Training in progress, step 1600, checkpoint

checkpoint-1600/adapter_config.json

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

checkpoint-1600/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a8c073bae865f487f625ae08fd9f8a2d31357feb7596fe0245504f4c776b9838
+oid sha256:d361249c57ef9dff0451dcb63ee6230ff14fd572453f4764913d3f4d5490e73c
 size 161533192

checkpoint-1600/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fc7827876d3d2d7ef56714027d202b42065acebca5461fa7338fd70f88bbd30d
+oid sha256:9be1392abd4d4a57fef895f7de979cfa39014edbbe0cc083116b056673715ab7
 size 323290986

checkpoint-1600/trainer_state.json

@@ -10,1378 +10,1378 @@
   "log_history": [
     {
       "epoch": 0.010604453870625663,
-      "grad_norm": 2.0663416385650635,
+      "grad_norm": 2.276311159133911,
       "learning_rate": 4.982325910215624e-05,
-      "loss": 1.91,
+      "loss": 1.9038,
       "step": 10
     },
     {
       "epoch": 0.021208907741251327,
-      "grad_norm": 0.7459995746612549,
+      "grad_norm": 0.7678110003471375,
       "learning_rate": 4.964651820431248e-05,
-      "loss": 1.0094,
+      "loss": 0.997,
       "step": 20
     },
     {
       "epoch": 0.031813361611876985,
-      "grad_norm": 0.656954824924469,
+      "grad_norm": 0.6726565361022949,
       "learning_rate": 4.946977730646872e-05,
-      "loss": 0.8052,
+      "loss": 0.8021,
       "step": 30
     },
     {
       "epoch": 0.042417815482502653,
-      "grad_norm": 0.7273707389831543,
+      "grad_norm": 0.7318501472473145,
       "learning_rate": 4.929303640862496e-05,
-      "loss": 0.7145,
+      "loss": 0.7122,
       "step": 40
     },
     {
       "epoch": 0.053022269353128315,
-      "grad_norm": 0.6689164042472839,
+      "grad_norm": 0.6708294749259949,
       "learning_rate": 4.91162955107812e-05,
-      "loss": 0.6797,
+      "loss": 0.6783,
       "step": 50
     },
     {
       "epoch": 0.053022269353128315,
-      "eval_loss": 0.8142186403274536,
-      "eval_runtime": 11.794,
-      "eval_samples_per_second": 40.699,
-      "eval_steps_per_second": 2.544,
+      "eval_loss": 0.8075584173202515,
+      "eval_runtime": 11.7244,
+      "eval_samples_per_second": 40.94,
+      "eval_steps_per_second": 2.559,
       "step": 50
     },
     {
       "epoch": 0.06362672322375397,
-      "grad_norm": 0.6781293749809265,
+      "grad_norm": 0.7024317979812622,
       "learning_rate": 4.893955461293744e-05,
-      "loss": 0.6249,
+      "loss": 0.6244,
       "step": 60
     },
     {
       "epoch": 0.07423117709437964,
-      "grad_norm": 0.6776543259620667,
+      "grad_norm": 0.6959078907966614,
       "learning_rate": 4.876281371509368e-05,
-      "loss": 0.612,
+      "loss": 0.6117,
       "step": 70
     },
     {
       "epoch": 0.08483563096500531,
-      "grad_norm": 0.7603946924209595,
+      "grad_norm": 0.7824214696884155,
       "learning_rate": 4.858607281724991e-05,
-      "loss": 0.5998,
+      "loss": 0.5989,
       "step": 80
     },
     {
       "epoch": 0.09544008483563096,
-      "grad_norm": 0.7581377029418945,
+      "grad_norm": 0.7792642116546631,
       "learning_rate": 4.8409331919406155e-05,
-      "loss": 0.6032,
+      "loss": 0.6033,
       "step": 90
     },
     {
       "epoch": 0.10604453870625663,
-      "grad_norm": 0.7402133345603943,
+      "grad_norm": 0.7486705780029297,
       "learning_rate": 4.823259102156239e-05,
-      "loss": 0.5912,
+      "loss": 0.5904,
       "step": 100
     },
     {
       "epoch": 0.10604453870625663,
-      "eval_loss": 0.7294260859489441,
-      "eval_runtime": 11.8272,
-      "eval_samples_per_second": 40.585,
-      "eval_steps_per_second": 2.537,
+      "eval_loss": 0.7230511903762817,
+      "eval_runtime": 11.7589,
+      "eval_samples_per_second": 40.82,
+      "eval_steps_per_second": 2.551,
       "step": 100
     },
     {
       "epoch": 0.11664899257688228,
-      "grad_norm": 0.8224169015884399,
+      "grad_norm": 0.8214134573936462,
       "learning_rate": 4.8055850123718633e-05,
-      "loss": 0.578,
+      "loss": 0.5778,
       "step": 110
     },
     {
       "epoch": 0.12725344644750794,
-      "grad_norm": 0.7843881249427795,
+      "grad_norm": 0.7937195897102356,
       "learning_rate": 4.7879109225874866e-05,
-      "loss": 0.5751,
+      "loss": 0.5748,
       "step": 120
     },
     {
       "epoch": 0.1378579003181336,
-      "grad_norm": 0.772152304649353,
+      "grad_norm": 0.7781085968017578,
       "learning_rate": 4.770236832803111e-05,
-      "loss": 0.5471,
+      "loss": 0.5469,
       "step": 130
     },
     {
       "epoch": 0.14846235418875928,
-      "grad_norm": 0.6972452998161316,
+      "grad_norm": 0.695203423500061,
       "learning_rate": 4.7525627430187344e-05,
-      "loss": 0.5516,
+      "loss": 0.5511,
       "step": 140
     },
     {
       "epoch": 0.15906680805938495,
-      "grad_norm": 0.7806265354156494,
+      "grad_norm": 0.7824985384941101,
       "learning_rate": 4.734888653234359e-05,
-      "loss": 0.5263,
+      "loss": 0.5258,
       "step": 150
     },
     {
       "epoch": 0.15906680805938495,
-      "eval_loss": 0.6928939819335938,
-      "eval_runtime": 11.8093,
-      "eval_samples_per_second": 40.646,
-      "eval_steps_per_second": 2.54,
+      "eval_loss": 0.6867014765739441,
+      "eval_runtime": 11.7852,
+      "eval_samples_per_second": 40.729,
+      "eval_steps_per_second": 2.546,
       "step": 150
     },
     {
       "epoch": 0.16967126193001061,
-      "grad_norm": 0.7372023463249207,
+      "grad_norm": 0.7392640113830566,
       "learning_rate": 4.717214563449982e-05,
-      "loss": 0.5715,
+      "loss": 0.5714,
       "step": 160
     },
     {
       "epoch": 0.18027571580063625,
-      "grad_norm": 0.8236038684844971,
+      "grad_norm": 0.8316007852554321,
       "learning_rate": 4.699540473665606e-05,
-      "loss": 0.4987,
+      "loss": 0.4982,
       "step": 170
     },
     {
       "epoch": 0.19088016967126192,
-      "grad_norm": 0.8057310581207275,
+      "grad_norm": 0.8199877738952637,
       "learning_rate": 4.681866383881231e-05,
-      "loss": 0.5097,
+      "loss": 0.5092,
       "step": 180
     },
     {
       "epoch": 0.2014846235418876,
-      "grad_norm": 0.7124797105789185,
+      "grad_norm": 0.7089737057685852,
       "learning_rate": 4.664192294096854e-05,
-      "loss": 0.4952,
+      "loss": 0.4953,
       "step": 190
     },
     {
