🍻 cheers

all_results.json

@@ -1,12 +1,12 @@
 {
     "epoch": 7.0,
-    "eval_accuracy": 0.6993638060991633,
-    "eval_loss": 1.0771288871765137,
-    "eval_runtime": 155.7222,
-    "eval_samples_per_second": 185.728,
-    "eval_steps_per_second": 23.221,
-    "train_loss": 1.079674410547284,
-    "train_runtime": 23712.848,
-    "train_samples_per_second": 68.302,
-    "train_steps_per_second": 2.135
+    "eval_accuracy": 0.6974621395477492,
+    "eval_loss": 1.0670689344406128,
+    "eval_runtime": 150.7024,
+    "eval_samples_per_second": 191.915,
+    "eval_steps_per_second": 23.994,
+    "train_loss": 1.126858214650741,
+    "train_runtime": 24456.4619,
+    "train_samples_per_second": 66.225,
+    "train_steps_per_second": 2.07
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 7.0,
-    "eval_accuracy": 0.6993638060991633,
-    "eval_loss": 1.0771288871765137,
-    "eval_runtime": 155.7222,
-    "eval_samples_per_second": 185.728,
-    "eval_steps_per_second": 23.221
+    "eval_accuracy": 0.6974621395477492,
+    "eval_loss": 1.0670689344406128,
+    "eval_runtime": 150.7024,
+    "eval_samples_per_second": 191.915,
+    "eval_steps_per_second": 23.994
 }

runs/Jun01_08-09-44_adrian-development/events.out.tfevents.1685624418.adrian-development.8928.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:56322dc9a143e5f266501aba5a58b69b3da1373c8b732c8711b04686aed248cf
+size 369

train_results.json

@@ -1,7 +1,7 @@
 {
     "epoch": 7.0,
-    "train_loss": 1.079674410547284,
-    "train_runtime": 23712.848,
-    "train_samples_per_second": 68.302,
-    "train_steps_per_second": 2.135
+    "train_loss": 1.126858214650741,
+    "train_runtime": 24456.4619,
+    "train_samples_per_second": 66.225,
+    "train_steps_per_second": 2.07
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": 1.0771288871765137,
-  "best_model_checkpoint": "Transformers/vit-base-clothing-leafs-example-full-simple/checkpoint-36000",
+  "best_metric": 1.0670689344406128,
+  "best_model_checkpoint": "Transformers/vit-base-clothing-leafs-example-full-simple/checkpoint-47000",
   "epoch": 7.0,
   "global_step": 50617,
   "is_hyper_param_search": false,
@@ -10,61 +10,61 @@
     {
       "epoch": 0.14,
       "learning_rate": 4.9012189580575696e-06,
-      "loss": 2.7949,
+      "loss": 2.8236,
       "step": 1000
     },
     {
       "epoch": 0.14,
-      "eval_accuracy": 0.4637646082566904,
-      "eval_loss": 2.3152551651000977,
-      "eval_runtime": 150.7767,
-      "eval_samples_per_second": 191.82,
-      "eval_steps_per_second": 23.982,
+      "eval_accuracy": 0.4710600926630247,
+      "eval_loss": 2.3487069606781006,
+      "eval_runtime": 155.6192,
+      "eval_samples_per_second": 185.851,
+      "eval_steps_per_second": 23.236,
       "step": 1000
     },
     {
       "epoch": 0.28,
       "learning_rate": 4.80243791611514e-06,
-      "loss": 2.1028,
+      "loss": 2.1379,
       "step": 2000
     },
     {
       "epoch": 0.28,
-      "eval_accuracy": 0.5636193900836733,
-      "eval_loss": 1.9398735761642456,
-      "eval_runtime": 149.8216,
-      "eval_samples_per_second": 193.043,
-      "eval_steps_per_second": 24.135,
+      "eval_accuracy": 0.5444644215476108,
+      "eval_loss": 1.9659085273742676,
+      "eval_runtime": 153.7302,
+      "eval_samples_per_second": 188.135,
+      "eval_steps_per_second": 23.522,
       "step": 2000
     },
     {
       "epoch": 0.41,
       "learning_rate": 4.703755655214652e-06,
-      "loss": 1.8063,
+      "loss": 1.8288,
       "step": 3000
     },
     {
       "epoch": 0.41,
-      "eval_accuracy": 0.601479842334555,
-      "eval_loss": 1.7228820323944092,
-      "eval_runtime": 150.2545,
-      "eval_samples_per_second": 192.487,
-      "eval_steps_per_second": 24.066,
+      "eval_accuracy": 0.6094322660950142,
+      "eval_loss": 1.736711025238037,
+      "eval_runtime": 154.9586,
+      "eval_samples_per_second": 186.643,
+      "eval_steps_per_second": 23.335,
       "step": 3000
     },
     {
       "epoch": 0.55,
       "learning_rate": 4.605073394314164e-06,
-      "loss": 1.6252,
+      "loss": 1.6449,
       "step": 4000
     },
     {
       "epoch": 0.55,
-      "eval_accuracy": 0.6286563861420372,
