Model save

README.md

@@ -19,7 +19,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on the generator dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.9361
+- Loss: 0.9355
 
 ## Model description
 
@@ -48,13 +48,14 @@ The following hyperparameters were used during training:
 - total_eval_batch_size: 64
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: cosine
+- lr_scheduler_warmup_ratio: 0.1
 - num_epochs: 1
 
 ### Training results
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 0.8975        | 1.0   | 1090 | 0.9361          |
+| 0.9077        | 1.0   | 1090 | 0.9355          |
 
 
 ### Framework versions

all_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 1.0,
-    "eval_loss": 0.9360852241516113,
-    "eval_runtime": 142.4913,
+    "eval_loss": 0.9354520440101624,
+    "eval_runtime": 141.6467,
     "eval_samples": 23110,
-    "eval_samples_per_second": 108.294,
-    "eval_steps_per_second": 1.698,
-    "train_loss": 0.07633398782222643,
-    "train_runtime": 563.604,
+    "eval_samples_per_second": 108.94,
+    "eval_steps_per_second": 1.708,
+    "train_loss": 0.977718904259008,
+    "train_runtime": 5540.1569,
     "train_samples": 207865,
-    "train_samples_per_second": 247.383,
-    "train_steps_per_second": 1.934
+    "train_samples_per_second": 25.166,
+    "train_steps_per_second": 0.197
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 1.0,
-    "eval_loss": 0.9360852241516113,
-    "eval_runtime": 142.4913,
+    "eval_loss": 0.9354520440101624,
+    "eval_runtime": 141.6467,
     "eval_samples": 23110,
-    "eval_samples_per_second": 108.294,
-    "eval_steps_per_second": 1.698
+    "eval_samples_per_second": 108.94,
+    "eval_steps_per_second": 1.708
 }

model-00001-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d962f5f669d3b46ec1ee3f616d374040c5b5341d94e31eb87c37b783850a6941
+oid sha256:95cac20d250ebe1af880f799daa7651f8a4cde980ef67f7f35ef5f737f4626b8
 size 4943162336

model-00002-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:70645b2075d65260c14c0cd5b6587092bf6ad7baf2ec58bacb170f20d06b8453
+oid sha256:ccdad8425df2b803dc7832086d05d28e923ad35019e463773d78567a15096eb6
 size 4999819336

model-00003-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:44584b41e975a724de8f368833fca905f60dd80bdc3307178aab1a1b9ca18e07
+oid sha256:2c0c5d235e699c06e411975e75c6a2b5bd15ed9b273c417d6baef9b04c6c218b
 size 4540516344

runs/Jan04_14-03-58_ip-26-0-163-58/events.out.tfevents.1704377069.ip-26-0-163-58.3278807.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8a9edec07680a7087d366c8ec17a8595bbca6ddcf0eba2680881eeea78f198b8
-size 35909
+oid sha256:58f5f119f271f171d1b9d736262f27102765c9826b7867a65b31958c527376a1
+size 39360

runs/Jan04_14-03-58_ip-26-0-163-58/events.out.tfevents.1704382750.ip-26-0-163-58.3278807.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5fc7fdb8733d06d9d26f73aacd1fd49b2bfd93440f9b4a7eda3ae33d1586c25b
+size 359

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 1.0,
-    "train_loss": 0.07633398782222643,
-    "train_runtime": 563.604,
+    "train_loss": 0.977718904259008,
+    "train_runtime": 5540.1569,
     "train_samples": 207865,
-    "train_samples_per_second": 247.383,
-    "train_steps_per_second": 1.934
+    "train_samples_per_second": 25.166,
+    "train_steps_per_second": 0.197
 }

