diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,9088 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.5696761816173477,
+  "global_step": 756000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.998269757295396e-05,
+      "loss": 2.8529,
+      "step": 500
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.996539514590791e-05,
+      "loss": 2.3953,
+      "step": 1000
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.994809271886186e-05,
+      "loss": 2.0905,
+      "step": 1500
+    },
+    {
+      "epoch": 0.0,
+      "learning_rate": 4.993079029181581e-05,
+      "loss": 1.672,
+      "step": 2000
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.991348786476977e-05,
+      "loss": 1.4671,
+      "step": 2500
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.989618543772373e-05,
+      "loss": 1.3692,
+      "step": 3000
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.987888301067767e-05,
+      "loss": 1.3064,
+      "step": 3500
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.986158058363163e-05,
+      "loss": 1.2556,
+      "step": 4000
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.984427815658558e-05,
+      "loss": 1.2133,
+      "step": 4500
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.982697572953954e-05,
+      "loss": 1.1793,
+      "step": 5000
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.9809673302493495e-05,
+      "loss": 1.1548,
+      "step": 5500
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.979237087544744e-05,
+      "loss": 1.1329,
+      "step": 6000
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.97750684484014e-05,
+      "loss": 1.1147,
+      "step": 6500
+    },
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.9757766021355354e-05,
+      "loss": 1.0988,
+      "step": 7000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9740463594309305e-05,
+      "loss": 1.0875,
+      "step": 7500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9723161167263256e-05,
+      "loss": 1.074,
+      "step": 8000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9705858740217207e-05,
+      "loss": 1.062,
+      "step": 8500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9688556313171164e-05,
+      "loss": 1.0505,
+      "step": 9000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.967125388612512e-05,
+      "loss": 1.0386,
+      "step": 9500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9653951459079066e-05,
+      "loss": 1.0334,
+      "step": 10000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.963664903203302e-05,
+      "loss": 1.0225,
+      "step": 10500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9619346604986974e-05,
+      "loss": 1.0151,
+      "step": 11000
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.960204417794093e-05,
+      "loss": 1.0073,
+      "step": 11500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.958474175089489e-05,
+      "loss": 0.9995,
+      "step": 12000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.956743932384883e-05,
+      "loss": 0.9931,
+      "step": 12500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.955013689680279e-05,
+      "loss": 0.987,
+      "step": 13000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.953283446975674e-05,
+      "loss": 0.9855,
+      "step": 13500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.95155320427107e-05,
+      "loss": 0.9781,
+      "step": 14000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.949822961566465e-05,
+      "loss": 0.975,
+      "step": 14500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.94809271886186e-05,
+      "loss": 0.9656,
+      "step": 15000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.946362476157256e-05,
+      "loss": 0.9622,
+      "step": 15500
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.944632233452651e-05,
+      "loss": 0.9595,
+      "step": 16000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.942901990748047e-05,
+      "loss": 0.9548,
+      "step": 16500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.941171748043442e-05,
+      "loss": 0.9504,
+      "step": 17000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.939441505338837e-05,
+      "loss": 0.946,
+      "step": 17500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9377112626342326e-05,
+      "loss": 0.9376,
+      "step": 18000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9359810199296284e-05,
+      "loss": 0.9376,
+      "step": 18500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.934250777225023e-05,
+      "loss": 0.9326,
+      "step": 19000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9325205345204185e-05,
+      "loss": 0.9279,
+      "step": 19500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9307902918158136e-05,
+      "loss": 0.9285,
+      "step": 20000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9290600491112093e-05,
+      "loss": 0.9256,
+      "step": 20500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9273298064066044e-05,
+      "loss": 0.9209,
+      "step": 21000
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.9255995637019995e-05,
+      "loss": 0.9163,
+      "step": 21500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.923869320997395e-05,
+      "loss": 0.9134,
+      "step": 22000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.9221390782927903e-05,
+      "loss": 0.9129,
+      "step": 22500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.920408835588186e-05,
+      "loss": 0.9094,
+      "step": 23000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.918678592883581e-05,
+      "loss": 0.9038,
+      "step": 23500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.916948350178976e-05,
+      "loss": 0.9096,
+      "step": 24000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.915218107474372e-05,
+      "loss": 0.9048,
+      "step": 24500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.913487864769767e-05,
+      "loss": 0.8989,
+      "step": 25000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.911757622065162e-05,
+      "loss": 0.8963,
+      "step": 25500
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.910027379360558e-05,
+      "loss": 0.8994,
+      "step": 26000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.908297136655953e-05,
+      "loss": 0.8913,
+      "step": 26500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.906566893951349e-05,
+      "loss": 0.8912,
+      "step": 27000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.904836651246744e-05,
+      "loss": 0.8877,
+      "step": 27500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.903106408542139e-05,
+      "loss": 0.8891,
+      "step": 28000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.901376165837535e-05,
+      "loss": 0.8837,
+      "step": 28500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.89964592313293e-05,
+      "loss": 0.8786,
+      "step": 29000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.8979156804283255e-05,
+      "loss": 0.8785,
+      "step": 29500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.8961854377237206e-05,
+      "loss": 0.8798,
+      "step": 30000
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.894455195019116e-05,
+      "loss": 0.8761,
+      "step": 30500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.8927249523145114e-05,
+      "loss": 0.8739,
+      "step": 31000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.8909947096099065e-05,
+      "loss": 0.8718,
+      "step": 31500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.8892644669053016e-05,
+      "loss": 0.8711,
+      "step": 32000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.8875342242006974e-05,
+      "loss": 0.8694,
+      "step": 32500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.8858039814960924e-05,
+      "loss": 0.8676,
+      "step": 33000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.884073738791488e-05,
+      "loss": 0.865,
+      "step": 33500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.882343496086883e-05,
+      "loss": 0.8607,
+      "step": 34000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.8806132533822784e-05,
+      "loss": 0.859,
+      "step": 34500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.878883010677674e-05,
+      "loss": 0.863,
+      "step": 35000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.877152767973069e-05,
+      "loss": 0.8618,
+      "step": 35500
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.875422525268465e-05,
+      "loss": 0.8585,
+      "step": 36000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.87369228256386e-05,
+      "loss": 0.8588,
+      "step": 36500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.871962039859255e-05,
+      "loss": 0.8581,
+      "step": 37000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.870231797154651e-05,
+      "loss": 0.8541,
+      "step": 37500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.868501554450046e-05,
+      "loss": 0.8515,
+      "step": 38000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.866771311745442e-05,
+      "loss": 0.8512,
+      "step": 38500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.865041069040837e-05,
+      "loss": 0.8508,
+      "step": 39000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.863310826336232e-05,
+      "loss": 0.8493,
+      "step": 39500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.8615805836316276e-05,
+      "loss": 0.8456,
+      "step": 40000
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.859850340927023e-05,
+      "loss": 0.845,
+      "step": 40500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.858120098222418e-05,
+      "loss": 0.8439,
+      "step": 41000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.8563898555178135e-05,
+      "loss": 0.843,
+      "step": 41500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.8546596128132086e-05,
+      "loss": 0.8414,
+      "step": 42000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.8529293701086044e-05,
+      "loss": 0.8401,
+      "step": 42500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.8511991274039995e-05,
+      "loss": 0.8395,
+      "step": 43000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.8494688846993945e-05,
+      "loss": 0.8404,
+      "step": 43500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.84773864199479e-05,
+      "loss": 0.8395,
+      "step": 44000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.8460083992901854e-05,
+      "loss": 0.8342,
+      "step": 44500
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.844278156585581e-05,
+      "loss": 0.837,
+      "step": 45000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.842547913880976e-05,
+      "loss": 0.8376,
+      "step": 45500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.840817671176371e-05,
+      "loss": 0.8338,
+      "step": 46000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.839087428471767e-05,
+      "loss": 0.829,
+      "step": 46500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.837357185767162e-05,
+      "loss": 0.831,
+      "step": 47000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.835626943062557e-05,
+      "loss": 0.8312,
+      "step": 47500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.833896700357953e-05,
+      "loss": 0.8275,
+      "step": 48000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.832166457653348e-05,
+      "loss": 0.8284,
+      "step": 48500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.830436214948744e-05,
+      "loss": 0.8291,
+      "step": 49000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.828705972244139e-05,
+      "loss": 0.8278,
+      "step": 49500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.826975729539534e-05,
+      "loss": 0.8224,
+      "step": 50000
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.82524548683493e-05,
+      "loss": 0.8266,
+      "step": 50500
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.823515244130325e-05,
+      "loss": 0.8224,
+      "step": 51000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.8217850014257206e-05,
+      "loss": 0.8193,
+      "step": 51500
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.8200547587211156e-05,
+      "loss": 0.8193,
+      "step": 52000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.818324516016511e-05,
+      "loss": 0.824,
+      "step": 52500
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.8165942733119065e-05,
+      "loss": 0.8245,
+      "step": 53000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.8148640306073015e-05,
+      "loss": 0.8204,
+      "step": 53500
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.8131337879026966e-05,
+      "loss": 0.8217,
+      "step": 54000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.8114035451980924e-05,
+      "loss": 0.8162,
+      "step": 54500
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.8096733024934875e-05,
+      "loss": 0.8187,
+      "step": 55000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.807943059788883e-05,
+      "loss": 0.8161,
+      "step": 55500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.806212817084278e-05,
+      "loss": 0.8158,
+      "step": 56000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.8044825743796734e-05,
+      "loss": 0.815,
+      "step": 56500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.802752331675069e-05,
+      "loss": 0.8144,
+      "step": 57000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.801022088970464e-05,
+      "loss": 0.8111,
+      "step": 57500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.79929184626586e-05,
+      "loss": 0.8122,
+      "step": 58000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.797561603561255e-05,
+      "loss": 0.8077,
+      "step": 58500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.79583136085665e-05,
+      "loss": 0.8111,
+      "step": 59000
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.794101118152046e-05,
+      "loss": 0.8113,
+      "step": 59500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.792370875447441e-05,
+      "loss": 0.81,
+      "step": 60000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.790640632742837e-05,
+      "loss": 0.8167,
+      "step": 60500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.788910390038232e-05,
+      "loss": 0.8092,
+      "step": 61000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.787180147333627e-05,
+      "loss": 0.8052,
+      "step": 61500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.7854499046290226e-05,
+      "loss": 0.8035,
+      "step": 62000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.783719661924418e-05,
+      "loss": 0.8066,
+      "step": 62500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.781989419219813e-05,
+      "loss": 0.8046,
+      "step": 63000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.7802591765152086e-05,
+      "loss": 0.8043,
+      "step": 63500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.7785289338106036e-05,
+      "loss": 0.8023,
+      "step": 64000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.7767986911059994e-05,
+      "loss": 0.8048,
+      "step": 64500
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.775068448401394e-05,
+      "loss": 0.8039,
+      "step": 65000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.7733382056967896e-05,
+      "loss": 0.8056,
+      "step": 65500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.771607962992185e-05,
+      "loss": 0.801,
+      "step": 66000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.7698777202875804e-05,
+      "loss": 0.7974,
+      "step": 66500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.768147477582976e-05,
+      "loss": 0.7979,
+      "step": 67000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.7664172348783706e-05,
+      "loss": 0.7978,
+      "step": 67500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.764686992173766e-05,
+      "loss": 0.8001,
+      "step": 68000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.762956749469162e-05,
+      "loss": 0.7976,
+      "step": 68500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.761226506764557e-05,
+      "loss": 0.8022,
+      "step": 69000
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.759496264059952e-05,
+      "loss": 0.7929,
+      "step": 69500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.757766021355348e-05,
+      "loss": 0.7953,
+      "step": 70000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.756035778650743e-05,
+      "loss": 0.7933,
+      "step": 70500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.754305535946139e-05,
+      "loss": 0.7926,
+      "step": 71000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.752575293241534e-05,
+      "loss": 0.7925,
+      "step": 71500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.750845050536929e-05,
+      "loss": 0.795,
+      "step": 72000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.749114807832325e-05,
+      "loss": 0.7901,
+      "step": 72500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.74738456512772e-05,
+      "loss": 0.7915,
+      "step": 73000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.7456543224231156e-05,
+      "loss": 0.7932,
+      "step": 73500
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.74392407971851e-05,
+      "loss": 0.7913,
+      "step": 74000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.742193837013906e-05,
+      "loss": 0.7921,
+      "step": 74500
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.7404635943093015e-05,
+      "loss": 0.7902,
+      "step": 75000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.7387333516046966e-05,
+      "loss": 0.7885,
+      "step": 75500
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.737003108900092e-05,
+      "loss": 0.7925,
+      "step": 76000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.735272866195487e-05,
+      "loss": 0.7888,
+      "step": 76500
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.7335426234908825e-05,
+      "loss": 0.7879,
+      "step": 77000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.731812380786278e-05,
+      "loss": 0.7894,
+      "step": 77500
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.730082138081673e-05,
+      "loss": 0.787,
+      "step": 78000
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.7283518953770684e-05,
+      "loss": 0.7846,
+      "step": 78500
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.7266216526724635e-05,
+      "loss": 0.7878,
+      "step": 79000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.724891409967859e-05,
+      "loss": 0.7859,
+      "step": 79500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.723161167263255e-05,
+      "loss": 0.784,
+      "step": 80000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.7214309245586494e-05,
+      "loss": 0.7876,
+      "step": 80500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.719700681854045e-05,
+      "loss": 0.7856,
+      "step": 81000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.717970439149441e-05,
+      "loss": 0.785,
+      "step": 81500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.716240196444836e-05,
+      "loss": 0.7831,
+      "step": 82000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.714509953740232e-05,
+      "loss": 0.7856,
+      "step": 82500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.712779711035626e-05,
+      "loss": 0.7805,
+      "step": 83000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.711049468331022e-05,
+      "loss": 0.7826,
+      "step": 83500
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.709319225626418e-05,
+      "loss": 0.7822,
+      "step": 84000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.707588982921813e-05,
+      "loss": 0.7811,
+      "step": 84500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.705858740217208e-05,
+      "loss": 0.781,
+      "step": 85000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.704128497512603e-05,
+      "loss": 0.7836,
+      "step": 85500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.702398254807999e-05,
+      "loss": 0.7795,
+      "step": 86000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.7006680121033944e-05,
+      "loss": 0.7805,
+      "step": 86500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.6989377693987895e-05,
+      "loss": 0.7783,
+      "step": 87000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.6972075266941846e-05,
+      "loss": 0.7777,
+      "step": 87500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.69547728398958e-05,
+      "loss": 0.7762,
+      "step": 88000
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.6937470412849754e-05,
+      "loss": 0.778,
+      "step": 88500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.692016798580371e-05,
+      "loss": 0.7771,
+      "step": 89000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.6902865558757656e-05,
+      "loss": 0.7751,
+      "step": 89500
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.6885563131711613e-05,
+      "loss": 0.7801,
+      "step": 90000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.6868260704665564e-05,
+      "loss": 0.7757,
+      "step": 90500
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.685095827761952e-05,
+      "loss": 0.774,
+      "step": 91000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.683365585057347e-05,
+      "loss": 0.774,
+      "step": 91500
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.681635342352742e-05,
+      "loss": 0.7757,
+      "step": 92000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.679905099648138e-05,
+      "loss": 0.7743,
+      "step": 92500
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.678174856943534e-05,
+      "loss": 0.7753,
+      "step": 93000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.676444614238929e-05,
+      "loss": 0.7702,
+      "step": 93500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.674714371534324e-05,
+      "loss": 0.7765,
+      "step": 94000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.672984128829719e-05,
+      "loss": 0.7732,
+      "step": 94500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.671253886125115e-05,
+      "loss": 0.7707,
+      "step": 95000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.6695236434205106e-05,
+      "loss": 0.7707,
