End of training

README.md

@@ -16,9 +16,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # sft_mc
 
-This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on the None dataset.
+This model is a fine-tuned version of [meta-llama/Meta-Llama-3-8B-Instruct](https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct) on the identity and the data_mc datasets.
 It achieves the following results on the evaluation set:
-- Loss: 2.3016
+- Loss: 2.3011
 
 ## Model description
 

all_results.json

@@ -1,12 +1,12 @@
 {
     "epoch": 10.0,
-    "eval_loss": 2.3273847103118896,
-    "eval_runtime": 0.9866,
-    "eval_samples_per_second": 22.299,
-    "eval_steps_per_second": 3.041,
+    "eval_loss": 2.3011481761932373,
+    "eval_runtime": 0.9906,
+    "eval_samples_per_second": 22.21,
+    "eval_steps_per_second": 3.029,
     "total_flos": 51054080163840.0,
-    "train_loss": 0.28787647419011414,
-    "train_runtime": 2129.2799,
-    "train_samples_per_second": 5.02,
-    "train_steps_per_second": 0.315
+    "train_loss": 0.28853574658174125,
+    "train_runtime": 2679.9258,
+    "train_samples_per_second": 3.989,
+    "train_steps_per_second": 0.25
 }

eval_results.json

@@ -1,7 +1,7 @@
 {
     "epoch": 10.0,
-    "eval_loss": 2.3273847103118896,
-    "eval_runtime": 0.9866,
-    "eval_samples_per_second": 22.299,
-    "eval_steps_per_second": 3.041
+    "eval_loss": 2.3011481761932373,
+    "eval_runtime": 0.9906,
+    "eval_samples_per_second": 22.21,
+    "eval_steps_per_second": 3.029
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 10.0,
     "total_flos": 51054080163840.0,
-    "train_loss": 0.28787647419011414,
-    "train_runtime": 2129.2799,
-    "train_samples_per_second": 5.02,
-    "train_steps_per_second": 0.315
+    "train_loss": 0.28853574658174125,
+    "train_runtime": 2679.9258,
+    "train_samples_per_second": 3.989,
+    "train_steps_per_second": 0.25
 }

trainer_state.json

@@ -10,585 +10,585 @@
   "log_history": [
     {
       "epoch": 0.14925373134328357,
-      "grad_norm": 35.90125075274422,
+      "grad_norm": 35.596871044953,
       "learning_rate": 1.4925373134328358e-06,
-      "loss": 1.921,
+      "loss": 1.9213,
       "step": 10
     },
     {
       "epoch": 0.29850746268656714,
-      "grad_norm": 21.611909970996276,
+      "grad_norm": 20.609759250234525,
       "learning_rate": 2.9850746268656716e-06,
-      "loss": 1.1511,
+      "loss": 1.151,
       "step": 20
     },
     {
       "epoch": 0.44776119402985076,
-      "grad_norm": 17.54986781724882,
+      "grad_norm": 17.51541675553714,
       "learning_rate": 4.477611940298508e-06,
-      "loss": 0.9965,
+      "loss": 0.9953,
       "step": 30
     },
     {
       "epoch": 0.5970149253731343,
-      "grad_norm": 12.86435917115879,
+      "grad_norm": 12.529056819699033,
       "learning_rate": 5.970149253731343e-06,
-      "loss": 1.0133,
+      "loss": 1.0143,
       "step": 40
     },
     {
       "epoch": 0.746268656716418,
-      "grad_norm": 12.804255629397026,
+      "grad_norm": 12.891928164600921,
       "learning_rate": 7.46268656716418e-06,
-      "loss": 1.053,
+      "loss": 1.0534,
       "step": 50
     },
     {
       "epoch": 0.746268656716418,
-      "eval_loss": 1.2667393684387207,
-      "eval_runtime": 0.9884,
-      "eval_samples_per_second": 22.259,
