm": 0.6367089152336121,
+      "learning_rate": 3.15595363540569e-05,
+      "loss": 0.2087,
       "step": 1050
     },
     {
-      "epoch": 1.1029411764705883,
-      "eval_loss": 0.33292290568351746,
-      "eval_runtime": 13.5611,
-      "eval_samples_per_second": 35.395,
-      "eval_steps_per_second": 2.212,
+      "epoch": 1.106427818756586,
+      "eval_loss": 0.3410273492336273,
+      "eval_runtime": 13.6191,
+      "eval_samples_per_second": 35.245,
+      "eval_steps_per_second": 2.203,
       "step": 1050
     },
     {
-      "epoch": 1.1134453781512605,
-      "grad_norm": 0.6113711595535278,
-      "learning_rate": 3.144257703081233e-05,
-      "loss": 0.2,
+      "epoch": 1.1169652265542676,
+      "grad_norm": 0.5743587017059326,
+      "learning_rate": 3.1383912890762205e-05,
+      "loss": 0.2063,
       "step": 1060
     },
     {
-      "epoch": 1.1239495798319328,
-      "grad_norm": 0.5068053603172302,
-      "learning_rate": 3.126750700280112e-05,
-      "loss": 0.1903,
+      "epoch": 1.1275026343519494,
+      "grad_norm": 0.5830729603767395,
+      "learning_rate": 3.120828942746751e-05,
+      "loss": 0.2021,
       "step": 1070
     },
     {
-      "epoch": 1.134453781512605,
-      "grad_norm": 0.6518192291259766,
-      "learning_rate": 3.1092436974789916e-05,
-      "loss": 0.2006,
+      "epoch": 1.1380400421496313,
+      "grad_norm": 0.585697591304779,
+      "learning_rate": 3.1032665964172816e-05,
+      "loss": 0.2003,
       "step": 1080
     },
     {
-      "epoch": 1.1449579831932772,
-      "grad_norm": 0.6932762861251831,
-      "learning_rate": 3.091736694677872e-05,
-      "loss": 0.2051,
+      "epoch": 1.148577449947313,
+      "grad_norm": 0.6112616062164307,
+      "learning_rate": 3.085704250087812e-05,
+      "loss": 0.2035,
       "step": 1090
     },
     {
-      "epoch": 1.1554621848739495,
-      "grad_norm": 0.5372537970542908,
-      "learning_rate": 3.074229691876751e-05,
-      "loss": 0.2037,
+      "epoch": 1.1591148577449948,
+      "grad_norm": 0.6068351864814758,
+      "learning_rate": 3.068141903758342e-05,
+      "loss": 0.198,
       "step": 1100
     },
     {
-      "epoch": 1.1554621848739495,
-      "eval_loss": 0.33518460392951965,
-      "eval_runtime": 13.5628,
-      "eval_samples_per_second": 35.391,
-      "eval_steps_per_second": 2.212,
+      "epoch": 1.1591148577449948,
+      "eval_loss": 0.33999618887901306,
+      "eval_runtime": 13.5917,
+      "eval_samples_per_second": 35.316,
+      "eval_steps_per_second": 2.207,
       "step": 1100
     },
     {
-      "epoch": 1.165966386554622,
-      "grad_norm": 0.6258675456047058,
-      "learning_rate": 3.0567226890756306e-05,
-      "loss": 0.1953,
+      "epoch": 1.1696522655426764,
+      "grad_norm": 0.6323215961456299,
+      "learning_rate": 3.0505795574288726e-05,
+      "loss": 0.204,
       "step": 1110
     },
     {
-      "epoch": 1.1764705882352942,
-      "grad_norm": 0.5910756587982178,
-      "learning_rate": 3.0392156862745097e-05,
-      "loss": 0.2098,
+      "epoch": 1.1801896733403583,
+      "grad_norm": 0.6767095923423767,
+      "learning_rate": 3.0330172110994032e-05,
+      "loss": 0.2,
       "step": 1120
     },
     {
-      "epoch": 1.1869747899159664,
-      "grad_norm": 0.6931313276290894,
-      "learning_rate": 3.0217086834733894e-05,
-      "loss": 0.2022,
+      "epoch": 1.1907270811380402,
+      "grad_norm": 0.6808314919471741,
+      "learning_rate": 3.015454864769933e-05,
+      "loss": 0.2052,
       "step": 1130
     },
     {
-      "epoch": 1.1974789915966386,
-      "grad_norm": 0.6102430820465088,
-      "learning_rate": 3.004201680672269e-05,
-      "loss": 0.1946,
+      "epoch": 1.2012644889357218,
+      "grad_norm": 0.7271966934204102,