-      "eval_loss": 1.5740585327148438,
-      "eval_runtime": 149.4266,
-      "eval_samples_per_second": 193.553,
-      "eval_steps_per_second": 24.199,
+      "eval_accuracy": 0.6326325980222668,
+      "eval_loss": 1.5849583148956299,
+      "eval_runtime": 154.5958,
+      "eval_samples_per_second": 187.081,
+      "eval_steps_per_second": 23.39,
       "step": 4000
     },
     {
@@ -75,696 +75,696 @@
     },
     {
       "epoch": 0.69,
-      "eval_accuracy": 0.6416222944471337,
-      "eval_loss": 1.475919246673584,
-      "eval_runtime": 150.2657,
-      "eval_samples_per_second": 192.472,
-      "eval_steps_per_second": 24.064,
+      "eval_accuracy": 0.6461517184150474,
+      "eval_loss": 1.4778083562850952,
+      "eval_runtime": 156.4372,
+      "eval_samples_per_second": 184.879,
+      "eval_steps_per_second": 23.115,
       "step": 5000
     },
     {
       "epoch": 0.83,
-      "learning_rate": 4.407610091471245e-06,
-      "loss": 1.4225,
+      "learning_rate": 4.407511310429303e-06,
+      "loss": 1.4122,
       "step": 6000
     },
     {
       "epoch": 0.83,
-      "eval_accuracy": 0.6508540211603624,
-      "eval_loss": 1.3999030590057373,
-      "eval_runtime": 149.8691,
-      "eval_samples_per_second": 192.982,
-      "eval_steps_per_second": 24.128,
+      "eval_accuracy": 0.6564898693036443,
+      "eval_loss": 1.399396300315857,
+      "eval_runtime": 154.7049,
+      "eval_samples_per_second": 186.949,
+      "eval_steps_per_second": 23.374,
       "step": 6000
     },
     {
       "epoch": 0.97,
       "learning_rate": 4.3088290495288145e-06,
-      "loss": 1.3573,
+      "loss": 1.3623,
       "step": 7000
     },
     {
       "epoch": 0.97,
-      "eval_accuracy": 0.659013899453703,
-      "eval_loss": 1.3376258611679077,
-      "eval_runtime": 149.2556,
-      "eval_samples_per_second": 193.775,
-      "eval_steps_per_second": 24.227,
+      "eval_accuracy": 0.6619874144250052,
+      "eval_loss": 1.3486990928649902,
+      "eval_runtime": 153.9773,
+      "eval_samples_per_second": 187.833,
+      "eval_steps_per_second": 23.484,
       "step": 7000
     },
     {
       "epoch": 1.11,
       "learning_rate": 4.2100480075863845e-06,
-      "loss": 1.2666,
+      "loss": 1.293,
       "step": 8000
     },
     {
       "epoch": 1.11,
-      "eval_accuracy": 0.665721596016873,
-      "eval_loss": 1.2908859252929688,
-      "eval_runtime": 149.4901,
-      "eval_samples_per_second": 193.471,
-      "eval_steps_per_second": 24.189,
+      "eval_accuracy": 0.6671392019915635,
+      "eval_loss": 1.299405813217163,
+      "eval_runtime": 154.5924,
+      "eval_samples_per_second": 187.086,
+      "eval_steps_per_second": 23.391,
       "step": 8000
     },
     {
       "epoch": 1.24,
-      "learning_rate": 4.111365746685896e-06,
-      "loss": 1.2226,
+      "learning_rate": 4.1112669656439545e-06,
+      "loss": 1.2382,
       "step": 9000
     },
     {
       "epoch": 1.24,
-      "eval_accuracy": 0.6698706866745039,
-      "eval_loss": 1.2563775777816772,
-      "eval_runtime": 149.7035,
-      "eval_samples_per_second": 193.195,
-      "eval_steps_per_second": 24.154,
+      "eval_accuracy": 0.6701818684738261,
+      "eval_loss": 1.27021062374115,
+      "eval_runtime": 153.5258,
+      "eval_samples_per_second": 188.385,
+      "eval_steps_per_second": 23.553,
       "step": 9000
     },
     {
       "epoch": 1.38,
       "learning_rate": 4.012683485785409e-06,
-      "loss": 1.1999,
+      "loss": 1.2186,
       "step": 10000
     },
     {
       "epoch": 1.38,
-      "eval_accuracy": 0.6753682317958647,
-      "eval_loss": 1.2273180484771729,
-      "eval_runtime": 149.9221,
-      "eval_samples_per_second": 192.914,
-      "eval_steps_per_second": 24.119,
+      "eval_accuracy": 0.6728787774012862,
+      "eval_loss": 1.2421326637268066,
+      "eval_runtime": 155.3885,
+      "eval_samples_per_second": 186.127,
+      "eval_steps_per_second": 23.271,
       "step": 10000
     },
     {
       "epoch": 1.52,
       "learning_rate": 3.913902443842978e-06,
-      "loss": 1.1858,
+      "loss": 1.1912,
       "step": 11000
     },
     {
       "epoch": 1.52,
-      "eval_accuracy": 0.6769932923034369,
