Training in progress, step 100000, checkpoint

checkpoint-100000/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a9d86e93883eb6d9a7d934e9d633f9ab1defad75e9d7ff1b2a1d2b10413bb7dc
+oid sha256:6d4d26485b6ed327d57df9f4d760544f21a09d022aa883b2d2a95b250937b184
 size 51007160

checkpoint-100000/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ee9e9b6b0cea88241797fa7178e8f017cc455d8d492f2405d32d95bafa25fb1c
+oid sha256:845f5b24aacb670446648033ee980df3778450d2877e1a36bdd116856bba4bd6
 size 102078202

checkpoint-100000/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2f44d3719722389c44e71ab43f2ad39b614d48138139b6fb77bca1f23caa7a15
+oid sha256:2a2636ac949507f96df73d1312bb88fdedc1bb70cdb75f3d49b25985bd439da1
 size 14308

checkpoint-100000/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ea11ef7fecdb8225bbd08827b61fefeb249d82daf67dcc74289d58dadd10a06a
+oid sha256:140bfad5bb944679e52f54db724e3c3699f36a0988c8abe780b4e5c75feff49c
 size 1000

checkpoint-100000/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 3.9348273277282715,
-  "best_model_checkpoint": "/home/p318482/babyLM_controlled/models_trained/fr_clm/childes_30/checkpoint-28000",
-  "epoch": 157.23270440251574,
+  "best_metric": 4.583011150360107,
+  "best_model_checkpoint": "/home/p318482/babyLM_controlled/models_trained/de_clm/childes_30/checkpoint-32000",
+  "epoch": 104.82180293501048,
   "eval_steps": 2000,
   "global_step": 100000,
   "is_hyper_param_search": false,
@@ -9,585 +9,585 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 3.1446540880503147,
-      "eval_loss": 6.67229700088501,
-      "eval_runtime": 0.6568,
-      "eval_samples_per_second": 1386.938,
-      "eval_steps_per_second": 86.779,
+      "epoch": 2.0964360587002098,
+      "eval_loss": 7.102903366088867,
+      "eval_runtime": 0.9708,
+      "eval_samples_per_second": 1416.286,
+      "eval_steps_per_second": 88.582,
       "step": 2000
     },
     {
-      "epoch": 6.289308176100629,
-      "grad_norm": 1.4630517959594727,
+      "epoch": 4.1928721174004195,
+      "grad_norm": 1.3964662551879883,
       "learning_rate": 1e-05,
-      "loss": 6.5359,
+      "loss": 6.9987,
       "step": 4000
     },
     {
-      "epoch": 6.289308176100629,
-      "eval_loss": 5.102311611175537,
-      "eval_runtime": 0.6383,
-      "eval_samples_per_second": 1427.13,
-      "eval_steps_per_second": 89.294,
+      "epoch": 4.1928721174004195,
+      "eval_loss": 5.884151935577393,
+      "eval_runtime": 0.966,
+      "eval_samples_per_second": 1423.408,
+      "eval_steps_per_second": 89.028,
       "step": 4000
     },
     {
-      "epoch": 9.433962264150944,
-      "eval_loss": 4.675341606140137,
-      "eval_runtime": 0.6396,
-      "eval_samples_per_second": 1424.427,
-      "eval_steps_per_second": 89.124,
+      "epoch": 6.289308176100629,
+      "eval_loss": 5.54873514175415,
+      "eval_runtime": 0.9657,
+      "eval_samples_per_second": 1423.84,
+      "eval_steps_per_second": 89.055,
       "step": 6000
     },
     {
-      "epoch": 12.578616352201259,
-      "grad_norm": 2.3778915405273438,
+      "epoch": 8.385744234800839,
+      "grad_norm": 2.7172107696533203,
       "learning_rate": 2e-05,
-      "loss": 4.34,
+      "loss": 5.2204,
       "step": 8000
     },
     {
-      "epoch": 12.578616352201259,
-      "eval_loss": 4.42986536026001,
-      "eval_runtime": 0.646,
-      "eval_samples_per_second": 1410.229,
-      "eval_steps_per_second": 88.236,
+      "epoch": 8.385744234800839,
+      "eval_loss": 5.2793288230896,
+      "eval_runtime": 0.9644,
+      "eval_samples_per_second": 1425.779,
+      "eval_steps_per_second": 89.176,
       "step": 8000
     },
     {
-      "epoch": 15.723270440251572,
-      "eval_loss": 4.2789154052734375,
