diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,6127 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 10.0,
+  "global_step": 508600,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.01,
+      "learning_rate": 4.995084545812033e-05,
+      "loss": 8.5269,
+      "step": 500
+    },
+    {
+      "epoch": 0.02,
+      "learning_rate": 4.9901690916240665e-05,
+      "loss": 3.887,
+      "step": 1000
+    },
+    {
+      "epoch": 0.03,
+      "learning_rate": 4.985253637436099e-05,
+      "loss": 3.3828,
+      "step": 1500
+    },
+    {
+      "epoch": 0.04,
+      "learning_rate": 4.980338183248132e-05,
+      "loss": 3.1234,
+      "step": 2000
+    },
+    {
+      "epoch": 0.05,
+      "learning_rate": 4.975422729060165e-05,
+      "loss": 2.9408,
+      "step": 2500
+    },
+    {
+      "epoch": 0.06,
+      "learning_rate": 4.970507274872199e-05,
+      "loss": 2.798,
+      "step": 3000
+    },
+    {
+      "epoch": 0.07,
+      "learning_rate": 4.965591820684232e-05,
+      "loss": 2.6747,
+      "step": 3500
+    },
+    {
+      "epoch": 0.08,
+      "learning_rate": 4.9606763664962646e-05,
+      "loss": 2.6112,
+      "step": 4000
+    },
+    {
+      "epoch": 0.09,
+      "learning_rate": 4.9557609123082973e-05,
+      "loss": 2.5234,
+      "step": 4500
+    },
+    {
+      "epoch": 0.1,
+      "learning_rate": 4.950845458120331e-05,
+      "loss": 2.4435,
+      "step": 5000
+    },
+    {
+      "epoch": 0.11,
+      "learning_rate": 4.9459300039323636e-05,
+      "loss": 2.3883,
+      "step": 5500
+    },
+    {
+      "epoch": 0.12,
+      "learning_rate": 4.9410145497443964e-05,
+      "loss": 2.3303,
+      "step": 6000
+    },
+    {
+      "epoch": 0.13,
+      "learning_rate": 4.93609909555643e-05,
+      "loss": 2.286,
+      "step": 6500
+    },
+    {
+      "epoch": 0.14,
+      "learning_rate": 4.9311836413684626e-05,
+      "loss": 2.2177,
+      "step": 7000
+    },
+    {
+      "epoch": 0.15,
+      "learning_rate": 4.9262681871804954e-05,
+      "loss": 2.203,
+      "step": 7500
+    },
+    {
+      "epoch": 0.16,
+      "learning_rate": 4.921352732992528e-05,
+      "loss": 2.1681,
+      "step": 8000
+    },
+    {
+      "epoch": 0.17,
+      "learning_rate": 4.916437278804562e-05,
+      "loss": 2.1236,
+      "step": 8500
+    },
+    {
+      "epoch": 0.18,
+      "learning_rate": 4.911521824616595e-05,
+      "loss": 2.0806,
+      "step": 9000
+    },
+    {
+      "epoch": 0.19,
+      "learning_rate": 4.906606370428628e-05,
+      "loss": 2.0495,
+      "step": 9500
+    },
+    {
+      "epoch": 0.2,
+      "learning_rate": 4.901690916240661e-05,
+      "loss": 2.0479,
+      "step": 10000
+    },
+    {
+      "epoch": 0.21,
+      "learning_rate": 4.896775462052694e-05,
+      "loss": 2.0075,
+      "step": 10500
+    },
+    {
+      "epoch": 0.22,
+      "learning_rate": 4.891860007864727e-05,
+      "loss": 1.9711,
+      "step": 11000
+    },
+    {
+      "epoch": 0.23,
+      "learning_rate": 4.88694455367676e-05,
+      "loss": 1.9652,
+      "step": 11500
+    },
+    {
+      "epoch": 0.24,
+      "learning_rate": 4.882029099488793e-05,
+      "loss": 1.9282,
+      "step": 12000
+    },
+    {
+      "epoch": 0.25,
+      "learning_rate": 4.877113645300826e-05,
+      "loss": 1.906,
+      "step": 12500
+    },
+    {
+      "epoch": 0.26,
+      "learning_rate": 4.872198191112859e-05,
+      "loss": 1.8884,
+      "step": 13000
+    },
+    {
+      "epoch": 0.27,
+      "learning_rate": 4.8672827369248916e-05,
+      "loss": 1.8655,
+      "step": 13500
+    },
+    {
+      "epoch": 0.28,
+      "learning_rate": 4.862367282736925e-05,
+      "loss": 1.8642,
+      "step": 14000
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.8574518285489585e-05,
+      "loss": 1.8338,
+      "step": 14500
+    },
+    {
+      "epoch": 0.29,
+      "learning_rate": 4.852536374360991e-05,
+      "loss": 1.8359,
+      "step": 15000
+    },
+    {
+      "epoch": 0.3,
+      "learning_rate": 4.847620920173024e-05,
+      "loss": 1.8034,
+      "step": 15500
+    },
+    {
+      "epoch": 0.31,
+      "learning_rate": 4.8427054659850575e-05,
+      "loss": 1.7909,
+      "step": 16000
+    },
+    {
+      "epoch": 0.32,
+      "learning_rate": 4.83779001179709e-05,
+      "loss": 1.7587,
+      "step": 16500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 4.832874557609123e-05,
+      "loss": 1.7676,
+      "step": 17000
+    },
+    {
+      "epoch": 0.34,
+      "learning_rate": 4.8279591034211566e-05,
+      "loss": 1.7414,
+      "step": 17500
+    },
+    {
+      "epoch": 0.35,
+      "learning_rate": 4.8230436492331893e-05,
+      "loss": 1.7267,
+      "step": 18000
+    },
+    {
+      "epoch": 0.36,
+      "learning_rate": 4.818128195045222e-05,
+      "loss": 1.7167,
+      "step": 18500
+    },
+    {
+      "epoch": 0.37,
+      "learning_rate": 4.8132127408572556e-05,
+      "loss": 1.7057,
+      "step": 19000
+    },
+    {
+      "epoch": 0.38,
+      "learning_rate": 4.8082972866692884e-05,
+      "loss": 1.6797,
+      "step": 19500
+    },
+    {
+      "epoch": 0.39,
+      "learning_rate": 4.803381832481321e-05,
+      "loss": 1.6798,
+      "step": 20000
+    },
+    {
+      "epoch": 0.4,
+      "learning_rate": 4.7984663782933546e-05,
+      "loss": 1.6704,
+      "step": 20500
+    },
+    {
+      "epoch": 0.41,
+      "learning_rate": 4.7935509241053874e-05,
+      "loss": 1.653,
+      "step": 21000
+    },
+    {
+      "epoch": 0.42,
+      "learning_rate": 4.788635469917421e-05,
+      "loss": 1.6541,
+      "step": 21500
+    },
+    {
+      "epoch": 0.43,
+      "learning_rate": 4.783720015729454e-05,
+      "loss": 1.6404,
+      "step": 22000
+    },
+    {
+      "epoch": 0.44,
+      "learning_rate": 4.7788045615414865e-05,
+      "loss": 1.6296,
+      "step": 22500
+    },
+    {
+      "epoch": 0.45,
+      "learning_rate": 4.77388910735352e-05,
+      "loss": 1.6205,
+      "step": 23000
+    },
+    {
+      "epoch": 0.46,
+      "learning_rate": 4.768973653165553e-05,
+      "loss": 1.5994,
+      "step": 23500
+    },
+    {
+      "epoch": 0.47,
+      "learning_rate": 4.7640581989775855e-05,
+      "loss": 1.5933,
+      "step": 24000
+    },
+    {
+      "epoch": 0.48,
+      "learning_rate": 4.759142744789619e-05,
+      "loss": 1.5998,
+      "step": 24500
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 4.754227290601652e-05,
+      "loss": 1.5913,
+      "step": 25000
+    },
+    {
+      "epoch": 0.5,
+      "learning_rate": 4.7493118364136845e-05,
+      "loss": 1.5682,
+      "step": 25500
+    },
+    {
+      "epoch": 0.51,
+      "learning_rate": 4.744396382225718e-05,
+      "loss": 1.5572,
+      "step": 26000
+    },
+    {
+      "epoch": 0.52,
+      "learning_rate": 4.7394809280377515e-05,
+      "loss": 1.5719,
+      "step": 26500
+    },
+    {
+      "epoch": 0.53,
+      "learning_rate": 4.734565473849784e-05,
+      "loss": 1.5551,
+      "step": 27000
+    },
+    {
+      "epoch": 0.54,
+      "learning_rate": 4.729650019661817e-05,
+      "loss": 1.5426,
+      "step": 27500
+    },
+    {
+      "epoch": 0.55,
+      "learning_rate": 4.72473456547385e-05,
+      "loss": 1.5426,
+      "step": 28000
+    },
+    {
+      "epoch": 0.56,
+      "learning_rate": 4.719819111285883e-05,
+      "loss": 1.5393,
+      "step": 28500
+    },
+    {
+      "epoch": 0.57,
+      "learning_rate": 4.714903657097916e-05,
+      "loss": 1.5167,
+      "step": 29000
+    },
+    {
+      "epoch": 0.58,
+      "learning_rate": 4.709988202909949e-05,
+      "loss": 1.5168,
+      "step": 29500
+    },
+    {
+      "epoch": 0.59,
+      "learning_rate": 4.705072748721982e-05,
+      "loss": 1.492,
+      "step": 30000
+    },
+    {
+      "epoch": 0.6,
+      "learning_rate": 4.700157294534015e-05,
+      "loss": 1.5187,
+      "step": 30500
+    },
+    {
+      "epoch": 0.61,
+      "learning_rate": 4.695241840346048e-05,
+      "loss": 1.491,
+      "step": 31000
+    },
+    {
+      "epoch": 0.62,
+      "learning_rate": 4.690326386158081e-05,
+      "loss": 1.4883,
+      "step": 31500
+    },
+    {
+      "epoch": 0.63,
+      "learning_rate": 4.685410931970115e-05,
+      "loss": 1.4935,
+      "step": 32000
+    },
+    {
+      "epoch": 0.64,
+      "learning_rate": 4.6804954777821476e-05,
+      "loss": 1.4733,
+      "step": 32500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 4.6755800235941804e-05,
+      "loss": 1.487,
+      "step": 33000
+    },
+    {
+      "epoch": 0.66,
+      "learning_rate": 4.670664569406213e-05,
+      "loss": 1.4625,
+      "step": 33500
+    },
+    {
+      "epoch": 0.67,
+      "learning_rate": 4.6657491152182466e-05,
+      "loss": 1.4589,
+      "step": 34000
+    },
+    {
+      "epoch": 0.68,
+      "learning_rate": 4.6608336610302794e-05,
+      "loss": 1.4498,
+      "step": 34500
+    },
+    {
+      "epoch": 0.69,
+      "learning_rate": 4.655918206842312e-05,
+      "loss": 1.4464,
+      "step": 35000
+    },
+    {
+      "epoch": 0.7,
+      "learning_rate": 4.651002752654346e-05,
+      "loss": 1.441,
+      "step": 35500
+    },
+    {
+      "epoch": 0.71,
+      "learning_rate": 4.6460872984663785e-05,
+      "loss": 1.4486,
+      "step": 36000
+    },
+    {
+      "epoch": 0.72,
+      "learning_rate": 4.641171844278411e-05,
+      "loss": 1.4283,
+      "step": 36500
+    },
+    {
+      "epoch": 0.73,
+      "learning_rate": 4.636256390090444e-05,
+      "loss": 1.4265,
+      "step": 37000
+    },
+    {
+      "epoch": 0.74,
+      "learning_rate": 4.631340935902478e-05,
+      "loss": 1.4173,
+      "step": 37500
+    },
+    {
+      "epoch": 0.75,
+      "learning_rate": 4.626425481714511e-05,
+      "loss": 1.4309,
+      "step": 38000
+    },
+    {
+      "epoch": 0.76,
+      "learning_rate": 4.621510027526544e-05,
+      "loss": 1.411,
+      "step": 38500
+    },
+    {
+      "epoch": 0.77,
+      "learning_rate": 4.6165945733385765e-05,
+      "loss": 1.4155,
+      "step": 39000
+    },
+    {
+      "epoch": 0.78,
+      "learning_rate": 4.61167911915061e-05,
+      "loss": 1.4097,
+      "step": 39500
+    },
+    {
+      "epoch": 0.79,
+      "learning_rate": 4.606763664962643e-05,
+      "loss": 1.3916,
+      "step": 40000
+    },
+    {
+      "epoch": 0.8,
+      "learning_rate": 4.6018482107746756e-05,
+      "loss": 1.399,
+      "step": 40500
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 4.596932756586709e-05,
+      "loss": 1.3804,
+      "step": 41000
+    },
+    {
+      "epoch": 0.82,
+      "learning_rate": 4.592017302398742e-05,
+      "loss": 1.3847,
+      "step": 41500
+    },
+    {
+      "epoch": 0.83,
+      "learning_rate": 4.5871018482107746e-05,
+      "loss": 1.3797,
+      "step": 42000
+    },
+    {
+      "epoch": 0.84,
+      "learning_rate": 4.5821863940228074e-05,
+      "loss": 1.3729,
+      "step": 42500
+    },
+    {
+      "epoch": 0.85,
+      "learning_rate": 4.5772709398348415e-05,
+      "loss": 1.3667,
+      "step": 43000
+    },
+    {
+      "epoch": 0.86,
+      "learning_rate": 4.572355485646874e-05,
+      "loss": 1.3765,
+      "step": 43500
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.567440031458907e-05,
+      "loss": 1.378,
+      "step": 44000
+    },
+    {
+      "epoch": 0.87,
+      "learning_rate": 4.56252457727094e-05,
+      "loss": 1.3574,
+      "step": 44500
+    },
+    {
+      "epoch": 0.88,
+      "learning_rate": 4.5576091230829733e-05,
+      "loss": 1.3604,
+      "step": 45000
+    },
+    {
+      "epoch": 0.89,
+      "learning_rate": 4.552693668895006e-05,
+      "loss": 1.3426,
+      "step": 45500
+    },
+    {
+      "epoch": 0.9,
+      "learning_rate": 4.547778214707039e-05,
+      "loss": 1.3398,
+      "step": 46000
+    },
+    {
+      "epoch": 0.91,
+      "learning_rate": 4.5428627605190724e-05,
+      "loss": 1.3466,
+      "step": 46500
+    },
+    {
+      "epoch": 0.92,
+      "learning_rate": 4.537947306331105e-05,
+      "loss": 1.3327,
+      "step": 47000
+    },
+    {
+      "epoch": 0.93,
+      "learning_rate": 4.533031852143138e-05,
+      "loss": 1.3253,
+      "step": 47500
+    },
+    {
+      "epoch": 0.94,
+      "learning_rate": 4.528116397955171e-05,
+      "loss": 1.337,
+      "step": 48000
+    },
+    {
+      "epoch": 0.95,
+      "learning_rate": 4.523200943767204e-05,
+      "loss": 1.3282,
+      "step": 48500
+    },
+    {
+      "epoch": 0.96,
+      "learning_rate": 4.518285489579238e-05,
+      "loss": 1.3378,
+      "step": 49000
+    },
+    {
+      "epoch": 0.97,
+      "learning_rate": 4.5133700353912705e-05,
+      "loss": 1.3306,
+      "step": 49500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 4.508454581203303e-05,
+      "loss": 1.3192,
+      "step": 50000
+    },
+    {
+      "epoch": 0.99,
+      "learning_rate": 4.503539127015337e-05,
+      "loss": 1.3191,
+      "step": 50500
+    },
+    {
+      "epoch": 1.0,
+      "learning_rate": 4.4986236728273695e-05,
+      "loss": 1.316,
+      "step": 51000
+    },
+    {
+      "epoch": 1.01,
+      "learning_rate": 4.493708218639402e-05,
+      "loss": 1.3035,
+      "step": 51500
+    },
+    {
+      "epoch": 1.02,
+      "learning_rate": 4.488792764451436e-05,
+      "loss": 1.2951,
+      "step": 52000
+    },
+    {
+      "epoch": 1.03,
+      "learning_rate": 4.4838773102634685e-05,
+      "loss": 1.2945,
+      "step": 52500
+    },
+    {
+      "epoch": 1.04,
+      "learning_rate": 4.478961856075501e-05,
+      "loss": 1.2912,
+      "step": 53000
+    },
+    {
+      "epoch": 1.05,
+      "learning_rate": 4.474046401887534e-05,
+      "loss": 1.2787,
+      "step": 53500
+    },
+    {
+      "epoch": 1.06,
+      "learning_rate": 4.4691309476995676e-05,
+      "loss": 1.268,
+      "step": 54000
+    },
+    {
+      "epoch": 1.07,
+      "learning_rate": 4.464215493511601e-05,
+      "loss": 1.2773,
+      "step": 54500
+    },
+    {
+      "epoch": 1.08,
+      "learning_rate": 4.459300039323634e-05,
+      "loss": 1.2789,
+      "step": 55000
+    },
+    {
+      "epoch": 1.09,
+      "learning_rate": 4.4543845851356666e-05,
+      "loss": 1.2872,
+      "step": 55500
+    },
+    {
+      "epoch": 1.1,
+      "learning_rate": 4.4494691309477e-05,
+      "loss": 1.2806,
+      "step": 56000
+    },
+    {
+      "epoch": 1.11,
+      "learning_rate": 4.444553676759733e-05,
+      "loss": 1.265,
+      "step": 56500
+    },
+    {
+      "epoch": 1.12,
+      "learning_rate": 4.4396382225717656e-05,
+      "loss": 1.2772,
+      "step": 57000
+    },
+    {
+      "epoch": 1.13,
+      "learning_rate": 4.434722768383799e-05,
+      "loss": 1.2578,
+      "step": 57500
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 4.429807314195832e-05,
+      "loss": 1.2662,
+      "step": 58000
+    },
+    {
+      "epoch": 1.15,
+      "learning_rate": 4.424891860007865e-05,
+      "loss": 1.2703,
+      "step": 58500
+    },
+    {
+      "epoch": 1.16,
+      "learning_rate": 4.419976405819898e-05,
+      "loss": 1.2594,
+      "step": 59000
+    },
+    {
+      "epoch": 1.17,
+      "learning_rate": 4.415060951631931e-05,
+      "loss": 1.2607,
+      "step": 59500
+    },
+    {
+      "epoch": 1.18,
+      "learning_rate": 4.410145497443964e-05,
+      "loss": 1.2501,
+      "step": 60000
+    },
+    {
+      "epoch": 1.19,
+      "learning_rate": 4.405230043255997e-05,
+      "loss": 1.2423,
+      "step": 60500
+    },
+    {
+      "epoch": 1.2,
+      "learning_rate": 4.40031458906803e-05,
+      "loss": 1.2518,
+      "step": 61000
+    },
+    {
+      "epoch": 1.21,
+      "learning_rate": 4.3953991348800634e-05,
+      "loss": 1.25,
+      "step": 61500
+    },
+    {
+      "epoch": 1.22,
+      "learning_rate": 4.390483680692096e-05,
+      "loss": 1.2464,
+      "step": 62000
+    },
+    {
+      "epoch": 1.23,
+      "learning_rate": 4.385568226504129e-05,
+      "loss": 1.2386,
+      "step": 62500
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 4.3806527723161625e-05,
+      "loss": 1.2492,
+      "step": 63000
+    },
+    {
+      "epoch": 1.25,
+      "learning_rate": 4.375737318128195e-05,
+      "loss": 1.2396,
+      "step": 63500
+    },
+    {
+      "epoch": 1.26,
+      "learning_rate": 4.370821863940228e-05,