+      "step": 95500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.667793400715905e-05,
+      "loss": 0.7742,
+      "step": 96000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.666063158011301e-05,
+      "loss": 0.7722,
+      "step": 96500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.664332915306696e-05,
+      "loss": 0.771,
+      "step": 97000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.6626026726020916e-05,
+      "loss": 0.7725,
+      "step": 97500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.6608724298974874e-05,
+      "loss": 0.7682,
+      "step": 98000
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.659142187192882e-05,
+      "loss": 0.7698,
+      "step": 98500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.6574119444882775e-05,
+      "loss": 0.7699,
+      "step": 99000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.6556817017836726e-05,
+      "loss": 0.7701,
+      "step": 99500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.6539514590790684e-05,
+      "loss": 0.7707,
+      "step": 100000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.6522212163744634e-05,
+      "loss": 0.7673,
+      "step": 100500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.6504909736698585e-05,
+      "loss": 0.7688,
+      "step": 101000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.648760730965254e-05,
+      "loss": 0.7683,
+      "step": 101500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.6470304882606494e-05,
+      "loss": 0.7664,
+      "step": 102000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.6453002455560444e-05,
+      "loss": 0.7686,
+      "step": 102500
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.64357000285144e-05,
+      "loss": 0.7631,
+      "step": 103000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.641839760146835e-05,
+      "loss": 0.766,
+      "step": 103500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.640109517442231e-05,
+      "loss": 0.763,
+      "step": 104000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.638379274737627e-05,
+      "loss": 0.7652,
+      "step": 104500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.636649032033021e-05,
+      "loss": 0.7643,
+      "step": 105000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.634918789328417e-05,
+      "loss": 0.7684,
+      "step": 105500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.633188546623812e-05,
+      "loss": 0.7648,
+      "step": 106000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.631458303919208e-05,
+      "loss": 0.7646,
+      "step": 106500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.629728061214603e-05,
+      "loss": 0.7632,
+      "step": 107000
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.627997818509998e-05,
+      "loss": 0.7637,
+      "step": 107500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.626267575805394e-05,
+      "loss": 0.7655,
+      "step": 108000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.624537333100789e-05,
+      "loss": 0.763,
+      "step": 108500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.6228070903961845e-05,
+      "loss": 0.7623,
+      "step": 109000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.6210768476915796e-05,
+      "loss": 0.7644,
+      "step": 109500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.619346604986975e-05,
+      "loss": 0.7602,
+      "step": 110000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.6176163622823704e-05,
+      "loss": 0.7603,
+      "step": 110500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.6158861195777655e-05,
+      "loss": 0.7602,
+      "step": 111000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.6141558768731606e-05,
+      "loss": 0.7619,
+      "step": 111500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.6124256341685564e-05,
+      "loss": 0.7615,
+      "step": 112000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.6106953914639514e-05,
+      "loss": 0.7613,
+      "step": 112500
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.608965148759347e-05,
+      "loss": 0.758,
+      "step": 113000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.607234906054742e-05,
+      "loss": 0.7649,
+      "step": 113500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.6055046633501374e-05,
+      "loss": 0.7627,
+      "step": 114000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.603774420645533e-05,
+      "loss": 0.7607,
+      "step": 114500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.602044177940928e-05,
+      "loss": 0.7582,
+      "step": 115000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.600313935236324e-05,
+      "loss": 0.7613,
+      "step": 115500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.598583692531719e-05,
+      "loss": 0.7597,
+      "step": 116000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.596853449827114e-05,
+      "loss": 0.7553,
+      "step": 116500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.59512320712251e-05,
+      "loss": 0.759,
+      "step": 117000
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.593392964417905e-05,
+      "loss": 0.7578,
+      "step": 117500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.5916627217133e-05,
+      "loss": 0.7563,
+      "step": 118000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.589932479008696e-05,
+      "loss": 0.7547,
+      "step": 118500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.588202236304091e-05,
+      "loss": 0.7567,
+      "step": 119000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.5864719935994866e-05,
+      "loss": 0.7622,
+      "step": 119500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.584741750894882e-05,
+      "loss": 0.7592,
+      "step": 120000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.583011508190277e-05,
+      "loss": 0.7524,
+      "step": 120500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.5812812654856725e-05,
+      "loss": 0.7589,
+      "step": 121000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.5795510227810676e-05,
+      "loss": 0.7546,
+      "step": 121500
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.5778207800764634e-05,
+      "loss": 0.7544,
+      "step": 122000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.5760905373718585e-05,
+      "loss": 0.7557,
+      "step": 122500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.5743602946672535e-05,
+      "loss": 0.7542,
+      "step": 123000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.572630051962649e-05,
+      "loss": 0.7563,
+      "step": 123500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.5708998092580444e-05,
+      "loss": 0.7552,
+      "step": 124000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.5691695665534395e-05,
+      "loss": 0.752,
+      "step": 124500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.567439323848835e-05,
+      "loss": 0.7533,
+      "step": 125000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.56570908114423e-05,
+      "loss": 0.7568,
+      "step": 125500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.563978838439626e-05,
+      "loss": 0.7557,
+      "step": 126000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.562248595735021e-05,
+      "loss": 0.7518,
+      "step": 126500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.560518353030416e-05,
+      "loss": 0.7511,
+      "step": 127000
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.558788110325812e-05,
+      "loss": 0.7521,
+      "step": 127500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.557057867621207e-05,
+      "loss": 0.7536,
+      "step": 128000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.555327624916603e-05,
+      "loss": 0.7497,
+      "step": 128500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.553597382211998e-05,
+      "loss": 0.7507,
+      "step": 129000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.551867139507393e-05,
+      "loss": 0.7483,
+      "step": 129500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.550136896802789e-05,
+      "loss": 0.7484,
+      "step": 130000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.548406654098184e-05,
+      "loss": 0.7508,
+      "step": 130500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.5466764113935796e-05,
+      "loss": 0.749,
+      "step": 131000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.5449461686889746e-05,
+      "loss": 0.7509,
+      "step": 131500
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.54321592598437e-05,
+      "loss": 0.7543,
+      "step": 132000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.5414856832797655e-05,
+      "loss": 0.7492,
+      "step": 132500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.5397554405751606e-05,
+      "loss": 0.7485,
+      "step": 133000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.5380251978705556e-05,
+      "loss": 0.7489,
+      "step": 133500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.5362949551659514e-05,
+      "loss": 0.7515,
+      "step": 134000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.5345647124613465e-05,
+      "loss": 0.7492,
+      "step": 134500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.532834469756742e-05,
+      "loss": 0.7486,
+      "step": 135000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.531104227052137e-05,
+      "loss": 0.7482,
+      "step": 135500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.5293739843475324e-05,
+      "loss": 0.7462,
+      "step": 136000
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.527643741642928e-05,
+      "loss": 0.7432,
+      "step": 136500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.525913498938323e-05,
+      "loss": 0.7454,
+      "step": 137000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.524183256233719e-05,
+      "loss": 0.7498,
+      "step": 137500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.522453013529114e-05,
+      "loss": 0.7431,
+      "step": 138000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.520722770824509e-05,
+      "loss": 0.7475,
+      "step": 138500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.518992528119905e-05,
+      "loss": 0.7482,
+      "step": 139000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.5172622854153e-05,
+      "loss": 0.741,
+      "step": 139500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.515532042710695e-05,
+      "loss": 0.7451,
+      "step": 140000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.513801800006091e-05,
+      "loss": 0.747,
+      "step": 140500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.512071557301486e-05,
+      "loss": 0.7433,
+      "step": 141000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.5103413145968817e-05,
+      "loss": 0.7445,
+      "step": 141500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.508611071892277e-05,
+      "loss": 0.7464,
+      "step": 142000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.506880829187672e-05,
+      "loss": 0.7445,
+      "step": 142500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.5051505864830676e-05,
+      "loss": 0.7465,
+      "step": 143000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.5034203437784627e-05,
+      "loss": 0.7442,
+      "step": 143500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.5016901010738584e-05,
+      "loss": 0.7444,
+      "step": 144000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.4999598583692535e-05,
+      "loss": 0.7452,
+      "step": 144500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.4982296156646486e-05,
+      "loss": 0.7492,
+      "step": 145000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.496499372960044e-05,
+      "loss": 0.7428,
+      "step": 145500
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.4947691302554394e-05,
+      "loss": 0.7416,
+      "step": 146000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.493038887550835e-05,
+      "loss": 0.7431,
+      "step": 146500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.49130864484623e-05,
+      "loss": 0.7451,
+      "step": 147000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.489578402141625e-05,
+      "loss": 0.7441,
+      "step": 147500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.487848159437021e-05,
+      "loss": 0.7434,
+      "step": 148000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.486117916732416e-05,
+      "loss": 0.7429,
+      "step": 148500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.484387674027811e-05,
+      "loss": 0.7402,
+      "step": 149000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.482657431323207e-05,
+      "loss": 0.7403,
+      "step": 149500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.480927188618602e-05,
+      "loss": 0.74,
+      "step": 150000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.479196945913998e-05,
+      "loss": 0.7417,
+      "step": 150500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.477466703209392e-05,
+      "loss": 0.7424,
+      "step": 151000
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.475736460504788e-05,
+      "loss": 0.7426,
+      "step": 151500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.474006217800184e-05,
+      "loss": 0.739,
+      "step": 152000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.472275975095579e-05,
+      "loss": 0.7382,
+      "step": 152500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.4705457323909746e-05,
+      "loss": 0.7391,
+      "step": 153000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.468815489686369e-05,
+      "loss": 0.7398,
+      "step": 153500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.467085246981765e-05,
+      "loss": 0.7412,
+      "step": 154000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.4653550042771605e-05,
+      "loss": 0.7394,
+      "step": 154500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.4636247615725556e-05,
+      "loss": 0.7393,
+      "step": 155000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.461894518867951e-05,
+      "loss": 0.7386,
+      "step": 155500
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.4601642761633464e-05,
+      "loss": 0.7384,
+      "step": 156000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.4584340334587415e-05,
+      "loss": 0.7416,
+      "step": 156500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.456703790754137e-05,
+      "loss": 0.7376,
+      "step": 157000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.454973548049532e-05,
+      "loss": 0.7368,
+      "step": 157500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.4532433053449274e-05,
+      "loss": 0.7374,
+      "step": 158000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.451513062640323e-05,
+      "loss": 0.7425,
+      "step": 158500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.449782819935718e-05,
+      "loss": 0.7384,
+      "step": 159000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.448052577231114e-05,
+      "loss": 0.7376,
+      "step": 159500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.4463223345265084e-05,
+      "loss": 0.738,
+      "step": 160000
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.444592091821904e-05,
+      "loss": 0.7409,
+      "step": 160500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.4428618491173e-05,
+      "loss": 0.7379,
+      "step": 161000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.441131606412695e-05,
+      "loss": 0.7364,
+      "step": 161500
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.43940136370809e-05,
+      "loss": 0.7382,
+      "step": 162000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.437671121003485e-05,
+      "loss": 0.7371,
+      "step": 162500
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.435940878298881e-05,
+      "loss": 0.7337,
+      "step": 163000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.434210635594277e-05,
+      "loss": 0.7358,
+      "step": 163500
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.432480392889672e-05,
+      "loss": 0.7358,
+      "step": 164000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.430750150185067e-05,
+      "loss": 0.7337,
+      "step": 164500
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.429019907480462e-05,
+      "loss": 0.7362,
+      "step": 165000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.427289664775858e-05,
+      "loss": 0.7326,
+      "step": 165500
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.4255594220712534e-05,
+      "loss": 0.7357,
+      "step": 166000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.423829179366648e-05,
+      "loss": 0.735,
+      "step": 166500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.4220989366620436e-05,
+      "loss": 0.7366,
+      "step": 167000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.4203686939574394e-05,
+      "loss": 0.7352,
+      "step": 167500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.4186384512528344e-05,
+      "loss": 0.7381,
+      "step": 168000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.41690820854823e-05,
+      "loss": 0.7354,
+      "step": 168500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.4151779658436246e-05,
+      "loss": 0.7362,
+      "step": 169000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.4134477231390203e-05,
+      "loss": 0.7346,
+      "step": 169500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.411717480434416e-05,
+      "loss": 0.7321,
+      "step": 170000
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.409987237729811e-05,
+      "loss": 0.732,
+      "step": 170500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.408256995025206e-05,
+      "loss": 0.7351,
+      "step": 171000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.4065267523206013e-05,
+      "loss": 0.738,
+      "step": 171500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.404796509615997e-05,
+      "loss": 0.735,
+      "step": 172000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.403066266911393e-05,
+      "loss": 0.7344,
+      "step": 172500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.401336024206787e-05,
+      "loss": 0.7299,
+      "step": 173000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.399605781502183e-05,
+      "loss": 0.7318,
+      "step": 173500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.397875538797578e-05,
+      "loss": 0.7329,
+      "step": 174000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.396145296092974e-05,
+      "loss": 0.7308,
+      "step": 174500
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.3944150533883696e-05,
+      "loss": 0.7307,
+      "step": 175000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.392684810683764e-05,
+      "loss": 0.731,
+      "step": 175500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.39095456797916e-05,
+      "loss": 0.7324,
+      "step": 176000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.389224325274555e-05,
+      "loss": 0.7361,
+      "step": 176500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.3874940825699506e-05,
+      "loss": 0.7347,
+      "step": 177000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.385763839865346e-05,
+      "loss": 0.7328,
+      "step": 177500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.384033597160741e-05,
+      "loss": 0.7287,
+      "step": 178000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.3823033544561365e-05,
+      "loss": 0.7317,
+      "step": 178500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.380573111751532e-05,
+      "loss": 0.7326,
+      "step": 179000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.3788428690469274e-05,
+      "loss": 0.7324,
+      "step": 179500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.3771126263423224e-05,
+      "loss": 0.7319,
+      "step": 180000
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.3753823836377175e-05,
+      "loss": 0.7338,
+      "step": 180500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.373652140933113e-05,
+      "loss": 0.7324,
+      "step": 181000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.371921898228509e-05,
+      "loss": 0.7341,
+      "step": 181500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.3701916555239034e-05,
+      "loss": 0.7282,
+      "step": 182000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.368461412819299e-05,
+      "loss": 0.7329,
+      "step": 182500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.366731170114694e-05,
+      "loss": 0.7307,
+      "step": 183000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.36500092741009e-05,
+      "loss": 0.7295,
+      "step": 183500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.363270684705485e-05,
+      "loss": 0.7281,
+      "step": 184000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.36154044200088e-05,
+      "loss": 0.73,
+      "step": 184500
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.359810199296276e-05,
+      "loss": 0.7307,
+      "step": 185000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.358079956591671e-05,
+      "loss": 0.7268,
+      "step": 185500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.356349713887067e-05,
+      "loss": 0.7281,
+      "step": 186000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.354619471182462e-05,
+      "loss": 0.7274,
+      "step": 186500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.352889228477857e-05,
+      "loss": 0.7291,
+      "step": 187000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.351158985773253e-05,
+      "loss": 0.7303,
+      "step": 187500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.349428743068648e-05,
+      "loss": 0.7282,
+      "step": 188000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.347698500364043e-05,
+      "loss": 0.7278,
+      "step": 188500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.3459682576594386e-05,
+      "loss": 0.7307,
+      "step": 189000
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.344238014954834e-05,
+      "loss": 0.7314,
+      "step": 189500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.3425077722502295e-05,
+      "loss": 0.7284,
+      "step": 190000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.340777529545625e-05,
+      "loss": 0.7318,
+      "step": 190500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.3390472868410196e-05,
+      "loss": 0.7323,