       "epoch": 0.21208907741251326,
-      "grad_norm": 0.7755710482597351,
+      "grad_norm": 0.7906930446624756,
       "learning_rate": 4.6465182043124786e-05,
-      "loss": 0.497,
+      "loss": 0.4974,
       "step": 200
     },
     {
       "epoch": 0.21208907741251326,
-      "eval_loss": 0.6748186349868774,
-      "eval_runtime": 11.8077,
-      "eval_samples_per_second": 40.651,
-      "eval_steps_per_second": 2.541,
+      "eval_loss": 0.669015645980835,
+      "eval_runtime": 11.8934,
+      "eval_samples_per_second": 40.358,
+      "eval_steps_per_second": 2.522,
       "step": 200
     },
     {
       "epoch": 0.22269353128313893,
-      "grad_norm": 0.912337064743042,
+      "grad_norm": 0.9085277318954468,
       "learning_rate": 4.628844114528102e-05,
-      "loss": 0.4997,
+      "loss": 0.4991,
       "step": 210
     },
     {
       "epoch": 0.23329798515376457,
-      "grad_norm": 0.771654486656189,
+      "grad_norm": 0.7775648832321167,
       "learning_rate": 4.6111700247437265e-05,
-      "loss": 0.5027,
+      "loss": 0.5024,
       "step": 220
     },
     {
       "epoch": 0.24390243902439024,
-      "grad_norm": 0.8241815567016602,
+      "grad_norm": 0.8324123620986938,
       "learning_rate": 4.59349593495935e-05,
-      "loss": 0.5196,
+      "loss": 0.5187,
       "step": 230
     },
     {
       "epoch": 0.2545068928950159,
-      "grad_norm": 0.7876004576683044,
+      "grad_norm": 0.8002191781997681,
       "learning_rate": 4.5758218451749736e-05,
-      "loss": 0.4791,
+      "loss": 0.4785,
       "step": 240
     },
     {
       "epoch": 0.2651113467656416,
-      "grad_norm": 0.7941911816596985,
+      "grad_norm": 0.8004704117774963,
       "learning_rate": 4.5581477553905975e-05,
-      "loss": 0.5078,
+      "loss": 0.5073,
       "step": 250
     },
     {
       "epoch": 0.2651113467656416,
-      "eval_loss": 0.653518795967102,
-      "eval_runtime": 11.7906,
-      "eval_samples_per_second": 40.71,
-      "eval_steps_per_second": 2.544,
+      "eval_loss": 0.6479213833808899,
+      "eval_runtime": 11.882,
+      "eval_samples_per_second": 40.397,
+      "eval_steps_per_second": 2.525,
       "step": 250
     },
     {
       "epoch": 0.2757158006362672,
-      "grad_norm": 0.8866105079650879,
+      "grad_norm": 0.893301248550415,
       "learning_rate": 4.5404736656062215e-05,
       "loss": 0.477,
       "step": 260
     },
     {
       "epoch": 0.2863202545068929,
-      "grad_norm": 0.8498968482017517,
+      "grad_norm": 0.8550834059715271,
       "learning_rate": 4.5227995758218454e-05,
-      "loss": 0.4926,
+      "loss": 0.4923,
       "step": 270
     },
     {
       "epoch": 0.29692470837751855,
-      "grad_norm": 0.8863232135772705,
+      "grad_norm": 0.8929152488708496,
       "learning_rate": 4.505125486037469e-05,
-      "loss": 0.463,
+      "loss": 0.4624,
       "step": 280
     },
     {
       "epoch": 0.3075291622481442,
-      "grad_norm": 0.9501034021377563,
+      "grad_norm": 0.9584974050521851,
       "learning_rate": 4.487451396253093e-05,
-      "loss": 0.4809,
+      "loss": 0.4801,
       "step": 290
     },
     {
       "epoch": 0.3181336161187699,
-      "grad_norm": 0.7933001518249512,
+      "grad_norm": 0.7924031615257263,
       "learning_rate": 4.469777306468717e-05,
-      "loss": 0.4941,
+      "loss": 0.4942,
       "step": 300
     },
     {
       "epoch": 0.3181336161187699,
-      "eval_loss": 0.6434358954429626,
-      "eval_runtime": 11.7969,
-      "eval_samples_per_second": 40.689,
-      "eval_steps_per_second": 2.543,
+      "eval_loss": 0.6371316313743591,
+      "eval_runtime": 11.8575,
+      "eval_samples_per_second": 40.481,
+      "eval_steps_per_second": 2.53,
       "step": 300
     },
     {
       "epoch": 0.32873806998939553,
-      "grad_norm": 0.9145284295082092,
+      "grad_norm": 0.9247742295265198,
       "learning_rate": 4.452103216684341e-05,
-      "loss": 0.5235,
+      "loss": 0.5234,
       "step": 310
     },
     {
       "epoch": 0.33934252386002123,
-      "grad_norm": 0.8557764887809753,
+      "grad_norm": 0.8617805242538452,
       "learning_rate": 4.434429126899965e-05,
-      "loss": 0.4809,
+      "loss": 0.4808,
       "step": 320
     },
     {
       "epoch": 0.34994697773064687,
-      "grad_norm": 0.8337067365646362,
+      "grad_norm": 0.8334285616874695,
       "learning_rate": 4.416755037115589e-05,
-      "loss": 0.4764,
+      "loss": 0.4758,
       "step": 330
     },
     {
       "epoch": 0.3605514316012725,
-      "grad_norm": 0.8928807377815247,
+      "grad_norm": 0.9079923629760742,
       "learning_rate": 4.399080947331213e-05,
-      "loss": 0.4634,
+      "loss": 0.463,
       "step": 340
     },
     {
       "epoch": 0.3711558854718982,
-      "grad_norm": 0.7639658451080322,
+      "grad_norm": 0.7632007002830505,
       "learning_rate": 4.381406857546837e-05,
-      "loss": 0.4829,
+      "loss": 0.4828,
       "step": 350
     },
     {
       "epoch": 0.3711558854718982,
-      "eval_loss": 0.6269502639770508,
-      "eval_runtime": 11.7838,
-      "eval_samples_per_second": 40.734,
-      "eval_steps_per_second": 2.546,
+      "eval_loss": 0.6212183237075806,
+      "eval_runtime": 11.8497,
+      "eval_samples_per_second": 40.507,
+      "eval_steps_per_second": 2.532,
       "step": 350
     },
     {
       "epoch": 0.38176033934252385,
-      "grad_norm": 0.8374601006507874,
+      "grad_norm": 0.8430343270301819,
       "learning_rate": 4.3637327677624606e-05,
-      "loss": 0.471,
+      "loss": 0.4711,
       "step": 360
     },
     {
       "epoch": 0.39236479321314954,
-      "grad_norm": 0.9471817016601562,
+      "grad_norm": 0.9495126605033875,
       "learning_rate": 4.3460586779780846e-05,
       "loss": 0.4528,
       "step": 370
     },
     {
       "epoch": 0.4029692470837752,
-      "grad_norm": 0.8860051035881042,
+      "grad_norm": 0.8903242945671082,
       "learning_rate": 4.328384588193708e-05,
-      "loss": 0.4684,
+      "loss": 0.468,
       "step": 380
     },
     {
       "epoch": 0.4135737009544008,
-      "grad_norm": 0.9099433422088623,
+      "grad_norm": 0.9306944608688354,
       "learning_rate": 4.3107104984093324e-05,
-      "loss": 0.4294,
+      "loss": 0.4299,
       "step": 390
     },
     {
       "epoch": 0.4241781548250265,
-      "grad_norm": 0.792826771736145,
+      "grad_norm": 0.7949309349060059,