-      "eval_runtime": 0.6485,
-      "eval_samples_per_second": 1404.856,
-      "eval_steps_per_second": 87.9,
+      "epoch": 10.482180293501049,
+      "eval_loss": 5.10486364364624,
+      "eval_runtime": 0.9641,
+      "eval_samples_per_second": 1426.204,
+      "eval_steps_per_second": 89.203,
       "step": 10000
     },
     {
-      "epoch": 18.867924528301888,
-      "grad_norm": 2.512312889099121,
-      "learning_rate": 2.9995e-05,
-      "loss": 3.8692,
+      "epoch": 12.578616352201259,
+      "grad_norm": 2.500443458557129,
+      "learning_rate": 2.99925e-05,
+      "loss": 4.7358,
       "step": 12000
     },
     {
-      "epoch": 18.867924528301888,
-      "eval_loss": 4.1580729484558105,
-      "eval_runtime": 0.6403,
-      "eval_samples_per_second": 1422.769,
-      "eval_steps_per_second": 89.021,
+      "epoch": 12.578616352201259,
+      "eval_loss": 4.983631134033203,
+      "eval_runtime": 0.9644,
+      "eval_samples_per_second": 1425.809,
+      "eval_steps_per_second": 89.178,
       "step": 12000
     },
     {
-      "epoch": 22.0125786163522,
-      "eval_loss": 4.082061767578125,
-      "eval_runtime": 0.6415,
-      "eval_samples_per_second": 1420.041,
-      "eval_steps_per_second": 88.85,
+      "epoch": 14.675052410901468,
+      "eval_loss": 4.882917404174805,
+      "eval_runtime": 0.9686,
+      "eval_samples_per_second": 1419.612,
+      "eval_steps_per_second": 88.79,
       "step": 14000
     },
     {
-      "epoch": 25.157232704402517,
-      "grad_norm": 2.4205257892608643,
+      "epoch": 16.771488469601678,
+      "grad_norm": 2.400749444961548,
       "learning_rate": 3.999e-05,
-      "loss": 3.5905,
+      "loss": 4.4216,
       "step": 16000
     },
     {
-      "epoch": 25.157232704402517,
-      "eval_loss": 4.033827781677246,
-      "eval_runtime": 0.6416,
-      "eval_samples_per_second": 1419.823,
-      "eval_steps_per_second": 88.836,
+      "epoch": 16.771488469601678,
+      "eval_loss": 4.802889823913574,
+      "eval_runtime": 0.9763,
+      "eval_samples_per_second": 1408.393,
+      "eval_steps_per_second": 88.089,
       "step": 16000
     },
     {
-      "epoch": 28.30188679245283,
-      "eval_loss": 3.9935402870178223,
-      "eval_runtime": 0.6614,
-      "eval_samples_per_second": 1377.343,
-      "eval_steps_per_second": 86.178,
+      "epoch": 18.867924528301888,
+      "eval_loss": 4.74226188659668,
+      "eval_runtime": 0.976,
+      "eval_samples_per_second": 1408.882,
+      "eval_steps_per_second": 88.119,
       "step": 18000
     },
     {
-      "epoch": 31.446540880503143,
-      "grad_norm": 2.5674731731414795,
+      "epoch": 20.964360587002098,
+      "grad_norm": 2.2613022327423096,
       "learning_rate": 4.9985e-05,
-      "loss": 3.3778,
+      "loss": 4.1842,
       "step": 20000
     },
     {
-      "epoch": 31.446540880503143,
-      "eval_loss": 3.9675967693328857,
-      "eval_runtime": 0.6406,
-      "eval_samples_per_second": 1422.059,
-      "eval_steps_per_second": 88.976,
+      "epoch": 20.964360587002098,
+      "eval_loss": 4.690371513366699,
+      "eval_runtime": 0.9669,
+      "eval_samples_per_second": 1422.037,
+      "eval_steps_per_second": 88.942,
       "step": 20000
     },
     {
-      "epoch": 34.59119496855346,
-      "eval_loss": 3.9494004249572754,
-      "eval_runtime": 0.6409,
-      "eval_samples_per_second": 1421.391,
-      "eval_steps_per_second": 88.934,
+      "epoch": 23.060796645702307,
+      "eval_loss": 4.645771503448486,
+      "eval_runtime": 0.966,
+      "eval_samples_per_second": 1423.404,
+      "eval_steps_per_second": 89.027,
       "step": 22000
     },
     {
-      "epoch": 37.735849056603776,
-      "grad_norm": 2.273855447769165,