+      "step": 191000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.3373170441364154e-05,
+      "loss": 0.726,
+      "step": 191500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.3355868014318105e-05,
+      "loss": 0.7276,
+      "step": 192000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.333856558727206e-05,
+      "loss": 0.7252,
+      "step": 192500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.332126316022601e-05,
+      "loss": 0.7254,
+      "step": 193000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.3303960733179964e-05,
+      "loss": 0.7282,
+      "step": 193500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.328665830613392e-05,
+      "loss": 0.7261,
+      "step": 194000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.326935587908787e-05,
+      "loss": 0.7287,
+      "step": 194500
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.325205345204182e-05,
+      "loss": 0.7259,
+      "step": 195000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.323475102499578e-05,
+      "loss": 0.7282,
+      "step": 195500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.321744859794973e-05,
+      "loss": 0.7254,
+      "step": 196000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.320014617090369e-05,
+      "loss": 0.7271,
+      "step": 196500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.318284374385764e-05,
+      "loss": 0.7235,
+      "step": 197000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.316554131681159e-05,
+      "loss": 0.7242,
+      "step": 197500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.314823888976555e-05,
+      "loss": 0.7259,
+      "step": 198000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.31309364627195e-05,
+      "loss": 0.7251,
+      "step": 198500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.3113634035673456e-05,
+      "loss": 0.7254,
+      "step": 199000
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.309633160862741e-05,
+      "loss": 0.7232,
+      "step": 199500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.307902918158136e-05,
+      "loss": 0.7235,
+      "step": 200000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.3061726754535316e-05,
+      "loss": 0.7233,
+      "step": 200500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.3044424327489266e-05,
+      "loss": 0.7229,
+      "step": 201000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.3027121900443224e-05,
+      "loss": 0.7228,
+      "step": 201500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.3009819473397175e-05,
+      "loss": 0.7254,
+      "step": 202000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.2992517046351125e-05,
+      "loss": 0.725,
+      "step": 202500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.297521461930508e-05,
+      "loss": 0.7247,
+      "step": 203000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.2957912192259034e-05,
+      "loss": 0.7229,
+      "step": 203500
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.2940609765212985e-05,
+      "loss": 0.7233,
+      "step": 204000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.292330733816694e-05,
+      "loss": 0.7203,
+      "step": 204500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.290600491112089e-05,
+      "loss": 0.7258,
+      "step": 205000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.288870248407485e-05,
+      "loss": 0.723,
+      "step": 205500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.28714000570288e-05,
+      "loss": 0.7224,
+      "step": 206000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.285409762998275e-05,
+      "loss": 0.7254,
+      "step": 206500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.283679520293671e-05,
+      "loss": 0.7237,
+      "step": 207000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.281949277589066e-05,
+      "loss": 0.7216,
+      "step": 207500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.280219034884462e-05,
+      "loss": 0.7216,
+      "step": 208000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.278488792179857e-05,
+      "loss": 0.7225,
+      "step": 208500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.276758549475252e-05,
+      "loss": 0.722,
+      "step": 209000
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.275028306770648e-05,
+      "loss": 0.7215,
+      "step": 209500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.273298064066043e-05,
+      "loss": 0.7224,
+      "step": 210000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.271567821361438e-05,
+      "loss": 0.7201,
+      "step": 210500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.2698375786568336e-05,
+      "loss": 0.7197,
+      "step": 211000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.268107335952229e-05,
+      "loss": 0.7205,
+      "step": 211500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.2663770932476245e-05,
+      "loss": 0.7191,
+      "step": 212000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.2646468505430196e-05,
+      "loss": 0.7233,
+      "step": 212500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.2629166078384146e-05,
+      "loss": 0.7181,
+      "step": 213000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.2611863651338104e-05,
+      "loss": 0.7211,
+      "step": 213500
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.2594561224292055e-05,
+      "loss": 0.7217,
+      "step": 214000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.257725879724601e-05,
+      "loss": 0.7223,
+      "step": 214500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.255995637019996e-05,
+      "loss": 0.7218,
+      "step": 215000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.2542653943153914e-05,
+      "loss": 0.7212,
+      "step": 215500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.252535151610787e-05,
+      "loss": 0.7201,
+      "step": 216000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.250804908906182e-05,
+      "loss": 0.7171,
+      "step": 216500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.249074666201578e-05,
+      "loss": 0.7244,
+      "step": 217000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.247344423496973e-05,
+      "loss": 0.7177,
+      "step": 217500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.245614180792368e-05,
+      "loss": 0.7195,
+      "step": 218000
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.243883938087764e-05,
+      "loss": 0.717,
+      "step": 218500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.242153695383159e-05,
+      "loss": 0.7194,
+      "step": 219000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.240423452678554e-05,
+      "loss": 0.72,
+      "step": 219500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.23869320997395e-05,
+      "loss": 0.7252,
+      "step": 220000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.236962967269345e-05,
+      "loss": 0.7195,
+      "step": 220500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.2352327245647407e-05,
+      "loss": 0.7186,
+      "step": 221000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.233502481860136e-05,
+      "loss": 0.7183,
+      "step": 221500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.231772239155531e-05,
+      "loss": 0.7199,
+      "step": 222000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.2300419964509266e-05,
+      "loss": 0.7185,
+      "step": 222500
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.2283117537463217e-05,
+      "loss": 0.7179,
+      "step": 223000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.2265815110417174e-05,
+      "loss": 0.7185,
+      "step": 223500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.2248512683371125e-05,
+      "loss": 0.7201,
+      "step": 224000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.2231210256325076e-05,
+      "loss": 0.7178,
+      "step": 224500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.221390782927903e-05,
+      "loss": 0.7179,
+      "step": 225000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.2196605402232984e-05,
+      "loss": 0.7196,
+      "step": 225500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.2179302975186935e-05,
+      "loss": 0.7172,
+      "step": 226000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.216200054814089e-05,
+      "loss": 0.7179,
+      "step": 226500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.214469812109484e-05,
+      "loss": 0.7194,
+      "step": 227000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.21273956940488e-05,
+      "loss": 0.7179,
+      "step": 227500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.211009326700275e-05,
+      "loss": 0.7179,
+      "step": 228000
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.20927908399567e-05,
+      "loss": 0.7139,
+      "step": 228500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.207548841291066e-05,
+      "loss": 0.7156,
+      "step": 229000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.205818598586461e-05,
+      "loss": 0.7157,
+      "step": 229500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.204088355881857e-05,
+      "loss": 0.7157,
+      "step": 230000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.202358113177252e-05,
+      "loss": 0.7201,
+      "step": 230500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.200627870472647e-05,
+      "loss": 0.7166,
+      "step": 231000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.198897627768043e-05,
+      "loss": 0.716,
+      "step": 231500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.197167385063438e-05,
+      "loss": 0.7194,
+      "step": 232000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.195437142358833e-05,
+      "loss": 0.717,
+      "step": 232500
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.193706899654229e-05,
+      "loss": 0.7136,
+      "step": 233000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.191976656949624e-05,
+      "loss": 0.7163,
+      "step": 233500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.1902464142450195e-05,
+      "loss": 0.7172,
+      "step": 234000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.1885161715404146e-05,
+      "loss": 0.7158,
+      "step": 234500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.18678592883581e-05,
+      "loss": 0.7176,
+      "step": 235000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.1850556861312054e-05,
+      "loss": 0.7139,
+      "step": 235500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.1833254434266005e-05,
+      "loss": 0.7185,
+      "step": 236000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.181595200721996e-05,
+      "loss": 0.7138,
+      "step": 236500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.179864958017391e-05,
+      "loss": 0.7171,
+      "step": 237000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.1781347153127864e-05,
+      "loss": 0.7193,
+      "step": 237500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.176404472608182e-05,
+      "loss": 0.7183,
+      "step": 238000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.174674229903577e-05,
+      "loss": 0.7184,
+      "step": 238500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.172943987198973e-05,
+      "loss": 0.7173,
+      "step": 239000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.1712137444943674e-05,
+      "loss": 0.717,
+      "step": 239500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.169483501789763e-05,
+      "loss": 0.7115,
+      "step": 240000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.167753259085159e-05,
+      "loss": 0.7124,
+      "step": 240500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.166023016380554e-05,
+      "loss": 0.7157,
+      "step": 241000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.164292773675949e-05,
+      "loss": 0.7142,
+      "step": 241500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.162562530971345e-05,
+      "loss": 0.7142,
+      "step": 242000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.16083228826674e-05,
+      "loss": 0.7148,
+      "step": 242500
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.159102045562136e-05,
+      "loss": 0.7122,
+      "step": 243000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.15737180285753e-05,
+      "loss": 0.7118,
+      "step": 243500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.155641560152926e-05,
+      "loss": 0.7149,
+      "step": 244000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.1539113174483216e-05,
+      "loss": 0.7145,
+      "step": 244500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.152181074743717e-05,
+      "loss": 0.7131,
+      "step": 245000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.1504508320391124e-05,
+      "loss": 0.7137,
+      "step": 245500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.148720589334507e-05,
+      "loss": 0.7131,
+      "step": 246000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.1469903466299026e-05,
+      "loss": 0.715,
+      "step": 246500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.1452601039252984e-05,
+      "loss": 0.7137,
+      "step": 247000
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.1435298612206934e-05,
+      "loss": 0.7131,
+      "step": 247500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.1417996185160885e-05,
+      "loss": 0.7131,
+      "step": 248000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.1400693758114836e-05,
+      "loss": 0.7128,
+      "step": 248500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.1383391331068794e-05,
+      "loss": 0.715,
+      "step": 249000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.136608890402275e-05,
+      "loss": 0.7149,
+      "step": 249500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.13487864769767e-05,
+      "loss": 0.7122,
+      "step": 250000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.133148404993065e-05,
+      "loss": 0.7136,
+      "step": 250500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.1314181622884603e-05,
+      "loss": 0.7108,
+      "step": 251000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.129687919583856e-05,
+      "loss": 0.7133,
+      "step": 251500
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.127957676879252e-05,
+      "loss": 0.7151,
+      "step": 252000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.126227434174646e-05,
+      "loss": 0.7166,
+      "step": 252500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.124497191470042e-05,
+      "loss": 0.7122,
+      "step": 253000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.122766948765438e-05,
+      "loss": 0.7154,
+      "step": 253500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.121036706060833e-05,
+      "loss": 0.7124,
+      "step": 254000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.119306463356228e-05,
+      "loss": 0.7121,
+      "step": 254500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.117576220651623e-05,
+      "loss": 0.7108,
+      "step": 255000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.115845977947019e-05,
+      "loss": 0.716,
+      "step": 255500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.1141157352424145e-05,
+      "loss": 0.7137,
+      "step": 256000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.1123854925378096e-05,
+      "loss": 0.712,
+      "step": 256500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.110655249833205e-05,
+      "loss": 0.7096,
+      "step": 257000
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.1089250071286e-05,
+      "loss": 0.7144,
+      "step": 257500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.1071947644239955e-05,
+      "loss": 0.7132,
+      "step": 258000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.105464521719391e-05,
+      "loss": 0.7105,
+      "step": 258500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.103734279014786e-05,
+      "loss": 0.7082,
+      "step": 259000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.1020040363101814e-05,
+      "loss": 0.7101,
+      "step": 259500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.1002737936055765e-05,
+      "loss": 0.7089,
+      "step": 260000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.098543550900972e-05,
+      "loss": 0.7088,
+      "step": 260500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.096813308196368e-05,
+      "loss": 0.7128,
+      "step": 261000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.0950830654917624e-05,
+      "loss": 0.7103,
+      "step": 261500
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.093352822787158e-05,
+      "loss": 0.7126,
+      "step": 262000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.091622580082553e-05,
+      "loss": 0.7137,
+      "step": 262500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.089892337377949e-05,
+      "loss": 0.7093,
+      "step": 263000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.088162094673344e-05,
+      "loss": 0.708,
+      "step": 263500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.086431851968739e-05,
+      "loss": 0.7103,
+      "step": 264000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.084701609264135e-05,
+      "loss": 0.71,
+      "step": 264500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.082971366559531e-05,
+      "loss": 0.7127,
+      "step": 265000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.081241123854925e-05,
+      "loss": 0.7088,
+      "step": 265500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.079510881150321e-05,
+      "loss": 0.7087,
+      "step": 266000
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.077780638445716e-05,
+      "loss": 0.7078,
+      "step": 266500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.076050395741112e-05,
+      "loss": 0.711,
+      "step": 267000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.0743201530365075e-05,
+      "loss": 0.7097,
+      "step": 267500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.072589910331902e-05,
+      "loss": 0.7068,
+      "step": 268000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.0708596676272976e-05,
+      "loss": 0.709,
+      "step": 268500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.069129424922693e-05,
+      "loss": 0.7079,
+      "step": 269000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.0673991822180885e-05,
+      "loss": 0.708,
+      "step": 269500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.0656689395134835e-05,
+      "loss": 0.7073,
+      "step": 270000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.0639386968088786e-05,
+      "loss": 0.7082,
+      "step": 270500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.0622084541042744e-05,
+      "loss": 0.7078,
+      "step": 271000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.0604782113996695e-05,
+      "loss": 0.7044,
+      "step": 271500
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.058747968695065e-05,
+      "loss": 0.7074,
+      "step": 272000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.05701772599046e-05,
+      "loss": 0.7086,
+      "step": 272500
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.0552874832858554e-05,
+      "loss": 0.707,
+      "step": 273000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.053557240581251e-05,
+      "loss": 0.7045,
+      "step": 273500
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.051826997876646e-05,
+      "loss": 0.7074,
+      "step": 274000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.050096755172041e-05,
+      "loss": 0.7056,
+      "step": 274500
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.048366512467437e-05,
+      "loss": 0.7061,
+      "step": 275000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.046636269762832e-05,
+      "loss": 0.705,
+      "step": 275500
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.044906027058228e-05,
+      "loss": 0.7063,
+      "step": 276000
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.0431757843536236e-05,
+      "loss": 0.7087,
+      "step": 276500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.041445541649018e-05,
+      "loss": 0.7069,
+      "step": 277000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.039715298944414e-05,
+      "loss": 0.707,
+      "step": 277500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.037985056239809e-05,
+      "loss": 0.7063,
+      "step": 278000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.0362548135352046e-05,
+      "loss": 0.7088,
+      "step": 278500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.0345245708306e-05,
+      "loss": 0.7041,
+      "step": 279000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.032794328125995e-05,
+      "loss": 0.7062,
+      "step": 279500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.0310640854213906e-05,
+      "loss": 0.7066,
+      "step": 280000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.0293338427167856e-05,
+      "loss": 0.7058,
+      "step": 280500
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.027603600012181e-05,
+      "loss": 0.7072,
+      "step": 281000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.0258733573075765e-05,
+      "loss": 0.7093,
+      "step": 281500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.0241431146029716e-05,
+      "loss": 0.7062,
+      "step": 282000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.022412871898367e-05,
+      "loss": 0.7063,
+      "step": 282500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.0206826291937624e-05,
+      "loss": 0.7064,
+      "step": 283000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.0189523864891575e-05,
+      "loss": 0.7047,
+      "step": 283500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.017222143784553e-05,
+      "loss": 0.7075,
+      "step": 284000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.015491901079948e-05,
+      "loss": 0.7065,
+      "step": 284500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.013761658375344e-05,
+      "loss": 0.701,
+      "step": 285000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.012031415670739e-05,
+      "loss": 0.7071,
+      "step": 285500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.010301172966134e-05,
+      "loss": 0.7043,
+      "step": 286000
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.00857093026153e-05,
+      "loss": 0.7066,