trainer_state.json

@@ -10,1334 +10,1334 @@
   "log_history": [
     {
       "epoch": 0.0,
-      "learning_rate": 1.9999958464785925e-05,
+      "learning_rate": 1.8348623853211012e-07,
       "loss": 1.6972,
       "step": 1
     },
     {
       "epoch": 0.0,
-      "learning_rate": 1.9998961636899736e-05,
-      "loss": 2.1281,
+      "learning_rate": 9.174311926605506e-07,
+      "loss": 1.6464,
       "step": 5
     },
     {
       "epoch": 0.01,
-      "learning_rate": 1.9995846763238514e-05,
-      "loss": 1.4126,
+      "learning_rate": 1.8348623853211011e-06,
+      "loss": 1.3918,
       "step": 10
     },
     {
       "epoch": 0.01,
-      "learning_rate": 1.9990656025890315e-05,
-      "loss": 1.2437,
+      "learning_rate": 2.7522935779816517e-06,
+      "loss": 1.214,
       "step": 15
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.9983390502829168e-05,
-      "loss": 1.1895,
+      "learning_rate": 3.6697247706422022e-06,
+      "loss": 1.1699,
       "step": 20
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.997405170290528e-05,
-      "loss": 1.1382,
+      "learning_rate": 4.587155963302753e-06,
+      "loss": 1.1172,
       "step": 25
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.9962641565531694e-05,
-      "loss": 1.0928,
+      "learning_rate": 5.504587155963303e-06,
+      "loss": 1.0704,
       "step": 30
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.994916246028154e-05,
-      "loss": 1.0789,
+      "learning_rate": 6.422018348623854e-06,
+      "loss": 1.0575,
       "step": 35
     },
     {
       "epoch": 0.04,
-      "learning_rate": 1.9933617186395917e-05,
-      "loss": 1.0751,
+      "learning_rate": 7.3394495412844045e-06,
+      "loss": 1.052,
       "step": 40
     },
     {
       "epoch": 0.04,
-      "learning_rate": 1.9916008972202586e-05,
-      "loss": 1.07,
+      "learning_rate": 8.256880733944956e-06,
+      "loss": 1.047,
       "step": 45
     },
     {
       "epoch": 0.05,
-      "learning_rate": 1.9896341474445526e-05,
-      "loss": 1.0541,
+      "learning_rate": 9.174311926605506e-06,
+      "loss": 1.0309,
       "step": 50
     },
     {
       "epoch": 0.05,
-      "learning_rate": 1.987461877752552e-05,
-      "loss": 1.0326,
+      "learning_rate": 1.0091743119266055e-05,
+      "loss": 1.0092,
       "step": 55
     },
     {
       "epoch": 0.06,
-      "learning_rate": 1.985084539265195e-05,
-      "loss": 1.0559,
+      "learning_rate": 1.1009174311926607e-05,
+      "loss": 1.0362,
       "step": 60
     },
     {
       "epoch": 0.06,
-      "learning_rate": 1.982502625690595e-05,
-      "loss": 1.0275,
+      "learning_rate": 1.1926605504587156e-05,
+      "loss": 1.0082,
       "step": 65
     },
     {
       "epoch": 0.06,
-      "learning_rate": 1.9797166732215078e-05,
-      "loss": 1.0187,
+      "learning_rate": 1.2844036697247708e-05,
+      "loss": 1.0003,
       "step": 70
     },
     {
       "epoch": 0.07,
-      "learning_rate": 1.9767272604239823e-05,
-      "loss": 1.017,
+      "learning_rate": 1.3761467889908258e-05,
+      "loss": 1.0003,
       "step": 75
     },
     {
       "epoch": 0.07,
-      "learning_rate": 1.973535008117207e-05,
-      "loss": 1.0065,
+      "learning_rate": 1.4678899082568809e-05,
+      "loss": 0.9919,
       "step": 80
     },
     {
       "epoch": 0.08,
-      "learning_rate": 1.9701405792445815e-05,
-      "loss": 1.0279,
+      "learning_rate": 1.559633027522936e-05,
+      "loss": 1.0151,
       "step": 85
     },
     {
       "epoch": 0.08,
-      "learning_rate": 1.9665446787360444e-05,
-      "loss": 1.0082,
+      "learning_rate": 1.6513761467889912e-05,
+      "loss": 0.9974,
       "step": 90
     },
     {
       "epoch": 0.09,
-      "learning_rate": 1.962748053361675e-05,
-      "loss": 1.0237,
+      "learning_rate": 1.743119266055046e-05,
+      "loss": 1.0154,
       "step": 95
     },
     {
       "epoch": 0.09,
-      "learning_rate": 1.9587514915766124e-05,
-      "loss": 1.0222,
+      "learning_rate": 1.834862385321101e-05,
+      "loss": 1.0163,
       "step": 100
     },
     {
       "epoch": 0.1,
-      "learning_rate": 1.9545558233573136e-05,
-      "loss": 0.9915,
+      "learning_rate": 1.9266055045871563e-05,
+      "loss": 0.9874,
       "step": 105
     },
     {
       "epoch": 0.1,
-      "learning_rate": 1.950161920029191e-05,
-      "loss": 1.0055,
+      "learning_rate": 1.999994872196626e-05,
+      "loss": 1.0056,
       "step": 110
     },
     {
       "epoch": 0.11,
-      "learning_rate": 1.9455706940856602e-05,