+      "epoch": 25.157232704402517,
+      "grad_norm": 2.2588465213775635,
       "learning_rate": 5.9980000000000005e-05,
-      "loss": 3.1983,
+      "loss": 3.9858,
       "step": 24000
     },
     {
-      "epoch": 37.735849056603776,
-      "eval_loss": 3.9382567405700684,
-      "eval_runtime": 0.6423,
-      "eval_samples_per_second": 1418.386,
-      "eval_steps_per_second": 88.746,
+      "epoch": 25.157232704402517,
+      "eval_loss": 4.623382568359375,
+      "eval_runtime": 0.9666,
+      "eval_samples_per_second": 1422.538,
+      "eval_steps_per_second": 88.973,
       "step": 24000
     },
     {
-      "epoch": 40.880503144654085,
-      "eval_loss": 3.936068296432495,
-      "eval_runtime": 0.6434,
-      "eval_samples_per_second": 1415.972,
-      "eval_steps_per_second": 88.595,
+      "epoch": 27.253668763102727,
+      "eval_loss": 4.6056084632873535,
+      "eval_runtime": 0.9825,
+      "eval_samples_per_second": 1399.422,
+      "eval_steps_per_second": 87.527,
       "step": 26000
     },
     {
-      "epoch": 44.0251572327044,
-      "grad_norm": 2.377107858657837,
+      "epoch": 29.350104821802937,
+      "grad_norm": 2.1845760345458984,
       "learning_rate": 6.997500000000001e-05,
-      "loss": 3.0374,
+      "loss": 3.8189,
       "step": 28000
     },
     {
-      "epoch": 44.0251572327044,
-      "eval_loss": 3.9348273277282715,
-      "eval_runtime": 0.6383,
-      "eval_samples_per_second": 1427.128,
-      "eval_steps_per_second": 89.293,
+      "epoch": 29.350104821802937,
+      "eval_loss": 4.590851783752441,
+      "eval_runtime": 1.0367,
+      "eval_samples_per_second": 1326.293,
+      "eval_steps_per_second": 82.954,
       "step": 28000
     },
     {
-      "epoch": 47.16981132075472,
-      "eval_loss": 3.9625139236450195,
-      "eval_runtime": 0.6461,
-      "eval_samples_per_second": 1409.936,
-      "eval_steps_per_second": 88.218,
+      "epoch": 31.446540880503143,
+      "eval_loss": 4.586838245391846,
+      "eval_runtime": 0.9705,
+      "eval_samples_per_second": 1416.743,
+      "eval_steps_per_second": 88.611,
       "step": 30000
     },
     {
-      "epoch": 50.314465408805034,
-      "grad_norm": 2.460918426513672,
+      "epoch": 33.542976939203356,
+      "grad_norm": 2.2118289470672607,
       "learning_rate": 7.997e-05,
-      "loss": 2.887,
+      "loss": 3.6763,
       "step": 32000
     },
     {
-      "epoch": 50.314465408805034,
-      "eval_loss": 3.9899542331695557,
-      "eval_runtime": 0.6383,
-      "eval_samples_per_second": 1427.149,
-      "eval_steps_per_second": 89.295,
+      "epoch": 33.542976939203356,
+      "eval_loss": 4.583011150360107,
+      "eval_runtime": 0.9706,
+      "eval_samples_per_second": 1416.604,
+      "eval_steps_per_second": 88.602,
       "step": 32000
     },
     {
-      "epoch": 53.459119496855344,
-      "eval_loss": 4.022701740264893,
-      "eval_runtime": 0.6416,
-      "eval_samples_per_second": 1419.9,
-      "eval_steps_per_second": 88.841,
+      "epoch": 35.63941299790356,
+      "eval_loss": 4.57816743850708,
+      "eval_runtime": 0.9657,
+      "eval_samples_per_second": 1423.818,
+      "eval_steps_per_second": 89.053,
       "step": 34000
     },
     {
-      "epoch": 56.60377358490566,
-      "grad_norm": 2.467155694961548,
+      "epoch": 37.735849056603776,
+      "grad_norm": 2.189404010772705,
       "learning_rate": 8.9965e-05,
-      "loss": 2.7426,
+      "loss": 3.5493,
       "step": 36000
     },
     {
-      "epoch": 56.60377358490566,
-      "eval_loss": 4.057339191436768,
-      "eval_runtime": 0.6431,
-      "eval_samples_per_second": 1416.6,
-      "eval_steps_per_second": 88.635,
+      "epoch": 37.735849056603776,
+      "eval_loss": 4.585381031036377,