+      "step": 286500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.006840687556925e-05,
+      "loss": 0.7066,
+      "step": 287000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.005110444852321e-05,
+      "loss": 0.7052,
+      "step": 287500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.003380202147716e-05,
+      "loss": 0.7077,
+      "step": 288000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.001649959443111e-05,
+      "loss": 0.7045,
+      "step": 288500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 3.999919716738507e-05,
+      "loss": 0.7046,
+      "step": 289000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 3.998189474033902e-05,
+      "loss": 0.7035,
+      "step": 289500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 3.996459231329297e-05,
+      "loss": 0.704,
+      "step": 290000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 3.9947289886246927e-05,
+      "loss": 0.7048,
+      "step": 290500
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 3.992998745920088e-05,
+      "loss": 0.7057,
+      "step": 291000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 3.9912685032154835e-05,
+      "loss": 0.7046,
+      "step": 291500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 3.9895382605108786e-05,
+      "loss": 0.7051,
+      "step": 292000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 3.9878080178062736e-05,
+      "loss": 0.7015,
+      "step": 292500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 3.9860777751016694e-05,
+      "loss": 0.7031,
+      "step": 293000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 3.9843475323970645e-05,
+      "loss": 0.706,
+      "step": 293500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 3.98261728969246e-05,
+      "loss": 0.7017,
+      "step": 294000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 3.980887046987855e-05,
+      "loss": 0.7027,
+      "step": 294500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 3.9791568042832504e-05,
+      "loss": 0.7043,
+      "step": 295000
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 3.977426561578646e-05,
+      "loss": 0.707,
+      "step": 295500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 3.975696318874041e-05,
+      "loss": 0.7039,
+      "step": 296000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 3.973966076169436e-05,
+      "loss": 0.7039,
+      "step": 296500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 3.972235833464832e-05,
+      "loss": 0.7056,
+      "step": 297000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 3.970505590760227e-05,
+      "loss": 0.7066,
+      "step": 297500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 3.968775348055623e-05,
+      "loss": 0.7069,
+      "step": 298000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 3.967045105351018e-05,
+      "loss": 0.705,
+      "step": 298500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 3.965314862646413e-05,
+      "loss": 0.7062,
+      "step": 299000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 3.963584619941809e-05,
+      "loss": 0.7031,
+      "step": 299500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 3.961854377237204e-05,
+      "loss": 0.7028,
+      "step": 300000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 3.9601241345326e-05,
+      "loss": 0.7007,
+      "step": 300500
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 3.958393891827995e-05,
+      "loss": 0.7035,
+      "step": 301000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 3.95666364912339e-05,
+      "loss": 0.6988,
+      "step": 301500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 3.9549334064187856e-05,
+      "loss": 0.6997,
+      "step": 302000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 3.953203163714181e-05,
+      "loss": 0.7025,
+      "step": 302500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 3.951472921009576e-05,
+      "loss": 0.7059,
+      "step": 303000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 3.9497426783049715e-05,
+      "loss": 0.7001,
+      "step": 303500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 3.9480124356003666e-05,
+      "loss": 0.7035,
+      "step": 304000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 3.946282192895762e-05,
+      "loss": 0.7031,
+      "step": 304500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 3.9445519501911574e-05,
+      "loss": 0.7032,
+      "step": 305000
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 3.9428217074865525e-05,
+      "loss": 0.7046,
+      "step": 305500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 3.941091464781948e-05,
+      "loss": 0.7014,
+      "step": 306000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 3.939361222077343e-05,
+      "loss": 0.7028,
+      "step": 306500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 3.937630979372739e-05,
+      "loss": 0.7017,
+      "step": 307000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 3.935900736668134e-05,
+      "loss": 0.704,
+      "step": 307500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 3.934170493963529e-05,
+      "loss": 0.7001,
+      "step": 308000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 3.932440251258925e-05,
+      "loss": 0.7021,
+      "step": 308500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 3.93071000855432e-05,
+      "loss": 0.7044,
+      "step": 309000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 3.928979765849716e-05,
+      "loss": 0.7034,
+      "step": 309500
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 3.927249523145111e-05,
+      "loss": 0.7024,
+      "step": 310000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 3.925519280440506e-05,
+      "loss": 0.7039,
+      "step": 310500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 3.923789037735902e-05,
+      "loss": 0.7054,
+      "step": 311000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 3.922058795031297e-05,
+      "loss": 0.7036,
+      "step": 311500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 3.920328552326692e-05,
+      "loss": 0.699,
+      "step": 312000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 3.918598309622088e-05,
+      "loss": 0.7025,
+      "step": 312500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 3.916868066917483e-05,
+      "loss": 0.7024,
+      "step": 313000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 3.9151378242128785e-05,
+      "loss": 0.703,
+      "step": 313500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 3.913407581508273e-05,
+      "loss": 0.6986,
+      "step": 314000
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 3.911677338803669e-05,
+      "loss": 0.7024,
+      "step": 314500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 3.9099470960990644e-05,
+      "loss": 0.7017,
+      "step": 315000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 3.9082168533944595e-05,
+      "loss": 0.7004,
+      "step": 315500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 3.906486610689855e-05,
+      "loss": 0.7005,
+      "step": 316000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 3.9047563679852503e-05,
+      "loss": 0.7025,
+      "step": 316500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 3.9030261252806454e-05,
+      "loss": 0.7,
+      "step": 317000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 3.901295882576041e-05,
+      "loss": 0.7022,
+      "step": 317500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 3.899565639871436e-05,
+      "loss": 0.6958,
+      "step": 318000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 3.8978353971668313e-05,
+      "loss": 0.7018,
+      "step": 318500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 3.896105154462227e-05,
+      "loss": 0.7028,
+      "step": 319000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 3.894374911757622e-05,
+      "loss": 0.6997,
+      "step": 319500
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 3.892644669053018e-05,
+      "loss": 0.7023,
+      "step": 320000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 3.890914426348413e-05,
+      "loss": 0.6995,
+      "step": 320500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 3.889184183643808e-05,
+      "loss": 0.6987,
+      "step": 321000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 3.887453940939204e-05,
+      "loss": 0.6987,
+      "step": 321500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 3.885723698234599e-05,
+      "loss": 0.7028,
+      "step": 322000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 3.883993455529995e-05,
+      "loss": 0.7013,
+      "step": 322500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 3.882263212825389e-05,
+      "loss": 0.6988,
+      "step": 323000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 3.880532970120785e-05,
+      "loss": 0.6989,
+      "step": 323500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 3.8788027274161806e-05,
+      "loss": 0.7001,
+      "step": 324000
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 3.877072484711576e-05,
+      "loss": 0.7,
+      "step": 324500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 3.875342242006971e-05,
+      "loss": 0.6978,
+      "step": 325000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 3.873611999302366e-05,
+      "loss": 0.6977,
+      "step": 325500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 3.8718817565977616e-05,
+      "loss": 0.7004,
+      "step": 326000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 3.8701515138931574e-05,
+      "loss": 0.7009,
+      "step": 326500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 3.8684212711885524e-05,
+      "loss": 0.6988,
+      "step": 327000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 3.8666910284839475e-05,
+      "loss": 0.7002,
+      "step": 327500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 3.864960785779343e-05,
+      "loss": 0.7003,
+      "step": 328000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 3.8632305430747384e-05,
+      "loss": 0.7012,
+      "step": 328500
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 3.861500300370134e-05,
+      "loss": 0.6975,
+      "step": 329000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 3.8597700576655285e-05,
+      "loss": 0.6992,
+      "step": 329500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 3.858039814960924e-05,
+      "loss": 0.6955,
+      "step": 330000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 3.85630957225632e-05,
+      "loss": 0.6982,
+      "step": 330500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 3.854579329551715e-05,
+      "loss": 0.6977,
+      "step": 331000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 3.852849086847111e-05,
+      "loss": 0.7023,
+      "step": 331500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 3.851118844142505e-05,
+      "loss": 0.7031,
+      "step": 332000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 3.849388601437901e-05,
+      "loss": 0.698,
+      "step": 332500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 3.847658358733297e-05,
+      "loss": 0.6974,
+      "step": 333000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 3.845928116028692e-05,
+      "loss": 0.6984,
+      "step": 333500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 3.844197873324087e-05,
+      "loss": 0.6949,
+      "step": 334000
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 3.842467630619482e-05,
+      "loss": 0.6974,
+      "step": 334500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 3.840737387914878e-05,
+      "loss": 0.696,
+      "step": 335000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 3.8390071452102735e-05,
+      "loss": 0.6978,
+      "step": 335500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 3.8372769025056686e-05,
+      "loss": 0.6987,
+      "step": 336000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 3.835546659801064e-05,
+      "loss": 0.6974,
+      "step": 336500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 3.833816417096459e-05,
+      "loss": 0.6989,
+      "step": 337000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 3.8320861743918545e-05,
+      "loss": 0.6997,
+      "step": 337500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 3.83035593168725e-05,
+      "loss": 0.6982,
+      "step": 338000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 3.828625688982645e-05,
+      "loss": 0.6989,
+      "step": 338500
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 3.8268954462780405e-05,
+      "loss": 0.6938,
+      "step": 339000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 3.825165203573436e-05,
+      "loss": 0.6975,
+      "step": 339500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 3.823434960868831e-05,
+      "loss": 0.6945,
+      "step": 340000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 3.8217047181642264e-05,
+      "loss": 0.698,
+      "step": 340500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 3.8199744754596215e-05,
+      "loss": 0.6962,
+      "step": 341000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 3.818244232755017e-05,
+      "loss": 0.6969,
+      "step": 341500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 3.816513990050413e-05,
+      "loss": 0.6971,
+      "step": 342000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 3.814783747345808e-05,
+      "loss": 0.6965,
+      "step": 342500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 3.813053504641203e-05,
+      "loss": 0.6893,
+      "step": 343000
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 3.811323261936598e-05,
+      "loss": 0.6975,
+      "step": 343500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 3.809593019231994e-05,
+      "loss": 0.6978,
+      "step": 344000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.80786277652739e-05,
+      "loss": 0.6972,
+      "step": 344500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.806132533822784e-05,
+      "loss": 0.6947,
+      "step": 345000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.80440229111818e-05,
+      "loss": 0.6951,
+      "step": 345500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.802672048413575e-05,
+      "loss": 0.6948,
+      "step": 346000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.800941805708971e-05,
+      "loss": 0.6982,
+      "step": 346500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.7992115630043665e-05,
+      "loss": 0.695,
+      "step": 347000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.797481320299761e-05,
+      "loss": 0.6963,
+      "step": 347500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.7957510775951566e-05,
+      "loss": 0.6958,
+      "step": 348000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.794020834890552e-05,
+      "loss": 0.6963,
+      "step": 348500
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 3.7922905921859475e-05,
+      "loss": 0.6944,
+      "step": 349000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.7905603494813425e-05,
+      "loss": 0.6924,
+      "step": 349500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.7888301067767376e-05,
+      "loss": 0.696,
+      "step": 350000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.7870998640721334e-05,
+      "loss": 0.6968,
+      "step": 350500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.785369621367529e-05,
+      "loss": 0.6939,
+      "step": 351000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.7836393786629235e-05,
+      "loss": 0.6961,
+      "step": 351500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.781909135958319e-05,
+      "loss": 0.6965,
+      "step": 352000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.7801788932537144e-05,
+      "loss": 0.6946,
+      "step": 352500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.77844865054911e-05,
+      "loss": 0.6949,
+      "step": 353000
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 3.776718407844506e-05,
+      "loss": 0.6943,
+      "step": 353500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.7749881651399e-05,
+      "loss": 0.6971,
+      "step": 354000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.773257922435296e-05,
+      "loss": 0.6944,
+      "step": 354500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.771527679730691e-05,
+      "loss": 0.696,
+      "step": 355000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.769797437026087e-05,
+      "loss": 0.6944,
+      "step": 355500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.768067194321482e-05,
+      "loss": 0.6947,
+      "step": 356000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.766336951616877e-05,
+      "loss": 0.6968,
+      "step": 356500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.764606708912273e-05,
+      "loss": 0.6951,
+      "step": 357000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.762876466207668e-05,
+      "loss": 0.6945,
+      "step": 357500
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.7611462235030636e-05,
+      "loss": 0.6945,
+      "step": 358000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 3.759415980798459e-05,
+      "loss": 0.6926,
+      "step": 358500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.757685738093854e-05,
+      "loss": 0.6965,
+      "step": 359000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.7559554953892496e-05,
+      "loss": 0.6948,
+      "step": 359500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.7542252526846446e-05,
+      "loss": 0.6941,
+      "step": 360000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.75249500998004e-05,
+      "loss": 0.6918,
+      "step": 360500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.7507647672754355e-05,
+      "loss": 0.6922,
+      "step": 361000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.7490345245708306e-05,
+      "loss": 0.6945,
+      "step": 361500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.747304281866226e-05,
+      "loss": 0.6972,
+      "step": 362000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.7455740391616214e-05,
+      "loss": 0.6923,
+      "step": 362500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.7438437964570165e-05,
+      "loss": 0.6946,
+      "step": 363000
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 3.742113553752412e-05,
+      "loss": 0.692,
+      "step": 363500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 3.740383311047807e-05,
+      "loss": 0.6934,
+      "step": 364000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 3.738653068343203e-05,
+      "loss": 0.6938,
+      "step": 364500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 3.736922825638598e-05,
+      "loss": 0.6944,
+      "step": 365000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 3.735192582933993e-05,
+      "loss": 0.6936,
+      "step": 365500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 3.733462340229389e-05,
+      "loss": 0.6951,
+      "step": 366000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 3.731732097524784e-05,
+      "loss": 0.6938,
+      "step": 366500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 3.730001854820179e-05,
+      "loss": 0.6928,
+      "step": 367000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 3.728271612115575e-05,
+      "loss": 0.6913,
+      "step": 367500
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 3.72654136941097e-05,
+      "loss": 0.6955,
+      "step": 368000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 3.724811126706366e-05,
+      "loss": 0.6886,
+      "step": 368500
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 3.723080884001761e-05,
+      "loss": 0.6944,
+      "step": 369000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 3.721350641297156e-05,
+      "loss": 0.6942,
+      "step": 369500
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 3.7196203985925517e-05,
+      "loss": 0.6946,
+      "step": 370000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 3.717890155887947e-05,
+      "loss": 0.6934,
+      "step": 370500
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 3.7161599131833425e-05,
+      "loss": 0.6929,
+      "step": 371000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 3.7144296704787376e-05,
+      "loss": 0.693,
+      "step": 371500
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 3.7126994277741327e-05,
+      "loss": 0.6918,
+      "step": 372000
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 3.7109691850695284e-05,
+      "loss": 0.6926,
+      "step": 372500
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 3.7092389423649235e-05,
+      "loss": 0.6937,
+      "step": 373000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.7075086996603186e-05,
+      "loss": 0.695,
+      "step": 373500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.705778456955714e-05,
+      "loss": 0.6922,
+      "step": 374000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.7040482142511094e-05,
+      "loss": 0.6917,
+      "step": 374500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.702317971546505e-05,
+      "loss": 0.6938,
+      "step": 375000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.7005877288419e-05,
+      "loss": 0.6909,
+      "step": 375500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.698857486137295e-05,
+      "loss": 0.6942,
+      "step": 376000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.697127243432691e-05,
+      "loss": 0.6931,
+      "step": 376500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.695397000728086e-05,
+      "loss": 0.6948,
+      "step": 377000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.693666758023482e-05,
+      "loss": 0.6937,
+      "step": 377500
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 3.691936515318877e-05,
+      "loss": 0.69,
+      "step": 378000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.690206272614272e-05,
+      "loss": 0.691,
+      "step": 378500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.688476029909668e-05,
+      "loss": 0.6935,
+      "step": 379000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.686745787205063e-05,
+      "loss": 0.6907,
+      "step": 379500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.685015544500459e-05,
+      "loss": 0.6969,
+      "step": 380000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.683285301795854e-05,
+      "loss": 0.6874,
+      "step": 380500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.681555059091249e-05,