+      "loss": 1.2515,
+      "step": 64000
+    },
+    {
+      "epoch": 1.27,
+      "learning_rate": 4.3659064097522615e-05,
+      "loss": 1.2206,
+      "step": 64500
+    },
+    {
+      "epoch": 1.28,
+      "learning_rate": 4.360990955564294e-05,
+      "loss": 1.2247,
+      "step": 65000
+    },
+    {
+      "epoch": 1.29,
+      "learning_rate": 4.356075501376327e-05,
+      "loss": 1.2272,
+      "step": 65500
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 4.3511600471883605e-05,
+      "loss": 1.2426,
+      "step": 66000
+    },
+    {
+      "epoch": 1.31,
+      "learning_rate": 4.346244593000394e-05,
+      "loss": 1.2121,
+      "step": 66500
+    },
+    {
+      "epoch": 1.32,
+      "learning_rate": 4.341329138812427e-05,
+      "loss": 1.2197,
+      "step": 67000
+    },
+    {
+      "epoch": 1.33,
+      "learning_rate": 4.3364136846244596e-05,
+      "loss": 1.2219,
+      "step": 67500
+    },
+    {
+      "epoch": 1.34,
+      "learning_rate": 4.3314982304364923e-05,
+      "loss": 1.1951,
+      "step": 68000
+    },
+    {
+      "epoch": 1.35,
+      "learning_rate": 4.326582776248526e-05,
+      "loss": 1.2068,
+      "step": 68500
+    },
+    {
+      "epoch": 1.36,
+      "learning_rate": 4.3216673220605586e-05,
+      "loss": 1.2066,
+      "step": 69000
+    },
+    {
+      "epoch": 1.37,
+      "learning_rate": 4.3167518678725914e-05,
+      "loss": 1.1981,
+      "step": 69500
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 4.311836413684625e-05,
+      "loss": 1.2024,
+      "step": 70000
+    },
+    {
+      "epoch": 1.39,
+      "learning_rate": 4.3069209594966576e-05,
+      "loss": 1.2009,
+      "step": 70500
+    },
+    {
+      "epoch": 1.4,
+      "learning_rate": 4.3020055053086904e-05,
+      "loss": 1.2027,
+      "step": 71000
+    },
+    {
+      "epoch": 1.41,
+      "learning_rate": 4.297090051120723e-05,
+      "loss": 1.2003,
+      "step": 71500
+    },
+    {
+      "epoch": 1.42,
+      "learning_rate": 4.2921745969327573e-05,
+      "loss": 1.1934,
+      "step": 72000
+    },
+    {
+      "epoch": 1.43,
+      "learning_rate": 4.28725914274479e-05,
+      "loss": 1.2002,
+      "step": 72500
+    },
+    {
+      "epoch": 1.44,
+      "learning_rate": 4.282343688556823e-05,
+      "loss": 1.1969,
+      "step": 73000
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 4.277428234368856e-05,
+      "loss": 1.1947,
+      "step": 73500
+    },
+    {
+      "epoch": 1.45,
+      "learning_rate": 4.272512780180889e-05,
+      "loss": 1.1822,
+      "step": 74000
+    },
+    {
+      "epoch": 1.46,
+      "learning_rate": 4.267597325992922e-05,
+      "loss": 1.1913,
+      "step": 74500
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 4.262681871804955e-05,
+      "loss": 1.1953,
+      "step": 75000
+    },
+    {
+      "epoch": 1.48,
+      "learning_rate": 4.257766417616988e-05,
+      "loss": 1.1846,
+      "step": 75500
+    },
+    {
+      "epoch": 1.49,
+      "learning_rate": 4.252850963429021e-05,
+      "loss": 1.1956,
+      "step": 76000
+    },
+    {
+      "epoch": 1.5,
+      "learning_rate": 4.247935509241054e-05,
+      "loss": 1.1806,
+      "step": 76500
+    },
+    {
+      "epoch": 1.51,
+      "learning_rate": 4.2430200550530866e-05,
+      "loss": 1.1779,
+      "step": 77000
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 4.238104600865121e-05,
+      "loss": 1.174,
+      "step": 77500
+    },
+    {
+      "epoch": 1.53,
+      "learning_rate": 4.2331891466771535e-05,
+      "loss": 1.1799,
+      "step": 78000
+    },
+    {
+      "epoch": 1.54,
+      "learning_rate": 4.228273692489186e-05,
+      "loss": 1.1598,
+      "step": 78500
+    },
+    {
+      "epoch": 1.55,
+      "learning_rate": 4.223358238301219e-05,
+      "loss": 1.1724,
+      "step": 79000
+    },
+    {
+      "epoch": 1.56,
+      "learning_rate": 4.2184427841132525e-05,
+      "loss": 1.1707,
+      "step": 79500
+    },
+    {
+      "epoch": 1.57,
+      "learning_rate": 4.213527329925285e-05,
+      "loss": 1.175,
+      "step": 80000
+    },
+    {
+      "epoch": 1.58,
+      "learning_rate": 4.208611875737318e-05,
+      "loss": 1.1588,
+      "step": 80500
+    },
+    {
+      "epoch": 1.59,
+      "learning_rate": 4.2036964215493516e-05,
+      "loss": 1.1796,
+      "step": 81000
+    },
+    {
+      "epoch": 1.6,
+      "learning_rate": 4.1987809673613843e-05,
+      "loss": 1.1678,
+      "step": 81500
+    },
+    {
+      "epoch": 1.61,
+      "learning_rate": 4.193865513173417e-05,
+      "loss": 1.1689,
+      "step": 82000
+    },
+    {
+      "epoch": 1.62,
+      "learning_rate": 4.18895005898545e-05,
+      "loss": 1.1536,
+      "step": 82500
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 4.1840346047974834e-05,
+      "loss": 1.1578,
+      "step": 83000
+    },
+    {
+      "epoch": 1.64,
+      "learning_rate": 4.179119150609517e-05,
+      "loss": 1.1559,
+      "step": 83500
+    },
+    {
+      "epoch": 1.65,
+      "learning_rate": 4.1742036964215496e-05,
+      "loss": 1.1559,
+      "step": 84000
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 4.1692882422335824e-05,
+      "loss": 1.1495,
+      "step": 84500
+    },
+    {
+      "epoch": 1.67,
+      "learning_rate": 4.164372788045616e-05,
+      "loss": 1.1531,
+      "step": 85000
+    },
+    {
+      "epoch": 1.68,
+      "learning_rate": 4.159457333857649e-05,
+      "loss": 1.1426,
+      "step": 85500
+    },
+    {
+      "epoch": 1.69,
+      "learning_rate": 4.1545418796696815e-05,
+      "loss": 1.1647,
+      "step": 86000
+    },
+    {
+      "epoch": 1.7,
+      "learning_rate": 4.149626425481715e-05,
+      "loss": 1.1457,
+      "step": 86500
+    },
+    {
+      "epoch": 1.71,
+      "learning_rate": 4.144710971293748e-05,
+      "loss": 1.1493,
+      "step": 87000
+    },
+    {
+      "epoch": 1.72,
+      "learning_rate": 4.1397955171057805e-05,
+      "loss": 1.1545,
+      "step": 87500
+    },
+    {
+      "epoch": 1.73,
+      "learning_rate": 4.134880062917813e-05,
+      "loss": 1.1392,
+      "step": 88000
+    },
+    {
+      "epoch": 1.74,
+      "learning_rate": 4.129964608729847e-05,
+      "loss": 1.1561,
+      "step": 88500
+    },
+    {
+      "epoch": 1.75,
+      "learning_rate": 4.12504915454188e-05,
+      "loss": 1.1363,
+      "step": 89000
+    },
+    {
+      "epoch": 1.76,
+      "learning_rate": 4.120133700353913e-05,
+      "loss": 1.1419,
+      "step": 89500
+    },
+    {
+      "epoch": 1.77,
+      "learning_rate": 4.115218246165946e-05,
+      "loss": 1.1408,
+      "step": 90000
+    },
+    {
+      "epoch": 1.78,
+      "learning_rate": 4.110302791977979e-05,
+      "loss": 1.1412,
+      "step": 90500
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 4.105387337790012e-05,
+      "loss": 1.1396,
+      "step": 91000
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 4.100471883602045e-05,
+      "loss": 1.1399,
+      "step": 91500
+    },
+    {
+      "epoch": 1.81,
+      "learning_rate": 4.095556429414078e-05,
+      "loss": 1.1394,
+      "step": 92000
+    },
+    {
+      "epoch": 1.82,
+      "learning_rate": 4.090640975226111e-05,
+      "loss": 1.1282,
+      "step": 92500
+    },
+    {
+      "epoch": 1.83,
+      "learning_rate": 4.085725521038144e-05,
+      "loss": 1.1286,
+      "step": 93000
+    },
+    {
+      "epoch": 1.84,
+      "learning_rate": 4.0808100668501766e-05,
+      "loss": 1.1295,
+      "step": 93500
+    },
+    {
+      "epoch": 1.85,
+      "learning_rate": 4.07589461266221e-05,
+      "loss": 1.132,
+      "step": 94000
+    },
+    {
+      "epoch": 1.86,
+      "learning_rate": 4.070979158474243e-05,
+      "loss": 1.1294,
+      "step": 94500
+    },
+    {
+      "epoch": 1.87,
+      "learning_rate": 4.0660637042862763e-05,
+      "loss": 1.1155,
+      "step": 95000
+    },
+    {
+      "epoch": 1.88,
+      "learning_rate": 4.061148250098309e-05,
+      "loss": 1.1194,
+      "step": 95500
+    },
+    {
+      "epoch": 1.89,
+      "learning_rate": 4.0562327959103426e-05,
+      "loss": 1.1236,
+      "step": 96000
+    },
+    {
+      "epoch": 1.9,
+      "learning_rate": 4.0513173417223754e-05,
+      "loss": 1.119,
+      "step": 96500
+    },
+    {
+      "epoch": 1.91,
+      "learning_rate": 4.046401887534408e-05,
+      "loss": 1.1285,
+      "step": 97000
+    },
+    {
+      "epoch": 1.92,
+      "learning_rate": 4.0414864333464416e-05,
+      "loss": 1.1138,
+      "step": 97500
+    },
+    {
+      "epoch": 1.93,
+      "learning_rate": 4.0365709791584744e-05,
+      "loss": 1.1138,
+      "step": 98000
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 4.031655524970507e-05,
+      "loss": 1.1217,
+      "step": 98500
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 4.026740070782541e-05,
+      "loss": 1.1228,
+      "step": 99000
+    },
+    {
+      "epoch": 1.96,
+      "learning_rate": 4.0218246165945735e-05,
+      "loss": 1.121,
+      "step": 99500
+    },
+    {
+      "epoch": 1.97,
+      "learning_rate": 4.016909162406606e-05,
+      "loss": 1.114,
+      "step": 100000
+    },
+    {
+      "epoch": 1.98,
+      "learning_rate": 4.01199370821864e-05,
+      "loss": 1.0939,
+      "step": 100500
+    },
+    {
+      "epoch": 1.99,
+      "learning_rate": 4.0070782540306725e-05,
+      "loss": 1.1102,
+      "step": 101000
+    },
+    {
+      "epoch": 2.0,
+      "learning_rate": 4.002162799842706e-05,
+      "loss": 1.1052,
+      "step": 101500
+    },
+    {
+      "epoch": 2.01,
+      "learning_rate": 3.997247345654739e-05,
+      "loss": 1.1097,
+      "step": 102000
+    },
+    {
+      "epoch": 2.02,
+      "learning_rate": 3.9923318914667715e-05,
+      "loss": 1.0889,
+      "step": 102500
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 3.987416437278805e-05,
+      "loss": 1.0865,
+      "step": 103000
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 3.982500983090838e-05,
+      "loss": 1.0952,
+      "step": 103500
+    },
+    {
+      "epoch": 2.04,
+      "learning_rate": 3.9775855289028706e-05,
+      "loss": 1.0877,
+      "step": 104000
+    },
+    {
+      "epoch": 2.05,
+      "learning_rate": 3.972670074714904e-05,
+      "loss": 1.0977,
+      "step": 104500
+    },
+    {
+      "epoch": 2.06,
+      "learning_rate": 3.967754620526937e-05,
+      "loss": 1.0931,
+      "step": 105000
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 3.9628391663389696e-05,
+      "loss": 1.0867,
+      "step": 105500
+    },
+    {
+      "epoch": 2.08,
+      "learning_rate": 3.957923712151003e-05,
+      "loss": 1.0886,
+      "step": 106000
+    },
+    {
+      "epoch": 2.09,
+      "learning_rate": 3.9530082579630365e-05,
+      "loss": 1.0922,
+      "step": 106500
+    },
+    {
+      "epoch": 2.1,
+      "learning_rate": 3.948092803775069e-05,
+      "loss": 1.0695,
+      "step": 107000
+    },
+    {
+      "epoch": 2.11,
+      "learning_rate": 3.943177349587102e-05,
+      "loss": 1.0791,
+      "step": 107500
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 3.938261895399135e-05,
+      "loss": 1.0952,
+      "step": 108000
+    },
+    {
+      "epoch": 2.13,
+      "learning_rate": 3.9333464412111683e-05,
+      "loss": 1.0809,
+      "step": 108500
+    },
+    {
+      "epoch": 2.14,
+      "learning_rate": 3.928430987023201e-05,
+      "loss": 1.0857,
+      "step": 109000
+    },
+    {
+      "epoch": 2.15,
+      "learning_rate": 3.923515532835234e-05,
+      "loss": 1.0904,
+      "step": 109500
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 3.9186000786472674e-05,
+      "loss": 1.075,
+      "step": 110000
+    },
+    {
+      "epoch": 2.17,
+      "learning_rate": 3.9136846244593e-05,
+      "loss": 1.0758,
+      "step": 110500
+    },
+    {
+      "epoch": 2.18,
+      "learning_rate": 3.908769170271333e-05,
+      "loss": 1.0709,
+      "step": 111000
+    },
+    {
+      "epoch": 2.19,
+      "learning_rate": 3.903853716083366e-05,
+      "loss": 1.0911,
+      "step": 111500
+    },
+    {
+      "epoch": 2.2,
+      "learning_rate": 3.8989382618954e-05,
+      "loss": 1.0792,
+      "step": 112000
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 3.894022807707433e-05,
+      "loss": 1.0647,
+      "step": 112500
+    },
+    {
+      "epoch": 2.22,
+      "learning_rate": 3.8891073535194655e-05,
+      "loss": 1.075,
+      "step": 113000
+    },
+    {
+      "epoch": 2.23,
+      "learning_rate": 3.884191899331498e-05,
+      "loss": 1.0679,
+      "step": 113500
+    },
+    {
+      "epoch": 2.24,
+      "learning_rate": 3.879276445143532e-05,
+      "loss": 1.0721,
+      "step": 114000
+    },
+    {
+      "epoch": 2.25,
+      "learning_rate": 3.8743609909555645e-05,
+      "loss": 1.0644,
+      "step": 114500
+    },
+    {
+      "epoch": 2.26,
+      "learning_rate": 3.869445536767597e-05,
+      "loss": 1.065,
+      "step": 115000
+    },
+    {
+      "epoch": 2.27,
+      "learning_rate": 3.864530082579631e-05,
+      "loss": 1.0658,
+      "step": 115500
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 3.8596146283916635e-05,
+      "loss": 1.0685,
+      "step": 116000
+    },
+    {
+      "epoch": 2.29,
+      "learning_rate": 3.854699174203696e-05,
+      "loss": 1.0833,
+      "step": 116500
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 3.849783720015729e-05,
+      "loss": 1.0584,
+      "step": 117000
+    },
+    {
+      "epoch": 2.31,
+      "learning_rate": 3.844868265827763e-05,
+      "loss": 1.0713,
+      "step": 117500
+    },
+    {
+      "epoch": 2.32,
+      "learning_rate": 3.839952811639796e-05,
+      "loss": 1.0549,
+      "step": 118000
+    },
+    {
+      "epoch": 2.33,
+      "learning_rate": 3.835037357451829e-05,
+      "loss": 1.0563,
+      "step": 118500
+    },
+    {
+      "epoch": 2.34,
+      "learning_rate": 3.8301219032638616e-05,
+      "loss": 1.0497,
+      "step": 119000
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 3.825206449075895e-05,
+      "loss": 1.062,
+      "step": 119500
+    },
+    {
+      "epoch": 2.36,
+      "learning_rate": 3.820290994887928e-05,
+      "loss": 1.0632,
+      "step": 120000
+    },
+    {
+      "epoch": 2.37,
+      "learning_rate": 3.8153755406999606e-05,
+      "loss": 1.0727,
+      "step": 120500
+    },
+    {
+      "epoch": 2.38,
+      "learning_rate": 3.810460086511994e-05,
+      "loss": 1.0598,
+      "step": 121000
+    },
+    {
+      "epoch": 2.39,
+      "learning_rate": 3.805544632324027e-05,
+      "loss": 1.0472,
+      "step": 121500
+    },
+    {
+      "epoch": 2.4,
+      "learning_rate": 3.80062917813606e-05,
+      "loss": 1.0476,
+      "step": 122000
+    },
+    {
+      "epoch": 2.41,
+      "learning_rate": 3.7957137239480925e-05,
+      "loss": 1.0586,
+      "step": 122500
+    },
+    {
+      "epoch": 2.42,
+      "learning_rate": 3.790798269760126e-05,
+      "loss": 1.0544,
+      "step": 123000
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 3.7858828155721594e-05,
+      "loss": 1.0625,
+      "step": 123500
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 3.780967361384192e-05,
+      "loss": 1.0521,
+      "step": 124000
+    },
+    {
+      "epoch": 2.45,
+      "learning_rate": 3.776051907196225e-05,
+      "loss": 1.0443,
+      "step": 124500
+    },
+    {
+      "epoch": 2.46,
+      "learning_rate": 3.7711364530082584e-05,
+      "loss": 1.0614,
+      "step": 125000
+    },
+    {
+      "epoch": 2.47,
+      "learning_rate": 3.766220998820291e-05,
+      "loss": 1.0455,
+      "step": 125500
+    },
+    {
+      "epoch": 2.48,
+      "learning_rate": 3.761305544632324e-05,
+      "loss": 1.0448,
+      "step": 126000
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 3.7563900904443575e-05,
+      "loss": 1.0527,
+      "step": 126500
+    },
+    {
+      "epoch": 2.5,
+      "learning_rate": 3.75147463625639e-05,
+      "loss": 1.0393,
+      "step": 127000
+    },
+    {
+      "epoch": 2.51,
+      "learning_rate": 3.746559182068423e-05,
+      "loss": 1.0405,
+      "step": 127500
+    },
+    {
+      "epoch": 2.52,
+      "learning_rate": 3.741643727880456e-05,
+      "loss": 1.0341,
+      "step": 128000
+    },
+    {
+      "epoch": 2.53,