-      "eval_steps_per_second": 3.035,
+      "eval_loss": 1.2635215520858765,
+      "eval_runtime": 0.9888,
+      "eval_samples_per_second": 22.25,
+      "eval_steps_per_second": 3.034,
       "step": 50
     },
     {
       "epoch": 0.8955223880597015,
-      "grad_norm": 13.275489475520203,
+      "grad_norm": 13.18139035026108,
       "learning_rate": 8.955223880597016e-06,
-      "loss": 1.0093,
+      "loss": 1.0096,
       "step": 60
     },
     {
       "epoch": 1.044776119402985,
-      "grad_norm": 10.227125744078247,
+      "grad_norm": 10.740181523405264,
       "learning_rate": 9.999389284703265e-06,
-      "loss": 0.9774,
+      "loss": 0.9688,
       "step": 70
     },
     {
       "epoch": 1.1940298507462686,
-      "grad_norm": 10.433290314777752,
+      "grad_norm": 10.70923590419392,
       "learning_rate": 9.988536273658876e-06,
-      "loss": 0.7151,
+      "loss": 0.711,
       "step": 80
     },
     {
       "epoch": 1.3432835820895521,
-      "grad_norm": 10.688663360597172,
+      "grad_norm": 11.258468673854201,
       "learning_rate": 9.964145714351633e-06,
-      "loss": 0.7292,
+      "loss": 0.7263,
       "step": 90
     },
     {
       "epoch": 1.4925373134328357,
-      "grad_norm": 14.363632603556267,
+      "grad_norm": 13.780566009646858,
       "learning_rate": 9.926283796211796e-06,
-      "loss": 0.8178,
+      "loss": 0.8118,
       "step": 100
     },
     {
       "epoch": 1.4925373134328357,
-      "eval_loss": 1.3229804039001465,
-      "eval_runtime": 0.9845,
-      "eval_samples_per_second": 22.346,
-      "eval_steps_per_second": 3.047,
+      "eval_loss": 1.3804577589035034,
+      "eval_runtime": 0.9931,
+      "eval_samples_per_second": 22.152,
+      "eval_steps_per_second": 3.021,
       "step": 100
     },
     {
       "epoch": 1.6417910447761193,
-      "grad_norm": 10.45476204806484,
+      "grad_norm": 9.18658078765986,
       "learning_rate": 9.87505326632108e-06,
-      "loss": 0.8059,
+      "loss": 0.8126,
       "step": 110
     },
     {
       "epoch": 1.7910447761194028,
-      "grad_norm": 9.865796302977193,
+      "grad_norm": 10.993066748048813,
       "learning_rate": 9.810593150584658e-06,
-      "loss": 0.8046,
+      "loss": 0.8099,
       "step": 120
     },
     {
       "epoch": 1.9402985074626866,
-      "grad_norm": 9.16685167044992,
+      "grad_norm": 8.891553991106202,
       "learning_rate": 9.733078376452172e-06,
-      "loss": 0.7796,
+      "loss": 0.804,
       "step": 130
     },
     {
       "epoch": 2.08955223880597,
-      "grad_norm": 8.577950674483414,
+      "grad_norm": 8.268316338255294,
       "learning_rate": 9.642719298211602e-06,
-      "loss": 0.5014,
+      "loss": 0.4978,
       "step": 140
     },
     {
       "epoch": 2.2388059701492535,
-      "grad_norm": 9.564526041646893,
+      "grad_norm": 8.48854187035418,
       "learning_rate": 9.539761126144193e-06,
-      "loss": 0.3803,
+      "loss": 0.3889,
       "step": 150
     },
     {
       "epoch": 2.2388059701492535,
-      "eval_loss": 1.5490957498550415,
-      "eval_runtime": 0.9879,
-      "eval_samples_per_second": 22.27,
-      "eval_steps_per_second": 3.037,
+      "eval_loss": 1.6006965637207031,
+      "eval_runtime": 0.9927,
+      "eval_samples_per_second": 22.162,
+      "eval_steps_per_second": 3.022,