-      "eval_loss": 1.204084038734436,
-      "eval_runtime": 149.2049,
-      "eval_samples_per_second": 193.841,
-      "eval_steps_per_second": 24.235,
+      "eval_accuracy": 0.6746767166862596,
+      "eval_loss": 1.2220091819763184,
+      "eval_runtime": 155.3267,
+      "eval_samples_per_second": 186.201,
+      "eval_steps_per_second": 23.28,
       "step": 11000
     },
     {
       "epoch": 1.66,
       "learning_rate": 3.815121401900548e-06,
-      "loss": 1.1457,
+      "loss": 1.1798,
       "step": 12000
     },
     {
       "epoch": 1.66,
-      "eval_accuracy": 0.6784108982781274,
-      "eval_loss": 1.190036416053772,
-      "eval_runtime": 153.2132,
-      "eval_samples_per_second": 188.77,
-      "eval_steps_per_second": 23.601,
+      "eval_accuracy": 0.6796556254754167,
+      "eval_loss": 1.1974430084228516,
+      "eval_runtime": 153.7759,
+      "eval_samples_per_second": 188.079,
+      "eval_steps_per_second": 23.515,
       "step": 12000
     },
     {
       "epoch": 1.8,
-      "learning_rate": 3.716340359958117e-06,
-      "loss": 1.1526,
+      "learning_rate": 3.7164391410000594e-06,
+      "loss": 1.1605,
       "step": 13000
     },
     {
       "epoch": 1.8,
-      "eval_accuracy": 0.6814881405158703,
-      "eval_loss": 1.1732572317123413,
-      "eval_runtime": 153.4358,
-      "eval_samples_per_second": 188.496,
-      "eval_steps_per_second": 23.567,
+      "eval_accuracy": 0.6826982919576793,
+      "eval_loss": 1.1833155155181885,
+      "eval_runtime": 153.6706,
+      "eval_samples_per_second": 188.208,
+      "eval_steps_per_second": 23.531,
       "step": 13000
     },
     {
       "epoch": 1.94,
       "learning_rate": 3.6176580990576294e-06,
-      "loss": 1.1285,
+      "loss": 1.1454,
       "step": 14000
     },
     {
       "epoch": 1.94,
-      "eval_accuracy": 0.6809349284281861,
-      "eval_loss": 1.1645420789718628,
-      "eval_runtime": 154.9231,
-      "eval_samples_per_second": 186.686,
-      "eval_steps_per_second": 23.341,
+      "eval_accuracy": 0.6838047161330475,
+      "eval_loss": 1.1689262390136719,
+      "eval_runtime": 155.3074,
+      "eval_samples_per_second": 186.224,
+      "eval_steps_per_second": 23.283,
       "step": 14000
     },
     {
       "epoch": 2.07,
-      "learning_rate": 3.5188770571151986e-06,
-      "loss": 1.0864,
+      "learning_rate": 3.518975838157141e-06,
+      "loss": 1.1076,
       "step": 15000
     },
     {
       "epoch": 2.07,
-      "eval_accuracy": 0.6872622916810732,
-      "eval_loss": 1.1494485139846802,
-      "eval_runtime": 153.8139,
-      "eval_samples_per_second": 188.032,
-      "eval_steps_per_second": 23.509,
+      "eval_accuracy": 0.6820759283590346,
+      "eval_loss": 1.1666451692581177,
+      "eval_runtime": 156.0771,
+      "eval_samples_per_second": 185.306,
+      "eval_steps_per_second": 23.168,
       "step": 15000
     },
     {
       "epoch": 2.21,
-      "learning_rate": 3.4200960151727686e-06,
-      "loss": 1.0623,
+      "learning_rate": 3.4201947962147106e-06,
+      "loss": 1.0882,
       "step": 16000
     },
     {
       "epoch": 2.21,
-      "eval_accuracy": 0.687089412903672,
-      "eval_loss": 1.142927646636963,
-      "eval_runtime": 150.573,
-      "eval_samples_per_second": 192.08,
-      "eval_steps_per_second": 24.015,
+      "eval_accuracy": 0.6835626858446857,
+      "eval_loss": 1.1561516523361206,
+      "eval_runtime": 154.5444,
+      "eval_samples_per_second": 187.144,
+      "eval_steps_per_second": 23.398,
       "step": 16000
     },
     {
       "epoch": 2.35,
-      "learning_rate": 3.3213149732303378e-06,
-      "loss": 1.0428,
+      "learning_rate": 3.3214137542722802e-06,
+      "loss": 1.0832,
       "step": 17000
     },
     {
       "epoch": 2.35,
-      "eval_accuracy": 0.6891293824770072,
-      "eval_loss": 1.1337573528289795,
-      "eval_runtime": 150.0663,
-      "eval_samples_per_second": 192.728,
-      "eval_steps_per_second": 24.096,
+      "eval_accuracy": 0.6874005947029943,
+      "eval_loss": 1.1425527334213257,
+      "eval_runtime": 155.3385,
+      "eval_samples_per_second": 186.187,
+      "eval_steps_per_second": 23.278,
       "step": 17000
     },
     {