+      "learning_rate": 3.736728273692489e-05,
+      "loss": 1.0404,
+      "step": 128500
+    },
+    {
+      "epoch": 2.54,
+      "learning_rate": 3.731812819504523e-05,
+      "loss": 1.0347,
+      "step": 129000
+    },
+    {
+      "epoch": 2.55,
+      "learning_rate": 3.7268973653165555e-05,
+      "loss": 1.0495,
+      "step": 129500
+    },
+    {
+      "epoch": 2.56,
+      "learning_rate": 3.721981911128588e-05,
+      "loss": 1.0613,
+      "step": 130000
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 3.717066456940622e-05,
+      "loss": 1.0439,
+      "step": 130500
+    },
+    {
+      "epoch": 2.58,
+      "learning_rate": 3.7121510027526546e-05,
+      "loss": 1.0346,
+      "step": 131000
+    },
+    {
+      "epoch": 2.59,
+      "learning_rate": 3.7072355485646873e-05,
+      "loss": 1.0467,
+      "step": 131500
+    },
+    {
+      "epoch": 2.6,
+      "learning_rate": 3.702320094376721e-05,
+      "loss": 1.0357,
+      "step": 132000
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 3.6974046401887536e-05,
+      "loss": 1.037,
+      "step": 132500
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 3.6924891860007864e-05,
+      "loss": 1.04,
+      "step": 133000
+    },
+    {
+      "epoch": 2.62,
+      "learning_rate": 3.687573731812819e-05,
+      "loss": 1.0377,
+      "step": 133500
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 3.6826582776248526e-05,
+      "loss": 1.037,
+      "step": 134000
+    },
+    {
+      "epoch": 2.64,
+      "learning_rate": 3.6777428234368854e-05,
+      "loss": 1.0349,
+      "step": 134500
+    },
+    {
+      "epoch": 2.65,
+      "learning_rate": 3.672827369248919e-05,
+      "loss": 1.0344,
+      "step": 135000
+    },
+    {
+      "epoch": 2.66,
+      "learning_rate": 3.667911915060952e-05,
+      "loss": 1.0398,
+      "step": 135500
+    },
+    {
+      "epoch": 2.67,
+      "learning_rate": 3.662996460872985e-05,
+      "loss": 1.0351,
+      "step": 136000
+    },
+    {
+      "epoch": 2.68,
+      "learning_rate": 3.658081006685018e-05,
+      "loss": 1.0319,
+      "step": 136500
+    },
+    {
+      "epoch": 2.69,
+      "learning_rate": 3.653165552497051e-05,
+      "loss": 1.0313,
+      "step": 137000
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 3.648250098309084e-05,
+      "loss": 1.0188,
+      "step": 137500
+    },
+    {
+      "epoch": 2.71,
+      "learning_rate": 3.643334644121117e-05,
+      "loss": 1.0337,
+      "step": 138000
+    },
+    {
+      "epoch": 2.72,
+      "learning_rate": 3.63841918993315e-05,
+      "loss": 1.0237,
+      "step": 138500
+    },
+    {
+      "epoch": 2.73,
+      "learning_rate": 3.633503735745183e-05,
+      "loss": 1.019,
+      "step": 139000
+    },
+    {
+      "epoch": 2.74,
+      "learning_rate": 3.628588281557216e-05,
+      "loss": 1.0349,
+      "step": 139500
+    },
+    {
+      "epoch": 2.75,
+      "learning_rate": 3.623672827369249e-05,
+      "loss": 1.0288,
+      "step": 140000
+    },
+    {
+      "epoch": 2.76,
+      "learning_rate": 3.618757373181282e-05,
+      "loss": 1.0249,
+      "step": 140500
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 3.613841918993316e-05,
+      "loss": 1.0196,
+      "step": 141000
+    },
+    {
+      "epoch": 2.78,
+      "learning_rate": 3.6089264648053485e-05,
+      "loss": 1.0194,
+      "step": 141500
+    },
+    {
+      "epoch": 2.79,
+      "learning_rate": 3.604011010617381e-05,
+      "loss": 1.0231,
+      "step": 142000
+    },
+    {
+      "epoch": 2.8,
+      "learning_rate": 3.599095556429414e-05,
+      "loss": 1.023,
+      "step": 142500
+    },
+    {
+      "epoch": 2.81,
+      "learning_rate": 3.5941801022414475e-05,
+      "loss": 1.0237,
+      "step": 143000
+    },
+    {
+      "epoch": 2.82,
+      "learning_rate": 3.58926464805348e-05,
+      "loss": 1.0143,
+      "step": 143500
+    },
+    {
+      "epoch": 2.83,
+      "learning_rate": 3.584349193865513e-05,
+      "loss": 1.0362,
+      "step": 144000
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 3.5794337396775466e-05,
+      "loss": 1.0177,
+      "step": 144500
+    },
+    {
+      "epoch": 2.85,
+      "learning_rate": 3.5745182854895793e-05,
+      "loss": 1.0137,
+      "step": 145000
+    },
+    {
+      "epoch": 2.86,
+      "learning_rate": 3.569602831301612e-05,
+      "loss": 1.0201,
+      "step": 145500
+    },
+    {
+      "epoch": 2.87,
+      "learning_rate": 3.564687377113645e-05,
+      "loss": 1.0108,
+      "step": 146000
+    },
+    {
+      "epoch": 2.88,
+      "learning_rate": 3.559771922925679e-05,
+      "loss": 1.0214,
+      "step": 146500
+    },
+    {
+      "epoch": 2.89,
+      "learning_rate": 3.554856468737712e-05,
+      "loss": 1.0192,
+      "step": 147000
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 3.5499410145497446e-05,
+      "loss": 1.0106,
+      "step": 147500
+    },
+    {
+      "epoch": 2.91,
+      "learning_rate": 3.5450255603617774e-05,
+      "loss": 1.007,
+      "step": 148000
+    },
+    {
+      "epoch": 2.92,
+      "learning_rate": 3.540110106173811e-05,
+      "loss": 1.0105,
+      "step": 148500
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 3.535194651985844e-05,
+      "loss": 1.01,
+      "step": 149000
+    },
+    {
+      "epoch": 2.94,
+      "learning_rate": 3.5302791977978765e-05,
+      "loss": 0.9991,
+      "step": 149500
+    },
+    {
+      "epoch": 2.95,
+      "learning_rate": 3.52536374360991e-05,
+      "loss": 1.0076,
+      "step": 150000
+    },
+    {
+      "epoch": 2.96,
+      "learning_rate": 3.520448289421943e-05,
+      "loss": 0.9889,
+      "step": 150500
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 3.5155328352339755e-05,
+      "loss": 1.0086,
+      "step": 151000
+    },
+    {
+      "epoch": 2.98,
+      "learning_rate": 3.510617381046008e-05,
+      "loss": 0.9982,
+      "step": 151500
+    },
+    {
+      "epoch": 2.99,
+      "learning_rate": 3.5057019268580424e-05,
+      "loss": 1.0239,
+      "step": 152000
+    },
+    {
+      "epoch": 3.0,
+      "learning_rate": 3.500786472670075e-05,
+      "loss": 1.0058,
+      "step": 152500
+    },
+    {
+      "epoch": 3.01,
+      "learning_rate": 3.495871018482108e-05,
+      "loss": 0.9919,
+      "step": 153000
+    },
+    {
+      "epoch": 3.02,
+      "learning_rate": 3.490955564294141e-05,
+      "loss": 0.9908,
+      "step": 153500
+    },
+    {
+      "epoch": 3.03,
+      "learning_rate": 3.486040110106174e-05,
+      "loss": 0.9909,
+      "step": 154000
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 3.481124655918207e-05,
+      "loss": 0.9875,
+      "step": 154500
+    },
+    {
+      "epoch": 3.05,
+      "learning_rate": 3.47620920173024e-05,
+      "loss": 0.9964,
+      "step": 155000
+    },
+    {
+      "epoch": 3.06,
+      "learning_rate": 3.471293747542273e-05,
+      "loss": 1.0046,
+      "step": 155500
+    },
+    {
+      "epoch": 3.07,
+      "learning_rate": 3.466378293354306e-05,
+      "loss": 0.9902,
+      "step": 156000
+    },
+    {
+      "epoch": 3.08,
+      "learning_rate": 3.461462839166339e-05,
+      "loss": 0.9969,
+      "step": 156500
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 3.4565473849783716e-05,
+      "loss": 0.9927,
+      "step": 157000
+    },
+    {
+      "epoch": 3.1,
+      "learning_rate": 3.451631930790405e-05,
+      "loss": 0.982,
+      "step": 157500
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 3.4467164766024386e-05,
+      "loss": 0.9925,
+      "step": 158000
+    },
+    {
+      "epoch": 3.12,
+      "learning_rate": 3.4418010224144713e-05,
+      "loss": 0.9998,
+      "step": 158500
+    },
+    {
+      "epoch": 3.13,
+      "learning_rate": 3.436885568226504e-05,
+      "loss": 0.9885,
+      "step": 159000
+    },
+    {
+      "epoch": 3.14,
+      "learning_rate": 3.4319701140385376e-05,
+      "loss": 0.9937,
+      "step": 159500
+    },
+    {
+      "epoch": 3.15,
+      "learning_rate": 3.4270546598505704e-05,
+      "loss": 0.994,
+      "step": 160000
+    },
+    {
+      "epoch": 3.16,
+      "learning_rate": 3.422139205662603e-05,
+      "loss": 0.9903,
+      "step": 160500
+    },
+    {
+      "epoch": 3.17,
+      "learning_rate": 3.4172237514746366e-05,
+      "loss": 0.9833,
+      "step": 161000
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 3.4123082972866694e-05,
+      "loss": 0.9885,
+      "step": 161500
+    },
+    {
+      "epoch": 3.19,
+      "learning_rate": 3.407392843098702e-05,
+      "loss": 0.9767,
+      "step": 162000
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 3.402477388910735e-05,
+      "loss": 0.9803,
+      "step": 162500
+    },
+    {
+      "epoch": 3.2,
+      "learning_rate": 3.3975619347227685e-05,
+      "loss": 0.9805,
+      "step": 163000
+    },
+    {
+      "epoch": 3.21,
+      "learning_rate": 3.392646480534802e-05,
+      "loss": 0.989,
+      "step": 163500
+    },
+    {
+      "epoch": 3.22,
+      "learning_rate": 3.387731026346835e-05,
+      "loss": 0.9912,
+      "step": 164000
+    },
+    {
+      "epoch": 3.23,
+      "learning_rate": 3.3828155721588675e-05,
+      "loss": 0.9862,
+      "step": 164500
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 3.377900117970901e-05,
+      "loss": 0.9749,
+      "step": 165000
+    },
+    {
+      "epoch": 3.25,
+      "learning_rate": 3.372984663782934e-05,
+      "loss": 0.9795,
+      "step": 165500
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 3.3680692095949665e-05,
+      "loss": 0.9771,
+      "step": 166000
+    },
+    {
+      "epoch": 3.27,
+      "learning_rate": 3.363153755407e-05,
+      "loss": 0.9764,
+      "step": 166500
+    },
+    {
+      "epoch": 3.28,
+      "learning_rate": 3.358238301219033e-05,
+      "loss": 0.9857,
+      "step": 167000
+    },
+    {
+      "epoch": 3.29,
+      "learning_rate": 3.3533228470310656e-05,
+      "loss": 0.9892,
+      "step": 167500
+    },
+    {
+      "epoch": 3.3,
+      "learning_rate": 3.3484073928430984e-05,
+      "loss": 0.9787,
+      "step": 168000
+    },
+    {
+      "epoch": 3.31,
+      "learning_rate": 3.343491938655132e-05,
+      "loss": 0.9915,
+      "step": 168500
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 3.338576484467165e-05,
+      "loss": 0.9666,
+      "step": 169000
+    },
+    {
+      "epoch": 3.33,
+      "learning_rate": 3.333661030279198e-05,
+      "loss": 0.9831,
+      "step": 169500
+    },
+    {
+      "epoch": 3.34,
+      "learning_rate": 3.328745576091231e-05,
+      "loss": 0.9747,
+      "step": 170000
+    },
+    {
+      "epoch": 3.35,
+      "learning_rate": 3.323830121903264e-05,
+      "loss": 0.9757,
+      "step": 170500
+    },
+    {
+      "epoch": 3.36,
+      "learning_rate": 3.318914667715297e-05,
+      "loss": 0.9724,
+      "step": 171000
+    },
+    {
+      "epoch": 3.37,
+      "learning_rate": 3.31399921352733e-05,
+      "loss": 0.9753,
+      "step": 171500
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 3.3090837593393633e-05,
+      "loss": 0.979,
+      "step": 172000
+    },
+    {
+      "epoch": 3.39,
+      "learning_rate": 3.304168305151396e-05,
+      "loss": 0.973,
+      "step": 172500
+    },
+    {
+      "epoch": 3.4,
+      "learning_rate": 3.299252850963429e-05,
+      "loss": 0.9694,
+      "step": 173000
+    },
+    {
+      "epoch": 3.41,
+      "learning_rate": 3.2943373967754624e-05,
+      "loss": 0.9577,
+      "step": 173500
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 3.289421942587495e-05,
+      "loss": 0.971,
+      "step": 174000
+    },
+    {
+      "epoch": 3.43,
+      "learning_rate": 3.284506488399528e-05,
+      "loss": 0.9624,
+      "step": 174500
+    },
+    {
+      "epoch": 3.44,
+      "learning_rate": 3.2795910342115614e-05,
+      "loss": 0.9716,
+      "step": 175000
+    },
+    {
+      "epoch": 3.45,
+      "learning_rate": 3.274675580023594e-05,
+      "loss": 0.9722,
+      "step": 175500
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 3.269760125835628e-05,
+      "loss": 0.9705,
+      "step": 176000
+    },
+    {
+      "epoch": 3.47,
+      "learning_rate": 3.2648446716476605e-05,
+      "loss": 0.9772,
+      "step": 176500
+    },
+    {
+      "epoch": 3.48,
+      "learning_rate": 3.259929217459693e-05,
+      "loss": 0.972,
+      "step": 177000
+    },
+    {
+      "epoch": 3.49,
+      "learning_rate": 3.255013763271727e-05,
+      "loss": 0.959,
+      "step": 177500
+    },
+    {
+      "epoch": 3.5,
+      "learning_rate": 3.2500983090837595e-05,
+      "loss": 0.9589,
+      "step": 178000
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 3.245182854895792e-05,
+      "loss": 0.9893,
+      "step": 178500
+    },
+    {
+      "epoch": 3.52,
+      "learning_rate": 3.240267400707826e-05,
+      "loss": 0.9634,
+      "step": 179000
+    },
+    {
+      "epoch": 3.53,
+      "learning_rate": 3.2353519465198585e-05,
+      "loss": 0.9634,
+      "step": 179500
+    },
+    {
+      "epoch": 3.54,
+      "learning_rate": 3.230436492331891e-05,
+      "loss": 0.9689,
+      "step": 180000
+    },
+    {
+      "epoch": 3.55,
+      "learning_rate": 3.225521038143925e-05,
+      "loss": 0.9702,
+      "step": 180500
+    },
+    {
+      "epoch": 3.56,
+      "learning_rate": 3.220605583955958e-05,
+      "loss": 0.9681,
+      "step": 181000
+    },
+    {
+      "epoch": 3.57,
+      "learning_rate": 3.215690129767991e-05,
+      "loss": 0.9618,
+      "step": 181500
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 3.210774675580024e-05,
+      "loss": 0.9647,
+      "step": 182000
+    },
+    {
+      "epoch": 3.59,
+      "learning_rate": 3.2058592213920566e-05,
+      "loss": 0.967,
+      "step": 182500
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 3.20094376720409e-05,
+      "loss": 0.9672,
+      "step": 183000
+    },
+    {
+      "epoch": 3.61,
+      "learning_rate": 3.196028313016123e-05,
+      "loss": 0.9629,
+      "step": 183500
+    },
+    {
+      "epoch": 3.62,
+      "learning_rate": 3.1911128588281556e-05,
+      "loss": 0.9462,
+      "step": 184000
+    },
+    {
+      "epoch": 3.63,
+      "learning_rate": 3.186197404640189e-05,
+      "loss": 0.9627,
+      "step": 184500
+    },
+    {
+      "epoch": 3.64,
+      "learning_rate": 3.181281950452222e-05,
+      "loss": 0.9598,
+      "step": 185000
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 3.176366496264255e-05,
+      "loss": 0.9721,
+      "step": 185500
+    },
+    {
+      "epoch": 3.66,
+      "learning_rate": 3.1714510420762875e-05,
+      "loss": 0.9611,
+      "step": 186000
+    },
+    {
+      "epoch": 3.67,
+      "learning_rate": 3.1665355878883216e-05,
+      "loss": 0.9586,
+      "step": 186500
+    },
+    {
+      "epoch": 3.68,
+      "learning_rate": 3.1616201337003544e-05,
+      "loss": 0.9655,
+      "step": 187000
+    },
+    {
+      "epoch": 3.69,
+      "learning_rate": 3.156704679512387e-05,
+      "loss": 0.961,
+      "step": 187500
+    },
+    {
+      "epoch": 3.7,
+      "learning_rate": 3.15178922532442e-05,
+      "loss": 0.9581,
+      "step": 188000
+    },
+    {
+      "epoch": 3.71,
+      "learning_rate": 3.1468737711364534e-05,
+      "loss": 0.9596,
+      "step": 188500
+    },
+    {
+      "epoch": 3.72,
+      "learning_rate": 3.141958316948486e-05,
+      "loss": 0.9639,
+      "step": 189000
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 3.137042862760519e-05,
+      "loss": 0.9501,
+      "step": 189500
+    },
+    {
+      "epoch": 3.74,
+      "learning_rate": 3.1321274085725525e-05,
+      "loss": 0.9481,
+      "step": 190000
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 3.127211954384585e-05,
+      "loss": 0.9452,
+      "step": 190500
+    },
+    {
+      "epoch": 3.76,
+      "learning_rate": 3.122296500196618e-05,
+      "loss": 0.9524,
+      "step": 191000
+    },
+    {
+      "epoch": 3.77,
+      "learning_rate": 3.117381046008651e-05,
+      "loss": 0.9552,
+      "step": 191500
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 3.112465591820685e-05,
+      "loss": 0.9461,
+      "step": 192000
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 3.107550137632718e-05,
+      "loss": 0.9513,
+      "step": 192500
+    },
+    {