+      "loss": 0.6898,
+      "step": 381000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.6798248163866446e-05,
+      "loss": 0.6952,
+      "step": 381500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.67809457368204e-05,
+      "loss": 0.6901,
+      "step": 382000
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 3.676364330977435e-05,
+      "loss": 0.6917,
+      "step": 382500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 3.6746340882728305e-05,
+      "loss": 0.6911,
+      "step": 383000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 3.6729038455682256e-05,
+      "loss": 0.6894,
+      "step": 383500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 3.6711736028636213e-05,
+      "loss": 0.6884,
+      "step": 384000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 3.6694433601590164e-05,
+      "loss": 0.6952,
+      "step": 384500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 3.6677131174544115e-05,
+      "loss": 0.6892,
+      "step": 385000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 3.665982874749807e-05,
+      "loss": 0.6908,
+      "step": 385500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 3.6642526320452023e-05,
+      "loss": 0.6895,
+      "step": 386000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 3.662522389340598e-05,
+      "loss": 0.6931,
+      "step": 386500
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 3.660792146635993e-05,
+      "loss": 0.6886,
+      "step": 387000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 3.659061903931388e-05,
+      "loss": 0.6914,
+      "step": 387500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 3.657331661226784e-05,
+      "loss": 0.693,
+      "step": 388000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 3.655601418522179e-05,
+      "loss": 0.6895,
+      "step": 388500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 3.653871175817574e-05,
+      "loss": 0.6906,
+      "step": 389000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 3.65214093311297e-05,
+      "loss": 0.6939,
+      "step": 389500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 3.650410690408365e-05,
+      "loss": 0.6899,
+      "step": 390000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 3.648680447703761e-05,
+      "loss": 0.6899,
+      "step": 390500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 3.646950204999156e-05,
+      "loss": 0.6866,
+      "step": 391000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 3.645219962294551e-05,
+      "loss": 0.6906,
+      "step": 391500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 3.643489719589947e-05,
+      "loss": 0.6913,
+      "step": 392000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 3.641759476885342e-05,
+      "loss": 0.6874,
+      "step": 392500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 3.6400292341807375e-05,
+      "loss": 0.6887,
+      "step": 393000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 3.6382989914761326e-05,
+      "loss": 0.6923,
+      "step": 393500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 3.636568748771528e-05,
+      "loss": 0.6912,
+      "step": 394000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 3.6348385060669234e-05,
+      "loss": 0.6916,
+      "step": 394500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 3.6331082633623185e-05,
+      "loss": 0.6883,
+      "step": 395000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 3.6313780206577136e-05,
+      "loss": 0.6907,
+      "step": 395500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 3.6296477779531094e-05,
+      "loss": 0.688,
+      "step": 396000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 3.6279175352485044e-05,
+      "loss": 0.6935,
+      "step": 396500
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 3.6261872925439e-05,
+      "loss": 0.6899,
+      "step": 397000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 3.624457049839295e-05,
+      "loss": 0.6912,
+      "step": 397500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 3.6227268071346904e-05,
+      "loss": 0.6924,
+      "step": 398000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 3.620996564430086e-05,
+      "loss": 0.6874,
+      "step": 398500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 3.619266321725481e-05,
+      "loss": 0.6869,
+      "step": 399000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 3.617536079020877e-05,
+      "loss": 0.6907,
+      "step": 399500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 3.6158058363162713e-05,
+      "loss": 0.6898,
+      "step": 400000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 3.614075593611667e-05,
+      "loss": 0.6914,
+      "step": 400500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 3.612345350907063e-05,
+      "loss": 0.6892,
+      "step": 401000
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 3.610615108202458e-05,
+      "loss": 0.6897,
+      "step": 401500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 3.608884865497854e-05,
+      "loss": 0.6891,
+      "step": 402000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 3.607154622793249e-05,
+      "loss": 0.6895,
+      "step": 402500
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 3.605424380088644e-05,
+      "loss": 0.6912,
+      "step": 403000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 3.6036941373840396e-05,
+      "loss": 0.6895,
+      "step": 403500
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 3.601963894679435e-05,
+      "loss": 0.6892,
+      "step": 404000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 3.60023365197483e-05,
+      "loss": 0.6862,
+      "step": 404500
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 3.5985034092702255e-05,
+      "loss": 0.6856,
+      "step": 405000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 3.5967731665656206e-05,
+      "loss": 0.6891,
+      "step": 405500
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 3.5950429238610164e-05,
+      "loss": 0.6886,
+      "step": 406000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 3.5933126811564114e-05,
+      "loss": 0.6884,
+      "step": 406500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.5915824384518065e-05,
+      "loss": 0.6897,
+      "step": 407000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.589852195747202e-05,
+      "loss": 0.6923,
+      "step": 407500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.5881219530425974e-05,
+      "loss": 0.6864,
+      "step": 408000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.586391710337993e-05,
+      "loss": 0.6882,
+      "step": 408500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.5846614676333875e-05,
+      "loss": 0.6919,
+      "step": 409000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.582931224928783e-05,
+      "loss": 0.6871,
+      "step": 409500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.581200982224179e-05,
+      "loss": 0.6887,
+      "step": 410000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.579470739519574e-05,
+      "loss": 0.6882,
+      "step": 410500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.577740496814969e-05,
+      "loss": 0.688,
+      "step": 411000
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 3.576010254110364e-05,
+      "loss": 0.6881,
+      "step": 411500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.57428001140576e-05,
+      "loss": 0.6882,
+      "step": 412000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.572549768701156e-05,
+      "loss": 0.6906,
+      "step": 412500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.570819525996551e-05,
+      "loss": 0.6882,
+      "step": 413000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.569089283291946e-05,
+      "loss": 0.687,
+      "step": 413500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.567359040587342e-05,
+      "loss": 0.6896,
+      "step": 414000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.565628797882737e-05,
+      "loss": 0.6911,
+      "step": 414500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.5638985551781325e-05,
+      "loss": 0.6879,
+      "step": 415000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.562168312473527e-05,
+      "loss": 0.6892,
+      "step": 415500
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.560438069768923e-05,
+      "loss": 0.6889,
+      "step": 416000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 3.5587078270643185e-05,
+      "loss": 0.685,
+      "step": 416500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 3.5569775843597135e-05,
+      "loss": 0.6835,
+      "step": 417000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 3.555247341655109e-05,
+      "loss": 0.6903,
+      "step": 417500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 3.553517098950504e-05,
+      "loss": 0.6837,
+      "step": 418000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 3.5517868562458995e-05,
+      "loss": 0.6899,
+      "step": 418500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 3.550056613541295e-05,
+      "loss": 0.6897,
+      "step": 419000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 3.54832637083669e-05,
+      "loss": 0.6879,
+      "step": 419500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 3.5465961281320854e-05,
+      "loss": 0.6874,
+      "step": 420000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 3.5448658854274805e-05,
+      "loss": 0.6864,
+      "step": 420500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 3.543135642722876e-05,
+      "loss": 0.6883,
+      "step": 421000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 3.541405400018272e-05,
+      "loss": 0.688,
+      "step": 421500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 3.5396751573136664e-05,
+      "loss": 0.683,
+      "step": 422000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 3.537944914609062e-05,
+      "loss": 0.6851,
+      "step": 422500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 3.536214671904457e-05,
+      "loss": 0.6862,
+      "step": 423000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 3.534484429199853e-05,
+      "loss": 0.6868,
+      "step": 423500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 3.532754186495249e-05,
+      "loss": 0.6852,
+      "step": 424000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 3.531023943790643e-05,
+      "loss": 0.6898,
+      "step": 424500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 3.529293701086039e-05,
+      "loss": 0.686,
+      "step": 425000
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 3.5275634583814346e-05,
+      "loss": 0.69,
+      "step": 425500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 3.52583321567683e-05,
+      "loss": 0.6898,
+      "step": 426000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 3.524102972972225e-05,
+      "loss": 0.6886,
+      "step": 426500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 3.52237273026762e-05,
+      "loss": 0.6886,
+      "step": 427000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 3.5206424875630156e-05,
+      "loss": 0.6862,
+      "step": 427500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 3.5189122448584114e-05,
+      "loss": 0.6848,
+      "step": 428000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 3.5171820021538065e-05,
+      "loss": 0.6854,
+      "step": 428500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 3.5154517594492016e-05,
+      "loss": 0.6862,
+      "step": 429000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 3.5137215167445966e-05,
+      "loss": 0.687,
+      "step": 429500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 3.5119912740399924e-05,
+      "loss": 0.6928,
+      "step": 430000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 3.510261031335388e-05,
+      "loss": 0.6883,
+      "step": 430500
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 3.5085307886307826e-05,
+      "loss": 0.6858,
+      "step": 431000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 3.506800545926178e-05,
+      "loss": 0.6884,
+      "step": 431500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 3.5050703032215734e-05,
+      "loss": 0.6864,
+      "step": 432000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 3.503340060516969e-05,
+      "loss": 0.6876,
+      "step": 432500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 3.501609817812364e-05,
+      "loss": 0.6845,
+      "step": 433000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 3.499879575107759e-05,
+      "loss": 0.6896,
+      "step": 433500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 3.498149332403155e-05,
+      "loss": 0.685,
+      "step": 434000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 3.49641908969855e-05,
+      "loss": 0.6892,
+      "step": 434500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 3.494688846993946e-05,
+      "loss": 0.6848,
+      "step": 435000
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 3.492958604289341e-05,
+      "loss": 0.686,
+      "step": 435500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.491228361584736e-05,
+      "loss": 0.6869,
+      "step": 436000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.489498118880132e-05,
+      "loss": 0.6859,
+      "step": 436500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.4877678761755276e-05,
+      "loss": 0.6869,
+      "step": 437000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.486037633470922e-05,
+      "loss": 0.6863,
+      "step": 437500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.484307390766318e-05,
+      "loss": 0.6865,
+      "step": 438000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.482577148061713e-05,
+      "loss": 0.6875,
+      "step": 438500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.4808469053571086e-05,
+      "loss": 0.6861,
+      "step": 439000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.479116662652504e-05,
+      "loss": 0.6882,
+      "step": 439500
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.477386419947899e-05,
+      "loss": 0.6852,
+      "step": 440000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 3.4756561772432945e-05,
+      "loss": 0.6836,
+      "step": 440500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.4739259345386896e-05,
+      "loss": 0.6832,
+      "step": 441000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.472195691834085e-05,
+      "loss": 0.6867,
+      "step": 441500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.4704654491294804e-05,
+      "loss": 0.685,
+      "step": 442000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.4687352064248755e-05,
+      "loss": 0.6858,
+      "step": 442500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.467004963720271e-05,
+      "loss": 0.6841,
+      "step": 443000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.465274721015666e-05,
+      "loss": 0.6847,
+      "step": 443500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.4635444783110614e-05,
+      "loss": 0.6858,
+      "step": 444000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.461814235606457e-05,
+      "loss": 0.6839,
+      "step": 444500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.460083992901852e-05,
+      "loss": 0.684,
+      "step": 445000
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 3.458353750197248e-05,
+      "loss": 0.6866,
+      "step": 445500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 3.456623507492643e-05,
+      "loss": 0.6864,
+      "step": 446000
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 3.454893264788038e-05,
+      "loss": 0.684,
+      "step": 446500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 3.453163022083434e-05,
+      "loss": 0.6844,
+      "step": 447000
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 3.451432779378829e-05,
+      "loss": 0.6871,
+      "step": 447500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 3.449702536674225e-05,
+      "loss": 0.6847,
+      "step": 448000
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 3.44797229396962e-05,
+      "loss": 0.6845,
+      "step": 448500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 3.446242051265015e-05,
+      "loss": 0.6842,
+      "step": 449000
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 3.444511808560411e-05,
+      "loss": 0.6866,
+      "step": 449500
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 3.442781565855806e-05,
+      "loss": 0.6845,
+      "step": 450000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 3.4410513231512015e-05,
+      "loss": 0.6835,
+      "step": 450500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 3.4393210804465966e-05,
+      "loss": 0.6835,
+      "step": 451000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 3.4375908377419917e-05,
+      "loss": 0.6858,
+      "step": 451500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 3.4358605950373874e-05,
+      "loss": 0.6825,
+      "step": 452000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 3.4341303523327825e-05,
+      "loss": 0.6871,
+      "step": 452500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 3.4324001096281776e-05,
+      "loss": 0.688,
+      "step": 453000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 3.430669866923573e-05,
+      "loss": 0.6829,
+      "step": 453500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 3.4289396242189684e-05,
+      "loss": 0.6863,
+      "step": 454000
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 3.427209381514364e-05,
+      "loss": 0.6852,
+      "step": 454500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 3.425479138809759e-05,
+      "loss": 0.6884,
+      "step": 455000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 3.423748896105154e-05,
+      "loss": 0.6869,
+      "step": 455500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 3.42201865340055e-05,
+      "loss": 0.6827,
+      "step": 456000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 3.420288410695945e-05,
+      "loss": 0.6814,
+      "step": 456500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 3.418558167991341e-05,
+      "loss": 0.6826,
+      "step": 457000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 3.416827925286736e-05,
+      "loss": 0.6849,
+      "step": 457500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 3.415097682582131e-05,
+      "loss": 0.6824,
+      "step": 458000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 3.413367439877527e-05,
+      "loss": 0.683,
+      "step": 458500
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 3.411637197172922e-05,
+      "loss": 0.682,
+      "step": 459000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 3.409906954468317e-05,
+      "loss": 0.6837,
+      "step": 459500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.408176711763713e-05,
+      "loss": 0.6811,
+      "step": 460000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.406446469059108e-05,
+      "loss": 0.6821,
+      "step": 460500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.4047162263545036e-05,
+      "loss": 0.6816,
+      "step": 461000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.402985983649899e-05,
+      "loss": 0.6826,
+      "step": 461500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.401255740945294e-05,
+      "loss": 0.6811,
+      "step": 462000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.3995254982406895e-05,
+      "loss": 0.6849,
+      "step": 462500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.3977952555360846e-05,
+      "loss": 0.6855,
+      "step": 463000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.3960650128314803e-05,
+      "loss": 0.6849,
+      "step": 463500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.3943347701268754e-05,
+      "loss": 0.6831,
+      "step": 464000
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 3.3926045274222705e-05,
+      "loss": 0.6838,
+      "step": 464500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.390874284717666e-05,
+      "loss": 0.6798,
+      "step": 465000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.3891440420130613e-05,
+      "loss": 0.6813,
+      "step": 465500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.3874137993084564e-05,
+      "loss": 0.6801,
+      "step": 466000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.385683556603852e-05,
+      "loss": 0.6828,
+      "step": 466500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.383953313899247e-05,
+      "loss": 0.6789,
+      "step": 467000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.382223071194643e-05,
+      "loss": 0.68,
+      "step": 467500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.380492828490038e-05,
+      "loss": 0.6828,
+      "step": 468000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.378762585785433e-05,
+      "loss": 0.6842,
+      "step": 468500
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.377032343080829e-05,
+      "loss": 0.6823,
+      "step": 469000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 3.375302100376224e-05,
+      "loss": 0.6833,
+      "step": 469500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.37357185767162e-05,
+      "loss": 0.684,
+      "step": 470000
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.371841614967015e-05,
+      "loss": 0.6805,
+      "step": 470500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.37011137226241e-05,
+      "loss": 0.6838,
+      "step": 471000
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.368381129557806e-05,
+      "loss": 0.6798,
+      "step": 471500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.366650886853201e-05,
+      "loss": 0.681,
+      "step": 472000
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.3649206441485965e-05,
+      "loss": 0.6826,
+      "step": 472500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.3631904014439916e-05,
+      "loss": 0.6838,
+      "step": 473000
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.361460158739387e-05,
+      "loss": 0.6811,
+      "step": 473500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 3.3597299160347824e-05,
+      "loss": 0.6824,
+      "step": 474000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.3579996733301775e-05,
+      "loss": 0.6816,
+      "step": 474500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.3562694306255726e-05,
+      "loss": 0.6807,
+      "step": 475000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.3545391879209684e-05,