+      "eval_runtime": 0.9726,
+      "eval_samples_per_second": 1413.783,
+      "eval_steps_per_second": 88.426,
       "step": 36000
     },
     {
-      "epoch": 59.74842767295598,
-      "eval_loss": 4.104013442993164,
-      "eval_runtime": 0.6393,
-      "eval_samples_per_second": 1424.887,
-      "eval_steps_per_second": 89.153,
+      "epoch": 39.83228511530398,
+      "eval_loss": 4.596414566040039,
+      "eval_runtime": 0.971,
+      "eval_samples_per_second": 1416.025,
+      "eval_steps_per_second": 88.566,
       "step": 38000
     },
     {
-      "epoch": 62.893081761006286,
-      "grad_norm": 2.5848844051361084,
+      "epoch": 41.928721174004195,
+      "grad_norm": 2.1832594871520996,
       "learning_rate": 9.996000000000001e-05,
-      "loss": 2.5994,
+      "loss": 3.4327,
       "step": 40000
     },
     {
-      "epoch": 62.893081761006286,
-      "eval_loss": 4.14555549621582,
-      "eval_runtime": 0.6397,
-      "eval_samples_per_second": 1424.163,
-      "eval_steps_per_second": 89.108,
+      "epoch": 41.928721174004195,
+      "eval_loss": 4.610367774963379,
+      "eval_runtime": 0.9687,
+      "eval_samples_per_second": 1419.473,
+      "eval_steps_per_second": 88.782,
       "step": 40000
     },
     {
-      "epoch": 66.0377358490566,
-      "eval_loss": 4.226268768310547,
-      "eval_runtime": 0.6407,
-      "eval_samples_per_second": 1421.964,
-      "eval_steps_per_second": 88.97,
+      "epoch": 44.0251572327044,
+      "eval_loss": 4.636893272399902,
+      "eval_runtime": 0.9651,
+      "eval_samples_per_second": 1424.742,
+      "eval_steps_per_second": 89.111,
       "step": 42000
     },
     {
-      "epoch": 69.18238993710692,
-      "grad_norm": 2.974116325378418,
+      "epoch": 46.121593291404615,
+      "grad_norm": 2.1686387062072754,
       "learning_rate": 9.336333333333334e-05,
-      "loss": 2.4407,
+      "loss": 3.3112,
       "step": 44000
     },
     {
-      "epoch": 69.18238993710692,
-      "eval_loss": 4.292743682861328,
-      "eval_runtime": 0.6391,
-      "eval_samples_per_second": 1425.506,
-      "eval_steps_per_second": 89.192,
+      "epoch": 46.121593291404615,
+      "eval_loss": 4.66969108581543,
+      "eval_runtime": 0.9642,
+      "eval_samples_per_second": 1426.082,
+      "eval_steps_per_second": 89.195,
       "step": 44000
     },
     {
-      "epoch": 72.32704402515724,
-      "eval_loss": 4.376113414764404,
-      "eval_runtime": 0.639,
-      "eval_samples_per_second": 1425.622,
-      "eval_steps_per_second": 89.199,
+      "epoch": 48.21802935010482,
+      "eval_loss": 4.69525146484375,
+      "eval_runtime": 0.966,
+      "eval_samples_per_second": 1423.408,
+      "eval_steps_per_second": 89.028,
       "step": 46000
     },
     {
-      "epoch": 75.47169811320755,
-      "grad_norm": 3.167375087738037,
+      "epoch": 50.314465408805034,
+      "grad_norm": 2.3040497303009033,
       "learning_rate": 8.67e-05,
-      "loss": 2.2824,
+      "loss": 3.1908,
       "step": 48000
     },
     {
-      "epoch": 75.47169811320755,
-      "eval_loss": 4.446501731872559,
-      "eval_runtime": 0.6391,
-      "eval_samples_per_second": 1425.43,
-      "eval_steps_per_second": 89.187,
+      "epoch": 50.314465408805034,
+      "eval_loss": 4.727965831756592,
+      "eval_runtime": 0.9829,
+      "eval_samples_per_second": 1398.863,
+      "eval_steps_per_second": 87.493,
       "step": 48000
     },
     {
-      "epoch": 78.61635220125787,
-      "eval_loss": 4.518404960632324,
-      "eval_runtime": 0.6421,
-      "eval_samples_per_second": 1418.867,
-      "eval_steps_per_second": 88.777,
+      "epoch": 52.41090146750524,
+      "eval_loss": 4.762918472290039,
+      "eval_runtime": 1.0048,