       "learning_rate": 4.2930364086249556e-05,
-      "loss": 0.4597,
+      "loss": 0.4589,
       "step": 400
     },
     {
       "epoch": 0.4241781548250265,
-      "eval_loss": 0.6197967529296875,
-      "eval_runtime": 11.8062,
-      "eval_samples_per_second": 40.657,
-      "eval_steps_per_second": 2.541,
+      "eval_loss": 0.6148895621299744,
+      "eval_runtime": 11.8536,
+      "eval_samples_per_second": 40.494,
+      "eval_steps_per_second": 2.531,
       "step": 400
     },
     {
       "epoch": 0.43478260869565216,
-      "grad_norm": 0.8295171856880188,
+      "grad_norm": 0.8351686000823975,
       "learning_rate": 4.27536231884058e-05,
-      "loss": 0.4842,
+      "loss": 0.4836,
       "step": 410
     },
     {
       "epoch": 0.44538706256627786,
-      "grad_norm": 0.9126204252243042,
+      "grad_norm": 0.9184524416923523,
       "learning_rate": 4.2576882290562035e-05,
-      "loss": 0.4688,
+      "loss": 0.4686,
       "step": 420
     },
     {
       "epoch": 0.4559915164369035,
-      "grad_norm": 0.9485805630683899,
+      "grad_norm": 0.9610297679901123,
       "learning_rate": 4.240014139271828e-05,
-      "loss": 0.4803,
+      "loss": 0.4804,
       "step": 430
     },
     {
       "epoch": 0.46659597030752914,
-      "grad_norm": 0.8756628036499023,
+      "grad_norm": 0.8793377876281738,
       "learning_rate": 4.222340049487451e-05,
-      "loss": 0.4572,
+      "loss": 0.4571,
       "step": 440
     },
     {
       "epoch": 0.47720042417815484,
-      "grad_norm": 0.907244086265564,
+      "grad_norm": 0.913858950138092,
       "learning_rate": 4.204665959703076e-05,
-      "loss": 0.4394,
+      "loss": 0.4391,
       "step": 450
     },
     {
       "epoch": 0.47720042417815484,
-      "eval_loss": 0.6146622896194458,
-      "eval_runtime": 11.8005,
-      "eval_samples_per_second": 40.676,
-      "eval_steps_per_second": 2.542,
+      "eval_loss": 0.6096637845039368,
+      "eval_runtime": 11.8533,
+      "eval_samples_per_second": 40.495,
+      "eval_steps_per_second": 2.531,
       "step": 450
     },
     {
       "epoch": 0.4878048780487805,
-      "grad_norm": 0.9201627969741821,
+      "grad_norm": 0.9159062504768372,
       "learning_rate": 4.186991869918699e-05,
       "loss": 0.4294,
       "step": 460
     },
     {
       "epoch": 0.4984093319194062,
-      "grad_norm": 0.9242031574249268,
+      "grad_norm": 0.9346657395362854,
       "learning_rate": 4.169317780134323e-05,
-      "loss": 0.4335,
+      "loss": 0.433,
       "step": 470
     },
     {
       "epoch": 0.5090137857900318,
-      "grad_norm": 0.840977132320404,
+      "grad_norm": 0.8480401635169983,
       "learning_rate": 4.151643690349947e-05,
       "loss": 0.4639,
       "step": 480
     },
     {
       "epoch": 0.5196182396606575,
-      "grad_norm": 0.9487953782081604,
+      "grad_norm": 0.9547140002250671,
       "learning_rate": 4.133969600565571e-05,
       "loss": 0.4173,
       "step": 490
     },
     {
       "epoch": 0.5302226935312832,
-      "grad_norm": 0.8343156576156616,
+      "grad_norm": 0.8660142421722412,
       "learning_rate": 4.116295510781195e-05,
-      "loss": 0.4647,
+      "loss": 0.4644,
       "step": 500
     },
     {
       "epoch": 0.5302226935312832,
-      "eval_loss": 0.6104319095611572,
-      "eval_runtime": 11.8139,
-      "eval_samples_per_second": 40.63,
-      "eval_steps_per_second": 2.539,
+      "eval_loss": 0.6053425073623657,
+      "eval_runtime": 11.8341,
+      "eval_samples_per_second": 40.561,
+      "eval_steps_per_second": 2.535,
       "step": 500
     },
     {
       "epoch": 0.5408271474019088,
-      "grad_norm": 0.9216362237930298,
+      "grad_norm": 0.9299489855766296,
       "learning_rate": 4.098621420996819e-05,
-      "loss": 0.4499,
+      "loss": 0.4488,
       "step": 510
     },
     {
       "epoch": 0.5514316012725344,
-      "grad_norm": 0.957593560218811,
+      "grad_norm": 0.9637112021446228,
       "learning_rate": 4.080947331212443e-05,
-      "loss": 0.4202,
+      "loss": 0.4197,
       "step": 520
     },
     {
       "epoch": 0.5620360551431601,
-      "grad_norm": 1.0103106498718262,
+      "grad_norm": 0.9725180864334106,
       "learning_rate": 4.0632732414280666e-05,
-      "loss": 0.4315,
+      "loss": 0.4308,
       "step": 530
     },
     {
       "epoch": 0.5726405090137858,
-      "grad_norm": 1.0909500122070312,
+      "grad_norm": 1.095697045326233,
       "learning_rate": 4.0455991516436905e-05,
-      "loss": 0.4476,
+      "loss": 0.4473,
       "step": 540
     },
     {
       "epoch": 0.5832449628844114,
-      "grad_norm": 0.8890084028244019,
+      "grad_norm": 0.8940134048461914,
       "learning_rate": 4.0279250618593144e-05,
-      "loss": 0.4408,
+      "loss": 0.441,
       "step": 550
     },
     {
       "epoch": 0.5832449628844114,
-      "eval_loss": 0.6050440073013306,
-      "eval_runtime": 11.7965,
-      "eval_samples_per_second": 40.69,
-      "eval_steps_per_second": 2.543,
+      "eval_loss": 0.6002687215805054,
+      "eval_runtime": 11.8218,
+      "eval_samples_per_second": 40.603,
+      "eval_steps_per_second": 2.538,
       "step": 550
     },
     {
       "epoch": 0.5938494167550371,
-      "grad_norm": 1.038818359375,
+      "grad_norm": 1.051446795463562,
       "learning_rate": 4.0102509720749383e-05,
-      "loss": 0.4529,
+      "loss": 0.4531,
       "step": 560
     },
     {
       "epoch": 0.6044538706256628,
-      "grad_norm": 1.0575263500213623,
+      "grad_norm": 1.055331826210022,
       "learning_rate": 3.992576882290562e-05,
-      "loss": 0.4409,
+      "loss": 0.4405,
       "step": 570
     },
     {
       "epoch": 0.6150583244962884,
-      "grad_norm": 0.9827993512153625,
+      "grad_norm": 0.9812480211257935,
       "learning_rate": 3.974902792506186e-05,
-      "loss": 0.4132,
+      "loss": 0.4128,
       "step": 580
     },
     {
       "epoch": 0.6256627783669141,
-      "grad_norm": 0.9469496011734009,
+      "grad_norm": 0.9506167769432068,
       "learning_rate": 3.95722870272181e-05,
-      "loss": 0.4075,
+      "loss": 0.4077,
       "step": 590
     },
     {
       "epoch": 0.6362672322375398,
-      "grad_norm": 0.9693032503128052,
+      "grad_norm": 0.9744945764541626,
       "learning_rate": 3.939554612937434e-05,
-      "loss": 0.4533,
+      "loss": 0.4531,
       "step": 600
     },
     {
       "epoch": 0.6362672322375398,