+      "epoch": 3.79,
+      "learning_rate": 3.1026346834447505e-05,
+      "loss": 0.9568,
+      "step": 193000
+    },
+    {
+      "epoch": 3.8,
+      "learning_rate": 3.097719229256783e-05,
+      "loss": 0.9367,
+      "step": 193500
+    },
+    {
+      "epoch": 3.81,
+      "learning_rate": 3.092803775068817e-05,
+      "loss": 0.9545,
+      "step": 194000
+    },
+    {
+      "epoch": 3.82,
+      "learning_rate": 3.0878883208808496e-05,
+      "loss": 0.9552,
+      "step": 194500
+    },
+    {
+      "epoch": 3.83,
+      "learning_rate": 3.0829728666928824e-05,
+      "loss": 0.963,
+      "step": 195000
+    },
+    {
+      "epoch": 3.84,
+      "learning_rate": 3.078057412504916e-05,
+      "loss": 0.9411,
+      "step": 195500
+    },
+    {
+      "epoch": 3.85,
+      "learning_rate": 3.0731419583169486e-05,
+      "loss": 0.9481,
+      "step": 196000
+    },
+    {
+      "epoch": 3.86,
+      "learning_rate": 3.0682265041289814e-05,
+      "loss": 0.9341,
+      "step": 196500
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 3.063311049941014e-05,
+      "loss": 0.9566,
+      "step": 197000
+    },
+    {
+      "epoch": 3.88,
+      "learning_rate": 3.0583955957530476e-05,
+      "loss": 0.9554,
+      "step": 197500
+    },
+    {
+      "epoch": 3.89,
+      "learning_rate": 3.053480141565081e-05,
+      "loss": 0.9637,
+      "step": 198000
+    },
+    {
+      "epoch": 3.9,
+      "learning_rate": 3.048564687377114e-05,
+      "loss": 0.9493,
+      "step": 198500
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 3.043649233189147e-05,
+      "loss": 0.9574,
+      "step": 199000
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 3.0387337790011798e-05,
+      "loss": 0.9536,
+      "step": 199500
+    },
+    {
+      "epoch": 3.93,
+      "learning_rate": 3.033818324813213e-05,
+      "loss": 0.9492,
+      "step": 200000
+    },
+    {
+      "epoch": 3.94,
+      "learning_rate": 3.028902870625246e-05,
+      "loss": 0.9514,
+      "step": 200500
+    },
+    {
+      "epoch": 3.95,
+      "learning_rate": 3.023987416437279e-05,
+      "loss": 0.9426,
+      "step": 201000
+    },
+    {
+      "epoch": 3.96,
+      "learning_rate": 3.019071962249312e-05,
+      "loss": 0.9471,
+      "step": 201500
+    },
+    {
+      "epoch": 3.97,
+      "learning_rate": 3.0141565080613447e-05,
+      "loss": 0.9454,
+      "step": 202000
+    },
+    {
+      "epoch": 3.98,
+      "learning_rate": 3.009241053873378e-05,
+      "loss": 0.9401,
+      "step": 202500
+    },
+    {
+      "epoch": 3.99,
+      "learning_rate": 3.004325599685411e-05,
+      "loss": 0.954,
+      "step": 203000
+    },
+    {
+      "epoch": 4.0,
+      "learning_rate": 2.9994101454974445e-05,
+      "loss": 0.9383,
+      "step": 203500
+    },
+    {
+      "epoch": 4.01,
+      "learning_rate": 2.9944946913094772e-05,
+      "loss": 0.9322,
+      "step": 204000
+    },
+    {
+      "epoch": 4.02,
+      "learning_rate": 2.9895792371215104e-05,
+      "loss": 0.9393,
+      "step": 204500
+    },
+    {
+      "epoch": 4.03,
+      "learning_rate": 2.984663782933543e-05,
+      "loss": 0.929,
+      "step": 205000
+    },
+    {
+      "epoch": 4.04,
+      "learning_rate": 2.9797483287455763e-05,
+      "loss": 0.937,
+      "step": 205500
+    },
+    {
+      "epoch": 4.05,
+      "learning_rate": 2.9748328745576094e-05,
+      "loss": 0.9368,
+      "step": 206000
+    },
+    {
+      "epoch": 4.06,
+      "learning_rate": 2.9699174203696422e-05,
+      "loss": 0.926,
+      "step": 206500
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 2.9650019661816753e-05,
+      "loss": 0.923,
+      "step": 207000
+    },
+    {
+      "epoch": 4.08,
+      "learning_rate": 2.960086511993708e-05,
+      "loss": 0.9243,
+      "step": 207500
+    },
+    {
+      "epoch": 4.09,
+      "learning_rate": 2.9551710578057412e-05,
+      "loss": 0.9342,
+      "step": 208000
+    },
+    {
+      "epoch": 4.1,
+      "learning_rate": 2.9502556036177744e-05,
+      "loss": 0.9409,
+      "step": 208500
+    },
+    {
+      "epoch": 4.11,
+      "learning_rate": 2.945340149429807e-05,
+      "loss": 0.9253,
+      "step": 209000
+    },
+    {
+      "epoch": 4.12,
+      "learning_rate": 2.9404246952418406e-05,
+      "loss": 0.9314,
+      "step": 209500
+    },
+    {
+      "epoch": 4.13,
+      "learning_rate": 2.9355092410538737e-05,
+      "loss": 0.9374,
+      "step": 210000
+    },
+    {
+      "epoch": 4.14,
+      "learning_rate": 2.930593786865907e-05,
+      "loss": 0.9401,
+      "step": 210500
+    },
+    {
+      "epoch": 4.15,
+      "learning_rate": 2.9256783326779396e-05,
+      "loss": 0.9256,
+      "step": 211000
+    },
+    {
+      "epoch": 4.16,
+      "learning_rate": 2.9207628784899728e-05,
+      "loss": 0.93,
+      "step": 211500
+    },
+    {
+      "epoch": 4.17,
+      "learning_rate": 2.9158474243020055e-05,
+      "loss": 0.9268,
+      "step": 212000
+    },
+    {
+      "epoch": 4.18,
+      "learning_rate": 2.9109319701140387e-05,
+      "loss": 0.9367,
+      "step": 212500
+    },
+    {
+      "epoch": 4.19,
+      "learning_rate": 2.9060165159260715e-05,
+      "loss": 0.9225,
+      "step": 213000
+    },
+    {
+      "epoch": 4.2,
+      "learning_rate": 2.9011010617381046e-05,
+      "loss": 0.9264,
+      "step": 213500
+    },
+    {
+      "epoch": 4.21,
+      "learning_rate": 2.8961856075501377e-05,
+      "loss": 0.9289,
+      "step": 214000
+    },
+    {
+      "epoch": 4.22,
+      "learning_rate": 2.8912701533621705e-05,
+      "loss": 0.9337,
+      "step": 214500
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 2.886354699174204e-05,
+      "loss": 0.925,
+      "step": 215000
+    },
+    {
+      "epoch": 4.24,
+      "learning_rate": 2.881439244986237e-05,
+      "loss": 0.9501,
+      "step": 215500
+    },
+    {
+      "epoch": 4.25,
+      "learning_rate": 2.8765237907982702e-05,
+      "loss": 0.9232,
+      "step": 216000
+    },
+    {
+      "epoch": 4.26,
+      "learning_rate": 2.871608336610303e-05,
+      "loss": 0.9317,
+      "step": 216500
+    },
+    {
+      "epoch": 4.27,
+      "learning_rate": 2.866692882422336e-05,
+      "loss": 0.9341,
+      "step": 217000
+    },
+    {
+      "epoch": 4.28,
+      "learning_rate": 2.861777428234369e-05,
+      "loss": 0.9314,
+      "step": 217500
+    },
+    {
+      "epoch": 4.29,
+      "learning_rate": 2.856861974046402e-05,
+      "loss": 0.9284,
+      "step": 218000
+    },
+    {
+      "epoch": 4.3,
+      "learning_rate": 2.8519465198584348e-05,
+      "loss": 0.9287,
+      "step": 218500
+    },
+    {
+      "epoch": 4.31,
+      "learning_rate": 2.847031065670468e-05,
+      "loss": 0.9219,
+      "step": 219000
+    },
+    {
+      "epoch": 4.32,
+      "learning_rate": 2.842115611482501e-05,
+      "loss": 0.9336,
+      "step": 219500
+    },
+    {
+      "epoch": 4.33,
+      "learning_rate": 2.837200157294534e-05,
+      "loss": 0.9202,
+      "step": 220000
+    },
+    {
+      "epoch": 4.34,
+      "learning_rate": 2.832284703106567e-05,
+      "loss": 0.9282,
+      "step": 220500
+    },
+    {
+      "epoch": 4.35,
+      "learning_rate": 2.8273692489186004e-05,
+      "loss": 0.9189,
+      "step": 221000
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 2.8224537947306336e-05,
+      "loss": 0.9112,
+      "step": 221500
+    },
+    {
+      "epoch": 4.36,
+      "learning_rate": 2.8175383405426664e-05,
+      "loss": 0.9356,
+      "step": 222000
+    },
+    {
+      "epoch": 4.37,
+      "learning_rate": 2.8126228863546995e-05,
+      "loss": 0.9301,
+      "step": 222500
+    },
+    {
+      "epoch": 4.38,
+      "learning_rate": 2.8077074321667323e-05,
+      "loss": 0.9275,
+      "step": 223000
+    },
+    {
+      "epoch": 4.39,
+      "learning_rate": 2.8027919779787654e-05,
+      "loss": 0.9387,
+      "step": 223500
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 2.7978765237907985e-05,
+      "loss": 0.9155,
+      "step": 224000
+    },
+    {
+      "epoch": 4.41,
+      "learning_rate": 2.7929610696028313e-05,
+      "loss": 0.9254,
+      "step": 224500
+    },
+    {
+      "epoch": 4.42,
+      "learning_rate": 2.7880456154148644e-05,
+      "loss": 0.9316,
+      "step": 225000
+    },
+    {
+      "epoch": 4.43,
+      "learning_rate": 2.7831301612268972e-05,
+      "loss": 0.9216,
+      "step": 225500
+    },
+    {
+      "epoch": 4.44,
+      "learning_rate": 2.7782147070389303e-05,
+      "loss": 0.9174,
+      "step": 226000
+    },
+    {
+      "epoch": 4.45,
+      "learning_rate": 2.7732992528509638e-05,
+      "loss": 0.9299,
+      "step": 226500
+    },
+    {
+      "epoch": 4.46,
+      "learning_rate": 2.768383798662997e-05,
+      "loss": 0.9068,
+      "step": 227000
+    },
+    {
+      "epoch": 4.47,
+      "learning_rate": 2.7634683444750297e-05,
+      "loss": 0.9211,
+      "step": 227500
+    },
+    {
+      "epoch": 4.48,
+      "learning_rate": 2.758552890287063e-05,
+      "loss": 0.9258,
+      "step": 228000
+    },
+    {
+      "epoch": 4.49,
+      "learning_rate": 2.7536374360990956e-05,
+      "loss": 0.9127,
+      "step": 228500
+    },
+    {
+      "epoch": 4.5,
+      "learning_rate": 2.7487219819111287e-05,
+      "loss": 0.9221,
+      "step": 229000
+    },
+    {
+      "epoch": 4.51,
+      "learning_rate": 2.743806527723162e-05,
+      "loss": 0.9136,
+      "step": 229500
+    },
+    {
+      "epoch": 4.52,
+      "learning_rate": 2.7388910735351947e-05,
+      "loss": 0.9226,
+      "step": 230000
+    },
+    {
+      "epoch": 4.53,
+      "learning_rate": 2.7339756193472278e-05,
+      "loss": 0.9116,
+      "step": 230500
+    },
+    {
+      "epoch": 4.54,
+      "learning_rate": 2.7290601651592606e-05,
+      "loss": 0.9146,
+      "step": 231000
+    },
+    {
+      "epoch": 4.55,
+      "learning_rate": 2.7241447109712937e-05,
+      "loss": 0.919,
+      "step": 231500
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 2.719229256783327e-05,
+      "loss": 0.9204,
+      "step": 232000
+    },
+    {
+      "epoch": 4.57,
+      "learning_rate": 2.7143138025953603e-05,
+      "loss": 0.9054,
+      "step": 232500
+    },
+    {
+      "epoch": 4.58,
+      "learning_rate": 2.709398348407393e-05,
+      "loss": 0.9276,
+      "step": 233000
+    },
+    {
+      "epoch": 4.59,
+      "learning_rate": 2.7044828942194262e-05,
+      "loss": 0.9172,
+      "step": 233500
+    },
+    {
+      "epoch": 4.6,
+      "learning_rate": 2.699567440031459e-05,
+      "loss": 0.9182,
+      "step": 234000
+    },
+    {
+      "epoch": 4.61,
+      "learning_rate": 2.694651985843492e-05,
+      "loss": 0.9263,
+      "step": 234500
+    },
+    {
+      "epoch": 4.62,
+      "learning_rate": 2.6897365316555252e-05,
+      "loss": 0.9243,
+      "step": 235000
+    },
+    {
+      "epoch": 4.63,
+      "learning_rate": 2.684821077467558e-05,
+      "loss": 0.9201,
+      "step": 235500
+    },
+    {
+      "epoch": 4.64,
+      "learning_rate": 2.679905623279591e-05,
+      "loss": 0.9056,
+      "step": 236000
+    },
+    {
+      "epoch": 4.65,
+      "learning_rate": 2.674990169091624e-05,
+      "loss": 0.9063,
+      "step": 236500
+    },
+    {
+      "epoch": 4.66,
+      "learning_rate": 2.670074714903657e-05,
+      "loss": 0.9132,
+      "step": 237000
+    },
+    {
+      "epoch": 4.67,
+      "learning_rate": 2.66515926071569e-05,
+      "loss": 0.9093,
+      "step": 237500
+    },
+    {
+      "epoch": 4.68,
+      "learning_rate": 2.6602438065277236e-05,
+      "loss": 0.9244,
+      "step": 238000
+    },
+    {
+      "epoch": 4.69,
+      "learning_rate": 2.6553283523397564e-05,
+      "loss": 0.9193,
+      "step": 238500
+    },
+    {
+      "epoch": 4.7,
+      "learning_rate": 2.6504128981517895e-05,
+      "loss": 0.9145,
+      "step": 239000
+    },
+    {
+      "epoch": 4.71,
+      "learning_rate": 2.6454974439638223e-05,
+      "loss": 0.9126,
+      "step": 239500
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 2.6405819897758555e-05,
+      "loss": 0.9158,
+      "step": 240000
+    },
+    {
+      "epoch": 4.73,
+      "learning_rate": 2.6356665355878886e-05,
+      "loss": 0.9091,
+      "step": 240500
+    },
+    {
+      "epoch": 4.74,
+      "learning_rate": 2.6307510813999214e-05,
+      "loss": 0.9016,
+      "step": 241000
+    },
+    {
+      "epoch": 4.75,
+      "learning_rate": 2.6258356272119545e-05,
+      "loss": 0.9103,
+      "step": 241500
+    },
+    {
+      "epoch": 4.76,
+      "learning_rate": 2.6209201730239873e-05,
+      "loss": 0.9037,
+      "step": 242000
+    },
+    {
+      "epoch": 4.77,
+      "learning_rate": 2.6160047188360204e-05,
+      "loss": 0.9009,
+      "step": 242500
+    },
+    {
+      "epoch": 4.78,
+      "learning_rate": 2.6110892646480535e-05,
+      "loss": 0.9105,
+      "step": 243000
+    },
+    {
+      "epoch": 4.79,
+      "learning_rate": 2.606173810460087e-05,
+      "loss": 0.8999,
+      "step": 243500
+    },
+    {
+      "epoch": 4.8,
+      "learning_rate": 2.6012583562721198e-05,
+      "loss": 0.8918,
+      "step": 244000
+    },
+    {
+      "epoch": 4.81,
+      "learning_rate": 2.596342902084153e-05,
+      "loss": 0.912,
+      "step": 244500
+    },
+    {
+      "epoch": 4.82,
+      "learning_rate": 2.5914274478961857e-05,
+      "loss": 0.9022,
+      "step": 245000
+    },
+    {
+      "epoch": 4.83,
+      "learning_rate": 2.5865119937082188e-05,
+      "loss": 0.9088,
+      "step": 245500
+    },
+    {
+      "epoch": 4.84,
+      "learning_rate": 2.581596539520252e-05,
+      "loss": 0.9229,
+      "step": 246000
+    },
+    {
+      "epoch": 4.85,
+      "learning_rate": 2.5766810853322847e-05,
+      "loss": 0.9173,
+      "step": 246500
+    },
+    {
+      "epoch": 4.86,
+      "learning_rate": 2.571765631144318e-05,
+      "loss": 0.9114,
+      "step": 247000
+    },
+    {
+      "epoch": 4.87,
+      "learning_rate": 2.5668501769563506e-05,
+      "loss": 0.919,
+      "step": 247500
+    },
+    {
+      "epoch": 4.88,
+      "learning_rate": 2.5619347227683838e-05,
+      "loss": 0.9007,
+      "step": 248000
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 2.557019268580417e-05,
+      "loss": 0.9098,
+      "step": 248500
+    },
+    {
+      "epoch": 4.9,
+      "learning_rate": 2.5521038143924497e-05,
+      "loss": 0.8994,
+      "step": 249000
+    },
+    {
+      "epoch": 4.91,
+      "learning_rate": 2.547188360204483e-05,
+      "loss": 0.9034,
+      "step": 249500
+    },
+    {
+      "epoch": 4.92,
+      "learning_rate": 2.5422729060165163e-05,
+      "loss": 0.9096,
+      "step": 250000
+    },
+    {
+      "epoch": 4.93,
+      "learning_rate": 2.5373574518285494e-05,
+      "loss": 0.9001,
+      "step": 250500
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 2.5324419976405822e-05,
+      "loss": 0.8986,
+      "step": 251000
+    },
+    {
+      "epoch": 4.94,
+      "learning_rate": 2.5275265434526153e-05,
+      "loss": 0.9168,
+      "step": 251500
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 2.522611089264648e-05,
+      "loss": 0.9086,
+      "step": 252000
+    },
+    {
+      "epoch": 4.96,
+      "learning_rate": 2.5176956350766812e-05,
+      "loss": 0.9067,
+      "step": 252500
+    },
+    {
+      "epoch": 4.97,
+      "learning_rate": 2.512780180888714e-05,
+      "loss": 0.9047,
+      "step": 253000
+    },
+    {
+      "epoch": 4.98,
+      "learning_rate": 2.507864726700747e-05,
+      "loss": 0.8987,
+      "step": 253500
+    },
+    {
+      "epoch": 4.99,
+      "learning_rate": 2.5029492725127802e-05,
+      "loss": 0.897,
+      "step": 254000
+    },
+    {
+      "epoch": 5.0,
+      "learning_rate": 2.4980338183248134e-05,
+      "loss": 0.8901,
+      "step": 254500
+    },
+    {
+      "epoch": 5.01,
+      "learning_rate": 2.4931183641368465e-05,
+      "loss": 0.9015,
+      "step": 255000
+    },
+    {
+      "epoch": 5.02,
+      "learning_rate": 2.4882029099488793e-05,
+      "loss": 0.9002,
+      "step": 255500
+    },
+    {
+      "epoch": 5.03,
+      "learning_rate": 2.4832874557609124e-05,
+      "loss": 0.9065,
+      "step": 256000
+    },
+    {
+      "epoch": 5.04,
+      "learning_rate": 2.4783720015729455e-05,