+      "eval_samples_per_second": 1368.397,
+      "eval_steps_per_second": 85.587,
       "step": 50000
     },
     {
-      "epoch": 81.76100628930817,
-      "grad_norm": 3.280864715576172,
+      "epoch": 54.507337526205454,
+      "grad_norm": 2.491647720336914,
       "learning_rate": 8.003666666666667e-05,
-      "loss": 2.1444,
+      "loss": 3.0857,
       "step": 52000
     },
     {
-      "epoch": 81.76100628930817,
-      "eval_loss": 4.594879150390625,
-      "eval_runtime": 0.6404,
-      "eval_samples_per_second": 1422.632,
-      "eval_steps_per_second": 89.012,
+      "epoch": 54.507337526205454,
+      "eval_loss": 4.792750835418701,
+      "eval_runtime": 0.9654,
+      "eval_samples_per_second": 1424.262,
+      "eval_steps_per_second": 89.081,
       "step": 52000
     },
     {
-      "epoch": 84.90566037735849,
-      "eval_loss": 4.654795169830322,
-      "eval_runtime": 0.6434,
-      "eval_samples_per_second": 1415.83,
-      "eval_steps_per_second": 88.586,
+      "epoch": 56.60377358490566,
+      "eval_loss": 4.819573879241943,
+      "eval_runtime": 0.9679,
+      "eval_samples_per_second": 1420.61,
+      "eval_steps_per_second": 88.853,
       "step": 54000
     },
     {
-      "epoch": 88.0503144654088,
-      "grad_norm": 3.498023509979248,
+      "epoch": 58.700209643605874,
+      "grad_norm": 2.5555057525634766,
       "learning_rate": 7.337333333333334e-05,
-      "loss": 2.0229,
+      "loss": 2.9936,
       "step": 56000
     },
     {
-      "epoch": 88.0503144654088,
-      "eval_loss": 4.731740474700928,
-      "eval_runtime": 0.6412,
-      "eval_samples_per_second": 1420.839,
-      "eval_steps_per_second": 88.9,
+      "epoch": 58.700209643605874,
+      "eval_loss": 4.8564348220825195,
+      "eval_runtime": 0.9696,
+      "eval_samples_per_second": 1418.079,
+      "eval_steps_per_second": 88.694,
       "step": 56000
     },
     {
-      "epoch": 91.19496855345912,
-      "eval_loss": 4.804233074188232,
-      "eval_runtime": 0.6438,
-      "eval_samples_per_second": 1415.109,
-      "eval_steps_per_second": 88.541,
+      "epoch": 60.79664570230608,
+      "eval_loss": 4.8890299797058105,
+      "eval_runtime": 0.9965,
+      "eval_samples_per_second": 1379.79,
+      "eval_steps_per_second": 86.3,
       "step": 58000
     },
     {
-      "epoch": 94.33962264150944,
-      "grad_norm": 3.759833335876465,
+      "epoch": 62.893081761006286,
+      "grad_norm": 2.6957740783691406,
       "learning_rate": 6.671e-05,
-      "loss": 1.9145,
+      "loss": 2.9113,
       "step": 60000
     },
     {
-      "epoch": 94.33962264150944,
-      "eval_loss": 4.86856746673584,
-      "eval_runtime": 0.6444,
-      "eval_samples_per_second": 1413.806,
-      "eval_steps_per_second": 88.46,
+      "epoch": 62.893081761006286,
+      "eval_loss": 4.919983386993408,
+      "eval_runtime": 0.9752,
+      "eval_samples_per_second": 1409.972,
+      "eval_steps_per_second": 88.187,
       "step": 60000
     },
     {
-      "epoch": 97.48427672955975,
-      "eval_loss": 4.92185640335083,
-      "eval_runtime": 0.6399,
-      "eval_samples_per_second": 1423.69,
-      "eval_steps_per_second": 89.078,
+      "epoch": 64.98951781970649,
+      "eval_loss": 4.953897476196289,
+      "eval_runtime": 0.9752,
+      "eval_samples_per_second": 1409.961,
+      "eval_steps_per_second": 88.187,
       "step": 62000
     },
     {
-      "epoch": 100.62893081761007,
-      "grad_norm": 3.9873037338256836,
-      "learning_rate": 6.004833333333334e-05,
-      "loss": 1.8236,
+      "epoch": 67.08595387840671,
+      "grad_norm": 2.9121437072753906,
+      "learning_rate": 6.0046666666666676e-05,
+      "loss": 2.8353,
       "step": 64000
     },
     {