       "epoch": 2.49,
-      "learning_rate": 3.2226327123298502e-06,
-      "loss": 1.0495,
+      "learning_rate": 3.2227314933717923e-06,
+      "loss": 1.0698,
       "step": 18000
     },
     {
       "epoch": 2.49,
-      "eval_accuracy": 0.6909618975174607,
-      "eval_loss": 1.1230798959732056,
-      "eval_runtime": 149.6696,
-      "eval_samples_per_second": 193.239,
-      "eval_steps_per_second": 24.16,
+      "eval_accuracy": 0.6872622916810732,
+      "eval_loss": 1.1318212747573853,
+      "eval_runtime": 155.6543,
+      "eval_samples_per_second": 185.809,
+      "eval_steps_per_second": 23.231,
       "step": 18000
     },
     {
       "epoch": 2.63,
       "learning_rate": 3.123950451429362e-06,
-      "loss": 1.0401,
+      "loss": 1.0752,
       "step": 19000
     },
     {
       "epoch": 2.63,
-      "eval_accuracy": 0.6924486550031118,
-      "eval_loss": 1.1154536008834839,
-      "eval_runtime": 150.8125,
-      "eval_samples_per_second": 191.775,
-      "eval_steps_per_second": 23.977,
+      "eval_accuracy": 0.6842542009542909,
+      "eval_loss": 1.1395872831344604,
+      "eval_runtime": 155.6622,
+      "eval_samples_per_second": 185.8,
+      "eval_steps_per_second": 23.23,
       "step": 19000
     },
     {
       "epoch": 2.77,
       "learning_rate": 3.0251694094869315e-06,
-      "loss": 1.0279,
+      "loss": 1.0659,
       "step": 20000
     },
     {
       "epoch": 2.77,
-      "eval_accuracy": 0.690581564207178,
-      "eval_loss": 1.1119346618652344,
-      "eval_runtime": 149.9854,
-      "eval_samples_per_second": 192.832,
-      "eval_steps_per_second": 24.109,
+      "eval_accuracy": 0.6903049581633358,
+      "eval_loss": 1.1166965961456299,
+      "eval_runtime": 154.6931,
+      "eval_samples_per_second": 186.964,
+      "eval_steps_per_second": 23.375,
       "step": 20000
     },
     {
       "epoch": 2.9,
-      "learning_rate": 2.926388367544501e-06,
-      "loss": 1.0205,
+      "learning_rate": 2.926487148586444e-06,
+      "loss": 1.0561,
       "step": 21000
     },
     {
       "epoch": 2.9,
-      "eval_accuracy": 0.694488624576447,
-      "eval_loss": 1.1036620140075684,
-      "eval_runtime": 150.0204,
-      "eval_samples_per_second": 192.787,
-      "eval_steps_per_second": 24.103,
+      "eval_accuracy": 0.6879538067906784,
+      "eval_loss": 1.1178348064422607,
+      "eval_runtime": 152.6144,
+      "eval_samples_per_second": 189.51,
+      "eval_steps_per_second": 23.694,
       "step": 21000
     },
     {
       "epoch": 3.04,
       "learning_rate": 2.827706106644013e-06,
-      "loss": 1.0102,
+      "loss": 1.0328,
       "step": 22000
     },
     {
       "epoch": 3.04,
-      "eval_accuracy": 0.6956296245072955,
-      "eval_loss": 1.1001828908920288,
-      "eval_runtime": 150.2658,
-      "eval_samples_per_second": 192.472,
-      "eval_steps_per_second": 24.064,
+      "eval_accuracy": 0.690581564207178,
+      "eval_loss": 1.1114201545715332,
+      "eval_runtime": 153.9011,
+      "eval_samples_per_second": 187.926,
+      "eval_steps_per_second": 23.496,
       "step": 22000
     },
     {
       "epoch": 3.18,
-      "learning_rate": 2.7289250647015827e-06,
-      "loss": 0.9516,
+      "learning_rate": 2.729023845743525e-06,
+      "loss": 1.0299,
       "step": 23000
     },
     {
       "epoch": 3.18,
-      "eval_accuracy": 0.6948343821312496,
-      "eval_loss": 1.0975209474563599,
-      "eval_runtime": 150.3336,
-      "eval_samples_per_second": 192.386,
-      "eval_steps_per_second": 24.053,
+      "eval_accuracy": 0.6917225641380265,
+      "eval_loss": 1.1057274341583252,
+      "eval_runtime": 155.3852,
+      "eval_samples_per_second": 186.131,
+      "eval_steps_per_second": 23.271,
       "step": 23000
     },
     {
       "epoch": 3.32,
       "learning_rate": 2.6302428038010947e-06,
-      "loss": 0.9526,
+      "loss": 0.9961,
       "step": 24000
     },
     {
       "epoch": 3.32,
-      "eval_accuracy": 0.6936242306894406,
-      "eval_loss": 1.1008094549179077,
-      "eval_runtime": 151.6663,
-      "eval_samples_per_second": 190.695,
-      "eval_steps_per_second": 23.842,
+      "eval_accuracy": 0.6912730793167831,
+      "eval_loss": 1.1056290864944458,