       "step": 150
     },
     {
       "epoch": 2.388059701492537,
-      "grad_norm": 6.6979250460920055,
+      "grad_norm": 6.609934011045634,
       "learning_rate": 9.424483261089584e-06,
-      "loss": 0.4,
+      "loss": 0.3846,
       "step": 160
     },
     {
       "epoch": 2.5373134328358207,
-      "grad_norm": 7.411435193851966,
+      "grad_norm": 7.50682908737009,
       "learning_rate": 9.297198536226927e-06,
-      "loss": 0.4118,
+      "loss": 0.4125,
       "step": 170
     },
     {
       "epoch": 2.6865671641791042,
-      "grad_norm": 8.596999087873268,
+      "grad_norm": 8.821510256598781,
       "learning_rate": 9.158252368129628e-06,
-      "loss": 0.4074,
+      "loss": 0.4261,
       "step": 180
     },
     {
       "epoch": 2.835820895522388,
-      "grad_norm": 7.821138995440488,
+      "grad_norm": 7.416255028966862,
       "learning_rate": 9.008021819397488e-06,
-      "loss": 0.438,
+      "loss": 0.4488,
       "step": 190
     },
     {
       "epoch": 2.9850746268656714,
-      "grad_norm": 8.079644227754834,
+      "grad_norm": 8.233603986842171,
       "learning_rate": 8.846914575410035e-06,
-      "loss": 0.436,
+      "loss": 0.4361,
       "step": 200
     },
     {
       "epoch": 2.9850746268656714,
-      "eval_loss": 1.5159422159194946,
-      "eval_runtime": 0.9922,
-      "eval_samples_per_second": 22.173,
-      "eval_steps_per_second": 3.024,
+      "eval_loss": 1.532719612121582,
+      "eval_runtime": 0.9841,
+      "eval_samples_per_second": 22.354,
+      "eval_steps_per_second": 3.048,
       "step": 200
     },
     {
       "epoch": 3.1343283582089554,
-      "grad_norm": 7.1129105753354125,
+      "grad_norm": 8.824258454974311,
       "learning_rate": 8.675367837977848e-06,
-      "loss": 0.2353,
+      "loss": 0.2399,
       "step": 210
     },
     {
       "epoch": 3.283582089552239,
-      "grad_norm": 7.345277153510642,
+      "grad_norm": 7.063462594451812,
       "learning_rate": 8.49384713889421e-06,
-      "loss": 0.2442,
+      "loss": 0.2246,
       "step": 220
     },
     {
       "epoch": 3.4328358208955225,
-      "grad_norm": 9.63151530312459,
+      "grad_norm": 7.331660049583603,
       "learning_rate": 8.302845076606786e-06,
-      "loss": 0.264,
+      "loss": 0.2415,
       "step": 230
     },
     {
       "epoch": 3.582089552238806,
-      "grad_norm": 6.18137195728966,
+      "grad_norm": 6.424375821012275,
       "learning_rate": 8.10287997943769e-06,
-      "loss": 0.2486,
+      "loss": 0.2615,
       "step": 240
     },
     {
       "epoch": 3.7313432835820897,
-      "grad_norm": 7.558744812430462,
+      "grad_norm": 9.043879070705094,
       "learning_rate": 7.894494498979558e-06,
-      "loss": 0.2529,
+      "loss": 0.265,
       "step": 250
     },
     {
       "epoch": 3.7313432835820897,
-      "eval_loss": 1.6716222763061523,
-      "eval_runtime": 0.9844,
-      "eval_samples_per_second": 22.349,
-      "eval_steps_per_second": 3.048,
+      "eval_loss": 1.6067496538162231,
+      "eval_runtime": 0.99,
+      "eval_samples_per_second": 22.223,
+      "eval_steps_per_second": 3.03,
       "step": 250
     },
     {
       "epoch": 3.8805970149253732,
-      "grad_norm": 7.589509313359255,
+      "grad_norm": 7.232856290764463,
       "learning_rate": 7.678254137484797e-06,