-      "epoch": 100.62893081761007,
-      "eval_loss": 4.991932392120361,
-      "eval_runtime": 0.6395,
-      "eval_samples_per_second": 1424.605,
-      "eval_steps_per_second": 89.136,
+      "epoch": 67.08595387840671,
+      "eval_loss": 4.993447303771973,
+      "eval_runtime": 0.9908,
+      "eval_samples_per_second": 1387.752,
+      "eval_steps_per_second": 86.798,
       "step": 64000
     },
     {
-      "epoch": 103.77358490566037,
-      "eval_loss": 5.044597625732422,
-      "eval_runtime": 0.6483,
-      "eval_samples_per_second": 1405.309,
-      "eval_steps_per_second": 87.928,
+      "epoch": 69.18238993710692,
+      "eval_loss": 5.029664993286133,
+      "eval_runtime": 0.9686,
+      "eval_samples_per_second": 1419.546,
+      "eval_steps_per_second": 88.786,
       "step": 66000
     },
     {
-      "epoch": 106.91823899371069,
-      "grad_norm": 4.110333442687988,
-      "learning_rate": 5.3385000000000005e-05,
-      "loss": 1.7439,
+      "epoch": 71.27882599580713,
+      "grad_norm": 2.9727697372436523,
+      "learning_rate": 5.338333333333334e-05,
+      "loss": 2.7673,
       "step": 68000
     },
     {
-      "epoch": 106.91823899371069,
-      "eval_loss": 5.092119216918945,
-      "eval_runtime": 0.6381,
-      "eval_samples_per_second": 1427.662,
-      "eval_steps_per_second": 89.327,
+      "epoch": 71.27882599580713,
+      "eval_loss": 5.061018943786621,
+      "eval_runtime": 0.976,
+      "eval_samples_per_second": 1408.755,
+      "eval_steps_per_second": 88.111,
       "step": 68000
     },
     {
-      "epoch": 110.062893081761,
-      "eval_loss": 5.164552688598633,
-      "eval_runtime": 0.6391,
-      "eval_samples_per_second": 1425.493,
-      "eval_steps_per_second": 89.191,
+      "epoch": 73.37526205450733,
+      "eval_loss": 5.080495834350586,
+      "eval_runtime": 0.9731,
+      "eval_samples_per_second": 1413.053,
+      "eval_steps_per_second": 88.38,
       "step": 70000
     },
     {
-      "epoch": 113.20754716981132,
-      "grad_norm": 4.256406784057617,
+      "epoch": 75.47169811320755,
+      "grad_norm": 3.07655930519104,
       "learning_rate": 4.672e-05,
-      "loss": 1.6698,
+      "loss": 2.7091,
       "step": 72000
     },
     {
-      "epoch": 113.20754716981132,
-      "eval_loss": 5.21901273727417,
-      "eval_runtime": 0.6465,
-      "eval_samples_per_second": 1409.16,
-      "eval_steps_per_second": 88.169,
+      "epoch": 75.47169811320755,
+      "eval_loss": 5.105360984802246,
+      "eval_runtime": 0.9653,
+      "eval_samples_per_second": 1424.397,
+      "eval_steps_per_second": 89.09,
       "step": 72000
     },
     {
-      "epoch": 116.35220125786164,
-      "eval_loss": 5.2676005363464355,
-      "eval_runtime": 0.6376,
-      "eval_samples_per_second": 1428.823,
-      "eval_steps_per_second": 89.399,
+      "epoch": 77.56813417190776,
+      "eval_loss": 5.1282548904418945,
+      "eval_runtime": 0.9738,
+      "eval_samples_per_second": 1412.066,
+      "eval_steps_per_second": 88.318,
       "step": 74000
     },
     {
-      "epoch": 119.49685534591195,
-      "grad_norm": 4.310237884521484,
+      "epoch": 79.66457023060796,
+      "grad_norm": 3.3577957153320312,
       "learning_rate": 4.005666666666667e-05,
-      "loss": 1.6065,
+      "loss": 2.6563,
       "step": 76000
     },
     {
-      "epoch": 119.49685534591195,
-      "eval_loss": 5.309641361236572,
-      "eval_runtime": 0.6416,
-      "eval_samples_per_second": 1419.883,
-      "eval_steps_per_second": 88.84,
+      "epoch": 79.66457023060796,
+      "eval_loss": 5.159351348876953,
+      "eval_runtime": 0.9725,
+      "eval_samples_per_second": 1413.916,
+      "eval_steps_per_second": 88.434,
       "step": 76000