+      "eval_runtime": 154.2366,
+      "eval_samples_per_second": 187.517,
+      "eval_steps_per_second": 23.445,
       "step": 24000
     },
     {
       "epoch": 3.46,
-      "learning_rate": 2.5314617618586647e-06,
-      "loss": 0.9694,
+      "learning_rate": 2.5315605429006068e-06,
+      "loss": 1.0128,
       "step": 25000
     },
     {
       "epoch": 3.46,
-      "eval_accuracy": 0.6935205034229998,
-      "eval_loss": 1.0990099906921387,
-      "eval_runtime": 150.2935,
-      "eval_samples_per_second": 192.437,
-      "eval_steps_per_second": 24.06,
+      "eval_accuracy": 0.6937971094668418,
+      "eval_loss": 1.0973228216171265,
+      "eval_runtime": 156.2066,
+      "eval_samples_per_second": 185.152,
+      "eval_steps_per_second": 23.149,
       "step": 25000
     },
     {
       "epoch": 3.6,
-      "learning_rate": 2.432680719916234e-06,
-      "loss": 0.9649,
+      "learning_rate": 2.4327795009581764e-06,
+      "loss": 1.0118,
       "step": 26000
     },
     {
       "epoch": 3.6,
-      "eval_accuracy": 0.6970126547265058,
-      "eval_loss": 1.0900559425354004,
-      "eval_runtime": 150.5398,
-      "eval_samples_per_second": 192.122,
-      "eval_steps_per_second": 24.02,
+      "eval_accuracy": 0.6942465942880852,
+      "eval_loss": 1.0930811166763306,
+      "eval_runtime": 155.695,
+      "eval_samples_per_second": 185.761,
+      "eval_steps_per_second": 23.225,
       "step": 26000
     },
     {
       "epoch": 3.73,
       "learning_rate": 2.333998459015746e-06,
-      "loss": 0.9522,
+      "loss": 1.0045,
       "step": 27000
     },
     {
       "epoch": 3.73,
-      "eval_accuracy": 0.6967014729271834,
-      "eval_loss": 1.088005542755127,
-      "eval_runtime": 151.5602,
-      "eval_samples_per_second": 190.828,
-      "eval_steps_per_second": 23.859,
+      "eval_accuracy": 0.6936933822004011,
+      "eval_loss": 1.089782953262329,
+      "eval_runtime": 155.3577,
+      "eval_samples_per_second": 186.164,
+      "eval_steps_per_second": 23.275,
       "step": 27000
     },
     {
       "epoch": 3.87,
-      "learning_rate": 2.2352174170733155e-06,
-      "loss": 0.9707,
+      "learning_rate": 2.2354149791572e-06,
+      "loss": 0.9923,
       "step": 28000
     },
     {
       "epoch": 3.87,
-      "eval_accuracy": 0.6988797455224397,
-      "eval_loss": 1.08329439163208,
-      "eval_runtime": 149.8845,
-      "eval_samples_per_second": 192.962,
-      "eval_steps_per_second": 24.125,
+      "eval_accuracy": 0.6958716547956573,
+      "eval_loss": 1.0858705043792725,
+      "eval_runtime": 155.5851,
+      "eval_samples_per_second": 185.892,
+      "eval_steps_per_second": 23.241,
       "step": 28000
     },
     {
       "epoch": 4.01,
-      "learning_rate": 2.136436375130885e-06,
-      "loss": 0.9533,
+      "learning_rate": 2.13663393721477e-06,
+      "loss": 0.9988,
       "step": 29000
     },
     {
       "epoch": 4.01,
-      "eval_accuracy": 0.69891432127792,
-      "eval_loss": 1.078162670135498,
-      "eval_runtime": 150.2294,
-      "eval_samples_per_second": 192.519,
-      "eval_steps_per_second": 24.07,
+      "eval_accuracy": 0.6943848973100062,
+      "eval_loss": 1.0852184295654297,
+      "eval_runtime": 154.9704,
+      "eval_samples_per_second": 186.629,
+      "eval_steps_per_second": 23.333,
       "step": 29000
     },
     {
       "epoch": 4.15,
-      "learning_rate": 2.0376553331884547e-06,
-      "loss": 0.9187,
+      "learning_rate": 2.0378528952723396e-06,
+      "loss": 0.9773,
       "step": 30000
     },
     {
       "epoch": 4.15,
-      "eval_accuracy": 0.6963557153723808,
-      "eval_loss": 1.0856797695159912,
-      "eval_runtime": 153.2337,
-      "eval_samples_per_second": 188.744,
-      "eval_steps_per_second": 23.598,
+      "eval_accuracy": 0.6930018670907959,
+      "eval_loss": 1.089293122291565,
+      "eval_runtime": 156.3077,
+      "eval_samples_per_second": 185.033,
+      "eval_steps_per_second": 23.134,
       "step": 30000
     },
     {
       "epoch": 4.29,
       "learning_rate": 1.9391706343718513e-06,
-      "loss": 0.9019,
+      "loss": 0.9577,
       "step": 31000
     },
     {
       "epoch": 4.29,