-      "loss": 0.2382,
+      "loss": 0.2371,
       "step": 260
     },
     {
       "epoch": 4.029850746268656,
-      "grad_norm": 4.292329479116102,
+      "grad_norm": 3.885849230377616,
       "learning_rate": 7.4547457132442895e-06,
-      "loss": 0.2137,
+      "loss": 0.2234,
       "step": 270
     },
     {
       "epoch": 4.17910447761194,
-      "grad_norm": 4.93005317817384,
+      "grad_norm": 8.201942272896055,
       "learning_rate": 7.2245757681200835e-06,
-      "loss": 0.1201,
+      "loss": 0.1273,
       "step": 280
     },
     {
       "epoch": 4.3283582089552235,
-      "grad_norm": 6.340158286084005,
+      "grad_norm": 6.742950931031516,
       "learning_rate": 6.988368921553601e-06,
-      "loss": 0.1246,
+      "loss": 0.1232,
       "step": 290
     },
     {
       "epoch": 4.477611940298507,
-      "grad_norm": 7.035143744951474,
+      "grad_norm": 7.18067667554869,
       "learning_rate": 6.746766175516159e-06,
-      "loss": 0.1555,
+      "loss": 0.1347,
       "step": 300
     },
     {
       "epoch": 4.477611940298507,
-      "eval_loss": 1.847908616065979,
-      "eval_runtime": 0.9862,
-      "eval_samples_per_second": 22.307,
-      "eval_steps_per_second": 3.042,
+      "eval_loss": 1.8177189826965332,
+      "eval_runtime": 0.9881,
+      "eval_samples_per_second": 22.265,
+      "eval_steps_per_second": 3.036,
       "step": 300
     },
     {
       "epoch": 4.6268656716417915,
-      "grad_norm": 5.665676328156013,
+      "grad_norm": 4.134448034439449,
       "learning_rate": 6.500423175001705e-06,
-      "loss": 0.1391,
+      "loss": 0.1326,
       "step": 310
     },
     {
       "epoch": 4.776119402985074,
-      "grad_norm": 6.086585562332805,
+      "grad_norm": 5.984462599094163,
       "learning_rate": 6.2500084287822925e-06,
-      "loss": 0.1576,
+      "loss": 0.1424,
       "step": 320
     },
     {
       "epoch": 4.925373134328359,
-      "grad_norm": 5.964957579344617,
+      "grad_norm": 7.603214916490548,
       "learning_rate": 5.996201495254757e-06,
-      "loss": 0.1441,
+      "loss": 0.1489,
       "step": 330
     },
     {
       "epoch": 5.074626865671641,
-      "grad_norm": 2.2272515089974716,
+      "grad_norm": 4.174840343630945,
       "learning_rate": 5.73969113830165e-06,
-      "loss": 0.0898,
+      "loss": 0.0922,
       "step": 340
     },
     {
       "epoch": 5.223880597014926,
-      "grad_norm": 4.927338192771822,
+      "grad_norm": 4.1502544333430995,
       "learning_rate": 5.481173458170952e-06,
-      "loss": 0.078,
+      "loss": 0.0857,
       "step": 350
     },
     {
       "epoch": 5.223880597014926,
-      "eval_loss": 2.03751277923584,
-      "eval_runtime": 0.989,
-      "eval_samples_per_second": 22.245,
-      "eval_steps_per_second": 3.033,
+      "eval_loss": 1.977053165435791,
+      "eval_runtime": 0.9844,
+      "eval_samples_per_second": 22.349,
+      "eval_steps_per_second": 3.048,
       "step": 350
     },
     {
       "epoch": 5.373134328358209,
-      "grad_norm": 4.519214108822612,
+      "grad_norm": 6.862487683858245,
       "learning_rate": 5.221350002446882e-06,
-      "loss": 0.0828,
+      "loss": 0.0723,
       "step": 360
     },
     {
       "epoch": 5.522388059701493,
-      "grad_norm": 3.298936360185895,
+      "grad_norm": 4.515093540059101,