     },
     {
-      "epoch": 122.64150943396227,
-      "eval_loss": 5.348501682281494,
-      "eval_runtime": 0.641,
-      "eval_samples_per_second": 1421.238,
-      "eval_steps_per_second": 88.925,
+      "epoch": 81.76100628930817,
+      "eval_loss": 5.183588027954102,
+      "eval_runtime": 0.9642,
+      "eval_samples_per_second": 1426.112,
+      "eval_steps_per_second": 89.197,
       "step": 78000
     },
     {
-      "epoch": 125.78616352201257,
-      "grad_norm": 4.249606609344482,
+      "epoch": 83.85744234800839,
+      "grad_norm": 3.3268089294433594,
       "learning_rate": 3.339333333333334e-05,
-      "loss": 1.55,
+      "loss": 2.6077,
       "step": 80000
     },
     {
-      "epoch": 125.78616352201257,
-      "eval_loss": 5.3839430809021,
-      "eval_runtime": 0.6402,
-      "eval_samples_per_second": 1422.97,
-      "eval_steps_per_second": 89.033,
+      "epoch": 83.85744234800839,
+      "eval_loss": 5.200902462005615,
+      "eval_runtime": 0.963,
+      "eval_samples_per_second": 1427.864,
+      "eval_steps_per_second": 89.306,
       "step": 80000
     },
     {
-      "epoch": 128.9308176100629,
-      "eval_loss": 5.418987274169922,
-      "eval_runtime": 0.6394,
-      "eval_samples_per_second": 1424.833,
-      "eval_steps_per_second": 89.15,
+      "epoch": 85.9538784067086,
+      "eval_loss": 5.2230424880981445,
+      "eval_runtime": 0.9661,
+      "eval_samples_per_second": 1423.259,
+      "eval_steps_per_second": 89.018,
       "step": 82000
     },
     {
-      "epoch": 132.0754716981132,
-      "grad_norm": 4.494133472442627,
-      "learning_rate": 2.6728333333333333e-05,
-      "loss": 1.4991,
+      "epoch": 88.0503144654088,
+      "grad_norm": 3.465266704559326,
+      "learning_rate": 2.673e-05,
+      "loss": 2.5635,
       "step": 84000
     },
     {
-      "epoch": 132.0754716981132,
-      "eval_loss": 5.464737415313721,
-      "eval_runtime": 0.646,
-      "eval_samples_per_second": 1410.197,
-      "eval_steps_per_second": 88.234,
+      "epoch": 88.0503144654088,
+      "eval_loss": 5.244382381439209,
+      "eval_runtime": 0.9696,
+      "eval_samples_per_second": 1418.144,
+      "eval_steps_per_second": 88.698,
       "step": 84000
     },
     {
-      "epoch": 135.22012578616352,
-      "eval_loss": 5.502508640289307,
-      "eval_runtime": 0.6383,
-      "eval_samples_per_second": 1427.206,
-      "eval_steps_per_second": 89.298,
+      "epoch": 90.14675052410901,
+      "eval_loss": 5.2631378173828125,
+      "eval_runtime": 0.962,
+      "eval_samples_per_second": 1429.319,
+      "eval_steps_per_second": 89.397,
       "step": 86000
     },
     {
-      "epoch": 138.36477987421384,
-      "grad_norm": 4.311509609222412,
-      "learning_rate": 2.0065000000000002e-05,
-      "loss": 1.4526,
+      "epoch": 92.24318658280923,
+      "grad_norm": 3.63198184967041,
+      "learning_rate": 2.0066666666666665e-05,
+      "loss": 2.5229,
       "step": 88000
     },
     {
-      "epoch": 138.36477987421384,
-      "eval_loss": 5.533146858215332,
-      "eval_runtime": 0.6381,
-      "eval_samples_per_second": 1427.706,
-      "eval_steps_per_second": 89.33,
+      "epoch": 92.24318658280923,
+      "eval_loss": 5.279770374298096,
+      "eval_runtime": 0.9687,
+      "eval_samples_per_second": 1419.359,
+      "eval_steps_per_second": 88.774,
       "step": 88000
     },
     {
-      "epoch": 141.50943396226415,
-      "eval_loss": 5.56229829788208,
-      "eval_runtime": 0.6677,
-      "eval_samples_per_second": 1364.444,
-      "eval_steps_per_second": 85.371,
+      "epoch": 94.33962264150944,
+      "eval_loss": 5.295066833496094,
+      "eval_runtime": 0.9675,
+      "eval_samples_per_second": 1421.221,