+      "loss": 0.9017,
+      "step": 256500
+    },
+    {
+      "epoch": 5.05,
+      "learning_rate": 2.4734565473849787e-05,
+      "loss": 0.8924,
+      "step": 257000
+    },
+    {
+      "epoch": 5.06,
+      "learning_rate": 2.4685410931970114e-05,
+      "loss": 0.8953,
+      "step": 257500
+    },
+    {
+      "epoch": 5.07,
+      "learning_rate": 2.4636256390090446e-05,
+      "loss": 0.8832,
+      "step": 258000
+    },
+    {
+      "epoch": 5.08,
+      "learning_rate": 2.4587101848210774e-05,
+      "loss": 0.8899,
+      "step": 258500
+    },
+    {
+      "epoch": 5.09,
+      "learning_rate": 2.4537947306331105e-05,
+      "loss": 0.9004,
+      "step": 259000
+    },
+    {
+      "epoch": 5.1,
+      "learning_rate": 2.4488792764451436e-05,
+      "loss": 0.8799,
+      "step": 259500
+    },
+    {
+      "epoch": 5.11,
+      "learning_rate": 2.4439638222571767e-05,
+      "loss": 0.8888,
+      "step": 260000
+    },
+    {
+      "epoch": 5.12,
+      "learning_rate": 2.43904836806921e-05,
+      "loss": 0.8905,
+      "step": 260500
+    },
+    {
+      "epoch": 5.13,
+      "learning_rate": 2.4341329138812426e-05,
+      "loss": 0.8936,
+      "step": 261000
+    },
+    {
+      "epoch": 5.14,
+      "learning_rate": 2.4292174596932758e-05,
+      "loss": 0.8914,
+      "step": 261500
+    },
+    {
+      "epoch": 5.15,
+      "learning_rate": 2.4243020055053085e-05,
+      "loss": 0.8862,
+      "step": 262000
+    },
+    {
+      "epoch": 5.16,
+      "learning_rate": 2.419386551317342e-05,
+      "loss": 0.8883,
+      "step": 262500
+    },
+    {
+      "epoch": 5.17,
+      "learning_rate": 2.4144710971293748e-05,
+      "loss": 0.8886,
+      "step": 263000
+    },
+    {
+      "epoch": 5.18,
+      "learning_rate": 2.409555642941408e-05,
+      "loss": 0.8875,
+      "step": 263500
+    },
+    {
+      "epoch": 5.19,
+      "learning_rate": 2.404640188753441e-05,
+      "loss": 0.8873,
+      "step": 264000
+    },
+    {
+      "epoch": 5.2,
+      "learning_rate": 2.399724734565474e-05,
+      "loss": 0.8994,
+      "step": 264500
+    },
+    {
+      "epoch": 5.21,
+      "learning_rate": 2.394809280377507e-05,
+      "loss": 0.8865,
+      "step": 265000
+    },
+    {
+      "epoch": 5.22,
+      "learning_rate": 2.38989382618954e-05,
+      "loss": 0.8915,
+      "step": 265500
+    },
+    {
+      "epoch": 5.23,
+      "learning_rate": 2.3849783720015732e-05,
+      "loss": 0.8944,
+      "step": 266000
+    },
+    {
+      "epoch": 5.24,
+      "learning_rate": 2.380062917813606e-05,
+      "loss": 0.8924,
+      "step": 266500
+    },
+    {
+      "epoch": 5.25,
+      "learning_rate": 2.375147463625639e-05,
+      "loss": 0.8834,
+      "step": 267000
+    },
+    {
+      "epoch": 5.26,
+      "learning_rate": 2.370232009437672e-05,
+      "loss": 0.8867,
+      "step": 267500
+    },
+    {
+      "epoch": 5.27,
+      "learning_rate": 2.3653165552497054e-05,
+      "loss": 0.8953,
+      "step": 268000
+    },
+    {
+      "epoch": 5.28,
+      "learning_rate": 2.360401101061738e-05,
+      "loss": 0.8916,
+      "step": 268500
+    },
+    {
+      "epoch": 5.29,
+      "learning_rate": 2.3554856468737713e-05,
+      "loss": 0.8916,
+      "step": 269000
+    },
+    {
+      "epoch": 5.3,
+      "learning_rate": 2.3505701926858044e-05,
+      "loss": 0.894,
+      "step": 269500
+    },
+    {
+      "epoch": 5.31,
+      "learning_rate": 2.3456547384978372e-05,
+      "loss": 0.8907,
+      "step": 270000
+    },
+    {
+      "epoch": 5.32,
+      "learning_rate": 2.3407392843098703e-05,
+      "loss": 0.8874,
+      "step": 270500
+    },
+    {
+      "epoch": 5.33,
+      "learning_rate": 2.3358238301219034e-05,
+      "loss": 0.88,
+      "step": 271000
+    },
+    {
+      "epoch": 5.34,
+      "learning_rate": 2.3309083759339366e-05,
+      "loss": 0.8946,
+      "step": 271500
+    },
+    {
+      "epoch": 5.35,
+      "learning_rate": 2.3259929217459694e-05,
+      "loss": 0.9026,
+      "step": 272000
+    },
+    {
+      "epoch": 5.36,
+      "learning_rate": 2.3210774675580025e-05,
+      "loss": 0.8955,
+      "step": 272500
+    },
+    {
+      "epoch": 5.37,
+      "learning_rate": 2.3161620133700353e-05,
+      "loss": 0.8811,
+      "step": 273000
+    },
+    {
+      "epoch": 5.38,
+      "learning_rate": 2.3112465591820687e-05,
+      "loss": 0.8919,
+      "step": 273500
+    },
+    {
+      "epoch": 5.39,
+      "learning_rate": 2.3063311049941015e-05,
+      "loss": 0.8863,
+      "step": 274000
+    },
+    {
+      "epoch": 5.4,
+      "learning_rate": 2.3014156508061346e-05,
+      "loss": 0.8876,
+      "step": 274500
+    },
+    {
+      "epoch": 5.41,
+      "learning_rate": 2.2965001966181678e-05,
+      "loss": 0.8878,
+      "step": 275000
+    },
+    {
+      "epoch": 5.42,
+      "learning_rate": 2.2915847424302005e-05,
+      "loss": 0.8874,
+      "step": 275500
+    },
+    {
+      "epoch": 5.43,
+      "learning_rate": 2.2866692882422337e-05,
+      "loss": 0.8857,
+      "step": 276000
+    },
+    {
+      "epoch": 5.44,
+      "learning_rate": 2.2817538340542668e-05,
+      "loss": 0.895,
+      "step": 276500
+    },
+    {
+      "epoch": 5.45,
+      "learning_rate": 2.2768383798663e-05,
+      "loss": 0.8881,
+      "step": 277000
+    },
+    {
+      "epoch": 5.46,
+      "learning_rate": 2.2719229256783327e-05,
+      "loss": 0.8941,
+      "step": 277500
+    },
+    {
+      "epoch": 5.47,
+      "learning_rate": 2.267007471490366e-05,
+      "loss": 0.8857,
+      "step": 278000
+    },
+    {
+      "epoch": 5.48,
+      "learning_rate": 2.2620920173023986e-05,
+      "loss": 0.8874,
+      "step": 278500
+    },
+    {
+      "epoch": 5.49,
+      "learning_rate": 2.2571765631144317e-05,
+      "loss": 0.8857,
+      "step": 279000
+    },
+    {
+      "epoch": 5.5,
+      "learning_rate": 2.252261108926465e-05,
+      "loss": 0.8805,
+      "step": 279500
+    },
+    {
+      "epoch": 5.51,
+      "learning_rate": 2.247345654738498e-05,
+      "loss": 0.8782,
+      "step": 280000
+    },
+    {
+      "epoch": 5.52,
+      "learning_rate": 2.242430200550531e-05,
+      "loss": 0.8921,
+      "step": 280500
+    },
+    {
+      "epoch": 5.52,
+      "learning_rate": 2.237514746362564e-05,
+      "loss": 0.8851,
+      "step": 281000
+    },
+    {
+      "epoch": 5.53,
+      "learning_rate": 2.232599292174597e-05,
+      "loss": 0.8806,
+      "step": 281500
+    },
+    {
+      "epoch": 5.54,
+      "learning_rate": 2.2276838379866298e-05,
+      "loss": 0.882,
+      "step": 282000
+    },
+    {
+      "epoch": 5.55,
+      "learning_rate": 2.2227683837986633e-05,
+      "loss": 0.888,
+      "step": 282500
+    },
+    {
+      "epoch": 5.56,
+      "learning_rate": 2.217852929610696e-05,
+      "loss": 0.8936,
+      "step": 283000
+    },
+    {
+      "epoch": 5.57,
+      "learning_rate": 2.2129374754227292e-05,
+      "loss": 0.8964,
+      "step": 283500
+    },
+    {
+      "epoch": 5.58,
+      "learning_rate": 2.2080220212347623e-05,
+      "loss": 0.8976,
+      "step": 284000
+    },
+    {
+      "epoch": 5.59,
+      "learning_rate": 2.203106567046795e-05,
+      "loss": 0.8836,
+      "step": 284500
+    },
+    {
+      "epoch": 5.6,
+      "learning_rate": 2.1981911128588282e-05,
+      "loss": 0.8904,
+      "step": 285000
+    },
+    {
+      "epoch": 5.61,
+      "learning_rate": 2.1932756586708614e-05,
+      "loss": 0.8879,
+      "step": 285500
+    },
+    {
+      "epoch": 5.62,
+      "learning_rate": 2.1883602044828945e-05,
+      "loss": 0.8894,
+      "step": 286000
+    },
+    {
+      "epoch": 5.63,
+      "learning_rate": 2.1834447502949273e-05,
+      "loss": 0.8799,
+      "step": 286500
+    },
+    {
+      "epoch": 5.64,
+      "learning_rate": 2.1785292961069604e-05,
+      "loss": 0.8779,
+      "step": 287000
+    },
+    {
+      "epoch": 5.65,
+      "learning_rate": 2.1736138419189932e-05,
+      "loss": 0.8816,
+      "step": 287500
+    },
+    {
+      "epoch": 5.66,
+      "learning_rate": 2.1686983877310266e-05,
+      "loss": 0.8815,
+      "step": 288000
+    },
+    {
+      "epoch": 5.67,
+      "learning_rate": 2.1637829335430594e-05,
+      "loss": 0.8853,
+      "step": 288500
+    },
+    {
+      "epoch": 5.68,
+      "learning_rate": 2.1588674793550925e-05,
+      "loss": 0.886,
+      "step": 289000
+    },
+    {
+      "epoch": 5.69,
+      "learning_rate": 2.1539520251671257e-05,
+      "loss": 0.8881,
+      "step": 289500
+    },
+    {
+      "epoch": 5.7,
+      "learning_rate": 2.1490365709791585e-05,
+      "loss": 0.8789,
+      "step": 290000
+    },
+    {
+      "epoch": 5.71,
+      "learning_rate": 2.1441211167911916e-05,
+      "loss": 0.8827,
+      "step": 290500
+    },
+    {
+      "epoch": 5.72,
+      "learning_rate": 2.1392056626032247e-05,
+      "loss": 0.8844,
+      "step": 291000
+    },
+    {
+      "epoch": 5.73,
+      "learning_rate": 2.134290208415258e-05,
+      "loss": 0.8807,
+      "step": 291500
+    },
+    {
+      "epoch": 5.74,
+      "learning_rate": 2.1293747542272906e-05,
+      "loss": 0.8789,
+      "step": 292000
+    },
+    {
+      "epoch": 5.75,
+      "learning_rate": 2.1244593000393237e-05,
+      "loss": 0.8808,
+      "step": 292500
+    },
+    {
+      "epoch": 5.76,
+      "learning_rate": 2.1195438458513565e-05,
+      "loss": 0.8763,
+      "step": 293000
+    },
+    {
+      "epoch": 5.77,
+      "learning_rate": 2.1146283916633897e-05,
+      "loss": 0.8827,
+      "step": 293500
+    },
+    {
+      "epoch": 5.78,
+      "learning_rate": 2.1097129374754228e-05,
+      "loss": 0.8749,
+      "step": 294000
+    },
+    {
+      "epoch": 5.79,
+      "learning_rate": 2.104797483287456e-05,
+      "loss": 0.8746,
+      "step": 294500
+    },
+    {
+      "epoch": 5.8,
+      "learning_rate": 2.099882029099489e-05,
+      "loss": 0.8833,
+      "step": 295000
+    },
+    {
+      "epoch": 5.81,
+      "learning_rate": 2.0949665749115218e-05,
+      "loss": 0.8855,
+      "step": 295500
+    },
+    {
+      "epoch": 5.82,
+      "learning_rate": 2.090051120723555e-05,
+      "loss": 0.8844,
+      "step": 296000
+    },
+    {
+      "epoch": 5.83,
+      "learning_rate": 2.085135666535588e-05,
+      "loss": 0.8829,
+      "step": 296500
+    },
+    {
+      "epoch": 5.84,
+      "learning_rate": 2.0802202123476212e-05,
+      "loss": 0.8842,
+      "step": 297000
+    },
+    {
+      "epoch": 5.85,
+      "learning_rate": 2.075304758159654e-05,
+      "loss": 0.8749,
+      "step": 297500
+    },
+    {
+      "epoch": 5.86,
+      "learning_rate": 2.070389303971687e-05,
+      "loss": 0.8741,
+      "step": 298000
+    },
+    {
+      "epoch": 5.87,
+      "learning_rate": 2.06547384978372e-05,
+      "loss": 0.8771,
+      "step": 298500
+    },
+    {
+      "epoch": 5.88,
+      "learning_rate": 2.060558395595753e-05,
+      "loss": 0.8662,
+      "step": 299000
+    },
+    {
+      "epoch": 5.89,
+      "learning_rate": 2.055642941407786e-05,
+      "loss": 0.8765,
+      "step": 299500
+    },
+    {
+      "epoch": 5.9,
+      "learning_rate": 2.0507274872198193e-05,
+      "loss": 0.8746,
+      "step": 300000
+    },
+    {
+      "epoch": 5.91,
+      "learning_rate": 2.0458120330318524e-05,
+      "loss": 0.8781,
+      "step": 300500
+    },
+    {
+      "epoch": 5.92,
+      "learning_rate": 2.0408965788438852e-05,
+      "loss": 0.8693,
+      "step": 301000
+    },
+    {
+      "epoch": 5.93,
+      "learning_rate": 2.0359811246559183e-05,
+      "loss": 0.8664,
+      "step": 301500
+    },
+    {
+      "epoch": 5.94,
+      "learning_rate": 2.031065670467951e-05,
+      "loss": 0.88,
+      "step": 302000
+    },
+    {
+      "epoch": 5.95,
+      "learning_rate": 2.0261502162799845e-05,
+      "loss": 0.8883,
+      "step": 302500
+    },
+    {
+      "epoch": 5.96,
+      "learning_rate": 2.0212347620920173e-05,
+      "loss": 0.8749,
+      "step": 303000
+    },
+    {
+      "epoch": 5.97,
+      "learning_rate": 2.0163193079040505e-05,
+      "loss": 0.8776,
+      "step": 303500
+    },
+    {
+      "epoch": 5.98,
+      "learning_rate": 2.0114038537160836e-05,
+      "loss": 0.8703,
+      "step": 304000
+    },
+    {
+      "epoch": 5.99,
+      "learning_rate": 2.0064883995281164e-05,
+      "loss": 0.8707,
+      "step": 304500
+    },
+    {
+      "epoch": 6.0,
+      "learning_rate": 2.0015729453401495e-05,
+      "loss": 0.8787,
+      "step": 305000
+    },
+    {
+      "epoch": 6.01,
+      "learning_rate": 1.9966574911521826e-05,
+      "loss": 0.851,
+      "step": 305500
+    },
+    {
+      "epoch": 6.02,
+      "learning_rate": 1.9917420369642157e-05,
+      "loss": 0.8621,
+      "step": 306000
+    },
+    {
+      "epoch": 6.03,
+      "learning_rate": 1.9868265827762485e-05,
+      "loss": 0.8738,
+      "step": 306500
+    },
+    {
+      "epoch": 6.04,
+      "learning_rate": 1.9819111285882817e-05,
+      "loss": 0.8811,
+      "step": 307000
+    },
+    {
+      "epoch": 6.05,
+      "learning_rate": 1.9769956744003144e-05,
+      "loss": 0.8708,
+      "step": 307500
+    },
+    {
+      "epoch": 6.06,
+      "learning_rate": 1.972080220212348e-05,
+      "loss": 0.8619,
+      "step": 308000
+    },
+    {
+      "epoch": 6.07,
+      "learning_rate": 1.9671647660243807e-05,
+      "loss": 0.8724,
+      "step": 308500
+    },
+    {
+      "epoch": 6.08,
+      "learning_rate": 1.9622493118364138e-05,
+      "loss": 0.8707,
+      "step": 309000
+    },
+    {
+      "epoch": 6.09,
+      "learning_rate": 1.957333857648447e-05,
+      "loss": 0.869,
+      "step": 309500
+    },
+    {
+      "epoch": 6.1,
+      "learning_rate": 1.9524184034604797e-05,
+      "loss": 0.868,
+      "step": 310000
+    },
+    {
+      "epoch": 6.1,
+      "learning_rate": 1.947502949272513e-05,
+      "loss": 0.8664,
+      "step": 310500
+    },
+    {
+      "epoch": 6.11,
+      "learning_rate": 1.942587495084546e-05,
+      "loss": 0.8612,
+      "step": 311000
+    },
+    {
+      "epoch": 6.12,
+      "learning_rate": 1.937672040896579e-05,
+      "loss": 0.8683,
+      "step": 311500
+    },
+    {
+      "epoch": 6.13,
+      "learning_rate": 1.932756586708612e-05,
+      "loss": 0.8689,
+      "step": 312000
+    },
+    {
+      "epoch": 6.14,
+      "learning_rate": 1.927841132520645e-05,
+      "loss": 0.8764,
+      "step": 312500
+    },
+    {
+      "epoch": 6.15,
+      "learning_rate": 1.9229256783326778e-05,
+      "loss": 0.8652,
+      "step": 313000
+    },
+    {
+      "epoch": 6.16,
+      "learning_rate": 1.918010224144711e-05,
+      "loss": 0.8635,
+      "step": 313500
+    },
+    {
+      "epoch": 6.17,
+      "learning_rate": 1.913094769956744e-05,
+      "loss": 0.8757,
+      "step": 314000
+    },
+    {
+      "epoch": 6.18,
+      "learning_rate": 1.9081793157687772e-05,
+      "loss": 0.8671,
+      "step": 314500
+    },
+    {
+      "epoch": 6.19,
+      "learning_rate": 1.9032638615808103e-05,
+      "loss": 0.8652,
+      "step": 315000
+    },
+    {
+      "epoch": 6.2,
+      "learning_rate": 1.898348407392843e-05,
+      "loss": 0.8696,
+      "step": 315500
+    },
+    {
+      "epoch": 6.21,
+      "learning_rate": 1.8934329532048762e-05,
+      "loss": 0.8532,
+      "step": 316000
+    },
+    {
+      "epoch": 6.22,
+      "learning_rate": 1.8885174990169093e-05,
+      "loss": 0.8579,
+      "step": 316500
+    },
+    {
+      "epoch": 6.23,
+      "learning_rate": 1.8836020448289425e-05,
+      "loss": 0.864,
+      "step": 317000
+    },
+    {
+      "epoch": 6.24,
+      "learning_rate": 1.8786865906409752e-05,
+      "loss": 0.8615,
+      "step": 317500
+    },
+    {
+      "epoch": 6.25,
+      "learning_rate": 1.8737711364530084e-05,
+      "loss": 0.8663,
+      "step": 318000
+    },
+    {
+      "epoch": 6.26,
+      "learning_rate": 1.868855682265041e-05,
+      "loss": 0.859,
+      "step": 318500
+    },
+    {
+      "epoch": 6.27,
+      "learning_rate": 1.8639402280770743e-05,
+      "loss": 0.8673,
+      "step": 319000
+    },
+    {
+      "epoch": 6.28,
+      "learning_rate": 1.8590247738891074e-05,
+      "loss": 0.8673,
+      "step": 319500
+    },
+    {
+      "epoch": 6.29,