-      "eval_loss": 0.5991437435150146,
-      "eval_runtime": 11.8081,
-      "eval_samples_per_second": 40.65,
-      "eval_steps_per_second": 2.541,
+      "eval_loss": 0.5941649079322815,
+      "eval_runtime": 11.8037,
+      "eval_samples_per_second": 40.665,
+      "eval_steps_per_second": 2.542,
       "step": 600
     },
     {
       "epoch": 0.6468716861081655,
-      "grad_norm": 0.9771609306335449,
+      "grad_norm": 0.9825407862663269,
       "learning_rate": 3.921880523153057e-05,
-      "loss": 0.4275,
+      "loss": 0.4281,
       "step": 610
     },
     {
       "epoch": 0.6574761399787911,
-      "grad_norm": 0.9394034743309021,
+      "grad_norm": 0.9444342851638794,
       "learning_rate": 3.904206433368682e-05,
-      "loss": 0.4066,
+      "loss": 0.4069,
       "step": 620
     },
     {
       "epoch": 0.6680805938494168,
-      "grad_norm": 0.885517954826355,
+      "grad_norm": 0.8893631100654602,
       "learning_rate": 3.886532343584305e-05,
-      "loss": 0.4071,
+      "loss": 0.4067,
       "step": 630
     },
     {
       "epoch": 0.6786850477200425,
-      "grad_norm": 0.9775238633155823,
+      "grad_norm": 0.9782839417457581,
       "learning_rate": 3.86885825379993e-05,
-      "loss": 0.4234,
+      "loss": 0.4237,
       "step": 640
     },
     {
       "epoch": 0.689289501590668,
-      "grad_norm": 0.9256358742713928,
+      "grad_norm": 0.9258303642272949,
       "learning_rate": 3.851184164015553e-05,
-      "loss": 0.4186,
+      "loss": 0.4181,
       "step": 650
     },
     {
       "epoch": 0.689289501590668,
-      "eval_loss": 0.5974680781364441,
-      "eval_runtime": 11.8062,
-      "eval_samples_per_second": 40.657,
-      "eval_steps_per_second": 2.541,
+      "eval_loss": 0.5927689671516418,
+      "eval_runtime": 11.8222,
+      "eval_samples_per_second": 40.601,
+      "eval_steps_per_second": 2.538,
       "step": 650
     },
     {
       "epoch": 0.6998939554612937,
-      "grad_norm": 0.9220837354660034,
+      "grad_norm": 0.9336357116699219,
       "learning_rate": 3.8335100742311775e-05,
       "loss": 0.4264,
       "step": 660
     },
     {
       "epoch": 0.7104984093319194,
-      "grad_norm": 1.0769790410995483,
+      "grad_norm": 1.0914337635040283,
       "learning_rate": 3.8158359844468015e-05,
-      "loss": 0.4166,
+      "loss": 0.4163,
       "step": 670
     },
     {
       "epoch": 0.721102863202545,
-      "grad_norm": 0.9368733763694763,
+      "grad_norm": 0.9416537284851074,
       "learning_rate": 3.7981618946624254e-05,
-      "loss": 0.4296,
+      "loss": 0.4292,
       "step": 680
     },
     {
       "epoch": 0.7317073170731707,
-      "grad_norm": 1.0023365020751953,
+      "grad_norm": 1.0038481950759888,
       "learning_rate": 3.780487804878049e-05,
-      "loss": 0.4534,
+      "loss": 0.4532,
       "step": 690
     },
     {
       "epoch": 0.7423117709437964,
-      "grad_norm": 1.062111258506775,
+      "grad_norm": 1.0638171434402466,
       "learning_rate": 3.7628137150936725e-05,
-      "loss": 0.4378,
+      "loss": 0.437,
       "step": 700
     },
     {
       "epoch": 0.7423117709437964,
-      "eval_loss": 0.5889919400215149,
-      "eval_runtime": 11.8116,
-      "eval_samples_per_second": 40.638,
-      "eval_steps_per_second": 2.54,
+      "eval_loss": 0.5840710997581482,
+      "eval_runtime": 11.8245,
+      "eval_samples_per_second": 40.594,
+      "eval_steps_per_second": 2.537,
       "step": 700
     },
     {
       "epoch": 0.7529162248144221,
-      "grad_norm": 1.0798842906951904,
+      "grad_norm": 1.091558814048767,
       "learning_rate": 3.745139625309297e-05,
-      "loss": 0.4168,
+      "loss": 0.4171,
       "step": 710
     },
     {
       "epoch": 0.7635206786850477,
-      "grad_norm": 1.0573209524154663,
+      "grad_norm": 1.052101731300354,
       "learning_rate": 3.7274655355249204e-05,
-      "loss": 0.4244,
+      "loss": 0.425,
       "step": 720
     },
     {
       "epoch": 0.7741251325556734,
-      "grad_norm": 0.9914817810058594,
+      "grad_norm": 0.9969995021820068,
       "learning_rate": 3.709791445740545e-05,
-      "loss": 0.4332,
+      "loss": 0.4329,
       "step": 730
     },
     {
       "epoch": 0.7847295864262991,
-      "grad_norm": 1.0201358795166016,
+      "grad_norm": 1.0250300168991089,
       "learning_rate": 3.692117355956168e-05,
-      "loss": 0.4225,
+      "loss": 0.4221,
       "step": 740
     },
     {
       "epoch": 0.7953340402969247,
-      "grad_norm": 1.011526346206665,
+      "grad_norm": 1.0140836238861084,
       "learning_rate": 3.674443266171793e-05,
-      "loss": 0.4069,
+      "loss": 0.4059,
       "step": 750
     },
     {
       "epoch": 0.7953340402969247,
-      "eval_loss": 0.5948487520217896,
-      "eval_runtime": 11.8194,
-      "eval_samples_per_second": 40.611,
-      "eval_steps_per_second": 2.538,
+      "eval_loss": 0.5901935696601868,
+      "eval_runtime": 11.8492,
+      "eval_samples_per_second": 40.509,
+      "eval_steps_per_second": 2.532,
       "step": 750
     },
     {
       "epoch": 0.8059384941675504,
-      "grad_norm": 0.9645175933837891,
+      "grad_norm": 0.9665971994400024,
       "learning_rate": 3.656769176387416e-05,
-      "loss": 0.3833,
+      "loss": 0.3824,
       "step": 760
     },
     {
       "epoch": 0.8165429480381761,
-      "grad_norm": 0.9826123118400574,
+      "grad_norm": 0.9899901747703552,
       "learning_rate": 3.63909508660304e-05,
-      "loss": 0.4019,
+      "loss": 0.402,
       "step": 770
     },
     {
       "epoch": 0.8271474019088016,
-      "grad_norm": 0.957275927066803,
+      "grad_norm": 0.9623307585716248,
       "learning_rate": 3.621420996818664e-05,
-      "loss": 0.4083,
+      "loss": 0.4078,
       "step": 780
     },
     {
       "epoch": 0.8377518557794273,
-      "grad_norm": 1.0057812929153442,
+      "grad_norm": 1.0048012733459473,
       "learning_rate": 3.603746907034288e-05,
-      "loss": 0.4218,
+      "loss": 0.4211,
       "step": 790
     },
     {
       "epoch": 0.848356309650053,
-      "grad_norm": 1.0474872589111328,
+      "grad_norm": 1.04786217212677,
       "learning_rate": 3.586072817249912e-05,
       "loss": 0.4224,
       "step": 800
     },
     {
       "epoch": 0.848356309650053,
-      "eval_loss": 0.5902166962623596,
-      "eval_runtime": 11.8122,
-      "eval_samples_per_second": 40.636,