-      "eval_accuracy": 0.6997095636539659,
-      "eval_loss": 1.085842251777649,
-      "eval_runtime": 155.7046,
-      "eval_samples_per_second": 185.749,
-      "eval_steps_per_second": 23.223,
+      "eval_accuracy": 0.6968052001936242,
+      "eval_loss": 1.0807169675827026,
+      "eval_runtime": 154.1329,
+      "eval_samples_per_second": 187.643,
+      "eval_steps_per_second": 23.46,
       "step": 31000
     },
     {
       "epoch": 4.43,
       "learning_rate": 1.840389592429421e-06,
-      "loss": 0.9074,
+      "loss": 0.9748,
       "step": 32000
     },
     {
       "epoch": 4.43,
-      "eval_accuracy": 0.6987760182559989,
-      "eval_loss": 1.083946943283081,
-      "eval_runtime": 154.7842,
-      "eval_samples_per_second": 186.854,
-      "eval_steps_per_second": 23.362,
+      "eval_accuracy": 0.6957333517737363,
+      "eval_loss": 1.0788837671279907,
+      "eval_runtime": 154.5313,
+      "eval_samples_per_second": 187.16,
+      "eval_steps_per_second": 23.4,
       "step": 32000
     },
     {
       "epoch": 4.56,
       "learning_rate": 1.7416085504869907e-06,
-      "loss": 0.903,
+      "loss": 0.9777,
       "step": 33000
     },
     {
       "epoch": 4.56,
-      "eval_accuracy": 0.6998132909204066,
-      "eval_loss": 1.0829977989196777,
-      "eval_runtime": 155.014,
-      "eval_samples_per_second": 186.577,
-      "eval_steps_per_second": 23.327,
+      "eval_accuracy": 0.6924486550031118,
+      "eval_loss": 1.0864237546920776,
+      "eval_runtime": 155.2284,
+      "eval_samples_per_second": 186.319,
+      "eval_steps_per_second": 23.295,
       "step": 33000
     },
     {
       "epoch": 4.7,
       "learning_rate": 1.6428275085445602e-06,
-      "loss": 0.8951,
+      "loss": 0.9536,
       "step": 34000
     },
     {
       "epoch": 4.7,
-      "eval_accuracy": 0.6996749878984856,
-      "eval_loss": 1.081098198890686,
-      "eval_runtime": 155.221,
-      "eval_samples_per_second": 186.328,
-      "eval_steps_per_second": 23.296,
+      "eval_accuracy": 0.6949035336422101,
+      "eval_loss": 1.0813385248184204,
+      "eval_runtime": 153.9831,
+      "eval_samples_per_second": 187.826,
+      "eval_steps_per_second": 23.483,
       "step": 34000
     },
     {
       "epoch": 4.84,
-      "learning_rate": 1.5440464666021298e-06,
-      "loss": 0.8925,
+      "learning_rate": 1.544145247644072e-06,
+      "loss": 0.9507,
       "step": 35000
     },
     {
       "epoch": 4.84,
-      "eval_accuracy": 0.7006431090519328,
-      "eval_loss": 1.0793355703353882,
-      "eval_runtime": 154.8555,
-      "eval_samples_per_second": 186.768,
-      "eval_steps_per_second": 23.351,
+      "eval_accuracy": 0.6950418366641311,
+      "eval_loss": 1.0795472860336304,
+      "eval_runtime": 156.0888,
+      "eval_samples_per_second": 185.292,
+      "eval_steps_per_second": 23.166,
       "step": 35000
     },
     {
       "epoch": 4.98,
       "learning_rate": 1.4453642057016417e-06,
-      "loss": 0.901,
+      "loss": 0.9627,
       "step": 36000
     },
     {
       "epoch": 4.98,
-      "eval_accuracy": 0.6993638060991633,
-      "eval_loss": 1.0771288871765137,
-      "eval_runtime": 154.7342,
-      "eval_samples_per_second": 186.914,
-      "eval_steps_per_second": 23.369,
+      "eval_accuracy": 0.6955258972408547,
+      "eval_loss": 1.0754951238632202,
+      "eval_runtime": 156.2467,
+      "eval_samples_per_second": 185.105,
+      "eval_steps_per_second": 23.143,
       "step": 36000
     },
     {
       "epoch": 5.12,
       "learning_rate": 1.3465831637592115e-06,
-      "loss": 0.8694,
+      "loss": 0.9399,
       "step": 37000
     },
     {
       "epoch": 5.12,
-      "eval_accuracy": 0.699190927321762,
-      "eval_loss": 1.0816415548324585,
-      "eval_runtime": 154.9133,
-      "eval_samples_per_second": 186.698,
-      "eval_steps_per_second": 23.342,
+      "eval_accuracy": 0.6960791093285388,
+      "eval_loss": 1.0770469903945923,
+      "eval_runtime": 153.8614,
+      "eval_samples_per_second": 187.974,
+      "eval_steps_per_second": 23.502,
       "step": 37000
     },
     {
       "epoch": 5.26,
       "learning_rate": 1.2479009028587235e-06,
-      "loss": 0.8709,
+      "loss": 0.9357,