+      "eval_steps_per_second": 88.891,
       "step": 90000
     },
     {
-      "epoch": 144.65408805031447,
-      "grad_norm": 4.507097244262695,
-      "learning_rate": 1.3401666666666667e-05,
-      "loss": 1.4134,
+      "epoch": 96.43605870020964,
+      "grad_norm": 3.696744203567505,
+      "learning_rate": 1.3403333333333334e-05,
+      "loss": 2.4886,
       "step": 92000
     },
     {
-      "epoch": 144.65408805031447,
-      "eval_loss": 5.574535846710205,
-      "eval_runtime": 0.6418,
-      "eval_samples_per_second": 1419.424,
-      "eval_steps_per_second": 88.811,
+      "epoch": 96.43605870020964,
+      "eval_loss": 5.310117721557617,
+      "eval_runtime": 0.9707,
+      "eval_samples_per_second": 1416.474,
+      "eval_steps_per_second": 88.594,
       "step": 92000
     },
     {
-      "epoch": 147.7987421383648,
-      "eval_loss": 5.601788520812988,
-      "eval_runtime": 0.6372,
-      "eval_samples_per_second": 1429.781,
-      "eval_steps_per_second": 89.459,
+      "epoch": 98.53249475890985,
+      "eval_loss": 5.318916320800781,
+      "eval_runtime": 0.9679,
+      "eval_samples_per_second": 1420.653,
+      "eval_steps_per_second": 88.855,
       "step": 94000
     },
     {
-      "epoch": 150.9433962264151,
-      "grad_norm": 4.453996181488037,
-      "learning_rate": 6.738333333333334e-06,
-      "loss": 1.3793,
+      "epoch": 100.62893081761007,
+      "grad_norm": 3.7400856018066406,
+      "learning_rate": 6.740000000000001e-06,
+      "loss": 2.4584,
       "step": 96000
     },
     {
-      "epoch": 150.9433962264151,
-      "eval_loss": 5.608455181121826,
-      "eval_runtime": 0.6397,
-      "eval_samples_per_second": 1424.187,
-      "eval_steps_per_second": 89.109,
+      "epoch": 100.62893081761007,
+      "eval_loss": 5.329978942871094,
+      "eval_runtime": 0.9671,
+      "eval_samples_per_second": 1421.725,
+      "eval_steps_per_second": 88.922,
       "step": 96000
     },
     {
-      "epoch": 154.08805031446542,
-      "eval_loss": 5.6233811378479,
-      "eval_runtime": 0.64,
-      "eval_samples_per_second": 1423.385,
-      "eval_steps_per_second": 89.059,
+      "epoch": 102.72536687631028,
+      "eval_loss": 5.332658767700195,
+      "eval_runtime": 0.9676,
+      "eval_samples_per_second": 1420.979,
+      "eval_steps_per_second": 88.876,
       "step": 98000
     },
     {
-      "epoch": 157.23270440251574,
-      "grad_norm": 4.54045295715332,
-      "learning_rate": 7.500000000000001e-08,
-      "loss": 1.3512,
+      "epoch": 104.82180293501048,
+      "grad_norm": 3.5724282264709473,
+      "learning_rate": 7.666666666666668e-08,
+      "loss": 2.4337,
       "step": 100000
     },
     {
-      "epoch": 157.23270440251574,
-      "eval_loss": 5.628963470458984,
-      "eval_runtime": 0.6422,
-      "eval_samples_per_second": 1418.625,
-      "eval_steps_per_second": 88.761,
+      "epoch": 104.82180293501048,
+      "eval_loss": 5.3366007804870605,
+      "eval_runtime": 0.9697,
+      "eval_samples_per_second": 1417.91,
+      "eval_steps_per_second": 88.684,
       "step": 100000
     }
   ],
   "logging_steps": 4000,
   "max_steps": 100000,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 158,
+  "num_train_epochs": 105,
   "save_steps": 4000,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -601,7 +601,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 2.5854087677362176e+16,
+  "total_flos": 2.5856622838726656e+16,
   "train_batch_size": 16,
   "trial_name": null,
   "trial_params": null

checkpoint-100000/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6af9de70859d62cbd3b61f71e8a5bc95702dbc6ddb62d2994641a31953e4ea9b
+oid sha256:0c779122676b54107edc62ae0b9293c062733193c5f82a36a2bc097bca192814
 size 5368