-      "eval_steps_per_second": 2.54,
+      "eval_loss": 0.5850113034248352,
+      "eval_runtime": 11.8518,
+      "eval_samples_per_second": 40.5,
+      "eval_steps_per_second": 2.531,
       "step": 800
     },
     {
       "epoch": 0.8589607635206787,
-      "grad_norm": 1.1121183633804321,
+      "grad_norm": 1.1098370552062988,
       "learning_rate": 3.5683987274655356e-05,
-      "loss": 0.4035,
+      "loss": 0.4033,
       "step": 810
     },
     {
       "epoch": 0.8695652173913043,
-      "grad_norm": 1.078808069229126,
+      "grad_norm": 1.0828533172607422,
       "learning_rate": 3.5507246376811596e-05,
-      "loss": 0.3987,
+      "loss": 0.3976,
       "step": 820
     },
     {
       "epoch": 0.88016967126193,
-      "grad_norm": 0.9879220724105835,
+      "grad_norm": 1.0007745027542114,
       "learning_rate": 3.5330505478967835e-05,
-      "loss": 0.4107,
+      "loss": 0.4101,
       "step": 830
     },
     {
       "epoch": 0.8907741251325557,
-      "grad_norm": 1.0545059442520142,
+      "grad_norm": 1.0621625185012817,
       "learning_rate": 3.5153764581124074e-05,
-      "loss": 0.412,
+      "loss": 0.4113,
       "step": 840
     },
     {
       "epoch": 0.9013785790031813,
-      "grad_norm": 1.0051729679107666,
+      "grad_norm": 1.0198171138763428,
       "learning_rate": 3.497702368328031e-05,
-      "loss": 0.3962,
+      "loss": 0.3957,
       "step": 850
     },
     {
       "epoch": 0.9013785790031813,
-      "eval_loss": 0.586068868637085,
-      "eval_runtime": 11.7963,
-      "eval_samples_per_second": 40.691,
-      "eval_steps_per_second": 2.543,
+      "eval_loss": 0.5811945199966431,
+      "eval_runtime": 11.8706,
+      "eval_samples_per_second": 40.436,
+      "eval_steps_per_second": 2.527,
       "step": 850
     },
     {
       "epoch": 0.911983032873807,
-      "grad_norm": 1.0309350490570068,
+      "grad_norm": 1.035551905632019,
       "learning_rate": 3.480028278543655e-05,
-      "loss": 0.405,
+      "loss": 0.4047,
       "step": 860
     },
     {
       "epoch": 0.9225874867444327,
-      "grad_norm": 1.063778281211853,
+      "grad_norm": 1.0746926069259644,
       "learning_rate": 3.462354188759279e-05,
-      "loss": 0.4034,
+      "loss": 0.4035,
       "step": 870
     },
     {
       "epoch": 0.9331919406150583,
-      "grad_norm": 1.0570268630981445,
+      "grad_norm": 1.0593372583389282,
       "learning_rate": 3.444680098974903e-05,
-      "loss": 0.4038,
+      "loss": 0.403,
       "step": 880
     },
     {
       "epoch": 0.943796394485684,
-      "grad_norm": 1.0920844078063965,
+      "grad_norm": 1.0974407196044922,
       "learning_rate": 3.427006009190527e-05,
-      "loss": 0.3996,
+      "loss": 0.3994,
       "step": 890
     },
     {
       "epoch": 0.9544008483563097,
-      "grad_norm": 0.9112839698791504,
+      "grad_norm": 0.924757182598114,
       "learning_rate": 3.409331919406151e-05,
-      "loss": 0.3855,
+      "loss": 0.3854,
       "step": 900
     },
     {
       "epoch": 0.9544008483563097,
-      "eval_loss": 0.5862192511558533,
-      "eval_runtime": 11.7865,
-      "eval_samples_per_second": 40.725,
-      "eval_steps_per_second": 2.545,
+      "eval_loss": 0.5818654894828796,
+      "eval_runtime": 11.8558,
+      "eval_samples_per_second": 40.486,
+      "eval_steps_per_second": 2.53,
       "step": 900
     },
     {
       "epoch": 0.9650053022269353,
-      "grad_norm": 1.0250121355056763,
+      "grad_norm": 1.0254701375961304,
       "learning_rate": 3.391657829621775e-05,
-      "loss": 0.4219,
+      "loss": 0.4214,
       "step": 910
     },
     {
       "epoch": 0.975609756097561,
-      "grad_norm": 1.045753836631775,
+      "grad_norm": 1.0513700246810913,
       "learning_rate": 3.373983739837399e-05,
-      "loss": 0.3936,
+      "loss": 0.3934,
       "step": 920
     },
     {
       "epoch": 0.9862142099681867,
-      "grad_norm": 1.0251036882400513,
+      "grad_norm": 1.0283734798431396,
       "learning_rate": 3.356309650053022e-05,
-      "loss": 0.3798,
+      "loss": 0.3792,
       "step": 930
     },
     {
       "epoch": 0.9968186638388123,
-      "grad_norm": 1.0625455379486084,
+      "grad_norm": 1.0671693086624146,
       "learning_rate": 3.3386355602686466e-05,
-      "loss": 0.3779,
+      "loss": 0.3777,
       "step": 940
     },
     {
       "epoch": 1.007423117709438,
-      "grad_norm": 1.050057291984558,
+      "grad_norm": 1.0414798259735107,
       "learning_rate": 3.32096147048427e-05,
-      "loss": 0.3676,
+      "loss": 0.3681,
       "step": 950
     },
     {
       "epoch": 1.007423117709438,
-      "eval_loss": 0.5891817212104797,
-      "eval_runtime": 11.7414,
-      "eval_samples_per_second": 40.881,
-      "eval_steps_per_second": 2.555,
+      "eval_loss": 0.5841290950775146,
+      "eval_runtime": 11.8527,
+      "eval_samples_per_second": 40.497,
+      "eval_steps_per_second": 2.531,
       "step": 950
     },
     {
       "epoch": 1.0180275715800637,
-      "grad_norm": 1.109527826309204,
+      "grad_norm": 1.1088826656341553,
       "learning_rate": 3.3032873806998944e-05,
-      "loss": 0.3532,
+      "loss": 0.3536,
       "step": 960
     },
     {
       "epoch": 1.0286320254506893,
-      "grad_norm": 1.2735576629638672,
+      "grad_norm": 1.2678877115249634,
       "learning_rate": 3.285613290915518e-05,
-      "loss": 0.3804,
+      "loss": 0.38,
       "step": 970
     },
     {
       "epoch": 1.039236479321315,
-      "grad_norm": 1.079765796661377,
+      "grad_norm": 1.0846420526504517,
       "learning_rate": 3.267939201131142e-05,
-      "loss": 0.352,
+      "loss": 0.3522,
       "step": 980
     },
     {
       "epoch": 1.0498409331919407,
-      "grad_norm": 1.069228172302246,
+      "grad_norm": 1.0741760730743408,
       "learning_rate": 3.2502651113467655e-05,
-      "loss": 0.3402,
+      "loss": 0.3392,
       "step": 990
     },
     {
       "epoch": 1.0604453870625663,
-      "grad_norm": 1.2578569650650024,
+      "grad_norm": 1.2564727067947388,
       "learning_rate": 3.2325910215623894e-05,
-      "loss": 0.3541,
+      "loss": 0.3538,
       "step": 1000
     },
     {
       "epoch": 1.0604453870625663,
-      "eval_loss": 0.5981893539428711,
-      "eval_runtime": 11.7382,
-      "eval_samples_per_second": 40.892,
-      "eval_steps_per_second": 2.556,
+      "eval_loss": 0.5931693911552429,
+      "eval_runtime": 11.8583,