       "learning_rate": 4.96092586223808e-06,
-      "loss": 0.0734,
+      "loss": 0.0693,
       "step": 370
     },
     {
       "epoch": 5.6716417910447765,
-      "grad_norm": 6.294477283779623,
+      "grad_norm": 5.571199681266628,
       "learning_rate": 4.700607758749626e-06,
-      "loss": 0.0873,
+      "loss": 0.0978,
       "step": 380
     },
     {
       "epoch": 5.82089552238806,
-      "grad_norm": 4.501734589626586,
+      "grad_norm": 4.718390811556805,
       "learning_rate": 4.441102125431398e-06,
-      "loss": 0.0647,
+      "loss": 0.0792,
       "step": 390
     },
     {
       "epoch": 5.970149253731344,
-      "grad_norm": 6.080468816741591,
+      "grad_norm": 3.7184414288353422,
       "learning_rate": 4.183113190907349e-06,
-      "loss": 0.0726,
+      "loss": 0.0709,
       "step": 400
     },
     {
       "epoch": 5.970149253731344,
-      "eval_loss": 1.8965630531311035,
-      "eval_runtime": 0.9838,
-      "eval_samples_per_second": 22.363,
-      "eval_steps_per_second": 3.049,
+      "eval_loss": 1.9007922410964966,
+      "eval_runtime": 0.9935,
+      "eval_samples_per_second": 22.145,
+      "eval_steps_per_second": 3.02,
       "step": 400
     },
     {
       "epoch": 6.119402985074627,
-      "grad_norm": 3.2916344262433874,
+      "grad_norm": 3.9943470043032603,
       "learning_rate": 3.927341067888065e-06,
-      "loss": 0.0471,
+      "loss": 0.0309,
       "step": 410
     },
     {
       "epoch": 6.268656716417911,
-      "grad_norm": 3.847276551426768,
+      "grad_norm": 2.122106221543816,
       "learning_rate": 3.6744798532528137e-06,
-      "loss": 0.0332,
+      "loss": 0.0318,
       "step": 420
     },
     {
       "epoch": 6.417910447761194,
-      "grad_norm": 2.2055620155467532,
+      "grad_norm": 5.125755419214468,
       "learning_rate": 3.4252157444569478e-06,
-      "loss": 0.0302,
+      "loss": 0.0388,
       "step": 430
     },
     {
       "epoch": 6.567164179104478,
-      "grad_norm": 7.540352868440021,
+      "grad_norm": 3.6926779235153173,
       "learning_rate": 3.1802251773762294e-06,
-      "loss": 0.048,
+      "loss": 0.0549,
       "step": 440
     },
     {
       "epoch": 6.7164179104477615,
-      "grad_norm": 3.3252060274641644,
+      "grad_norm": 3.2890545865060345,
       "learning_rate": 2.9401729906414385e-06,
-      "loss": 0.0411,
+      "loss": 0.0474,
       "step": 450
     },
     {
       "epoch": 6.7164179104477615,
-      "eval_loss": 2.1223175525665283,
-      "eval_runtime": 0.9929,
-      "eval_samples_per_second": 22.157,
-      "eval_steps_per_second": 3.021,
+      "eval_loss": 2.131742238998413,
+      "eval_runtime": 0.9906,
+      "eval_samples_per_second": 22.208,
+      "eval_steps_per_second": 3.028,
       "step": 450
     },
     {
       "epoch": 6.865671641791045,
-      "grad_norm": 1.3769658356628893,
+      "grad_norm": 2.1811442357317303,
       "learning_rate": 2.7057106214448216e-06,
-      "loss": 0.0388,
+      "loss": 0.0502,
       "step": 460
     },
     {
       "epoch": 7.014925373134329,
-      "grad_norm": 0.5978691678502914,
+      "grad_norm": 1.2788531958401737,
       "learning_rate": 2.4774743377144265e-06,
-      "loss": 0.0395,
+      "loss": 0.0228,
       "step": 470
     },
     {
       "epoch": 7.164179104477612,