+      "learning_rate": 2.997892518440464e-05,
+      "loss": 0.2035,
       "step": 1140
     },
     {
-      "epoch": 1.2079831932773109,
-      "grad_norm": 0.6068236827850342,
-      "learning_rate": 2.9866946778711486e-05,
-      "loss": 0.2089,
+      "epoch": 1.2118018967334037,
+      "grad_norm": 0.5914655327796936,
+      "learning_rate": 2.9803301721109945e-05,
+      "loss": 0.1999,
       "step": 1150
     },
     {
-      "epoch": 1.2079831932773109,
-      "eval_loss": 0.3338312804698944,
-      "eval_runtime": 13.5402,
-      "eval_samples_per_second": 35.45,
-      "eval_steps_per_second": 2.216,
+      "epoch": 1.2118018967334037,
+      "eval_loss": 0.3386387825012207,
+      "eval_runtime": 13.6001,
+      "eval_samples_per_second": 35.294,
+      "eval_steps_per_second": 2.206,
       "step": 1150
     },
     {
-      "epoch": 1.2184873949579833,
-      "grad_norm": 0.6477882862091064,
-      "learning_rate": 2.969187675070028e-05,
-      "loss": 0.2065,
+      "epoch": 1.2223393045310853,
+      "grad_norm": 0.6497891545295715,
+      "learning_rate": 2.9627678257815244e-05,
+      "loss": 0.2059,
       "step": 1160
     },
     {
-      "epoch": 1.2289915966386555,
-      "grad_norm": 0.6269820928573608,
-      "learning_rate": 2.9516806722689078e-05,
-      "loss": 0.1926,
+      "epoch": 1.2328767123287672,
+      "grad_norm": 0.6975721120834351,
+      "learning_rate": 2.945205479452055e-05,
+      "loss": 0.1968,
       "step": 1170
     },
     {
-      "epoch": 1.2394957983193278,
-      "grad_norm": 0.6708040833473206,
-      "learning_rate": 2.9341736694677872e-05,
+      "epoch": 1.2434141201264488,
+      "grad_norm": 0.6115730404853821,
+      "learning_rate": 2.9276431331225852e-05,
       "loss": 0.1983,
       "step": 1180
     },
     {
-      "epoch": 1.25,
-      "grad_norm": 0.6625474095344543,
-      "learning_rate": 2.916666666666667e-05,
-      "loss": 0.1967,
+      "epoch": 1.2539515279241307,
+      "grad_norm": 0.6213293075561523,
+      "learning_rate": 2.9100807867931158e-05,
+      "loss": 0.2077,
       "step": 1190
     },
     {
-      "epoch": 1.2605042016806722,
-      "grad_norm": 0.5640371441841125,
-      "learning_rate": 2.8991596638655467e-05,
-      "loss": 0.205,
+      "epoch": 1.2644889357218125,
+      "grad_norm": 0.6509750485420227,
+      "learning_rate": 2.8925184404636463e-05,
+      "loss": 0.194,
       "step": 1200
     },
     {
-      "epoch": 1.2605042016806722,
-      "eval_loss": 0.33764752745628357,
-      "eval_runtime": 13.5241,
-      "eval_samples_per_second": 35.492,
-      "eval_steps_per_second": 2.218,
+      "epoch": 1.2644889357218125,
+      "eval_loss": 0.34055641293525696,
+      "eval_runtime": 13.997,
+      "eval_samples_per_second": 34.293,
+      "eval_steps_per_second": 2.143,
       "step": 1200
     },
     {
-      "epoch": 1.2710084033613445,
-      "grad_norm": 0.590160608291626,
-      "learning_rate": 2.8816526610644258e-05,
-      "loss": 0.2008,
+      "epoch": 1.2750263435194942,
+      "grad_norm": 0.4847603142261505,
+      "learning_rate": 2.8749560941341762e-05,
+      "loss": 0.1976,
       "step": 1210
     },
     {
-      "epoch": 1.2815126050420167,
-      "grad_norm": 0.6975618004798889,
-      "learning_rate": 2.8641456582633052e-05,
-      "loss": 0.1956,
+      "epoch": 1.285563751317176,
+      "grad_norm": 0.6689492464065552,
+      "learning_rate": 2.8573937478047068e-05,
+      "loss": 0.19,
       "step": 1220
     },
     {
-      "epoch": 1.2920168067226891,
-      "grad_norm": 0.6742042899131775,
-      "learning_rate": 2.846638655462185e-05,
-      "loss": 0.2089,
+      "epoch": 1.2961011591148577,
+      "grad_norm": 0.6299355626106262,