+      "eval_samples_per_second": 40.478,
+      "eval_steps_per_second": 2.53,
       "step": 1000
     },
     {
       "epoch": 1.0710498409331919,
-      "grad_norm": 1.1604409217834473,
+      "grad_norm": 1.1698198318481445,
       "learning_rate": 3.2149169317780133e-05,
       "loss": 0.3274,
       "step": 1010
     },
     {
       "epoch": 1.0816542948038177,
-      "grad_norm": 1.135129690170288,
+      "grad_norm": 1.1250603199005127,
       "learning_rate": 3.197242841993637e-05,
-      "loss": 0.3528,
+      "loss": 0.3519,
       "step": 1020
     },
     {
       "epoch": 1.0922587486744433,
-      "grad_norm": 1.1334470510482788,
+      "grad_norm": 1.1354424953460693,
       "learning_rate": 3.179568752209262e-05,
-      "loss": 0.3472,
+      "loss": 0.3474,
       "step": 1030
     },
     {
       "epoch": 1.1028632025450689,
-      "grad_norm": 1.1980904340744019,
+      "grad_norm": 1.2070807218551636,
       "learning_rate": 3.161894662424885e-05,
-      "loss": 0.343,
+      "loss": 0.3426,
       "step": 1040
     },
     {
       "epoch": 1.1134676564156947,
-      "grad_norm": 1.144959568977356,
+      "grad_norm": 1.147916555404663,
       "learning_rate": 3.14422057264051e-05,
-      "loss": 0.3447,
+      "loss": 0.3449,
       "step": 1050
     },
     {
       "epoch": 1.1134676564156947,
-      "eval_loss": 0.5998801589012146,
-      "eval_runtime": 11.7594,
-      "eval_samples_per_second": 40.818,
-      "eval_steps_per_second": 2.551,
+      "eval_loss": 0.5951507687568665,
+      "eval_runtime": 11.8382,
+      "eval_samples_per_second": 40.547,
+      "eval_steps_per_second": 2.534,
       "step": 1050
     },
     {
       "epoch": 1.1240721102863203,
-      "grad_norm": 1.1883800029754639,
+      "grad_norm": 1.1941380500793457,
       "learning_rate": 3.126546482856133e-05,
-      "loss": 0.3353,
+      "loss": 0.3348,
       "step": 1060
     },
     {
       "epoch": 1.1346765641569458,
-      "grad_norm": 1.0962867736816406,
+      "grad_norm": 1.0945191383361816,
       "learning_rate": 3.108872393071757e-05,
-      "loss": 0.3726,
+      "loss": 0.3724,
       "step": 1070
     },
     {
       "epoch": 1.1452810180275717,
-      "grad_norm": 1.3056061267852783,
+      "grad_norm": 1.322505235671997,
       "learning_rate": 3.091198303287381e-05,
-      "loss": 0.3317,
+      "loss": 0.3312,
       "step": 1080
     },
     {
       "epoch": 1.1558854718981972,
-      "grad_norm": 1.1041945219039917,
+      "grad_norm": 1.0995663404464722,
       "learning_rate": 3.073524213503005e-05,
-      "loss": 0.3509,
+      "loss": 0.351,
       "step": 1090
     },
     {
       "epoch": 1.1664899257688228,
-      "grad_norm": 1.1359084844589233,
+      "grad_norm": 1.1347137689590454,
       "learning_rate": 3.0558501237186286e-05,
-      "loss": 0.3424,
+      "loss": 0.343,
       "step": 1100
     },
     {
       "epoch": 1.1664899257688228,
-      "eval_loss": 0.6033491492271423,
-      "eval_runtime": 11.8171,
-      "eval_samples_per_second": 40.619,
-      "eval_steps_per_second": 2.539,
+      "eval_loss": 0.59864342212677,
+      "eval_runtime": 11.8201,
+      "eval_samples_per_second": 40.609,
+      "eval_steps_per_second": 2.538,
       "step": 1100
     },
     {
       "epoch": 1.1770943796394486,
-      "grad_norm": 1.2678911685943604,
+      "grad_norm": 1.2727549076080322,
       "learning_rate": 3.0381760339342525e-05,
-      "loss": 0.3421,
+      "loss": 0.342,
       "step": 1110
     },
     {
       "epoch": 1.1876988335100742,
-      "grad_norm": 1.2048425674438477,
+      "grad_norm": 1.1988881826400757,
       "learning_rate": 3.0205019441498768e-05,
       "loss": 0.3238,
       "step": 1120
     },
     {
       "epoch": 1.1983032873806998,
-      "grad_norm": 1.120171308517456,
+      "grad_norm": 1.1220091581344604,
       "learning_rate": 3.0028278543655004e-05,
-      "loss": 0.328,
+      "loss": 0.3281,
       "step": 1130
     },
     {
       "epoch": 1.2089077412513256,
-      "grad_norm": 1.0923452377319336,
+      "grad_norm": 1.1029518842697144,
       "learning_rate": 2.985153764581124e-05,
-      "loss": 0.3567,
+      "loss": 0.3566,
       "step": 1140
     },
     {
       "epoch": 1.2195121951219512,
-      "grad_norm": 1.0582720041275024,
+      "grad_norm": 1.0658305883407593,
       "learning_rate": 2.9674796747967482e-05,
       "loss": 0.3431,
       "step": 1150
     },
     {
       "epoch": 1.2195121951219512,
-      "eval_loss": 0.6016908288002014,
-      "eval_runtime": 11.8175,
-      "eval_samples_per_second": 40.618,
-      "eval_steps_per_second": 2.539,
+      "eval_loss": 0.5974192023277283,
+      "eval_runtime": 11.8028,
+      "eval_samples_per_second": 40.668,
+      "eval_steps_per_second": 2.542,
       "step": 1150
     },
     {
       "epoch": 1.2301166489925768,
-      "grad_norm": 1.304275393486023,
+      "grad_norm": 1.306966781616211,
       "learning_rate": 2.9498055850123718e-05,
-      "loss": 0.3363,
+      "loss": 0.3355,
       "step": 1160
     },
     {
       "epoch": 1.2407211028632026,
-      "grad_norm": 1.3723812103271484,
+      "grad_norm": 1.370133638381958,
       "learning_rate": 2.932131495227996e-05,
-      "loss": 0.3487,
+      "loss": 0.3475,
       "step": 1170
     },
     {
       "epoch": 1.2513255567338282,
-      "grad_norm": 1.195857286453247,
+      "grad_norm": 1.2001569271087646,
       "learning_rate": 2.9144574054436196e-05,
-      "loss": 0.3313,
+      "loss": 0.331,
       "step": 1180
     },
     {
       "epoch": 1.2619300106044538,
-      "grad_norm": 1.2628227472305298,
+      "grad_norm": 1.27125883102417,
       "learning_rate": 2.896783315659244e-05,
-      "loss": 0.3469,
+      "loss": 0.3461,
       "step": 1190
     },
     {
       "epoch": 1.2725344644750796,
-      "grad_norm": 1.398179531097412,
+      "grad_norm": 1.417680025100708,
       "learning_rate": 2.8791092258748675e-05,
-      "loss": 0.3304,
+      "loss": 0.3298,
       "step": 1200
     },
     {
       "epoch": 1.2725344644750796,
-      "eval_loss": 0.6067756414413452,
-      "eval_runtime": 11.8179,
-      "eval_samples_per_second": 40.616,
-      "eval_steps_per_second": 2.539,
+      "eval_loss": 0.6026219129562378,
+      "eval_runtime": 11.8206,
+      "eval_samples_per_second": 40.607,
+      "eval_steps_per_second": 2.538,
       "step": 1200
     },
     {