       "step": 38000
     },
     {
       "epoch": 5.26,
-      "eval_accuracy": 0.6991563515662818,
-      "eval_loss": 1.083854079246521,
-      "eval_runtime": 154.8522,
-      "eval_samples_per_second": 186.772,
-      "eval_steps_per_second": 23.351,
+      "eval_accuracy": 0.6960791093285388,
+      "eval_loss": 1.0759111642837524,
+      "eval_runtime": 153.1358,
+      "eval_samples_per_second": 188.865,
+      "eval_steps_per_second": 23.613,
       "step": 38000
     },
     {
       "epoch": 5.39,
       "learning_rate": 1.1491198609162931e-06,
-      "loss": 0.8557,
+      "loss": 0.943,
       "step": 39000
     },
     {
       "epoch": 5.39,
-      "eval_accuracy": 0.6984648364566766,
-      "eval_loss": 1.0835996866226196,
-      "eval_runtime": 156.6539,
-      "eval_samples_per_second": 184.623,
-      "eval_steps_per_second": 23.083,
+      "eval_accuracy": 0.6965977456607427,
+      "eval_loss": 1.0720691680908203,
+      "eval_runtime": 154.5633,
+      "eval_samples_per_second": 187.121,
+      "eval_steps_per_second": 23.395,
       "step": 39000
     },
     {
       "epoch": 5.53,
-      "learning_rate": 1.0503388189738627e-06,
-      "loss": 0.8583,
+      "learning_rate": 1.0504376000158052e-06,
+      "loss": 0.9244,
       "step": 40000
     },
     {
       "epoch": 5.53,
-      "eval_accuracy": 0.6977041698361109,
-      "eval_loss": 1.0821908712387085,
-      "eval_runtime": 154.2155,
-      "eval_samples_per_second": 187.543,
-      "eval_steps_per_second": 23.448,
+      "eval_accuracy": 0.696908927460065,
+      "eval_loss": 1.0704323053359985,
+      "eval_runtime": 154.315,
+      "eval_samples_per_second": 187.422,
+      "eval_steps_per_second": 23.433,
       "step": 40000
     },
     {
       "epoch": 5.67,
-      "learning_rate": 9.515577770314322e-07,
-      "loss": 0.8533,
+      "learning_rate": 9.516565580733748e-07,
+      "loss": 0.9231,
       "step": 41000
     },
     {
       "epoch": 5.67,
-      "eval_accuracy": 0.6983611091902358,
-      "eval_loss": 1.0834639072418213,
-      "eval_runtime": 155.654,
-      "eval_samples_per_second": 185.809,
-      "eval_steps_per_second": 23.231,
+      "eval_accuracy": 0.6960445335730586,
+      "eval_loss": 1.0727081298828125,
+      "eval_runtime": 153.587,
+      "eval_samples_per_second": 188.31,
+      "eval_steps_per_second": 23.544,
       "step": 41000
     },
     {
       "epoch": 5.81,
       "learning_rate": 8.528755161309442e-07,
-      "loss": 0.8545,
+      "loss": 0.9294,
       "step": 42000
     },
     {
       "epoch": 5.81,
-      "eval_accuracy": 0.6993983818546435,
-      "eval_loss": 1.0837448835372925,
-      "eval_runtime": 155.1115,
-      "eval_samples_per_second": 186.459,
-      "eval_steps_per_second": 23.312,
+      "eval_accuracy": 0.6969780789710255,
+      "eval_loss": 1.0715699195861816,
+      "eval_runtime": 153.1528,
+      "eval_samples_per_second": 188.844,
+      "eval_steps_per_second": 23.61,
       "step": 42000
     },
     {
       "epoch": 5.95,
       "learning_rate": 7.540944741885138e-07,
-      "loss": 0.8608,
+      "loss": 0.9416,
       "step": 43000
     },
     {
       "epoch": 5.95,
-      "eval_accuracy": 0.6997441394094461,
-      "eval_loss": 1.0805258750915527,
-      "eval_runtime": 153.7674,
-      "eval_samples_per_second": 188.089,
-      "eval_steps_per_second": 23.516,
+      "eval_accuracy": 0.6980845031463937,
+      "eval_loss": 1.0693832635879517,
+      "eval_runtime": 151.6066,
+      "eval_samples_per_second": 190.77,
+      "eval_steps_per_second": 23.851,
       "step": 43000
     },
     {
       "epoch": 6.08,
       "learning_rate": 6.554122132880259e-07,
-      "loss": 0.8292,
+      "loss": 0.9248,
       "step": 44000
     },
     {
       "epoch": 6.08,
-      "eval_accuracy": 0.6999515939423276,
-      "eval_loss": 1.0849448442459106,
-      "eval_runtime": 155.0279,
-      "eval_samples_per_second": 186.56,
-      "eval_steps_per_second": 23.325,
+      "eval_accuracy": 0.6991217758108015,
+      "eval_loss": 1.0678476095199585,
+      "eval_runtime": 151.367,
+      "eval_samples_per_second": 191.072,
+      "eval_steps_per_second": 23.889,
       "step": 44000
     },