+      "learning_rate": 2.839831401475237e-05,
+      "loss": 0.2025,
       "step": 1230
     },
     {
-      "epoch": 1.3025210084033614,
-      "grad_norm": 0.5806481838226318,
-      "learning_rate": 2.8291316526610644e-05,
-      "loss": 0.1937,
+      "epoch": 1.3066385669125395,
+      "grad_norm": 0.5940089821815491,
+      "learning_rate": 2.8222690551457676e-05,
+      "loss": 0.1917,
       "step": 1240
     },
     {
-      "epoch": 1.3130252100840336,
-      "grad_norm": 0.9894171953201294,
-      "learning_rate": 2.8116246498599442e-05,
-      "loss": 0.2098,
+      "epoch": 1.3171759747102212,
+      "grad_norm": 0.8054115772247314,
+      "learning_rate": 2.804706708816298e-05,
+      "loss": 0.1928,
       "step": 1250
     },
     {
-      "epoch": 1.3130252100840336,
-      "eval_loss": 0.3376815617084503,
-      "eval_runtime": 13.503,
-      "eval_samples_per_second": 35.548,
-      "eval_steps_per_second": 2.222,
+      "epoch": 1.3171759747102212,
+      "eval_loss": 0.34280383586883545,
+      "eval_runtime": 13.6317,
+      "eval_samples_per_second": 35.212,
+      "eval_steps_per_second": 2.201,
       "step": 1250
     },
     {
-      "epoch": 1.3235294117647058,
-      "grad_norm": 0.8022045493125916,
-      "learning_rate": 2.7941176470588236e-05,
-      "loss": 0.2043,
+      "epoch": 1.327713382507903,
+      "grad_norm": 0.5501131415367126,
+      "learning_rate": 2.7871443624868284e-05,
+      "loss": 0.1933,
       "step": 1260
     },
     {
-      "epoch": 1.334033613445378,
-      "grad_norm": 0.687003493309021,
-      "learning_rate": 2.7766106442577034e-05,
-      "loss": 0.2023,
+      "epoch": 1.338250790305585,
+      "grad_norm": 0.6450643539428711,
+      "learning_rate": 2.769582016157359e-05,
+      "loss": 0.1994,
       "step": 1270
     },
     {
-      "epoch": 1.3445378151260505,
-      "grad_norm": 0.7270589470863342,
-      "learning_rate": 2.7591036414565828e-05,
-      "loss": 0.1994,
+      "epoch": 1.3487881981032666,
+      "grad_norm": 0.6347033381462097,
+      "learning_rate": 2.7520196698278895e-05,
+      "loss": 0.2004,
       "step": 1280
     },
     {
-      "epoch": 1.3550420168067228,
-      "grad_norm": 0.5873326063156128,
-      "learning_rate": 2.7415966386554626e-05,
-      "loss": 0.1904,
+      "epoch": 1.3593256059009484,
+      "grad_norm": 0.6968681216239929,
+      "learning_rate": 2.7344573234984194e-05,
+      "loss": 0.214,
       "step": 1290
     },
     {
-      "epoch": 1.365546218487395,
-      "grad_norm": 0.6190339922904968,
-      "learning_rate": 2.7240896358543417e-05,
-      "loss": 0.198,
+      "epoch": 1.36986301369863,
+      "grad_norm": 0.5916845202445984,
+      "learning_rate": 2.71689497716895e-05,
+      "loss": 0.1849,
       "step": 1300
     },
     {
-      "epoch": 1.365546218487395,
-      "eval_loss": 0.3378330171108246,
-      "eval_runtime": 13.4926,
-      "eval_samples_per_second": 35.575,
-      "eval_steps_per_second": 2.223,
+      "epoch": 1.36986301369863,
+      "eval_loss": 0.34068360924720764,
+      "eval_runtime": 13.6631,
+      "eval_samples_per_second": 35.131,
+      "eval_steps_per_second": 2.196,
       "step": 1300
     }
   ],
   "logging_steps": 10,
-  "max_steps": 2856,
+  "max_steps": 2847,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 3,
   "save_steps": 100,
@@ -1144,7 +1144,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 1.5571083863025254e+17,
+  "total_flos": 1.5601386046262477e+17,
   "train_batch_size": 16,
   "trial_name": null,
   "trial_params": null

checkpoint-1300/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d5894a90f0aacad19de132730666f8b4647a0c4aa14309866a5f87d3723ce6a7
+oid sha256:b5fa9d8b62d1ebe6967a504a7decdb5eeee2bb4aac96e7414f3930f9adcff095
 size 5880