       "epoch": 1.2831389183457051,
-      "grad_norm": 1.351558804512024,
+      "grad_norm": 1.3502525091171265,
       "learning_rate": 2.8614351360904917e-05,
-      "loss": 0.3349,
+      "loss": 0.3348,
       "step": 1210
     },
     {
       "epoch": 1.2937433722163307,
-      "grad_norm": 1.0469820499420166,
+      "grad_norm": 1.0501817464828491,
       "learning_rate": 2.8437610463061153e-05,
       "loss": 0.3253,
       "step": 1220
     },
     {
       "epoch": 1.3043478260869565,
-      "grad_norm": 1.205822467803955,
+      "grad_norm": 1.197717308998108,
       "learning_rate": 2.826086956521739e-05,
-      "loss": 0.3245,
+      "loss": 0.3251,
       "step": 1230
     },
     {
       "epoch": 1.3149522799575821,
-      "grad_norm": 1.1349923610687256,
+      "grad_norm": 1.1236610412597656,
       "learning_rate": 2.808412866737363e-05,
-      "loss": 0.3225,
+      "loss": 0.3229,
       "step": 1240
     },
     {
       "epoch": 1.3255567338282077,
-      "grad_norm": 1.2688688039779663,
+      "grad_norm": 1.172448754310608,
       "learning_rate": 2.790738776952987e-05,
-      "loss": 0.341,
+      "loss": 0.3408,
       "step": 1250
     },
     {
       "epoch": 1.3255567338282077,
-      "eval_loss": 0.6084305047988892,
-      "eval_runtime": 11.8165,
-      "eval_samples_per_second": 40.621,
-      "eval_steps_per_second": 2.539,
+      "eval_loss": 0.6030397415161133,
+      "eval_runtime": 11.8274,
+      "eval_samples_per_second": 40.584,
+      "eval_steps_per_second": 2.536,
       "step": 1250
     },
     {
       "epoch": 1.3361611876988335,
-      "grad_norm": 1.3045439720153809,
+      "grad_norm": 1.33034086227417,
       "learning_rate": 2.773064687168611e-05,
-      "loss": 0.3352,
+      "loss": 0.3351,
       "step": 1260
     },
     {
       "epoch": 1.346765641569459,
-      "grad_norm": 1.314754843711853,
+      "grad_norm": 1.3041718006134033,
       "learning_rate": 2.755390597384235e-05,
-      "loss": 0.3388,
+      "loss": 0.339,
       "step": 1270
     },
     {
       "epoch": 1.357370095440085,
-      "grad_norm": 1.3320605754852295,
+      "grad_norm": 1.3249897956848145,
       "learning_rate": 2.7377165075998588e-05,
-      "loss": 0.3353,
+      "loss": 0.3357,
       "step": 1280
     },
     {
       "epoch": 1.3679745493107105,
-      "grad_norm": 1.239930272102356,
+      "grad_norm": 1.2230838537216187,
       "learning_rate": 2.7200424178154827e-05,
-      "loss": 0.3447,
+      "loss": 0.3445,
       "step": 1290
     },
     {
       "epoch": 1.378579003181336,
-      "grad_norm": 1.227774739265442,
+      "grad_norm": 1.249778389930725,
       "learning_rate": 2.7023683280311063e-05,
-      "loss": 0.332,
+      "loss": 0.3321,
       "step": 1300
     },
     {
       "epoch": 1.378579003181336,
-      "eval_loss": 0.6087238192558289,
-      "eval_runtime": 11.8173,
-      "eval_samples_per_second": 40.618,
-      "eval_steps_per_second": 2.539,
+      "eval_loss": 0.6035539507865906,
+      "eval_runtime": 11.8513,
+      "eval_samples_per_second": 40.502,
+      "eval_steps_per_second": 2.531,
       "step": 1300
     },
     {
       "epoch": 1.389183457051962,
-      "grad_norm": 1.2262154817581177,
+      "grad_norm": 1.2172768115997314,
       "learning_rate": 2.6846942382467306e-05,
-      "loss": 0.326,
+      "loss": 0.3265,
       "step": 1310
     },
     {
       "epoch": 1.3997879109225875,
-      "grad_norm": 1.4143425226211548,
+      "grad_norm": 1.420792579650879,
       "learning_rate": 2.667020148462354e-05,
-      "loss": 0.329,
+      "loss": 0.3291,
       "step": 1320
     },
     {
       "epoch": 1.4103923647932133,
-      "grad_norm": 1.548729658126831,
+      "grad_norm": 1.5471575260162354,
       "learning_rate": 2.6493460586779784e-05,
-      "loss": 0.3286,
+      "loss": 0.3282,
       "step": 1330
     },
     {
       "epoch": 1.4209968186638389,
-      "grad_norm": 1.2641792297363281,
+      "grad_norm": 1.1769243478775024,
       "learning_rate": 2.631671968893602e-05,
-      "loss": 0.3134,
+      "loss": 0.314,
       "step": 1340
     },
     {
       "epoch": 1.4316012725344645,
-      "grad_norm": 1.302757978439331,
+      "grad_norm": 1.299428105354309,
       "learning_rate": 2.6139978791092262e-05,
-      "loss": 0.3056,
+      "loss": 0.3055,
       "step": 1350
     },
     {
       "epoch": 1.4316012725344645,
-      "eval_loss": 0.6126307249069214,
-      "eval_runtime": 11.796,
-      "eval_samples_per_second": 40.692,
-      "eval_steps_per_second": 2.543,
+      "eval_loss": 0.6074148416519165,
+      "eval_runtime": 11.8791,
+      "eval_samples_per_second": 40.407,
+      "eval_steps_per_second": 2.525,
       "step": 1350
     },
     {
       "epoch": 1.4422057264050903,
-      "grad_norm": 1.2385451793670654,
+      "grad_norm": 1.2306607961654663,
       "learning_rate": 2.5963237893248498e-05,
-      "loss": 0.3352,
+      "loss": 0.3355,
       "step": 1360
     },
     {
       "epoch": 1.4528101802757158,
-      "grad_norm": 1.226826786994934,
+      "grad_norm": 1.236029863357544,
       "learning_rate": 2.5786496995404734e-05,
-      "loss": 0.3293,
+      "loss": 0.3299,
       "step": 1370
     },
     {
       "epoch": 1.4634146341463414,
-      "grad_norm": 1.1468507051467896,
+      "grad_norm": 1.1422163248062134,
       "learning_rate": 2.5609756097560977e-05,
       "loss": 0.3271,
       "step": 1380
     },
     {
       "epoch": 1.4740190880169672,
-      "grad_norm": 1.4662965536117554,
+      "grad_norm": 1.4650191068649292,
       "learning_rate": 2.5433015199717212e-05,
-      "loss": 0.3285,
+      "loss": 0.3298,
       "step": 1390
     },
     {
       "epoch": 1.4846235418875928,
-      "grad_norm": 1.3143432140350342,
+      "grad_norm": 1.3182368278503418,
       "learning_rate": 2.5256274301873455e-05,
-      "loss": 0.3271,
+      "loss": 0.3273,
       "step": 1400
     },
     {
       "epoch": 1.4846235418875928,
-      "eval_loss": 0.6134337186813354,
-      "eval_runtime": 11.777,
-      "eval_samples_per_second": 40.757,
-      "eval_steps_per_second": 2.547,
+      "eval_loss": 0.6101020574569702,
+      "eval_runtime": 11.8749,
+      "eval_samples_per_second": 40.421,
+      "eval_steps_per_second": 2.526,
       "step": 1400
     },
     {
       "epoch": 1.4952279957582184,
-      "grad_norm": 1.2217501401901245,
+      "grad_norm": 1.2115826606750488,