+      "loss": 0.6834,
+      "step": 475500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.3528089452163634e-05,
+      "loss": 0.6829,
+      "step": 476000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.351078702511759e-05,
+      "loss": 0.6812,
+      "step": 476500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.349348459807154e-05,
+      "loss": 0.6841,
+      "step": 477000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.3476182171025494e-05,
+      "loss": 0.6838,
+      "step": 477500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.345887974397945e-05,
+      "loss": 0.681,
+      "step": 478000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.34415773169334e-05,
+      "loss": 0.6794,
+      "step": 478500
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 3.342427488988736e-05,
+      "loss": 0.6809,
+      "step": 479000
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.340697246284131e-05,
+      "loss": 0.6832,
+      "step": 479500
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.338967003579526e-05,
+      "loss": 0.6812,
+      "step": 480000
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.337236760874922e-05,
+      "loss": 0.683,
+      "step": 480500
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.335506518170317e-05,
+      "loss": 0.6796,
+      "step": 481000
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.333776275465712e-05,
+      "loss": 0.6828,
+      "step": 481500
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.332046032761108e-05,
+      "loss": 0.6783,
+      "step": 482000
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.330315790056503e-05,
+      "loss": 0.6784,
+      "step": 482500
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.3285855473518986e-05,
+      "loss": 0.6806,
+      "step": 483000
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.326855304647294e-05,
+      "loss": 0.6821,
+      "step": 483500
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 3.325125061942689e-05,
+      "loss": 0.678,
+      "step": 484000
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 3.3233948192380845e-05,
+      "loss": 0.6828,
+      "step": 484500
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 3.3216645765334796e-05,
+      "loss": 0.6796,
+      "step": 485000
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 3.3199343338288754e-05,
+      "loss": 0.6801,
+      "step": 485500
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 3.31820409112427e-05,
+      "loss": 0.6773,
+      "step": 486000
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 3.3164738484196655e-05,
+      "loss": 0.676,
+      "step": 486500
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 3.314743605715061e-05,
+      "loss": 0.676,
+      "step": 487000
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 3.3130133630104564e-05,
+      "loss": 0.681,
+      "step": 487500
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 3.311283120305852e-05,
+      "loss": 0.6802,
+      "step": 488000
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 3.309552877601247e-05,
+      "loss": 0.6765,
+      "step": 488500
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.307822634896642e-05,
+      "loss": 0.6775,
+      "step": 489000
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.306092392192038e-05,
+      "loss": 0.6824,
+      "step": 489500
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.304362149487433e-05,
+      "loss": 0.6765,
+      "step": 490000
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.302631906782828e-05,
+      "loss": 0.6804,
+      "step": 490500
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.300901664078224e-05,
+      "loss": 0.6786,
+      "step": 491000
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.299171421373619e-05,
+      "loss": 0.6803,
+      "step": 491500
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.297441178669015e-05,
+      "loss": 0.6787,
+      "step": 492000
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.295710935964409e-05,
+      "loss": 0.682,
+      "step": 492500
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.293980693259805e-05,
+      "loss": 0.6748,
+      "step": 493000
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 3.292250450555201e-05,
+      "loss": 0.6849,
+      "step": 493500
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 3.290520207850596e-05,
+      "loss": 0.6794,
+      "step": 494000
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 3.2887899651459916e-05,
+      "loss": 0.6777,
+      "step": 494500
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 3.287059722441386e-05,
+      "loss": 0.6768,
+      "step": 495000
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 3.285329479736782e-05,
+      "loss": 0.6816,
+      "step": 495500
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 3.2835992370321775e-05,
+      "loss": 0.6757,
+      "step": 496000
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 3.2818689943275725e-05,
+      "loss": 0.6771,
+      "step": 496500
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 3.2801387516229676e-05,
+      "loss": 0.6794,
+      "step": 497000
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 3.278408508918363e-05,
+      "loss": 0.6757,
+      "step": 497500
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 3.2766782662137585e-05,
+      "loss": 0.6783,
+      "step": 498000
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 3.274948023509154e-05,
+      "loss": 0.6833,
+      "step": 498500
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 3.273217780804549e-05,
+      "loss": 0.6767,
+      "step": 499000
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 3.2714875380999444e-05,
+      "loss": 0.6809,
+      "step": 499500
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 3.26975729539534e-05,
+      "loss": 0.6789,
+      "step": 500000
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 3.268027052690735e-05,
+      "loss": 0.6788,
+      "step": 500500
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 3.266296809986131e-05,
+      "loss": 0.6774,
+      "step": 501000
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 3.2645665672815254e-05,
+      "loss": 0.6786,
+      "step": 501500
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 3.262836324576921e-05,
+      "loss": 0.676,
+      "step": 502000
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 3.261106081872317e-05,
+      "loss": 0.6776,
+      "step": 502500
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 3.259375839167712e-05,
+      "loss": 0.6806,
+      "step": 503000
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.257645596463107e-05,
+      "loss": 0.6774,
+      "step": 503500
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.255915353758502e-05,
+      "loss": 0.6787,
+      "step": 504000
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.254185111053898e-05,
+      "loss": 0.6807,
+      "step": 504500
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.2524548683492936e-05,
+      "loss": 0.6822,
+      "step": 505000
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.250724625644689e-05,
+      "loss": 0.6801,
+      "step": 505500
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.248994382940084e-05,
+      "loss": 0.6781,
+      "step": 506000
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.247264140235479e-05,
+      "loss": 0.6755,
+      "step": 506500
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.2455338975308746e-05,
+      "loss": 0.6797,
+      "step": 507000
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.2438036548262704e-05,
+      "loss": 0.6782,
+      "step": 507500
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 3.242073412121665e-05,
+      "loss": 0.6748,
+      "step": 508000
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 3.2403431694170606e-05,
+      "loss": 0.6828,
+      "step": 508500
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 3.2386129267124556e-05,
+      "loss": 0.6777,
+      "step": 509000
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 3.2368826840078514e-05,
+      "loss": 0.6759,
+      "step": 509500
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 3.235152441303247e-05,
+      "loss": 0.6751,
+      "step": 510000
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 3.2334221985986416e-05,
+      "loss": 0.6766,
+      "step": 510500
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 3.231691955894037e-05,
+      "loss": 0.6756,
+      "step": 511000
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 3.229961713189433e-05,
+      "loss": 0.6813,
+      "step": 511500
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 3.228231470484828e-05,
+      "loss": 0.6765,
+      "step": 512000
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 3.226501227780223e-05,
+      "loss": 0.6756,
+      "step": 512500
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 3.224770985075618e-05,
+      "loss": 0.6762,
+      "step": 513000
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 3.223040742371014e-05,
+      "loss": 0.674,
+      "step": 513500
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 3.22131049966641e-05,
+      "loss": 0.677,
+      "step": 514000
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 3.219580256961804e-05,
+      "loss": 0.6753,
+      "step": 514500
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 3.2178500142572e-05,
+      "loss": 0.6803,
+      "step": 515000
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 3.216119771552595e-05,
+      "loss": 0.6768,
+      "step": 515500
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 3.214389528847991e-05,
+      "loss": 0.6785,
+      "step": 516000
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 3.2126592861433866e-05,
+      "loss": 0.6769,
+      "step": 516500
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 3.210929043438781e-05,
+      "loss": 0.6724,
+      "step": 517000
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 3.209198800734177e-05,
+      "loss": 0.6771,
+      "step": 517500
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 3.207468558029572e-05,
+      "loss": 0.6779,
+      "step": 518000
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 3.2057383153249676e-05,
+      "loss": 0.6746,
+      "step": 518500
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 3.2040080726203627e-05,
+      "loss": 0.6778,
+      "step": 519000
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 3.202277829915758e-05,
+      "loss": 0.6775,
+      "step": 519500
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 3.2005475872111535e-05,
+      "loss": 0.676,
+      "step": 520000
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 3.1988173445065486e-05,
+      "loss": 0.6789,
+      "step": 520500
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 3.197087101801944e-05,
+      "loss": 0.6768,
+      "step": 521000
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 3.1953568590973394e-05,
+      "loss": 0.6766,
+      "step": 521500
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 3.1936266163927345e-05,
+      "loss": 0.6744,
+      "step": 522000
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 3.19189637368813e-05,
+      "loss": 0.6772,
+      "step": 522500
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 3.190166130983526e-05,
+      "loss": 0.6766,
+      "step": 523000
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 3.1884358882789204e-05,
+      "loss": 0.6766,
+      "step": 523500
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 3.186705645574316e-05,
+      "loss": 0.675,
+      "step": 524000
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 3.184975402869711e-05,
+      "loss": 0.6759,
+      "step": 524500
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 3.183245160165107e-05,
+      "loss": 0.6753,
+      "step": 525000
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 3.181514917460502e-05,
+      "loss": 0.6825,
+      "step": 525500
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 3.179784674755897e-05,
+      "loss": 0.6761,
+      "step": 526000
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 3.178054432051293e-05,
+      "loss": 0.6761,
+      "step": 526500
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 3.176324189346688e-05,
+      "loss": 0.678,
+      "step": 527000
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.174593946642084e-05,
+      "loss": 0.6751,
+      "step": 527500
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.172863703937479e-05,
+      "loss": 0.6766,
+      "step": 528000
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.171133461232874e-05,
+      "loss": 0.678,
+      "step": 528500
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.16940321852827e-05,
+      "loss": 0.6762,
+      "step": 529000
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.167672975823665e-05,
+      "loss": 0.6756,
+      "step": 529500
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.16594273311906e-05,
+      "loss": 0.6774,
+      "step": 530000
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.1642124904144556e-05,
+      "loss": 0.6753,
+      "step": 530500
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.162482247709851e-05,
+      "loss": 0.6745,
+      "step": 531000
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.1607520050052464e-05,
+      "loss": 0.677,
+      "step": 531500
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 3.1590217623006415e-05,
+      "loss": 0.677,
+      "step": 532000
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.1572915195960366e-05,
+      "loss": 0.6769,
+      "step": 532500
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.1555612768914323e-05,
+      "loss": 0.6753,
+      "step": 533000
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.1538310341868274e-05,
+      "loss": 0.6764,
+      "step": 533500
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.152100791482223e-05,
+      "loss": 0.677,
+      "step": 534000
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.150370548777618e-05,
+      "loss": 0.6775,
+      "step": 534500
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.148640306073013e-05,
+      "loss": 0.6759,
+      "step": 535000
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.146910063368409e-05,
+      "loss": 0.6793,
+      "step": 535500
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.145179820663804e-05,
+      "loss": 0.6747,
+      "step": 536000
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.1434495779592e-05,
+      "loss": 0.6757,
+      "step": 536500
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 3.141719335254595e-05,
+      "loss": 0.6767,
+      "step": 537000
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 3.13998909254999e-05,
+      "loss": 0.6725,
+      "step": 537500
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 3.138258849845386e-05,
+      "loss": 0.6763,
+      "step": 538000
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 3.136528607140781e-05,
+      "loss": 0.678,
+      "step": 538500
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 3.134798364436176e-05,
+      "loss": 0.6779,
+      "step": 539000
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 3.133068121731572e-05,
+      "loss": 0.6758,
+      "step": 539500
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 3.131337879026967e-05,
+      "loss": 0.6775,
+      "step": 540000
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 3.1296076363223626e-05,
+      "loss": 0.673,
+      "step": 540500
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 3.127877393617758e-05,
+      "loss": 0.6738,
+      "step": 541000
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 3.126147150913153e-05,
+      "loss": 0.6734,
+      "step": 541500
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.1244169082085485e-05,
+      "loss": 0.6748,
+      "step": 542000
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.1226866655039436e-05,
+      "loss": 0.6771,
+      "step": 542500
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.1209564227993394e-05,
+      "loss": 0.6727,
+      "step": 543000
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.1192261800947344e-05,
+      "loss": 0.6752,
+      "step": 543500
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.1174959373901295e-05,
+      "loss": 0.6759,
+      "step": 544000
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.115765694685525e-05,
+      "loss": 0.6774,
+      "step": 544500
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.1140354519809204e-05,
+      "loss": 0.676,
+      "step": 545000
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.1123052092763154e-05,
+      "loss": 0.6754,
+      "step": 545500
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.110574966571711e-05,
+      "loss": 0.6734,
+      "step": 546000
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 3.108844723867106e-05,
+      "loss": 0.6748,
+      "step": 546500
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.107114481162502e-05,
+      "loss": 0.6757,
+      "step": 547000
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.105384238457897e-05,
+      "loss": 0.6767,
+      "step": 547500
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.103653995753292e-05,
+      "loss": 0.6741,
+      "step": 548000
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.101923753048688e-05,
+      "loss": 0.6732,
+      "step": 548500
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.100193510344083e-05,
+      "loss": 0.6744,
+      "step": 549000
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.098463267639479e-05,
+      "loss": 0.6791,
+      "step": 549500
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.096733024934874e-05,
+      "loss": 0.6769,
+      "step": 550000
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.095002782230269e-05,
+      "loss": 0.6755,
+      "step": 550500
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 3.093272539525665e-05,
+      "loss": 0.6746,
+      "step": 551000
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.09154229682106e-05,
+      "loss": 0.6745,
+      "step": 551500
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.089812054116455e-05,
+      "loss": 0.6774,
+      "step": 552000
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.0880818114118506e-05,
+      "loss": 0.6768,
+      "step": 552500
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.086351568707246e-05,
+      "loss": 0.6749,
+      "step": 553000
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.0846213260026415e-05,
+      "loss": 0.6747,
+      "step": 553500
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.0828910832980365e-05,
+      "loss": 0.6756,
+      "step": 554000
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.0811608405934316e-05,
+      "loss": 0.6739,
+      "step": 554500
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.0794305978888274e-05,
+      "loss": 0.6742,
+      "step": 555000
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.0777003551842224e-05,
+      "loss": 0.6781,
+      "step": 555500
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 3.075970112479618e-05,
+      "loss": 0.6733,
+      "step": 556000
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.074239869775013e-05,
+      "loss": 0.6763,
+      "step": 556500
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.0725096270704084e-05,
+      "loss": 0.6769,
+      "step": 557000
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.070779384365804e-05,
+      "loss": 0.6755,
+      "step": 557500
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.069049141661199e-05,
+      "loss": 0.6722,
+      "step": 558000
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.067318898956595e-05,
+      "loss": 0.6759,
+      "step": 558500
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.06558865625199e-05,
+      "loss": 0.6716,
+      "step": 559000
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.063858413547385e-05,
+      "loss": 0.6743,
+      "step": 559500
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.062128170842781e-05,
+      "loss": 0.675,
+      "step": 560000
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.060397928138176e-05,
+      "loss": 0.6733,
+      "step": 560500
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 3.058667685433571e-05,
+      "loss": 0.6749,
+      "step": 561000
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.056937442728967e-05,
+      "loss": 0.6767,
+      "step": 561500
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.055207200024362e-05,
+      "loss": 0.6743,
+      "step": 562000
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.0534769573197576e-05,
+      "loss": 0.675,
+      "step": 562500
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.0517467146151524e-05,
+      "loss": 0.6742,
+      "step": 563000
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.0500164719105478e-05,
+      "loss": 0.6753,
+      "step": 563500
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.0482862292059432e-05,
+      "loss": 0.6729,
+      "step": 564000
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.0465559865013386e-05,
+      "loss": 0.6728,
+      "step": 564500
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.0448257437967344e-05,
+      "loss": 0.6763,
+      "step": 565000
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 3.043095501092129e-05,
+      "loss": 0.6706,
+      "step": 565500
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.0413652583875245e-05,
+      "loss": 0.6726,
+      "step": 566000
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.03963501568292e-05,
+      "loss": 0.6739,
+      "step": 566500
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.0379047729783157e-05,
+      "loss": 0.6711,
+      "step": 567000
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.0361745302737105e-05,
+      "loss": 0.6759,
+      "step": 567500
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.034444287569106e-05,
+      "loss": 0.6739,
+      "step": 568000
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.0327140448645013e-05,
+      "loss": 0.6722,
+      "step": 568500
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.0309838021598967e-05,
+      "loss": 0.6746,
+      "step": 569000
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.0292535594552925e-05,