+      "learning_rate": 1.8541093197011405e-05,
+      "loss": 0.8693,
+      "step": 320000
+    },
+    {
+      "epoch": 6.3,
+      "learning_rate": 1.8491938655131737e-05,
+      "loss": 0.867,
+      "step": 320500
+    },
+    {
+      "epoch": 6.31,
+      "learning_rate": 1.8442784113252064e-05,
+      "loss": 0.8635,
+      "step": 321000
+    },
+    {
+      "epoch": 6.32,
+      "learning_rate": 1.8393629571372396e-05,
+      "loss": 0.8611,
+      "step": 321500
+    },
+    {
+      "epoch": 6.33,
+      "learning_rate": 1.8344475029492724e-05,
+      "loss": 0.8657,
+      "step": 322000
+    },
+    {
+      "epoch": 6.34,
+      "learning_rate": 1.8295320487613058e-05,
+      "loss": 0.8595,
+      "step": 322500
+    },
+    {
+      "epoch": 6.35,
+      "learning_rate": 1.8246165945733386e-05,
+      "loss": 0.8592,
+      "step": 323000
+    },
+    {
+      "epoch": 6.36,
+      "learning_rate": 1.8197011403853717e-05,
+      "loss": 0.8648,
+      "step": 323500
+    },
+    {
+      "epoch": 6.37,
+      "learning_rate": 1.814785686197405e-05,
+      "loss": 0.8611,
+      "step": 324000
+    },
+    {
+      "epoch": 6.38,
+      "learning_rate": 1.8098702320094376e-05,
+      "loss": 0.869,
+      "step": 324500
+    },
+    {
+      "epoch": 6.39,
+      "learning_rate": 1.8049547778214708e-05,
+      "loss": 0.8639,
+      "step": 325000
+    },
+    {
+      "epoch": 6.4,
+      "learning_rate": 1.800039323633504e-05,
+      "loss": 0.8619,
+      "step": 325500
+    },
+    {
+      "epoch": 6.41,
+      "learning_rate": 1.795123869445537e-05,
+      "loss": 0.8499,
+      "step": 326000
+    },
+    {
+      "epoch": 6.42,
+      "learning_rate": 1.7902084152575698e-05,
+      "loss": 0.8649,
+      "step": 326500
+    },
+    {
+      "epoch": 6.43,
+      "learning_rate": 1.785292961069603e-05,
+      "loss": 0.8617,
+      "step": 327000
+    },
+    {
+      "epoch": 6.44,
+      "learning_rate": 1.7803775068816357e-05,
+      "loss": 0.8692,
+      "step": 327500
+    },
+    {
+      "epoch": 6.45,
+      "learning_rate": 1.7754620526936692e-05,
+      "loss": 0.8529,
+      "step": 328000
+    },
+    {
+      "epoch": 6.46,
+      "learning_rate": 1.770546598505702e-05,
+      "loss": 0.8672,
+      "step": 328500
+    },
+    {
+      "epoch": 6.47,
+      "learning_rate": 1.765631144317735e-05,
+      "loss": 0.8536,
+      "step": 329000
+    },
+    {
+      "epoch": 6.48,
+      "learning_rate": 1.7607156901297682e-05,
+      "loss": 0.8471,
+      "step": 329500
+    },
+    {
+      "epoch": 6.49,
+      "learning_rate": 1.755800235941801e-05,
+      "loss": 0.8562,
+      "step": 330000
+    },
+    {
+      "epoch": 6.5,
+      "learning_rate": 1.750884781753834e-05,
+      "loss": 0.856,
+      "step": 330500
+    },
+    {
+      "epoch": 6.51,
+      "learning_rate": 1.7459693275658672e-05,
+      "loss": 0.87,
+      "step": 331000
+    },
+    {
+      "epoch": 6.52,
+      "learning_rate": 1.7410538733779004e-05,
+      "loss": 0.8662,
+      "step": 331500
+    },
+    {
+      "epoch": 6.53,
+      "learning_rate": 1.736138419189933e-05,
+      "loss": 0.867,
+      "step": 332000
+    },
+    {
+      "epoch": 6.54,
+      "learning_rate": 1.7312229650019663e-05,
+      "loss": 0.8535,
+      "step": 332500
+    },
+    {
+      "epoch": 6.55,
+      "learning_rate": 1.726307510813999e-05,
+      "loss": 0.8589,
+      "step": 333000
+    },
+    {
+      "epoch": 6.56,
+      "learning_rate": 1.7213920566260322e-05,
+      "loss": 0.8549,
+      "step": 333500
+    },
+    {
+      "epoch": 6.57,
+      "learning_rate": 1.7164766024380653e-05,
+      "loss": 0.8739,
+      "step": 334000
+    },
+    {
+      "epoch": 6.58,
+      "learning_rate": 1.7115611482500984e-05,
+      "loss": 0.849,
+      "step": 334500
+    },
+    {
+      "epoch": 6.59,
+      "learning_rate": 1.7066456940621316e-05,
+      "loss": 0.8613,
+      "step": 335000
+    },
+    {
+      "epoch": 6.6,
+      "learning_rate": 1.7017302398741644e-05,
+      "loss": 0.862,
+      "step": 335500
+    },
+    {
+      "epoch": 6.61,
+      "learning_rate": 1.6968147856861975e-05,
+      "loss": 0.8658,
+      "step": 336000
+    },
+    {
+      "epoch": 6.62,
+      "learning_rate": 1.6918993314982306e-05,
+      "loss": 0.8595,
+      "step": 336500
+    },
+    {
+      "epoch": 6.63,
+      "learning_rate": 1.6869838773102637e-05,
+      "loss": 0.8511,
+      "step": 337000
+    },
+    {
+      "epoch": 6.64,
+      "learning_rate": 1.6820684231222965e-05,
+      "loss": 0.8568,
+      "step": 337500
+    },
+    {
+      "epoch": 6.65,
+      "learning_rate": 1.6771529689343296e-05,
+      "loss": 0.8589,
+      "step": 338000
+    },
+    {
+      "epoch": 6.66,
+      "learning_rate": 1.6722375147463624e-05,
+      "loss": 0.859,
+      "step": 338500
+    },
+    {
+      "epoch": 6.67,
+      "learning_rate": 1.6673220605583956e-05,
+      "loss": 0.8545,
+      "step": 339000
+    },
+    {
+      "epoch": 6.68,
+      "learning_rate": 1.6624066063704287e-05,
+      "loss": 0.8615,
+      "step": 339500
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 1.6574911521824618e-05,
+      "loss": 0.8763,
+      "step": 340000
+    },
+    {
+      "epoch": 6.69,
+      "learning_rate": 1.652575697994495e-05,
+      "loss": 0.856,
+      "step": 340500
+    },
+    {
+      "epoch": 6.7,
+      "learning_rate": 1.6476602438065277e-05,
+      "loss": 0.8626,
+      "step": 341000
+    },
+    {
+      "epoch": 6.71,
+      "learning_rate": 1.642744789618561e-05,
+      "loss": 0.8596,
+      "step": 341500
+    },
+    {
+      "epoch": 6.72,
+      "learning_rate": 1.6378293354305936e-05,
+      "loss": 0.8643,
+      "step": 342000
+    },
+    {
+      "epoch": 6.73,
+      "learning_rate": 1.632913881242627e-05,
+      "loss": 0.8606,
+      "step": 342500
+    },
+    {
+      "epoch": 6.74,
+      "learning_rate": 1.62799842705466e-05,
+      "loss": 0.8622,
+      "step": 343000
+    },
+    {
+      "epoch": 6.75,
+      "learning_rate": 1.623082972866693e-05,
+      "loss": 0.8629,
+      "step": 343500
+    },
+    {
+      "epoch": 6.76,
+      "learning_rate": 1.618167518678726e-05,
+      "loss": 0.8544,
+      "step": 344000
+    },
+    {
+      "epoch": 6.77,
+      "learning_rate": 1.613252064490759e-05,
+      "loss": 0.865,
+      "step": 344500
+    },
+    {
+      "epoch": 6.78,
+      "learning_rate": 1.608336610302792e-05,
+      "loss": 0.8574,
+      "step": 345000
+    },
+    {
+      "epoch": 6.79,
+      "learning_rate": 1.603421156114825e-05,
+      "loss": 0.8657,
+      "step": 345500
+    },
+    {
+      "epoch": 6.8,
+      "learning_rate": 1.5985057019268583e-05,
+      "loss": 0.8479,
+      "step": 346000
+    },
+    {
+      "epoch": 6.81,
+      "learning_rate": 1.593590247738891e-05,
+      "loss": 0.8548,
+      "step": 346500
+    },
+    {
+      "epoch": 6.82,
+      "learning_rate": 1.5886747935509242e-05,
+      "loss": 0.8485,
+      "step": 347000
+    },
+    {
+      "epoch": 6.83,
+      "learning_rate": 1.583759339362957e-05,
+      "loss": 0.8518,
+      "step": 347500
+    },
+    {
+      "epoch": 6.84,
+      "learning_rate": 1.5788438851749904e-05,
+      "loss": 0.8678,
+      "step": 348000
+    },
+    {
+      "epoch": 6.85,
+      "learning_rate": 1.5739284309870232e-05,
+      "loss": 0.8575,
+      "step": 348500
+    },
+    {
+      "epoch": 6.86,
+      "learning_rate": 1.5690129767990564e-05,
+      "loss": 0.8553,
+      "step": 349000
+    },
+    {
+      "epoch": 6.87,
+      "learning_rate": 1.5640975226110895e-05,
+      "loss": 0.865,
+      "step": 349500
+    },
+    {
+      "epoch": 6.88,
+      "learning_rate": 1.5591820684231223e-05,
+      "loss": 0.852,
+      "step": 350000
+    },
+    {
+      "epoch": 6.89,
+      "learning_rate": 1.5542666142351554e-05,
+      "loss": 0.8435,
+      "step": 350500
+    },
+    {
+      "epoch": 6.9,
+      "learning_rate": 1.5493511600471885e-05,
+      "loss": 0.8605,
+      "step": 351000
+    },
+    {
+      "epoch": 6.91,
+      "learning_rate": 1.5444357058592216e-05,
+      "loss": 0.87,
+      "step": 351500
+    },
+    {
+      "epoch": 6.92,
+      "learning_rate": 1.5395202516712544e-05,
+      "loss": 0.853,
+      "step": 352000
+    },
+    {
+      "epoch": 6.93,
+      "learning_rate": 1.5346047974832876e-05,
+      "loss": 0.8627,
+      "step": 352500
+    },
+    {
+      "epoch": 6.94,
+      "learning_rate": 1.5296893432953203e-05,
+      "loss": 0.8525,
+      "step": 353000
+    },
+    {
+      "epoch": 6.95,
+      "learning_rate": 1.5247738891073535e-05,
+      "loss": 0.854,
+      "step": 353500
+    },
+    {
+      "epoch": 6.96,
+      "learning_rate": 1.5198584349193868e-05,
+      "loss": 0.8404,
+      "step": 354000
+    },
+    {
+      "epoch": 6.97,
+      "learning_rate": 1.5149429807314197e-05,
+      "loss": 0.8652,
+      "step": 354500
+    },
+    {
+      "epoch": 6.98,
+      "learning_rate": 1.5100275265434527e-05,
+      "loss": 0.8651,
+      "step": 355000
+    },
+    {
+      "epoch": 6.99,
+      "learning_rate": 1.5051120723554856e-05,
+      "loss": 0.859,
+      "step": 355500
+    },
+    {
+      "epoch": 7.0,
+      "learning_rate": 1.5001966181675187e-05,
+      "loss": 0.865,
+      "step": 356000
+    },
+    {
+      "epoch": 7.01,
+      "learning_rate": 1.4952811639795517e-05,
+      "loss": 0.8454,
+      "step": 356500
+    },
+    {
+      "epoch": 7.02,
+      "learning_rate": 1.4903657097915848e-05,
+      "loss": 0.8597,
+      "step": 357000
+    },
+    {
+      "epoch": 7.03,
+      "learning_rate": 1.485450255603618e-05,
+      "loss": 0.8388,
+      "step": 357500
+    },
+    {
+      "epoch": 7.04,
+      "learning_rate": 1.4805348014156509e-05,
+      "loss": 0.8396,
+      "step": 358000
+    },
+    {
+      "epoch": 7.05,
+      "learning_rate": 1.4756193472276839e-05,
+      "loss": 0.8523,
+      "step": 358500
+    },
+    {
+      "epoch": 7.06,
+      "learning_rate": 1.4707038930397168e-05,
+      "loss": 0.8509,
+      "step": 359000
+    },
+    {
+      "epoch": 7.07,
+      "learning_rate": 1.4657884388517501e-05,
+      "loss": 0.8461,
+      "step": 359500
+    },
+    {
+      "epoch": 7.08,
+      "learning_rate": 1.460872984663783e-05,
+      "loss": 0.8475,
+      "step": 360000
+    },
+    {
+      "epoch": 7.09,
+      "learning_rate": 1.455957530475816e-05,
+      "loss": 0.8356,
+      "step": 360500
+    },
+    {
+      "epoch": 7.1,
+      "learning_rate": 1.451042076287849e-05,
+      "loss": 0.8534,
+      "step": 361000
+    },
+    {
+      "epoch": 7.11,
+      "learning_rate": 1.4461266220998821e-05,
+      "loss": 0.8501,
+      "step": 361500
+    },
+    {
+      "epoch": 7.12,
+      "learning_rate": 1.441211167911915e-05,
+      "loss": 0.8398,
+      "step": 362000
+    },
+    {
+      "epoch": 7.13,
+      "learning_rate": 1.4362957137239484e-05,
+      "loss": 0.8452,
+      "step": 362500
+    },
+    {
+      "epoch": 7.14,
+      "learning_rate": 1.4313802595359813e-05,
+      "loss": 0.8563,
+      "step": 363000
+    },
+    {
+      "epoch": 7.15,
+      "learning_rate": 1.4264648053480143e-05,
+      "loss": 0.8546,
+      "step": 363500
+    },
+    {
+      "epoch": 7.16,
+      "learning_rate": 1.4215493511600472e-05,
+      "loss": 0.8523,
+      "step": 364000
+    },
+    {
+      "epoch": 7.17,
+      "learning_rate": 1.4166338969720802e-05,
+      "loss": 0.8452,
+      "step": 364500
+    },
+    {
+      "epoch": 7.18,
+      "learning_rate": 1.4117184427841131e-05,
+      "loss": 0.8492,
+      "step": 365000
+    },
+    {
+      "epoch": 7.19,
+      "learning_rate": 1.4068029885961464e-05,
+      "loss": 0.8446,
+      "step": 365500
+    },
+    {
+      "epoch": 7.2,
+      "learning_rate": 1.4018875344081794e-05,
+      "loss": 0.8487,
+      "step": 366000
+    },
+    {
+      "epoch": 7.21,
+      "learning_rate": 1.3969720802202123e-05,
+      "loss": 0.8431,
+      "step": 366500
+    },
+    {
+      "epoch": 7.22,
+      "learning_rate": 1.3920566260322455e-05,
+      "loss": 0.8465,
+      "step": 367000
+    },
+    {
+      "epoch": 7.23,
+      "learning_rate": 1.3871411718442784e-05,
+      "loss": 0.8453,
+      "step": 367500
+    },
+    {
+      "epoch": 7.24,
+      "learning_rate": 1.3822257176563117e-05,
+      "loss": 0.8546,
+      "step": 368000
+    },
+    {
+      "epoch": 7.25,
+      "learning_rate": 1.3773102634683447e-05,
+      "loss": 0.8458,
+      "step": 368500
+    },
+    {
+      "epoch": 7.26,
+      "learning_rate": 1.3723948092803776e-05,
+      "loss": 0.8333,
+      "step": 369000
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 1.3674793550924106e-05,
+      "loss": 0.8514,
+      "step": 369500
+    },
+    {
+      "epoch": 7.27,
+      "learning_rate": 1.3625639009044435e-05,
+      "loss": 0.8466,
+      "step": 370000
+    },
+    {
+      "epoch": 7.28,
+      "learning_rate": 1.3576484467164765e-05,
+      "loss": 0.84,
+      "step": 370500
+    },
+    {
+      "epoch": 7.29,
+      "learning_rate": 1.3527329925285098e-05,
+      "loss": 0.858,
+      "step": 371000
+    },
+    {
+      "epoch": 7.3,
+      "learning_rate": 1.3478175383405427e-05,
+      "loss": 0.8472,
+      "step": 371500
+    },
+    {
+      "epoch": 7.31,
+      "learning_rate": 1.3429020841525759e-05,
+      "loss": 0.8519,
+      "step": 372000
+    },
+    {
+      "epoch": 7.32,
+      "learning_rate": 1.3379866299646088e-05,
+      "loss": 0.8513,
+      "step": 372500
+    },
+    {
+      "epoch": 7.33,
+      "learning_rate": 1.3330711757766418e-05,
+      "loss": 0.8448,
+      "step": 373000
+    },
+    {
+      "epoch": 7.34,
+      "learning_rate": 1.3281557215886747e-05,
+      "loss": 0.8454,
+      "step": 373500
+    },
+    {
+      "epoch": 7.35,
+      "learning_rate": 1.323240267400708e-05,
+      "loss": 0.8542,
+      "step": 374000
+    },
+    {
+      "epoch": 7.36,
+      "learning_rate": 1.318324813212741e-05,
+      "loss": 0.8522,
+      "step": 374500
+    },
+    {
+      "epoch": 7.37,
+      "learning_rate": 1.313409359024774e-05,
+      "loss": 0.8451,
+      "step": 375000
+    },
+    {
+      "epoch": 7.38,
+      "learning_rate": 1.3084939048368069e-05,
+      "loss": 0.8394,
+      "step": 375500
+    },
+    {
+      "epoch": 7.39,
+      "learning_rate": 1.30357845064884e-05,
+      "loss": 0.8375,
+      "step": 376000
+    },
+    {
+      "epoch": 7.4,
+      "learning_rate": 1.298662996460873e-05,
+      "loss": 0.8379,
+      "step": 376500
+    },
+    {
+      "epoch": 7.41,
+      "learning_rate": 1.2937475422729061e-05,
+      "loss": 0.8507,
+      "step": 377000
+    },
+    {
+      "epoch": 7.42,
+      "learning_rate": 1.2888320880849392e-05,
+      "loss": 0.8513,
+      "step": 377500
+    },
+    {
+      "epoch": 7.43,
+      "learning_rate": 1.2839166338969722e-05,
+      "loss": 0.8489,
+      "step": 378000
+    },
+    {
+      "epoch": 7.44,
+      "learning_rate": 1.2790011797090051e-05,
+      "loss": 0.8523,
+      "step": 378500
+    },
+    {
+      "epoch": 7.45,
+      "learning_rate": 1.2740857255210381e-05,
+      "loss": 0.8444,
+      "step": 379000
+    },
+    {
+      "epoch": 7.46,
+      "learning_rate": 1.2691702713330714e-05,
+      "loss": 0.8427,
+      "step": 379500
+    },
+    {
+      "epoch": 7.47,
+      "learning_rate": 1.2642548171451043e-05,
+      "loss": 0.8465,
+      "step": 380000
+    },
+    {
+      "epoch": 7.48,
+      "learning_rate": 1.2593393629571373e-05,
+      "loss": 0.8317,
+      "step": 380500
+    },
+    {
+      "epoch": 7.49,
+      "learning_rate": 1.2544239087691702e-05,
+      "loss": 0.8336,
+      "step": 381000
+    },
+    {
+      "epoch": 7.5,
+      "learning_rate": 1.2495084545812034e-05,
+      "loss": 0.8303,
+      "step": 381500
+    },
+    {
+      "epoch": 7.51,
+      "learning_rate": 1.2445930003932365e-05,
+      "loss": 0.8456,
+      "step": 382000
+    },
+    {
+      "epoch": 7.52,
+      "learning_rate": 1.2396775462052695e-05,
+      "loss": 0.8575,
+      "step": 382500
+    },
+    {
+      "epoch": 7.53,
+      "learning_rate": 1.2347620920173024e-05,
+      "loss": 0.8468,
+      "step": 383000
+    },
+    {