-      "grad_norm": 0.6955205479840721,
+      "grad_norm": 5.299430914553839,
       "learning_rate": 2.256083511453747e-06,
-      "loss": 0.0106,
+      "loss": 0.0145,
       "step": 480
     },
     {
       "epoch": 7.313432835820896,
-      "grad_norm": 5.0523793409280495,
+      "grad_norm": 2.916687972196201,
       "learning_rate": 2.042138937932388e-06,
-      "loss": 0.0252,
+      "loss": 0.019,
       "step": 490
     },
     {
       "epoch": 7.462686567164179,
-      "grad_norm": 1.442828007683923,
+      "grad_norm": 0.1327510881497756,
       "learning_rate": 1.8362212052889827e-06,
-      "loss": 0.0127,
+      "loss": 0.0286,
       "step": 500
     },
     {
       "epoch": 7.462686567164179,
-      "eval_loss": 2.1807539463043213,
-      "eval_runtime": 0.9861,
-      "eval_samples_per_second": 22.31,
-      "eval_steps_per_second": 3.042,
+      "eval_loss": 2.2198660373687744,
+      "eval_runtime": 0.9905,
+      "eval_samples_per_second": 22.21,
+      "eval_steps_per_second": 3.029,
       "step": 500
     },
     {
       "epoch": 7.611940298507463,
-      "grad_norm": 3.231146614611382,
+      "grad_norm": 2.6883922474481623,
       "learning_rate": 1.63888911897084e-06,
-      "loss": 0.017,
+      "loss": 0.0177,
       "step": 510
     },
     {
       "epoch": 7.7611940298507465,
-      "grad_norm": 0.9690088812134732,
+      "grad_norm": 2.30912712845238,
       "learning_rate": 1.4506781852859836e-06,
-      "loss": 0.0192,
+      "loss": 0.0247,
       "step": 520
     },
     {
       "epoch": 7.91044776119403,
-      "grad_norm": 0.15130005364803242,
+      "grad_norm": 0.9713338532013991,
       "learning_rate": 1.2720991581827852e-06,
-      "loss": 0.0102,
+      "loss": 0.0129,
       "step": 530
     },
     {
       "epoch": 8.059701492537313,
-      "grad_norm": 0.06663781297906822,
+      "grad_norm": 0.21817578611988062,
       "learning_rate": 1.1036366532008552e-06,
-      "loss": 0.0119,
+      "loss": 0.0118,
       "step": 540
     },
     {
       "epoch": 8.208955223880597,
-      "grad_norm": 0.904607664250542,
+      "grad_norm": 0.9407606204330533,
       "learning_rate": 9.457478323545749e-07,
-      "loss": 0.0083,
+      "loss": 0.0091,
       "step": 550
     },
     {
       "epoch": 8.208955223880597,
-      "eval_loss": 2.2392308712005615,
-      "eval_runtime": 0.9906,
-      "eval_samples_per_second": 22.209,
-      "eval_steps_per_second": 3.029,
+      "eval_loss": 2.2086477279663086,
+      "eval_runtime": 0.9909,
+      "eval_samples_per_second": 22.202,
+      "eval_steps_per_second": 3.028,
       "step": 550
     },
     {
       "epoch": 8.35820895522388,
-      "grad_norm": 0.4615723348056157,
+      "grad_norm": 0.49920720779292554,
       "learning_rate": 7.988611635181099e-07,
-      "loss": 0.0059,
+      "loss": 0.0065,
       "step": 560
     },
     {
       "epoch": 8.507462686567164,
-      "grad_norm": 0.4272952034809819,
+      "grad_norm": 0.4774032163274442,
       "learning_rate": 6.633752576786251e-07,
-      "loss": 0.012,
+      "loss": 0.0104,
       "step": 570
     },
     {
       "epoch": 8.656716417910447,
-      "grad_norm": 0.5145191101097174,
+      "grad_norm": 0.5916059752548383,
       "learning_rate": 5.396577872130676e-07,
-      "loss": 0.0071,
+      "loss": 0.006,
       "step": 580
     },