+      "loss": 0.6759,
+      "step": 569500
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.0275233167506872e-05,
+      "loss": 0.6742,
+      "step": 570000
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 3.0257930740460826e-05,
+      "loss": 0.6727,
+      "step": 570500
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.024062831341478e-05,
+      "loss": 0.6747,
+      "step": 571000
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.0223325886368735e-05,
+      "loss": 0.6721,
+      "step": 571500
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.0206023459322685e-05,
+      "loss": 0.6725,
+      "step": 572000
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.018872103227664e-05,
+      "loss": 0.6743,
+      "step": 572500
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.0171418605230594e-05,
+      "loss": 0.6743,
+      "step": 573000
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.0154116178184548e-05,
+      "loss": 0.6712,
+      "step": 573500
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.01368137511385e-05,
+      "loss": 0.6744,
+      "step": 574000
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.0119511324092453e-05,
+      "loss": 0.6753,
+      "step": 574500
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.0102208897046407e-05,
+      "loss": 0.6773,
+      "step": 575000
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 3.008490647000036e-05,
+      "loss": 0.6741,
+      "step": 575500
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3.0067604042954316e-05,
+      "loss": 0.6728,
+      "step": 576000
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3.0050301615908266e-05,
+      "loss": 0.6735,
+      "step": 576500
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3.003299918886222e-05,
+      "loss": 0.6704,
+      "step": 577000
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 3.0015696761816175e-05,
+      "loss": 0.6722,
+      "step": 577500
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 2.999839433477013e-05,
+      "loss": 0.6726,
+      "step": 578000
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 2.998109190772408e-05,
+      "loss": 0.6737,
+      "step": 578500
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 2.9963789480678034e-05,
+      "loss": 0.6704,
+      "step": 579000
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 2.9946487053631988e-05,
+      "loss": 0.6691,
+      "step": 579500
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 2.9929184626585942e-05,
+      "loss": 0.6715,
+      "step": 580000
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 2.9911882199539896e-05,
+      "loss": 0.6752,
+      "step": 580500
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 2.9894579772493847e-05,
+      "loss": 0.6714,
+      "step": 581000
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 2.98772773454478e-05,
+      "loss": 0.6705,
+      "step": 581500
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 2.9859974918401756e-05,
+      "loss": 0.6708,
+      "step": 582000
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 2.984267249135571e-05,
+      "loss": 0.6747,
+      "step": 582500
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 2.982537006430966e-05,
+      "loss": 0.6695,
+      "step": 583000
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 2.9808067637263615e-05,
+      "loss": 0.6729,
+      "step": 583500
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 2.979076521021757e-05,
+      "loss": 0.6741,
+      "step": 584000
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 2.9773462783171523e-05,
+      "loss": 0.6695,
+      "step": 584500
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 2.9756160356125474e-05,
+      "loss": 0.6723,
+      "step": 585000
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 2.9738857929079428e-05,
+      "loss": 0.6745,
+      "step": 585500
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 2.9721555502033382e-05,
+      "loss": 0.6723,
+      "step": 586000
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 2.9704253074987337e-05,
+      "loss": 0.6718,
+      "step": 586500
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 2.968695064794129e-05,
+      "loss": 0.6698,
+      "step": 587000
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 2.966964822089524e-05,
+      "loss": 0.6734,
+      "step": 587500
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 2.9652345793849196e-05,
+      "loss": 0.6706,
+      "step": 588000
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 2.963504336680315e-05,
+      "loss": 0.6716,
+      "step": 588500
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 2.9617740939757104e-05,
+      "loss": 0.6751,
+      "step": 589000
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 2.9600438512711055e-05,
+      "loss": 0.6726,
+      "step": 589500
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 2.958313608566501e-05,
+      "loss": 0.6703,
+      "step": 590000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 2.9565833658618963e-05,
+      "loss": 0.6726,
+      "step": 590500
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 2.9548531231572917e-05,
+      "loss": 0.6756,
+      "step": 591000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 2.953122880452687e-05,
+      "loss": 0.671,
+      "step": 591500
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 2.9513926377480822e-05,
+      "loss": 0.6724,
+      "step": 592000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 2.9496623950434777e-05,
+      "loss": 0.6687,
+      "step": 592500
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 2.947932152338873e-05,
+      "loss": 0.6694,
+      "step": 593000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 2.9462019096342685e-05,
+      "loss": 0.6742,
+      "step": 593500
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 2.9444716669296636e-05,
+      "loss": 0.6729,
+      "step": 594000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 2.942741424225059e-05,
+      "loss": 0.6678,
+      "step": 594500
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 2.9410111815204544e-05,
+      "loss": 0.6731,
+      "step": 595000
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 2.9392809388158498e-05,
+      "loss": 0.6756,
+      "step": 595500
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 2.937550696111245e-05,
+      "loss": 0.671,
+      "step": 596000
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 2.9358204534066403e-05,
+      "loss": 0.6738,
+      "step": 596500
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 2.9340902107020357e-05,
+      "loss": 0.6728,
+      "step": 597000
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 2.932359967997431e-05,
+      "loss": 0.6713,
+      "step": 597500
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 2.9306297252928266e-05,
+      "loss": 0.6709,
+      "step": 598000
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 2.9288994825882217e-05,
+      "loss": 0.6685,
+      "step": 598500
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 2.927169239883617e-05,
+      "loss": 0.6708,
+      "step": 599000
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 2.9254389971790125e-05,
+      "loss": 0.6712,
+      "step": 599500
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 2.923708754474408e-05,
+      "loss": 0.6741,
+      "step": 600000
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 2.921978511769803e-05,
+      "loss": 0.671,
+      "step": 600500
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 2.9202482690651984e-05,
+      "loss": 0.669,
+      "step": 601000
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 2.918518026360594e-05,
+      "loss": 0.6739,
+      "step": 601500
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 2.9167877836559893e-05,
+      "loss": 0.6737,
+      "step": 602000
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 2.9150575409513847e-05,
+      "loss": 0.6731,
+      "step": 602500
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 2.9133272982467798e-05,
+      "loss": 0.6736,
+      "step": 603000
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 2.911597055542175e-05,
+      "loss": 0.6722,
+      "step": 603500
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 2.9098668128375706e-05,
+      "loss": 0.6697,
+      "step": 604000
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 2.908136570132966e-05,
+      "loss": 0.6733,
+      "step": 604500
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 2.906406327428361e-05,
+      "loss": 0.6717,
+      "step": 605000
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 2.9046760847237565e-05,
+      "loss": 0.6723,
+      "step": 605500
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 2.902945842019152e-05,
+      "loss": 0.6717,
+      "step": 606000
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 2.9012155993145473e-05,
+      "loss": 0.6708,
+      "step": 606500
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 2.8994853566099428e-05,
+      "loss": 0.6729,
+      "step": 607000
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 2.897755113905338e-05,
+      "loss": 0.6695,
+      "step": 607500
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 2.8960248712007333e-05,
+      "loss": 0.6697,
+      "step": 608000
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 2.8942946284961287e-05,
+      "loss": 0.6692,
+      "step": 608500
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 2.892564385791524e-05,
+      "loss": 0.6699,
+      "step": 609000
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 2.8908341430869192e-05,
+      "loss": 0.6711,
+      "step": 609500
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 2.8891039003823146e-05,
+      "loss": 0.6692,
+      "step": 610000
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 2.88737365767771e-05,
+      "loss": 0.6701,
+      "step": 610500
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 2.8856434149731054e-05,
+      "loss": 0.6695,
+      "step": 611000
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 2.8839131722685002e-05,
+      "loss": 0.6667,
+      "step": 611500
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 2.882182929563896e-05,
+      "loss": 0.671,
+      "step": 612000
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 2.8804526868592913e-05,
+      "loss": 0.6711,
+      "step": 612500
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 2.8787224441546868e-05,
+      "loss": 0.6733,
+      "step": 613000
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 2.8769922014500822e-05,
+      "loss": 0.6711,
+      "step": 613500
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 2.8752619587454773e-05,
+      "loss": 0.6698,
+      "step": 614000
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 2.8735317160408727e-05,
+      "loss": 0.6689,
+      "step": 614500
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 2.871801473336268e-05,
+      "loss": 0.6709,
+      "step": 615000
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 2.8700712306316635e-05,
+      "loss": 0.6736,
+      "step": 615500
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 2.8683409879270583e-05,
+      "loss": 0.6688,
+      "step": 616000
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 2.866610745222454e-05,
+      "loss": 0.6725,
+      "step": 616500
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 2.8648805025178494e-05,
+      "loss": 0.6678,
+      "step": 617000
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 2.863150259813245e-05,
+      "loss": 0.6691,
+      "step": 617500
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 2.8614200171086403e-05,
+      "loss": 0.6712,
+      "step": 618000
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 2.859689774404035e-05,
+      "loss": 0.6697,
+      "step": 618500
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 2.8579595316994308e-05,
+      "loss": 0.6696,
+      "step": 619000
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 2.8562292889948262e-05,
+      "loss": 0.6706,
+      "step": 619500
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 2.8544990462902216e-05,
+      "loss": 0.6697,
+      "step": 620000
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 2.8527688035856163e-05,
+      "loss": 0.6713,
+      "step": 620500
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 2.851038560881012e-05,
+      "loss": 0.669,
+      "step": 621000
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 2.8493083181764075e-05,
+      "loss": 0.6664,
+      "step": 621500
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 2.847578075471803e-05,
+      "loss": 0.6703,
+      "step": 622000
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 2.8458478327671977e-05,
+      "loss": 0.6706,
+      "step": 622500
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 2.844117590062593e-05,
+      "loss": 0.6714,
+      "step": 623000
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 2.842387347357989e-05,
+      "loss": 0.6744,
+      "step": 623500
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 2.8406571046533843e-05,
+      "loss": 0.6678,
+      "step": 624000
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 2.8389268619487797e-05,
+      "loss": 0.6706,
+      "step": 624500
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 2.8371966192441744e-05,
+      "loss": 0.669,
+      "step": 625000
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 2.8354663765395702e-05,
+      "loss": 0.6683,
+      "step": 625500
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 2.8337361338349656e-05,
+      "loss": 0.6704,
+      "step": 626000
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 2.832005891130361e-05,
+      "loss": 0.6676,
+      "step": 626500
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 2.8302756484257558e-05,
+      "loss": 0.6738,
+      "step": 627000
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 2.8285454057211512e-05,
+      "loss": 0.6685,
+      "step": 627500
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 2.826815163016547e-05,
+      "loss": 0.6716,
+      "step": 628000
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 2.8250849203119424e-05,
+      "loss": 0.6706,
+      "step": 628500
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 2.8233546776073378e-05,
+      "loss": 0.6694,
+      "step": 629000
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 2.8216244349027325e-05,
+      "loss": 0.6694,
+      "step": 629500
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 2.8198941921981283e-05,
+      "loss": 0.6707,
+      "step": 630000
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 2.8181639494935237e-05,
+      "loss": 0.6693,
+      "step": 630500
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 2.816433706788919e-05,
+      "loss": 0.6702,
+      "step": 631000
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 2.814703464084314e-05,
+      "loss": 0.6695,
+      "step": 631500
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 2.8129732213797093e-05,
+      "loss": 0.6705,
+      "step": 632000
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 2.811242978675105e-05,
+      "loss": 0.6704,
+      "step": 632500
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 2.8095127359705005e-05,
+      "loss": 0.6676,
+      "step": 633000
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 2.8077824932658952e-05,
+      "loss": 0.6676,
+      "step": 633500
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 2.8060522505612906e-05,
+      "loss": 0.6684,
+      "step": 634000
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 2.804322007856686e-05,
+      "loss": 0.6723,
+      "step": 634500
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 2.8025917651520818e-05,
+      "loss": 0.6689,
+      "step": 635000
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 2.8008615224474772e-05,
+      "loss": 0.6693,
+      "step": 635500
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 2.799131279742872e-05,
+      "loss": 0.6708,
+      "step": 636000
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 2.7974010370382674e-05,
+      "loss": 0.6688,
+      "step": 636500
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 2.795670794333663e-05,
+      "loss": 0.6645,
+      "step": 637000
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 2.7939405516290585e-05,
+      "loss": 0.6724,
+      "step": 637500
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 2.7922103089244533e-05,
+      "loss": 0.6693,
+      "step": 638000
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 2.7904800662198487e-05,
+      "loss": 0.6676,
+      "step": 638500
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 2.788749823515244e-05,
+      "loss": 0.6684,
+      "step": 639000
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 2.78701958081064e-05,
+      "loss": 0.6688,
+      "step": 639500
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 2.7852893381060353e-05,
+      "loss": 0.6695,
+      "step": 640000
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 2.78355909540143e-05,
+      "loss": 0.6698,
+      "step": 640500
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 2.7818288526968255e-05,
+      "loss": 0.6697,
+      "step": 641000
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 2.7800986099922212e-05,
+      "loss": 0.67,
+      "step": 641500
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 2.7783683672876166e-05,
+      "loss": 0.6692,
+      "step": 642000
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 2.7766381245830114e-05,
+      "loss": 0.6666,
+      "step": 642500
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 2.7749078818784068e-05,
+      "loss": 0.6676,
+      "step": 643000
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 2.7731776391738022e-05,
+      "loss": 0.6663,
+      "step": 643500
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 2.771447396469198e-05,
+      "loss": 0.6679,
+      "step": 644000
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 2.7697171537645927e-05,
+      "loss": 0.6683,
+      "step": 644500
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 2.767986911059988e-05,
+      "loss": 0.6692,
+      "step": 645000
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 2.7662566683553835e-05,
+      "loss": 0.6701,
+      "step": 645500
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 2.764526425650779e-05,
+      "loss": 0.6681,
+      "step": 646000
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 2.7627961829461747e-05,
+      "loss": 0.6703,
+      "step": 646500
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 2.7610659402415695e-05,
+      "loss": 0.6687,
+      "step": 647000
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 2.759335697536965e-05,
+      "loss": 0.6677,
+      "step": 647500
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 2.7576054548323603e-05,
+      "loss": 0.6666,
+      "step": 648000
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 2.755875212127756e-05,
+      "loss": 0.6666,
+      "step": 648500
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 2.7541449694231508e-05,
+      "loss": 0.6692,
+      "step": 649000
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 2.7524147267185462e-05,
+      "loss": 0.6687,
+      "step": 649500
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 2.7506844840139416e-05,
+      "loss": 0.6703,
+      "step": 650000
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 2.748954241309337e-05,
+      "loss": 0.6716,
+      "step": 650500
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 2.7472239986047328e-05,
+      "loss": 0.6694,
+      "step": 651000
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 2.7454937559001276e-05,
+      "loss": 0.6703,
+      "step": 651500
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 2.743763513195523e-05,
+      "loss": 0.6683,
+      "step": 652000
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 2.7420332704909184e-05,
+      "loss": 0.6657,
+      "step": 652500
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.740303027786314e-05,
+      "loss": 0.6673,
+      "step": 653000
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.738572785081709e-05,
+      "loss": 0.666,
+      "step": 653500
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.7368425423771043e-05,
+      "loss": 0.6688,
+      "step": 654000
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.7351122996724997e-05,
+      "loss": 0.6688,
+      "step": 654500
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.733382056967895e-05,
+      "loss": 0.6676,
+      "step": 655000
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.7316518142632902e-05,
+      "loss": 0.67,
+      "step": 655500
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.7299215715586856e-05,
+      "loss": 0.6707,
+      "step": 656000
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.728191328854081e-05,
+      "loss": 0.6689,
+      "step": 656500
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 2.7264610861494765e-05,
+      "loss": 0.6674,
+      "step": 657000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.724730843444872e-05,
+      "loss": 0.6671,
+      "step": 657500
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.723000600740267e-05,
+      "loss": 0.6707,
+      "step": 658000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.7212703580356624e-05,
+      "loss": 0.6682,
+      "step": 658500
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.7195401153310578e-05,
+      "loss": 0.6634,
+      "step": 659000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.7178098726264532e-05,
+      "loss": 0.668,
+      "step": 659500
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.7160796299218483e-05,
+      "loss": 0.6663,
+      "step": 660000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.7143493872172437e-05,
+      "loss": 0.6673,
+      "step": 660500
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.712619144512639e-05,
+      "loss": 0.668,
+      "step": 661000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.7108889018080346e-05,
+      "loss": 0.6675,
+      "step": 661500
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 2.70915865910343e-05,
+      "loss": 0.669,
+      "step": 662000
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.707428416398825e-05,
+      "loss": 0.665,