       "learning_rate": 2.507953340402969e-05,
-      "loss": 0.3093,
+      "loss": 0.3087,
       "step": 1410
     },
     {
       "epoch": 1.5058324496288442,
-      "grad_norm": 1.479797124862671,
+      "grad_norm": 1.4292011260986328,
       "learning_rate": 2.4902792506185933e-05,
-      "loss": 0.3044,
+      "loss": 0.3048,
       "step": 1420
     },
     {
       "epoch": 1.5164369034994698,
-      "grad_norm": 1.2942014932632446,
+      "grad_norm": 1.291587471961975,
       "learning_rate": 2.4726051608342173e-05,
-      "loss": 0.3279,
+      "loss": 0.3277,
       "step": 1430
     },
     {
       "epoch": 1.5270413573700954,
-      "grad_norm": 1.194472074508667,
+      "grad_norm": 1.21229088306427,
       "learning_rate": 2.4549310710498412e-05,
-      "loss": 0.3124,
+      "loss": 0.3119,
       "step": 1440
     },
     {
       "epoch": 1.5376458112407212,
-      "grad_norm": 1.3158146142959595,
+      "grad_norm": 1.3125064373016357,
       "learning_rate": 2.437256981265465e-05,
-      "loss": 0.3174,
+      "loss": 0.3179,
       "step": 1450
     },
     {
       "epoch": 1.5376458112407212,
-      "eval_loss": 0.612784206867218,
-      "eval_runtime": 11.7301,
-      "eval_samples_per_second": 40.92,
-      "eval_steps_per_second": 2.558,
+      "eval_loss": 0.6085401773452759,
+      "eval_runtime": 11.8623,
+      "eval_samples_per_second": 40.464,
+      "eval_steps_per_second": 2.529,
       "step": 1450
     },
     {
       "epoch": 1.5482502651113468,
-      "grad_norm": 1.2772356271743774,
+      "grad_norm": 1.279478907585144,
       "learning_rate": 2.419582891481089e-05,
       "loss": 0.3233,
       "step": 1460
     },
     {
       "epoch": 1.5588547189819724,
-      "grad_norm": 1.2930830717086792,
+      "grad_norm": 1.2751704454421997,
       "learning_rate": 2.401908801696713e-05,
-      "loss": 0.3193,
+      "loss": 0.3192,
       "step": 1470
     },
     {
       "epoch": 1.5694591728525982,
-      "grad_norm": 1.3090952634811401,
+      "grad_norm": 1.305484652519226,
       "learning_rate": 2.3842347119123365e-05,
-      "loss": 0.3169,
+      "loss": 0.3175,
       "step": 1480
     },
     {
       "epoch": 1.5800636267232238,
-      "grad_norm": 1.2926676273345947,
+      "grad_norm": 1.276348352432251,
       "learning_rate": 2.3665606221279604e-05,
-      "loss": 0.3046,
+      "loss": 0.3043,
       "step": 1490
     },
     {
       "epoch": 1.5906680805938493,
-      "grad_norm": 1.358901023864746,
+      "grad_norm": 1.3701218366622925,
       "learning_rate": 2.3488865323435844e-05,
-      "loss": 0.318,
+      "loss": 0.3185,
       "step": 1500
     },
     {
       "epoch": 1.5906680805938493,
-      "eval_loss": 0.6163415908813477,
-      "eval_runtime": 11.7301,
-      "eval_samples_per_second": 40.92,
-      "eval_steps_per_second": 2.558,
+      "eval_loss": 0.6127668619155884,
+      "eval_runtime": 11.8543,
+      "eval_samples_per_second": 40.492,
+      "eval_steps_per_second": 2.531,
       "step": 1500
     },
     {
       "epoch": 1.6012725344644752,
-      "grad_norm": 1.4036508798599243,
+      "grad_norm": 1.39030122756958,
       "learning_rate": 2.3312124425592083e-05,
-      "loss": 0.3163,
+      "loss": 0.316,
       "step": 1510
     },
     {
       "epoch": 1.6118769883351007,
-      "grad_norm": 1.2453142404556274,
+      "grad_norm": 1.2342959642410278,
       "learning_rate": 2.3135383527748322e-05,
-      "loss": 0.2934,
+      "loss": 0.2928,
       "step": 1520
     },
     {
       "epoch": 1.6224814422057263,
-      "grad_norm": 1.5171256065368652,
+      "grad_norm": 1.546506404876709,
       "learning_rate": 2.295864262990456e-05,
-      "loss": 0.3186,
+      "loss": 0.3195,
       "step": 1530
     },
     {
       "epoch": 1.6330858960763521,
-      "grad_norm": 1.4850494861602783,
+      "grad_norm": 1.483469009399414,
       "learning_rate": 2.27819017320608e-05,
-      "loss": 0.3206,
+      "loss": 0.3208,
       "step": 1540
     },
     {
       "epoch": 1.6436903499469777,
-      "grad_norm": 1.475541353225708,
+      "grad_norm": 1.4679921865463257,
       "learning_rate": 2.260516083421704e-05,
-      "loss": 0.3218,
+      "loss": 0.3217,
       "step": 1550
     },
     {
       "epoch": 1.6436903499469777,
-      "eval_loss": 0.616825520992279,
-      "eval_runtime": 11.7336,
-      "eval_samples_per_second": 40.908,
-      "eval_steps_per_second": 2.557,
+      "eval_loss": 0.6129384636878967,
+      "eval_runtime": 11.8494,
+      "eval_samples_per_second": 40.508,
+      "eval_steps_per_second": 2.532,
       "step": 1550
     },
     {
       "epoch": 1.6542948038176033,
-      "grad_norm": 1.401365041732788,
+      "grad_norm": 1.4147011041641235,
       "learning_rate": 2.2428419936373275e-05,
-      "loss": 0.3423,
+      "loss": 0.3424,
       "step": 1560
     },
     {
       "epoch": 1.664899257688229,
-      "grad_norm": 1.2514119148254395,
+      "grad_norm": 1.2510008811950684,
       "learning_rate": 2.2251679038529514e-05,
-      "loss": 0.3146,
+      "loss": 0.3153,
       "step": 1570
     },
     {
       "epoch": 1.6755037115588547,
-      "grad_norm": 1.2970105409622192,
+      "grad_norm": 1.3089756965637207,
       "learning_rate": 2.2074938140685754e-05,
-      "loss": 0.3229,
+      "loss": 0.3235,
       "step": 1580
     },
     {
       "epoch": 1.6861081654294803,
-      "grad_norm": 1.398328185081482,
+      "grad_norm": 1.4045295715332031,
       "learning_rate": 2.1898197242841996e-05,
       "loss": 0.3194,
       "step": 1590
     },
     {
       "epoch": 1.696712619300106,
-      "grad_norm": 1.5763297080993652,
+      "grad_norm": 1.552812099456787,
       "learning_rate": 2.1721456344998235e-05,
-      "loss": 0.3169,
+      "loss": 0.3172,
       "step": 1600
     },
     {
       "epoch": 1.696712619300106,
-      "eval_loss": 0.6218429207801819,
-      "eval_runtime": 11.7407,
-      "eval_samples_per_second": 40.883,
-      "eval_steps_per_second": 2.555,
+      "eval_loss": 0.6166751980781555,
+      "eval_runtime": 11.8313,
+      "eval_samples_per_second": 40.57,
+      "eval_steps_per_second": 2.536,
       "step": 1600
     }
   ],

checkpoint-1600/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:872835407be2a0a95214d9cc3b7966660b3e25cbfda85f13ed9adad96c071793
+oid sha256:53a8c021c0523ab617f7d132f6e1d21f2075e97fdb7689e840960f38ef5cb15e
 size 5880