     {
       "epoch": 8.805970149253731,
-      "grad_norm": 0.12341593830842175,
+      "grad_norm": 0.2966687083499466,
       "learning_rate": 4.2804448812404754e-07,
-      "loss": 0.0036,
+      "loss": 0.006,
       "step": 590
     },
     {
       "epoch": 8.955223880597014,
-      "grad_norm": 0.7095059409594088,
+      "grad_norm": 0.7007110357286702,
       "learning_rate": 3.288382489424502e-07,
-      "loss": 0.0049,
+      "loss": 0.0054,
       "step": 600
     },
     {
       "epoch": 8.955223880597014,
-      "eval_loss": 2.3008365631103516,
-      "eval_runtime": 0.9934,
-      "eval_samples_per_second": 22.146,
-      "eval_steps_per_second": 3.02,
+      "eval_loss": 2.2865116596221924,
+      "eval_runtime": 0.9893,
+      "eval_samples_per_second": 22.239,
+      "eval_steps_per_second": 3.033,
       "step": 600
     },
     {
       "epoch": 9.104477611940299,
-      "grad_norm": 1.4842604841678675,
+      "grad_norm": 1.48416033550746,
       "learning_rate": 2.4230828876927293e-07,
-      "loss": 0.0051,
+      "loss": 0.006,
       "step": 610
     },
     {
       "epoch": 9.253731343283581,
-      "grad_norm": 1.0220912222856482,
+      "grad_norm": 1.0330803943081195,
       "learning_rate": 1.6868942668726408e-07,
-      "loss": 0.0032,
+      "loss": 0.0033,
       "step": 620
     },
     {
       "epoch": 9.402985074626866,
-      "grad_norm": 0.8790141026127328,
+      "grad_norm": 0.9190962016297656,
       "learning_rate": 1.0818144452496293e-07,
-      "loss": 0.0063,
+      "loss": 0.0064,
       "step": 630
     },
     {
       "epoch": 9.552238805970148,
-      "grad_norm": 0.04185510117018065,
+      "grad_norm": 0.03995791087011875,
       "learning_rate": 6.094854470245326e-08,
-      "loss": 0.0024,
+      "loss": 0.0023,
       "step": 640
     },
     {
       "epoch": 9.701492537313433,
-      "grad_norm": 0.05125416943229584,
+      "grad_norm": 0.04136050995455753,
       "learning_rate": 2.711890463007405e-08,
-      "loss": 0.0035,
+      "loss": 0.0038,
       "step": 650
     },
     {
       "epoch": 9.701492537313433,
-      "eval_loss": 2.3232414722442627,
-      "eval_runtime": 0.9838,
-      "eval_samples_per_second": 22.362,
-      "eval_steps_per_second": 3.049,
+      "eval_loss": 2.301581621170044,
+      "eval_runtime": 0.9856,
+      "eval_samples_per_second": 22.322,
+      "eval_steps_per_second": 3.044,
       "step": 650
     },
     {
       "epoch": 9.850746268656717,
-      "grad_norm": 0.8613851942534393,
+      "grad_norm": 0.875079509506855,
       "learning_rate": 6.784328869339218e-09,
-      "loss": 0.0035,
+      "loss": 0.0037,
       "step": 660
     },
     {
       "epoch": 10.0,
-      "grad_norm": 0.1820002975830766,
+      "grad_norm": 0.27404452578171096,
       "learning_rate": 0.0,
-      "loss": 0.0021,
+      "loss": 0.0034,
       "step": 670
     },
     {
       "epoch": 10.0,
       "step": 670,
       "total_flos": 51054080163840.0,
-      "train_loss": 0.28787647419011414,
-      "train_runtime": 2129.2799,
-      "train_samples_per_second": 5.02,
-      "train_steps_per_second": 0.315
+      "train_loss": 0.28853574658174125,
+      "train_runtime": 2679.9258,
+      "train_samples_per_second": 3.989,
+      "train_steps_per_second": 0.25
     }
   ],
   "logging_steps": 10,