+      "step": 662500
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.7056981736942205e-05,
+      "loss": 0.6649,
+      "step": 663000
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.703967930989616e-05,
+      "loss": 0.6661,
+      "step": 663500
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.7022376882850113e-05,
+      "loss": 0.668,
+      "step": 664000
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.7005074455804064e-05,
+      "loss": 0.6703,
+      "step": 664500
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.6987772028758018e-05,
+      "loss": 0.6681,
+      "step": 665000
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.6970469601711972e-05,
+      "loss": 0.6698,
+      "step": 665500
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.6953167174665927e-05,
+      "loss": 0.669,
+      "step": 666000
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.6935864747619877e-05,
+      "loss": 0.6676,
+      "step": 666500
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 2.691856232057383e-05,
+      "loss": 0.6683,
+      "step": 667000
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.6901259893527786e-05,
+      "loss": 0.6678,
+      "step": 667500
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.688395746648174e-05,
+      "loss": 0.666,
+      "step": 668000
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.6866655039435694e-05,
+      "loss": 0.6662,
+      "step": 668500
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.6849352612389645e-05,
+      "loss": 0.6681,
+      "step": 669000
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.68320501853436e-05,
+      "loss": 0.6633,
+      "step": 669500
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.6814747758297553e-05,
+      "loss": 0.6659,
+      "step": 670000
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.6797445331251507e-05,
+      "loss": 0.6671,
+      "step": 670500
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.6780142904205458e-05,
+      "loss": 0.6668,
+      "step": 671000
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 2.6762840477159412e-05,
+      "loss": 0.6688,
+      "step": 671500
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.6745538050113367e-05,
+      "loss": 0.6635,
+      "step": 672000
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.672823562306732e-05,
+      "loss": 0.666,
+      "step": 672500
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.6710933196021275e-05,
+      "loss": 0.6654,
+      "step": 673000
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.6693630768975226e-05,
+      "loss": 0.6655,
+      "step": 673500
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.667632834192918e-05,
+      "loss": 0.6675,
+      "step": 674000
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.6659025914883134e-05,
+      "loss": 0.6703,
+      "step": 674500
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.664172348783709e-05,
+      "loss": 0.6675,
+      "step": 675000
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.662442106079104e-05,
+      "loss": 0.6666,
+      "step": 675500
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.6607118633744993e-05,
+      "loss": 0.6653,
+      "step": 676000
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 2.6589816206698948e-05,
+      "loss": 0.6669,
+      "step": 676500
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.6572513779652902e-05,
+      "loss": 0.663,
+      "step": 677000
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.6555211352606856e-05,
+      "loss": 0.6671,
+      "step": 677500
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.6537908925560807e-05,
+      "loss": 0.6677,
+      "step": 678000
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.652060649851476e-05,
+      "loss": 0.6676,
+      "step": 678500
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.6503304071468715e-05,
+      "loss": 0.666,
+      "step": 679000
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.648600164442267e-05,
+      "loss": 0.6649,
+      "step": 679500
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.646869921737662e-05,
+      "loss": 0.6663,
+      "step": 680000
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.6451396790330574e-05,
+      "loss": 0.6657,
+      "step": 680500
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.643409436328453e-05,
+      "loss": 0.6678,
+      "step": 681000
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 2.6416791936238483e-05,
+      "loss": 0.6685,
+      "step": 681500
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.6399489509192433e-05,
+      "loss": 0.6645,
+      "step": 682000
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.6382187082146388e-05,
+      "loss": 0.6694,
+      "step": 682500
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.6364884655100342e-05,
+      "loss": 0.6635,
+      "step": 683000
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.6347582228054296e-05,
+      "loss": 0.6664,
+      "step": 683500
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.633027980100825e-05,
+      "loss": 0.6685,
+      "step": 684000
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.63129773739622e-05,
+      "loss": 0.6666,
+      "step": 684500
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.6295674946916155e-05,
+      "loss": 0.6664,
+      "step": 685000
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.627837251987011e-05,
+      "loss": 0.6692,
+      "step": 685500
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 2.6261070092824063e-05,
+      "loss": 0.6666,
+      "step": 686000
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.6243767665778014e-05,
+      "loss": 0.6659,
+      "step": 686500
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.622646523873197e-05,
+      "loss": 0.6655,
+      "step": 687000
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.6209162811685923e-05,
+      "loss": 0.6686,
+      "step": 687500
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.6191860384639877e-05,
+      "loss": 0.6666,
+      "step": 688000
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.617455795759383e-05,
+      "loss": 0.6678,
+      "step": 688500
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.6157255530547782e-05,
+      "loss": 0.664,
+      "step": 689000
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.6139953103501736e-05,
+      "loss": 0.6666,
+      "step": 689500
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.612265067645569e-05,
+      "loss": 0.6665,
+      "step": 690000
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.6105348249409644e-05,
+      "loss": 0.6639,
+      "step": 690500
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 2.6088045822363595e-05,
+      "loss": 0.6666,
+      "step": 691000
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.607074339531755e-05,
+      "loss": 0.6668,
+      "step": 691500
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.6053440968271504e-05,
+      "loss": 0.6672,
+      "step": 692000
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.6036138541225458e-05,
+      "loss": 0.6676,
+      "step": 692500
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.6018836114179405e-05,
+      "loss": 0.6652,
+      "step": 693000
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.6001533687133363e-05,
+      "loss": 0.6656,
+      "step": 693500
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.5984231260087317e-05,
+      "loss": 0.6672,
+      "step": 694000
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.596692883304127e-05,
+      "loss": 0.6662,
+      "step": 694500
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.5949626405995225e-05,
+      "loss": 0.6678,
+      "step": 695000
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 2.5932323978949176e-05,
+      "loss": 0.6641,
+      "step": 695500
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.591502155190313e-05,
+      "loss": 0.6655,
+      "step": 696000
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.5897719124857084e-05,
+      "loss": 0.6616,
+      "step": 696500
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.588041669781104e-05,
+      "loss": 0.669,
+      "step": 697000
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.5863114270764986e-05,
+      "loss": 0.6655,
+      "step": 697500
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.5845811843718944e-05,
+      "loss": 0.6651,
+      "step": 698000
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.5828509416672898e-05,
+      "loss": 0.6644,
+      "step": 698500
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.5811206989626852e-05,
+      "loss": 0.6666,
+      "step": 699000
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.5793904562580806e-05,
+      "loss": 0.6675,
+      "step": 699500
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.5776602135534757e-05,
+      "loss": 0.6653,
+      "step": 700000
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 2.575929970848871e-05,
+      "loss": 0.6676,
+      "step": 700500
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.5741997281442665e-05,
+      "loss": 0.6657,
+      "step": 701000
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.572469485439662e-05,
+      "loss": 0.6643,
+      "step": 701500
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.5707392427350567e-05,
+      "loss": 0.6646,
+      "step": 702000
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.5690090000304524e-05,
+      "loss": 0.6668,
+      "step": 702500
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.567278757325848e-05,
+      "loss": 0.6649,
+      "step": 703000
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.5655485146212433e-05,
+      "loss": 0.6683,
+      "step": 703500
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.563818271916638e-05,
+      "loss": 0.6649,
+      "step": 704000
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.5620880292120334e-05,
+      "loss": 0.6634,
+      "step": 704500
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.5603577865074292e-05,
+      "loss": 0.6651,
+      "step": 705000
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 2.5586275438028246e-05,
+      "loss": 0.6652,
+      "step": 705500
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.55689730109822e-05,
+      "loss": 0.6681,
+      "step": 706000
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.5551670583936148e-05,
+      "loss": 0.6656,
+      "step": 706500
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.5534368156890105e-05,
+      "loss": 0.6674,
+      "step": 707000
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.551706572984406e-05,
+      "loss": 0.666,
+      "step": 707500
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.5499763302798014e-05,
+      "loss": 0.6658,
+      "step": 708000
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.548246087575196e-05,
+      "loss": 0.6636,
+      "step": 708500
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.5465158448705915e-05,
+      "loss": 0.6637,
+      "step": 709000
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.5447856021659873e-05,
+      "loss": 0.6679,
+      "step": 709500
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 2.5430553594613827e-05,
+      "loss": 0.6648,
+      "step": 710000
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.541325116756778e-05,
+      "loss": 0.6635,
+      "step": 710500
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.539594874052173e-05,
+      "loss": 0.6647,
+      "step": 711000
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.5378646313475686e-05,
+      "loss": 0.6641,
+      "step": 711500
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.536134388642964e-05,
+      "loss": 0.6659,
+      "step": 712000
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.5344041459383595e-05,
+      "loss": 0.666,
+      "step": 712500
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.5326739032337542e-05,
+      "loss": 0.6667,
+      "step": 713000
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.5309436605291496e-05,
+      "loss": 0.6675,
+      "step": 713500
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.5292134178245454e-05,
+      "loss": 0.6668,
+      "step": 714000
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.5274831751199408e-05,
+      "loss": 0.6658,
+      "step": 714500
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 2.5257529324153355e-05,
+      "loss": 0.6658,
+      "step": 715000
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.524022689710731e-05,
+      "loss": 0.6637,
+      "step": 715500
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.5222924470061267e-05,
+      "loss": 0.6621,
+      "step": 716000
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.520562204301522e-05,
+      "loss": 0.6656,
+      "step": 716500
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.5188319615969176e-05,
+      "loss": 0.6604,
+      "step": 717000
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.5171017188923123e-05,
+      "loss": 0.6641,
+      "step": 717500
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.5153714761877077e-05,
+      "loss": 0.6653,
+      "step": 718000
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.5136412334831035e-05,
+      "loss": 0.6614,
+      "step": 718500
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.511910990778499e-05,
+      "loss": 0.6643,
+      "step": 719000
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.5101807480738936e-05,
+      "loss": 0.6644,
+      "step": 719500
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 2.508450505369289e-05,
+      "loss": 0.6678,
+      "step": 720000
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.5067202626646845e-05,
+      "loss": 0.6655,
+      "step": 720500
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.5049900199600802e-05,
+      "loss": 0.6613,
+      "step": 721000
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.5032597772554756e-05,
+      "loss": 0.664,
+      "step": 721500
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.5015295345508704e-05,
+      "loss": 0.6645,
+      "step": 722000
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.4997992918462658e-05,
+      "loss": 0.6618,
+      "step": 722500
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.4980690491416616e-05,
+      "loss": 0.6672,
+      "step": 723000
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.4963388064370566e-05,
+      "loss": 0.6673,
+      "step": 723500
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.494608563732452e-05,
+      "loss": 0.667,
+      "step": 724000
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 2.492878321027847e-05,
+      "loss": 0.6616,
+      "step": 724500
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.4911480783232426e-05,
+      "loss": 0.6639,
+      "step": 725000
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.489417835618638e-05,
+      "loss": 0.6665,
+      "step": 725500
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.4876875929140334e-05,
+      "loss": 0.6634,
+      "step": 726000
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.4859573502094288e-05,
+      "loss": 0.6666,
+      "step": 726500
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.484227107504824e-05,
+      "loss": 0.6633,
+      "step": 727000
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.4824968648002196e-05,
+      "loss": 0.6636,
+      "step": 727500
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.4807666220956147e-05,
+      "loss": 0.6653,
+      "step": 728000
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.47903637939101e-05,
+      "loss": 0.6638,
+      "step": 728500
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.4773061366864052e-05,
+      "loss": 0.6662,
+      "step": 729000
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 2.4755758939818006e-05,
+      "loss": 0.6648,
+      "step": 729500
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.473845651277196e-05,
+      "loss": 0.6651,
+      "step": 730000
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.4721154085725915e-05,
+      "loss": 0.6649,
+      "step": 730500
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.4703851658679866e-05,
+      "loss": 0.6626,
+      "step": 731000
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.468654923163382e-05,
+      "loss": 0.6608,
+      "step": 731500
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.4669246804587774e-05,
+      "loss": 0.6616,
+      "step": 732000
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.4651944377541728e-05,
+      "loss": 0.6607,
+      "step": 732500
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.4634641950495682e-05,
+      "loss": 0.6615,
+      "step": 733000
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.4617339523449633e-05,
+      "loss": 0.6642,
+      "step": 733500
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 2.4600037096403587e-05,
+      "loss": 0.6612,
+      "step": 734000
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.458273466935754e-05,
+      "loss": 0.6625,
+      "step": 734500
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.4565432242311496e-05,
+      "loss": 0.6619,
+      "step": 735000
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.4548129815265446e-05,
+      "loss": 0.6632,
+      "step": 735500
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.45308273882194e-05,
+      "loss": 0.6647,
+      "step": 736000
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.4513524961173355e-05,
+      "loss": 0.6672,
+      "step": 736500
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.449622253412731e-05,
+      "loss": 0.663,
+      "step": 737000
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.4478920107081263e-05,
+      "loss": 0.6659,
+      "step": 737500
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.4461617680035214e-05,
+      "loss": 0.663,
+      "step": 738000
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.4444315252989168e-05,
+      "loss": 0.6674,
+      "step": 738500
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 2.4427012825943122e-05,
+      "loss": 0.6651,
+      "step": 739000
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.4409710398897077e-05,
+      "loss": 0.662,
+      "step": 739500
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.4392407971851027e-05,
+      "loss": 0.6635,
+      "step": 740000
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.437510554480498e-05,
+      "loss": 0.6647,
+      "step": 740500
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.4357803117758936e-05,
+      "loss": 0.6635,
+      "step": 741000
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.434050069071289e-05,
+      "loss": 0.6634,
+      "step": 741500
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.432319826366684e-05,
+      "loss": 0.6627,
+      "step": 742000
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.4305895836620795e-05,
+      "loss": 0.6632,
+      "step": 742500
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.428859340957475e-05,
+      "loss": 0.6622,
+      "step": 743000
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.4271290982528703e-05,
+      "loss": 0.665,
+      "step": 743500
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 2.4253988555482657e-05,
+      "loss": 0.6626,
+      "step": 744000
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.4236686128436608e-05,
+      "loss": 0.6632,
+      "step": 744500
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.4219383701390562e-05,
+      "loss": 0.6652,
+      "step": 745000
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.4202081274344517e-05,
+      "loss": 0.6609,
+      "step": 745500
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.418477884729847e-05,
+      "loss": 0.6665,
+      "step": 746000
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.416747642025242e-05,
+      "loss": 0.6638,
+      "step": 746500
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.4150173993206376e-05,
+      "loss": 0.6657,
+      "step": 747000
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.413287156616033e-05,
+      "loss": 0.6614,
+      "step": 747500
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.4115569139114284e-05,
+      "loss": 0.6635,
+      "step": 748000
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 2.409826671206824e-05,
+      "loss": 0.6637,
+      "step": 748500
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.408096428502219e-05,
+      "loss": 0.6648,
+      "step": 749000
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.4063661857976143e-05,
+      "loss": 0.6601,
+      "step": 749500
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.4046359430930098e-05,
+      "loss": 0.6636,
+      "step": 750000
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.4029057003884052e-05,
+      "loss": 0.6637,
+      "step": 750500
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.4011754576838003e-05,
+      "loss": 0.6631,
+      "step": 751000
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.3994452149791957e-05,
+      "loss": 0.6647,
+      "step": 751500
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.3977149722745907e-05,
+      "loss": 0.6634,
+      "step": 752000
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.3959847295699865e-05,
+      "loss": 0.6644,
+      "step": 752500
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.3942544868653816e-05,
+      "loss": 0.6613,
+      "step": 753000
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 2.392524244160777e-05,
+      "loss": 0.6618,
+      "step": 753500
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.3907940014561724e-05,
+      "loss": 0.6634,
+      "step": 754000
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.389063758751568e-05,
+      "loss": 0.6586,
+      "step": 754500
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.3873335160469633e-05,
+      "loss": 0.6655,
+      "step": 755000
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.3856032733423583e-05,
+      "loss": 0.6656,
+      "step": 755500
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 2.3838730306377538e-05,
+      "loss": 0.6636,
+      "step": 756000
+    }
+  ],
+  "max_steps": 1444884,
+  "num_train_epochs": 3,
+  "total_flos": 6.321173702639616e+18,
+  "trial_name": null,
+  "trial_params": null
+}