     {
       "epoch": 6.22,
       "learning_rate": 5.566311713455954e-07,
-      "loss": 0.8385,
+      "loss": 0.9137,
       "step": 45000
     },
     {
       "epoch": 6.22,
-      "eval_accuracy": 0.6998478666758868,
-      "eval_loss": 1.084898591041565,
-      "eval_runtime": 154.6263,
-      "eval_samples_per_second": 187.044,
-      "eval_steps_per_second": 23.385,
+      "eval_accuracy": 0.6976350183251504,
+      "eval_loss": 1.0700552463531494,
+      "eval_runtime": 152.3071,
+      "eval_samples_per_second": 189.893,
+      "eval_steps_per_second": 23.742,
       "step": 45000
     },
     {
       "epoch": 6.36,
       "learning_rate": 4.5785012940316495e-07,
-      "loss": 0.826,
+      "loss": 0.91,
       "step": 46000
     },
     {
       "epoch": 6.36,
-      "eval_accuracy": 0.699329230343683,
-      "eval_loss": 1.0853327512741089,
-      "eval_runtime": 155.3096,
-      "eval_samples_per_second": 186.222,
-      "eval_steps_per_second": 23.283,
+      "eval_accuracy": 0.6971855335039071,
+      "eval_loss": 1.0688731670379639,
+      "eval_runtime": 151.0293,
+      "eval_samples_per_second": 191.499,
+      "eval_steps_per_second": 23.942,
       "step": 46000
     },
     {
       "epoch": 6.5,
       "learning_rate": 3.59167868502677e-07,
-      "loss": 0.829,
+      "loss": 0.9256,
       "step": 47000
     },
     {
       "epoch": 6.5,
-      "eval_accuracy": 0.6992946545882027,
-      "eval_loss": 1.0856379270553589,
-      "eval_runtime": 153.5481,
-      "eval_samples_per_second": 188.358,
-      "eval_steps_per_second": 23.55,
+      "eval_accuracy": 0.6974621395477492,
+      "eval_loss": 1.0670689344406128,
+      "eval_runtime": 150.9586,
+      "eval_samples_per_second": 191.589,
+      "eval_steps_per_second": 23.954,
       "step": 47000
     },
     {
       "epoch": 6.64,
-      "learning_rate": 2.603868265602466e-07,
-      "loss": 0.8345,
+      "learning_rate": 2.6048560760218905e-07,
+      "loss": 0.9085,
       "step": 48000
     },
     {
       "epoch": 6.64,
-      "eval_accuracy": 0.699190927321762,
-      "eval_loss": 1.0844392776489258,
-      "eval_runtime": 154.6451,
-      "eval_samples_per_second": 187.022,
-      "eval_steps_per_second": 23.383,
+      "eval_accuracy": 0.6985339879676371,
+      "eval_loss": 1.067813754081726,
+      "eval_runtime": 151.0599,
+      "eval_samples_per_second": 191.46,
+      "eval_steps_per_second": 23.938,
       "step": 48000
     },
     {
       "epoch": 6.78,
-      "learning_rate": 1.6160578461781615e-07,
-      "loss": 0.8347,
+      "learning_rate": 1.617045656597586e-07,
+      "loss": 0.9169,
       "step": 49000
     },
     {
       "epoch": 6.78,
-      "eval_accuracy": 0.6992600788327225,
-      "eval_loss": 1.0855780839920044,
-      "eval_runtime": 154.3431,
-      "eval_samples_per_second": 187.388,
-      "eval_steps_per_second": 23.428,
+      "eval_accuracy": 0.6984302607011963,
+      "eval_loss": 1.0689929723739624,
+      "eval_runtime": 151.6049,
+      "eval_samples_per_second": 190.772,
+      "eval_steps_per_second": 23.851,
       "step": 49000
     },
     {
       "epoch": 6.91,
-      "learning_rate": 6.282474267538574e-08,
-      "loss": 0.8266,
+      "learning_rate": 6.292352371732817e-08,
+      "loss": 0.9087,
       "step": 50000
     },
     {
       "epoch": 6.91,
-      "eval_accuracy": 0.6998478666758868,
-      "eval_loss": 1.085789680480957,
-      "eval_runtime": 154.1691,
-      "eval_samples_per_second": 187.599,
-      "eval_steps_per_second": 23.455,
+      "eval_accuracy": 0.6972546850148675,
+      "eval_loss": 1.069164752960205,
+      "eval_runtime": 151.0946,
+      "eval_samples_per_second": 191.417,
+      "eval_steps_per_second": 23.932,
       "step": 50000
     },
     {
       "epoch": 7.0,
       "step": 50617,
       "total_flos": 1.2555854407514107e+20,
-      "train_loss": 1.079674410547284,
-      "train_runtime": 23712.848,
-      "train_samples_per_second": 68.302,
-      "train_steps_per_second": 2.135
+      "train_loss": 1.126858214650741,
+      "train_runtime": 24456.4619,
+      "train_samples_per_second": 66.225,
+      "train_steps_per_second": 2.07
     }
   ],
   "max_steps": 50617,