+      "epoch": 7.54,
+      "learning_rate": 1.2298466378293355e-05,
+      "loss": 0.8435,
+      "step": 383500
+    },
+    {
+      "epoch": 7.55,
+      "learning_rate": 1.2249311836413685e-05,
+      "loss": 0.8425,
+      "step": 384000
+    },
+    {
+      "epoch": 7.56,
+      "learning_rate": 1.2200157294534016e-05,
+      "loss": 0.8421,
+      "step": 384500
+    },
+    {
+      "epoch": 7.57,
+      "learning_rate": 1.2151002752654346e-05,
+      "loss": 0.8518,
+      "step": 385000
+    },
+    {
+      "epoch": 7.58,
+      "learning_rate": 1.2101848210774675e-05,
+      "loss": 0.839,
+      "step": 385500
+    },
+    {
+      "epoch": 7.59,
+      "learning_rate": 1.2052693668895006e-05,
+      "loss": 0.8344,
+      "step": 386000
+    },
+    {
+      "epoch": 7.6,
+      "learning_rate": 1.2003539127015336e-05,
+      "loss": 0.8446,
+      "step": 386500
+    },
+    {
+      "epoch": 7.61,
+      "learning_rate": 1.1954384585135667e-05,
+      "loss": 0.8314,
+      "step": 387000
+    },
+    {
+      "epoch": 7.62,
+      "learning_rate": 1.1905230043255999e-05,
+      "loss": 0.8382,
+      "step": 387500
+    },
+    {
+      "epoch": 7.63,
+      "learning_rate": 1.1856075501376328e-05,
+      "loss": 0.8527,
+      "step": 388000
+    },
+    {
+      "epoch": 7.64,
+      "learning_rate": 1.1806920959496658e-05,
+      "loss": 0.8491,
+      "step": 388500
+    },
+    {
+      "epoch": 7.65,
+      "learning_rate": 1.1757766417616989e-05,
+      "loss": 0.834,
+      "step": 389000
+    },
+    {
+      "epoch": 7.66,
+      "learning_rate": 1.1708611875737318e-05,
+      "loss": 0.8415,
+      "step": 389500
+    },
+    {
+      "epoch": 7.67,
+      "learning_rate": 1.1659457333857648e-05,
+      "loss": 0.8423,
+      "step": 390000
+    },
+    {
+      "epoch": 7.68,
+      "learning_rate": 1.161030279197798e-05,
+      "loss": 0.8454,
+      "step": 390500
+    },
+    {
+      "epoch": 7.69,
+      "learning_rate": 1.1561148250098309e-05,
+      "loss": 0.8562,
+      "step": 391000
+    },
+    {
+      "epoch": 7.7,
+      "learning_rate": 1.151199370821864e-05,
+      "loss": 0.8516,
+      "step": 391500
+    },
+    {
+      "epoch": 7.71,
+      "learning_rate": 1.1462839166338971e-05,
+      "loss": 0.8434,
+      "step": 392000
+    },
+    {
+      "epoch": 7.72,
+      "learning_rate": 1.1413684624459301e-05,
+      "loss": 0.8353,
+      "step": 392500
+    },
+    {
+      "epoch": 7.73,
+      "learning_rate": 1.136453008257963e-05,
+      "loss": 0.8447,
+      "step": 393000
+    },
+    {
+      "epoch": 7.74,
+      "learning_rate": 1.1315375540699962e-05,
+      "loss": 0.8374,
+      "step": 393500
+    },
+    {
+      "epoch": 7.75,
+      "learning_rate": 1.1266220998820291e-05,
+      "loss": 0.8374,
+      "step": 394000
+    },
+    {
+      "epoch": 7.76,
+      "learning_rate": 1.1217066456940622e-05,
+      "loss": 0.8487,
+      "step": 394500
+    },
+    {
+      "epoch": 7.77,
+      "learning_rate": 1.1167911915060952e-05,
+      "loss": 0.8475,
+      "step": 395000
+    },
+    {
+      "epoch": 7.78,
+      "learning_rate": 1.1118757373181282e-05,
+      "loss": 0.8415,
+      "step": 395500
+    },
+    {
+      "epoch": 7.79,
+      "learning_rate": 1.1069602831301613e-05,
+      "loss": 0.8424,
+      "step": 396000
+    },
+    {
+      "epoch": 7.8,
+      "learning_rate": 1.1020448289421944e-05,
+      "loss": 0.8415,
+      "step": 396500
+    },
+    {
+      "epoch": 7.81,
+      "learning_rate": 1.0971293747542274e-05,
+      "loss": 0.8466,
+      "step": 397000
+    },
+    {
+      "epoch": 7.82,
+      "learning_rate": 1.0922139205662605e-05,
+      "loss": 0.834,
+      "step": 397500
+    },
+    {
+      "epoch": 7.83,
+      "learning_rate": 1.0872984663782934e-05,
+      "loss": 0.8413,
+      "step": 398000
+    },
+    {
+      "epoch": 7.84,
+      "learning_rate": 1.0823830121903264e-05,
+      "loss": 0.8476,
+      "step": 398500
+    },
+    {
+      "epoch": 7.85,
+      "learning_rate": 1.0774675580023595e-05,
+      "loss": 0.8419,
+      "step": 399000
+    },
+    {
+      "epoch": 7.85,
+      "learning_rate": 1.0725521038143925e-05,
+      "loss": 0.853,
+      "step": 399500
+    },
+    {
+      "epoch": 7.86,
+      "learning_rate": 1.0676366496264254e-05,
+      "loss": 0.8368,
+      "step": 400000
+    },
+    {
+      "epoch": 7.87,
+      "learning_rate": 1.0627211954384586e-05,
+      "loss": 0.8426,
+      "step": 400500
+    },
+    {
+      "epoch": 7.88,
+      "learning_rate": 1.0578057412504915e-05,
+      "loss": 0.8566,
+      "step": 401000
+    },
+    {
+      "epoch": 7.89,
+      "learning_rate": 1.0528902870625246e-05,
+      "loss": 0.8405,
+      "step": 401500
+    },
+    {
+      "epoch": 7.9,
+      "learning_rate": 1.0479748328745578e-05,
+      "loss": 0.8338,
+      "step": 402000
+    },
+    {
+      "epoch": 7.91,
+      "learning_rate": 1.0430593786865907e-05,
+      "loss": 0.8411,
+      "step": 402500
+    },
+    {
+      "epoch": 7.92,
+      "learning_rate": 1.0381439244986237e-05,
+      "loss": 0.8457,
+      "step": 403000
+    },
+    {
+      "epoch": 7.93,
+      "learning_rate": 1.0332284703106568e-05,
+      "loss": 0.8472,
+      "step": 403500
+    },
+    {
+      "epoch": 7.94,
+      "learning_rate": 1.0283130161226898e-05,
+      "loss": 0.8283,
+      "step": 404000
+    },
+    {
+      "epoch": 7.95,
+      "learning_rate": 1.0233975619347227e-05,
+      "loss": 0.8459,
+      "step": 404500
+    },
+    {
+      "epoch": 7.96,
+      "learning_rate": 1.0184821077467558e-05,
+      "loss": 0.8401,
+      "step": 405000
+    },
+    {
+      "epoch": 7.97,
+      "learning_rate": 1.0135666535587888e-05,
+      "loss": 0.8307,
+      "step": 405500
+    },
+    {
+      "epoch": 7.98,
+      "learning_rate": 1.008651199370822e-05,
+      "loss": 0.8379,
+      "step": 406000
+    },
+    {
+      "epoch": 7.99,
+      "learning_rate": 1.003735745182855e-05,
+      "loss": 0.8398,
+      "step": 406500
+    },
+    {
+      "epoch": 8.0,
+      "learning_rate": 9.98820290994888e-06,
+      "loss": 0.8283,
+      "step": 407000
+    },
+    {
+      "epoch": 8.01,
+      "learning_rate": 9.939048368069211e-06,
+      "loss": 0.8299,
+      "step": 407500
+    },
+    {
+      "epoch": 8.02,
+      "learning_rate": 9.88989382618954e-06,
+      "loss": 0.8323,
+      "step": 408000
+    },
+    {
+      "epoch": 8.03,
+      "learning_rate": 9.84073928430987e-06,
+      "loss": 0.8346,
+      "step": 408500
+    },
+    {
+      "epoch": 8.04,
+      "learning_rate": 9.791584742430202e-06,
+      "loss": 0.8382,
+      "step": 409000
+    },
+    {
+      "epoch": 8.05,
+      "learning_rate": 9.742430200550531e-06,
+      "loss": 0.8312,
+      "step": 409500
+    },
+    {
+      "epoch": 8.06,
+      "learning_rate": 9.69327565867086e-06,
+      "loss": 0.8264,
+      "step": 410000
+    },
+    {
+      "epoch": 8.07,
+      "learning_rate": 9.644121116791192e-06,
+      "loss": 0.8448,
+      "step": 410500
+    },
+    {
+      "epoch": 8.08,
+      "learning_rate": 9.594966574911521e-06,
+      "loss": 0.8387,
+      "step": 411000
+    },
+    {
+      "epoch": 8.09,
+      "learning_rate": 9.545812033031853e-06,
+      "loss": 0.8427,
+      "step": 411500
+    },
+    {
+      "epoch": 8.1,
+      "learning_rate": 9.496657491152184e-06,
+      "loss": 0.8295,
+      "step": 412000
+    },
+    {
+      "epoch": 8.11,
+      "learning_rate": 9.447502949272514e-06,
+      "loss": 0.8314,
+      "step": 412500
+    },
+    {
+      "epoch": 8.12,
+      "learning_rate": 9.398348407392843e-06,
+      "loss": 0.8331,
+      "step": 413000
+    },
+    {
+      "epoch": 8.13,
+      "learning_rate": 9.349193865513174e-06,
+      "loss": 0.8328,
+      "step": 413500
+    },
+    {
+      "epoch": 8.14,
+      "learning_rate": 9.300039323633504e-06,
+      "loss": 0.8389,
+      "step": 414000
+    },
+    {
+      "epoch": 8.15,
+      "learning_rate": 9.250884781753833e-06,
+      "loss": 0.8361,
+      "step": 414500
+    },
+    {
+      "epoch": 8.16,
+      "learning_rate": 9.201730239874165e-06,
+      "loss": 0.8297,
+      "step": 415000
+    },
+    {
+      "epoch": 8.17,
+      "learning_rate": 9.152575697994494e-06,
+      "loss": 0.8363,
+      "step": 415500
+    },
+    {
+      "epoch": 8.18,
+      "learning_rate": 9.103421156114826e-06,
+      "loss": 0.8236,
+      "step": 416000
+    },
+    {
+      "epoch": 8.19,
+      "learning_rate": 9.054266614235157e-06,
+      "loss": 0.8253,
+      "step": 416500
+    },
+    {
+      "epoch": 8.2,
+      "learning_rate": 9.005112072355486e-06,
+      "loss": 0.8374,
+      "step": 417000
+    },
+    {
+      "epoch": 8.21,
+      "learning_rate": 8.955957530475818e-06,
+      "loss": 0.8392,
+      "step": 417500
+    },
+    {
+      "epoch": 8.22,
+      "learning_rate": 8.906802988596147e-06,
+      "loss": 0.8299,
+      "step": 418000
+    },
+    {
+      "epoch": 8.23,
+      "learning_rate": 8.857648446716477e-06,
+      "loss": 0.8358,
+      "step": 418500
+    },
+    {
+      "epoch": 8.24,
+      "learning_rate": 8.808493904836808e-06,
+      "loss": 0.8382,
+      "step": 419000
+    },
+    {
+      "epoch": 8.25,
+      "learning_rate": 8.759339362957137e-06,
+      "loss": 0.8266,
+      "step": 419500
+    },
+    {
+      "epoch": 8.26,
+      "learning_rate": 8.710184821077467e-06,
+      "loss": 0.8376,
+      "step": 420000
+    },
+    {
+      "epoch": 8.27,
+      "learning_rate": 8.661030279197798e-06,
+      "loss": 0.8373,
+      "step": 420500
+    },
+    {
+      "epoch": 8.28,
+      "learning_rate": 8.611875737318128e-06,
+      "loss": 0.8358,
+      "step": 421000
+    },
+    {
+      "epoch": 8.29,
+      "learning_rate": 8.562721195438459e-06,
+      "loss": 0.8406,
+      "step": 421500
+    },
+    {
+      "epoch": 8.3,
+      "learning_rate": 8.51356665355879e-06,
+      "loss": 0.8322,
+      "step": 422000
+    },
+    {
+      "epoch": 8.31,
+      "learning_rate": 8.46441211167912e-06,
+      "loss": 0.8391,
+      "step": 422500
+    },
+    {
+      "epoch": 8.32,
+      "learning_rate": 8.41525756979945e-06,
+      "loss": 0.8487,
+      "step": 423000
+    },
+    {
+      "epoch": 8.33,
+      "learning_rate": 8.36610302791978e-06,
+      "loss": 0.8206,
+      "step": 423500
+    },
+    {
+      "epoch": 8.34,
+      "learning_rate": 8.31694848604011e-06,
+      "loss": 0.8305,
+      "step": 424000
+    },
+    {
+      "epoch": 8.35,
+      "learning_rate": 8.26779394416044e-06,
+      "loss": 0.8382,
+      "step": 424500
+    },
+    {
+      "epoch": 8.36,
+      "learning_rate": 8.218639402280771e-06,
+      "loss": 0.8278,
+      "step": 425000
+    },
+    {
+      "epoch": 8.37,
+      "learning_rate": 8.1694848604011e-06,
+      "loss": 0.8358,
+      "step": 425500
+    },
+    {
+      "epoch": 8.38,
+      "learning_rate": 8.120330318521432e-06,
+      "loss": 0.8298,
+      "step": 426000
+    },
+    {
+      "epoch": 8.39,
+      "learning_rate": 8.071175776641763e-06,
+      "loss": 0.8257,
+      "step": 426500
+    },
+    {
+      "epoch": 8.4,
+      "learning_rate": 8.022021234762093e-06,
+      "loss": 0.8395,
+      "step": 427000
+    },
+    {
+      "epoch": 8.41,
+      "learning_rate": 7.972866692882424e-06,
+      "loss": 0.8355,
+      "step": 427500
+    },
+    {
+      "epoch": 8.42,
+      "learning_rate": 7.923712151002753e-06,
+      "loss": 0.8289,
+      "step": 428000
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 7.874557609123083e-06,
+      "loss": 0.844,
+      "step": 428500
+    },
+    {
+      "epoch": 8.43,
+      "learning_rate": 7.825403067243414e-06,
+      "loss": 0.8313,
+      "step": 429000
+    },
+    {
+      "epoch": 8.44,
+      "learning_rate": 7.776248525363744e-06,
+      "loss": 0.8367,
+      "step": 429500
+    },
+    {
+      "epoch": 8.45,
+      "learning_rate": 7.727093983484073e-06,
+      "loss": 0.8315,
+      "step": 430000
+    },
+    {
+      "epoch": 8.46,
+      "learning_rate": 7.677939441604405e-06,
+      "loss": 0.8373,
+      "step": 430500
+    },
+    {
+      "epoch": 8.47,
+      "learning_rate": 7.628784899724735e-06,
+      "loss": 0.8369,
+      "step": 431000
+    },
+    {
+      "epoch": 8.48,
+      "learning_rate": 7.579630357845065e-06,
+      "loss": 0.8307,
+      "step": 431500
+    },
+    {
+      "epoch": 8.49,
+      "learning_rate": 7.530475815965396e-06,
+      "loss": 0.8268,
+      "step": 432000
+    },
+    {
+      "epoch": 8.5,
+      "learning_rate": 7.481321274085725e-06,
+      "loss": 0.8307,
+      "step": 432500
+    },
+    {
+      "epoch": 8.51,
+      "learning_rate": 7.432166732206056e-06,
+      "loss": 0.8254,
+      "step": 433000
+    },
+    {
+      "epoch": 8.52,
+      "learning_rate": 7.383012190326387e-06,
+      "loss": 0.8402,
+      "step": 433500
+    },
+    {
+      "epoch": 8.53,
+      "learning_rate": 7.333857648446717e-06,
+      "loss": 0.8292,
+      "step": 434000
+    },
+    {
+      "epoch": 8.54,
+      "learning_rate": 7.284703106567046e-06,
+      "loss": 0.8263,
+      "step": 434500
+    },
+    {
+      "epoch": 8.55,
+      "learning_rate": 7.235548564687377e-06,
+      "loss": 0.8418,
+      "step": 435000
+    },
+    {
+      "epoch": 8.56,
+      "learning_rate": 7.186394022807708e-06,
+      "loss": 0.8297,
+      "step": 435500
+    },
+    {
+      "epoch": 8.57,
+      "learning_rate": 7.137239480928037e-06,
+      "loss": 0.8327,
+      "step": 436000
+    },
+    {
+      "epoch": 8.58,
+      "learning_rate": 7.088084939048369e-06,
+      "loss": 0.8392,
+      "step": 436500
+    },
+    {
+      "epoch": 8.59,
+      "learning_rate": 7.038930397168698e-06,
+      "loss": 0.8328,
+      "step": 437000
+    },
+    {
+      "epoch": 8.6,
+      "learning_rate": 6.989775855289029e-06,
+      "loss": 0.8278,
+      "step": 437500
+    },
+    {
+      "epoch": 8.61,
+      "learning_rate": 6.94062131340936e-06,
+      "loss": 0.8306,
+      "step": 438000
+    },
+    {
+      "epoch": 8.62,
+      "learning_rate": 6.891466771529689e-06,
+      "loss": 0.8278,
+      "step": 438500
+    },
+    {
+      "epoch": 8.63,
+      "learning_rate": 6.842312229650021e-06,
+      "loss": 0.8322,
+      "step": 439000
+    },
+    {
+      "epoch": 8.64,
+      "learning_rate": 6.79315768777035e-06,
+      "loss": 0.8287,
+      "step": 439500
+    },
+    {
+      "epoch": 8.65,
+      "learning_rate": 6.7440031458906806e-06,
+      "loss": 0.8357,
+      "step": 440000
+    },
+    {
+      "epoch": 8.66,
+      "learning_rate": 6.694848604011012e-06,
+      "loss": 0.8366,
+      "step": 440500
+    },
+    {
+      "epoch": 8.67,
+      "learning_rate": 6.645694062131341e-06,
+      "loss": 0.8278,
+      "step": 441000
+    },
+    {
+      "epoch": 8.68,
+      "learning_rate": 6.596539520251671e-06,
+      "loss": 0.8388,
+      "step": 441500
+    },
+    {
+      "epoch": 8.69,
+      "learning_rate": 6.547384978372002e-06,
+      "loss": 0.8326,
+      "step": 442000
+    },
+    {
+      "epoch": 8.7,
+      "learning_rate": 6.4982304364923326e-06,
+      "loss": 0.8309,
+      "step": 442500
+    },
+    {
+      "epoch": 8.71,
+      "learning_rate": 6.449075894612662e-06,
+      "loss": 0.826,
+      "step": 443000
+    },
+    {
+      "epoch": 8.72,
+      "learning_rate": 6.399921352732993e-06,
+      "loss": 0.838,
+      "step": 443500
+    },
+    {
+      "epoch": 8.73,
+      "learning_rate": 6.350766810853323e-06,
+      "loss": 0.8418,
+      "step": 444000
+    },
+    {
+      "epoch": 8.74,
+      "learning_rate": 6.3016122689736525e-06,
+      "loss": 0.827,
+      "step": 444500
+    },
+    {
+      "epoch": 8.75,
+      "learning_rate": 6.252457727093984e-06,
+      "loss": 0.8258,
+      "step": 445000
+    },
+    {
+      "epoch": 8.76,
+      "learning_rate": 6.203303185214314e-06,
+      "loss": 0.8359,
+      "step": 445500
+    },
+    {
+      "epoch": 8.77,
+      "learning_rate": 6.1541486433346445e-06,
+      "loss": 0.8356,
+      "step": 446000
+    },
+    {
+      "epoch": 8.78,
+      "learning_rate": 6.104994101454975e-06,
+      "loss": 0.8164,
+      "step": 446500
+    },
+    {
+      "epoch": 8.79,
+      "learning_rate": 6.0558395595753045e-06,
+      "loss": 0.8447,
+      "step": 447000
+    },
+    {
+      "epoch": 8.8,
+      "learning_rate": 6.006685017695636e-06,
+      "loss": 0.8411,
+      "step": 447500
+    },
+    {
+      "epoch": 8.81,
+      "learning_rate": 5.957530475815966e-06,
+      "loss": 0.8361,
+      "step": 448000
+    },
+    {
+      "epoch": 8.82,
+      "learning_rate": 5.908375933936296e-06,
+      "loss": 0.8258,
+      "step": 448500
+    },
+    {
+      "epoch": 8.83,
+      "learning_rate": 5.859221392056626e-06,
+      "loss": 0.8248,
+      "step": 449000
+    },
+    {
+      "epoch": 8.84,
+      "learning_rate": 5.8100668501769565e-06,
+      "loss": 0.8292,
+      "step": 449500
+    },
+    {
+      "epoch": 8.85,
+      "learning_rate": 5.760912308297287e-06,
+      "loss": 0.8235,
+      "step": 450000
+    },
+    {
+      "epoch": 8.86,
+      "learning_rate": 5.711757766417617e-06,
+      "loss": 0.8326,
+      "step": 450500
+    },
+    {
+      "epoch": 8.87,
+      "learning_rate": 5.662603224537948e-06,
+      "loss": 0.8177,
+      "step": 451000
+    },
+    {
+      "epoch": 8.88,
+      "learning_rate": 5.613448682658278e-06,
+      "loss": 0.8496,
+      "step": 451500
+    },
+    {
+      "epoch": 8.89,
+      "learning_rate": 5.564294140778608e-06,
+      "loss": 0.8375,
+      "step": 452000
+    },
+    {
+      "epoch": 8.9,
+      "learning_rate": 5.515139598898939e-06,
+      "loss": 0.8365,
+      "step": 452500
+    },
+    {
+      "epoch": 8.91,
+      "learning_rate": 5.465985057019269e-06,
+      "loss": 0.8371,
+      "step": 453000
+    },
+    {
+      "epoch": 8.92,
+      "learning_rate": 5.416830515139599e-06,
+      "loss": 0.8284,
+      "step": 453500
+    },
+    {
+      "epoch": 8.93,
+      "learning_rate": 5.367675973259929e-06,
+      "loss": 0.8398,
+      "step": 454000
+    },
+    {
+      "epoch": 8.94,
+      "learning_rate": 5.31852143138026e-06,
+      "loss": 0.8371,
+      "step": 454500
+    },
+    {
+      "epoch": 8.95,
+      "learning_rate": 5.26936688950059e-06,
+      "loss": 0.8339,
+      "step": 455000
+    },
+    {
+      "epoch": 8.96,
+      "learning_rate": 5.2202123476209205e-06,
+      "loss": 0.8288,
+      "step": 455500
+    },
+    {
+      "epoch": 8.97,
+      "learning_rate": 5.171057805741251e-06,
+      "loss": 0.8191,
+      "step": 456000
+    },
+    {
+      "epoch": 8.98,
+      "learning_rate": 5.121903263861581e-06,
+      "loss": 0.8308,
+      "step": 456500
+    },
+    {
+      "epoch": 8.99,
+      "learning_rate": 5.072748721981911e-06,
+      "loss": 0.8308,
+      "step": 457000
+    },
+    {
+      "epoch": 9.0,
+      "learning_rate": 5.023594180102242e-06,
+      "loss": 0.8238,
+      "step": 457500
+    },
+    {
+      "epoch": 9.01,
+      "learning_rate": 4.9744396382225725e-06,
+      "loss": 0.8237,
+      "step": 458000
+    },
+    {
+      "epoch": 9.01,
+      "learning_rate": 4.925285096342902e-06,
+      "loss": 0.8287,
+      "step": 458500
+    },
+    {
+      "epoch": 9.02,
+      "learning_rate": 4.8761305544632324e-06,
+      "loss": 0.8243,
+      "step": 459000
+    },
+    {
+      "epoch": 9.03,
+      "learning_rate": 4.826976012583563e-06,
+      "loss": 0.8379,
+      "step": 459500
+    },
+    {
+      "epoch": 9.04,
+      "learning_rate": 4.777821470703893e-06,
+      "loss": 0.8295,
+      "step": 460000
+    },
+    {
+      "epoch": 9.05,
+      "learning_rate": 4.728666928824224e-06,
+      "loss": 0.824,
+      "step": 460500
+    },
+    {
+      "epoch": 9.06,
+      "learning_rate": 4.679512386944554e-06,
+      "loss": 0.8207,
+      "step": 461000
+    },
+    {
+      "epoch": 9.07,
+      "learning_rate": 4.6303578450648845e-06,
+      "loss": 0.825,
+      "step": 461500
+    },
+    {
+      "epoch": 9.08,
+      "learning_rate": 4.581203303185214e-06,
+      "loss": 0.8292,
+      "step": 462000
+    },
+    {
+      "epoch": 9.09,
+      "learning_rate": 4.532048761305545e-06,
+      "loss": 0.8146,
+      "step": 462500
+    },
+    {
+      "epoch": 9.1,
+      "learning_rate": 4.482894219425876e-06,
+      "loss": 0.825,
+      "step": 463000
+    },
+    {
+      "epoch": 9.11,
+      "learning_rate": 4.433739677546205e-06,
+      "loss": 0.8277,
+      "step": 463500
+    },
+    {
+      "epoch": 9.12,
+      "learning_rate": 4.384585135666536e-06,
+      "loss": 0.8252,
+      "step": 464000
+    },
+    {
+      "epoch": 9.13,
+      "learning_rate": 4.335430593786866e-06,
+      "loss": 0.8263,
+      "step": 464500
+    },
+    {
+      "epoch": 9.14,
+      "learning_rate": 4.286276051907196e-06,
+      "loss": 0.8302,
+      "step": 465000
+    },
+    {
+      "epoch": 9.15,
+      "learning_rate": 4.237121510027527e-06,
+      "loss": 0.8321,
+      "step": 465500
+    },
+    {
+      "epoch": 9.16,
+      "learning_rate": 4.187966968147857e-06,
+      "loss": 0.8288,
+      "step": 466000
+    },
+    {
+      "epoch": 9.17,
+      "learning_rate": 4.138812426268188e-06,
+      "loss": 0.8346,
+      "step": 466500
+    },
+    {
+      "epoch": 9.18,
+      "learning_rate": 4.089657884388517e-06,
+      "loss": 0.8262,
+      "step": 467000
+    },
+    {
+      "epoch": 9.19,
+      "learning_rate": 4.0405033425088484e-06,
+      "loss": 0.8265,
+      "step": 467500
+    },
+    {
+      "epoch": 9.2,
+      "learning_rate": 3.991348800629179e-06,
+      "loss": 0.8337,
+      "step": 468000
+    },
+    {
+      "epoch": 9.21,
+      "learning_rate": 3.942194258749508e-06,
+      "loss": 0.8469,
+      "step": 468500
+    },
+    {
+      "epoch": 9.22,
+      "learning_rate": 3.893039716869839e-06,
+      "loss": 0.8349,
+      "step": 469000
+    },
+    {
+      "epoch": 9.23,
+      "learning_rate": 3.843885174990169e-06,
+      "loss": 0.8192,
+      "step": 469500
+    },
+    {
+      "epoch": 9.24,
+      "learning_rate": 3.794730633110499e-06,
+      "loss": 0.8288,
+      "step": 470000
+    },
+    {
+      "epoch": 9.25,
+      "learning_rate": 3.74557609123083e-06,
+      "loss": 0.8289,
+      "step": 470500
+    },
+    {
+      "epoch": 9.26,
+      "learning_rate": 3.6964215493511604e-06,
+      "loss": 0.8261,
+      "step": 471000
+    },
+    {
+      "epoch": 9.27,
+      "learning_rate": 3.647267007471491e-06,
+      "loss": 0.8309,
+      "step": 471500
+    },
+    {
+      "epoch": 9.28,
+      "learning_rate": 3.5981124655918208e-06,
+      "loss": 0.8227,
+      "step": 472000
+    },
+    {
+      "epoch": 9.29,
+      "learning_rate": 3.548957923712151e-06,
+      "loss": 0.8222,
+      "step": 472500
+    },
+    {
+      "epoch": 9.3,
+      "learning_rate": 3.4998033818324816e-06,
+      "loss": 0.8214,
+      "step": 473000
+    },
+    {
+      "epoch": 9.31,
+      "learning_rate": 3.4506488399528116e-06,
+      "loss": 0.8295,
+      "step": 473500
+    },
+    {
+      "epoch": 9.32,
+      "learning_rate": 3.401494298073142e-06,
+      "loss": 0.8313,
+      "step": 474000
+    },
+    {
+      "epoch": 9.33,
+      "learning_rate": 3.3523397561934728e-06,
+      "loss": 0.8416,
+      "step": 474500
+    },
+    {
+      "epoch": 9.34,
+      "learning_rate": 3.3031852143138023e-06,
+      "loss": 0.8233,
+      "step": 475000
+    },
+    {
+      "epoch": 9.35,
+      "learning_rate": 3.254030672434133e-06,
+      "loss": 0.8229,
+      "step": 475500
+    },
+    {
+      "epoch": 9.36,
+      "learning_rate": 3.2048761305544636e-06,
+      "loss": 0.8294,
+      "step": 476000
+    },
+    {
+      "epoch": 9.37,
+      "learning_rate": 3.155721588674794e-06,
+      "loss": 0.8313,
+      "step": 476500
+    },
+    {
+      "epoch": 9.38,
+      "learning_rate": 3.1065670467951244e-06,
+      "loss": 0.8167,
+      "step": 477000
+    },
+    {
+      "epoch": 9.39,
+      "learning_rate": 3.0574125049154543e-06,
+      "loss": 0.8252,
+      "step": 477500
+    },
+    {
+      "epoch": 9.4,
+      "learning_rate": 3.0082579630357847e-06,
+      "loss": 0.824,
+      "step": 478000
+    },
+    {
+      "epoch": 9.41,
+      "learning_rate": 2.959103421156115e-06,
+      "loss": 0.8238,
+      "step": 478500
+    },
+    {
+      "epoch": 9.42,
+      "learning_rate": 2.909948879276445e-06,
+      "loss": 0.8217,
+      "step": 479000
+    },
+    {
+      "epoch": 9.43,
+      "learning_rate": 2.860794337396776e-06,
+      "loss": 0.8427,
+      "step": 479500
+    },
+    {
+      "epoch": 9.44,
+      "learning_rate": 2.811639795517106e-06,
+      "loss": 0.8242,
+      "step": 480000
+    },
+    {
+      "epoch": 9.45,
+      "learning_rate": 2.7624852536374363e-06,
+      "loss": 0.8186,
+      "step": 480500
+    },
+    {
+      "epoch": 9.46,
+      "learning_rate": 2.7133307117577667e-06,
+      "loss": 0.8238,
+      "step": 481000
+    },
+    {
+      "epoch": 9.47,
+      "learning_rate": 2.6641761698780967e-06,
+      "loss": 0.8291,
+      "step": 481500
+    },
+    {
+      "epoch": 9.48,
+      "learning_rate": 2.615021627998427e-06,
+      "loss": 0.8278,
+      "step": 482000
+    },
+    {
+      "epoch": 9.49,
+      "learning_rate": 2.5658670861187575e-06,
+      "loss": 0.8383,
+      "step": 482500
+    },
+    {
+      "epoch": 9.5,
+      "learning_rate": 2.516712544239088e-06,
+      "loss": 0.8377,
+      "step": 483000
+    },
+    {
+      "epoch": 9.51,
+      "learning_rate": 2.4675580023594183e-06,
+      "loss": 0.8205,
+      "step": 483500
+    },
+    {
+      "epoch": 9.52,
+      "learning_rate": 2.4184034604797483e-06,
+      "loss": 0.8293,
+      "step": 484000
+    },
+    {
+      "epoch": 9.53,
+      "learning_rate": 2.3692489186000787e-06,
+      "loss": 0.8206,
+      "step": 484500
+    },
+    {
+      "epoch": 9.54,
+      "learning_rate": 2.320094376720409e-06,
+      "loss": 0.826,
+      "step": 485000
+    },
+    {
+      "epoch": 9.55,
+      "learning_rate": 2.2709398348407395e-06,
+      "loss": 0.8192,
+      "step": 485500
+    },
+    {
+      "epoch": 9.56,
+      "learning_rate": 2.22178529296107e-06,
+      "loss": 0.8271,
+      "step": 486000
+    },
+    {
+      "epoch": 9.57,
+      "learning_rate": 2.1726307510814e-06,
+      "loss": 0.8296,
+      "step": 486500
+    },
+    {
+      "epoch": 9.58,
+      "learning_rate": 2.1234762092017303e-06,
+      "loss": 0.8286,
+      "step": 487000
+    },
+    {
+      "epoch": 9.59,
+      "learning_rate": 2.0743216673220607e-06,
+      "loss": 0.8282,
+      "step": 487500
+    },
+    {
+      "epoch": 9.59,
+      "learning_rate": 2.025167125442391e-06,
+      "loss": 0.8143,
+      "step": 488000
+    },
+    {
+      "epoch": 9.6,
+      "learning_rate": 1.9760125835627215e-06,
+      "loss": 0.829,
+      "step": 488500
+    },
+    {
+      "epoch": 9.61,
+      "learning_rate": 1.9268580416830515e-06,
+      "loss": 0.8236,
+      "step": 489000
+    },
+    {
+      "epoch": 9.62,
+      "learning_rate": 1.8777034998033819e-06,
+      "loss": 0.8227,
+      "step": 489500
+    },
+    {
+      "epoch": 9.63,
+      "learning_rate": 1.8285489579237123e-06,
+      "loss": 0.8212,
+      "step": 490000
+    },
+    {
+      "epoch": 9.64,
+      "learning_rate": 1.7793944160440425e-06,
+      "loss": 0.8176,
+      "step": 490500
+    },
+    {
+      "epoch": 9.65,
+      "learning_rate": 1.730239874164373e-06,
+      "loss": 0.823,
+      "step": 491000
+    },
+    {
+      "epoch": 9.66,
+      "learning_rate": 1.6810853322847033e-06,
+      "loss": 0.8264,
+      "step": 491500
+    },
+    {
+      "epoch": 9.67,
+      "learning_rate": 1.6319307904050335e-06,
+      "loss": 0.8226,
+      "step": 492000
+    },
+    {
+      "epoch": 9.68,
+      "learning_rate": 1.5827762485253639e-06,
+      "loss": 0.8084,
+      "step": 492500
+    },
+    {
+      "epoch": 9.69,
+      "learning_rate": 1.533621706645694e-06,
+      "loss": 0.8197,
+      "step": 493000
+    },
+    {
+      "epoch": 9.7,
+      "learning_rate": 1.4844671647660245e-06,
+      "loss": 0.8289,
+      "step": 493500
+    },
+    {
+      "epoch": 9.71,
+      "learning_rate": 1.4353126228863549e-06,
+      "loss": 0.8274,
+      "step": 494000
+    },
+    {
+      "epoch": 9.72,
+      "learning_rate": 1.386158081006685e-06,
+      "loss": 0.8411,
+      "step": 494500
+    },
+    {
+      "epoch": 9.73,
+      "learning_rate": 1.3370035391270154e-06,
+      "loss": 0.8231,
+      "step": 495000
+    },
+    {
+      "epoch": 9.74,
+      "learning_rate": 1.2878489972473456e-06,
+      "loss": 0.8238,
+      "step": 495500
+    },
+    {
+      "epoch": 9.75,
+      "learning_rate": 1.238694455367676e-06,
+      "loss": 0.8203,
+      "step": 496000
+    },
+    {
+      "epoch": 9.76,
+      "learning_rate": 1.1895399134880064e-06,
+      "loss": 0.8342,
+      "step": 496500
+    },
+    {
+      "epoch": 9.77,
+      "learning_rate": 1.1403853716083366e-06,
+      "loss": 0.8391,
+      "step": 497000
+    },
+    {
+      "epoch": 9.78,
+      "learning_rate": 1.091230829728667e-06,
+      "loss": 0.8203,
+      "step": 497500
+    },
+    {
+      "epoch": 9.79,
+      "learning_rate": 1.0420762878489972e-06,
+      "loss": 0.8204,
+      "step": 498000
+    },
+    {
+      "epoch": 9.8,
+      "learning_rate": 9.929217459693276e-07,
+      "loss": 0.8269,
+      "step": 498500
+    },
+    {
+      "epoch": 9.81,
+      "learning_rate": 9.437672040896579e-07,
+      "loss": 0.8294,
+      "step": 499000
+    },
+    {
+      "epoch": 9.82,
+      "learning_rate": 8.946126622099883e-07,
+      "loss": 0.8278,
+      "step": 499500
+    },
+    {
+      "epoch": 9.83,
+      "learning_rate": 8.454581203303186e-07,
+      "loss": 0.8291,
+      "step": 500000
+    },
+    {
+      "epoch": 9.84,
+      "learning_rate": 7.963035784506488e-07,
+      "loss": 0.8261,
+      "step": 500500
+    },
+    {
+      "epoch": 9.85,
+      "learning_rate": 7.471490365709792e-07,
+      "loss": 0.8279,
+      "step": 501000
+    },
+    {
+      "epoch": 9.86,
+      "learning_rate": 6.979944946913095e-07,
+      "loss": 0.8283,
+      "step": 501500
+    },
+    {
+      "epoch": 9.87,
+      "learning_rate": 6.488399528116398e-07,
+      "loss": 0.8279,
+      "step": 502000
+    },
+    {
+      "epoch": 9.88,
+      "learning_rate": 5.996854109319701e-07,
+      "loss": 0.8286,
+      "step": 502500
+    },
+    {
+      "epoch": 9.89,
+      "learning_rate": 5.505308690523005e-07,
+      "loss": 0.8287,
+      "step": 503000
+    },
+    {
+      "epoch": 9.9,
+      "learning_rate": 5.013763271726308e-07,
+      "loss": 0.8136,
+      "step": 503500
+    },
+    {
+      "epoch": 9.91,
+      "learning_rate": 4.522217852929611e-07,
+      "loss": 0.8294,
+      "step": 504000
+    },
+    {
+      "epoch": 9.92,
+      "learning_rate": 4.030672434132914e-07,
+      "loss": 0.8238,
+      "step": 504500
+    },
+    {
+      "epoch": 9.93,
+      "learning_rate": 3.5391270153362173e-07,
+      "loss": 0.8198,
+      "step": 505000
+    },
+    {
+      "epoch": 9.94,
+      "learning_rate": 3.0475815965395203e-07,
+      "loss": 0.8169,
+      "step": 505500
+    },
+    {
+      "epoch": 9.95,
+      "learning_rate": 2.556036177742823e-07,
+      "loss": 0.8261,
+      "step": 506000
+    },
+    {
+      "epoch": 9.96,
+      "learning_rate": 2.0644907589461265e-07,
+      "loss": 0.8285,
+      "step": 506500
+    },
+    {
+      "epoch": 9.97,
+      "learning_rate": 1.57294534014943e-07,
+      "loss": 0.8292,
+      "step": 507000
+    },
+    {
+      "epoch": 9.98,
+      "learning_rate": 1.0813999213527332e-07,
+      "loss": 0.8324,
+      "step": 507500
+    },
+    {
+      "epoch": 9.99,
+      "learning_rate": 5.898545025560362e-08,
+      "loss": 0.8205,
+      "step": 508000
+    },
+    {
+      "epoch": 10.0,
+      "learning_rate": 9.830908375933937e-09,
+      "loss": 0.8153,
+      "step": 508500
+    },
+    {
+      "epoch": 10.0,
+      "step": 508600,
+      "total_flos": 4.55385938558976e+17,
+      "train_loss": 1.0178077638950866,
+      "train_runtime": 115805.9671,
+      "train_samples_per_second": 52.702,
+      "train_steps_per_second": 4.392
+    }
+  ],
+  "max_steps": 508600,
+  "num_train_epochs": 10,
+  "total_flos": 4.55385938558976e+17,
+  "trial_name": null,
+  "trial_params": null
+}