-      "loss": 1.0128,
+      "learning_rate": 1.9998154046002822e-05,
+      "loss": 1.0142,
       "step": 115
     },
     {
       "epoch": 0.11,
-      "learning_rate": 1.940783098998643e-05,
-      "loss": 1.0022,
+      "learning_rate": 1.999379599421534e-05,
+      "loss": 1.0058,
       "step": 120
     },
     {
       "epoch": 0.11,
-      "learning_rate": 1.9358001290205542e-05,
-      "loss": 1.0152,
+      "learning_rate": 1.9986875683942535e-05,
+      "loss": 1.0201,
       "step": 125
     },
     {
       "epoch": 0.12,
-      "learning_rate": 1.9306228189778255e-05,
-      "loss": 1.0172,
+      "learning_rate": 1.9977394889447526e-05,
+      "loss": 1.0239,
       "step": 130
     },
     {
       "epoch": 0.12,
-      "learning_rate": 1.925252244055998e-05,
-      "loss": 1.0245,
+      "learning_rate": 1.9965356041462954e-05,
+      "loss": 1.0305,
       "step": 135
     },
     {
       "epoch": 0.13,
-      "learning_rate": 1.9196895195764363e-05,
-      "loss": 1.021,
+      "learning_rate": 1.9950762226567783e-05,
+      "loss": 1.0284,
       "step": 140
     },
     {
       "epoch": 0.13,
-      "learning_rate": 1.9139358007647085e-05,
-      "loss": 0.9968,
+      "learning_rate": 1.9933617186395917e-05,
+      "loss": 1.0047,
       "step": 145
     },
     {
       "epoch": 0.14,
-      "learning_rate": 1.907992282510675e-05,
-      "loss": 0.9947,
+      "learning_rate": 1.9913925316676946e-05,
+      "loss": 1.0028,
       "step": 150
     },
     {
       "epoch": 0.14,
-      "learning_rate": 1.901860199120344e-05,
-      "loss": 1.0144,
+      "learning_rate": 1.9891691666109112e-05,
+      "loss": 1.023,
       "step": 155
     },
     {
       "epoch": 0.15,
-      "learning_rate": 1.8955408240595396e-05,
-      "loss": 1.0033,
+      "learning_rate": 1.9866921935064907e-05,
+      "loss": 1.015,
       "step": 160
     },
     {
       "epoch": 0.15,
-      "learning_rate": 1.8890354696894374e-05,
-      "loss": 1.0049,
+      "learning_rate": 1.9839622474129595e-05,
+      "loss": 1.0164,
       "step": 165
     },
     {
       "epoch": 0.16,
-      "learning_rate": 1.8823454869940243e-05,
-      "loss": 0.9873,
+      "learning_rate": 1.9809800282473014e-05,
+      "loss": 0.9977,
       "step": 170
     },
     {
       "epoch": 0.16,
-      "learning_rate": 1.8754722652995346e-05,
-      "loss": 1.0093,
+      "learning_rate": 1.977746300605507e-05,
+      "loss": 1.0214,
       "step": 175
     },
     {
       "epoch": 0.17,
-      "learning_rate": 1.8684172319859258e-05,
-      "loss": 0.9972,
+      "learning_rate": 1.9742618935665478e-05,
+      "loss": 1.0083,
       "step": 180
     },
     {
       "epoch": 0.17,
-      "learning_rate": 1.861181852190451e-05,
-      "loss": 0.9873,
+      "learning_rate": 1.9705277004798072e-05,
+      "loss": 0.998,
       "step": 185
     },
     {
       "epoch": 0.17,
-      "learning_rate": 1.8537676285033886e-05,
-      "loss": 1.0001,
+      "learning_rate": 1.9665446787360444e-05,
+      "loss": 1.0102,
       "step": 190
     },
     {
       "epoch": 0.18,
-      "learning_rate": 1.8461761006559982e-05,
-      "loss": 0.9967,
+      "learning_rate": 1.9623138495219292e-05,
+      "loss": 1.0084,
       "step": 195
     },
     {
       "epoch": 0.18,
-      "learning_rate": 1.838408845200758e-05,
-      "loss": 1.0163,
+      "learning_rate": 1.957836297558229e-05,
+      "loss": 1.0272,
       "step": 200
     },
     {
       "epoch": 0.19,
-      "learning_rate": 1.8304674751839583e-05,
-      "loss": 0.9769,
+      "learning_rate": 1.9531131708217005e-05,
+      "loss": 0.989,
       "step": 205
     },
     {
       "epoch": 0.19,
-      "learning_rate": 1.8223536398107177e-05,
-      "loss": 0.9999,
+      "learning_rate": 1.948145680250766e-05,
+      "loss": 1.0116,
       "step": 210
     },
     {
       "epoch": 0.2,
-      "learning_rate": 1.8140690241024872e-05,
-      "loss": 0.9799,
+      "learning_rate": 1.9429350994350483e-05,
+      "loss": 0.9931,
       "step": 215
     },
     {
       "epoch": 0.2,
-      "learning_rate": 1.8056153485471167e-05,
-      "loss": 1.0117,
+      "learning_rate": 1.93748276428884e-05,
+      "loss": 1.0231,
       "step": 220
     },
     {
       "epoch": 0.21,
-      "learning_rate": 1.7969943687415575e-05,
-      "loss": 1.0062,
+      "learning_rate": 1.931790072708596e-05,
+      "loss": 1.0191,
       "step": 225
     },
     {
       "epoch": 0.21,
-      "learning_rate": 1.788207875027274e-05,
-      "loss": 0.9884,
+      "learning_rate": 1.9258584842145342e-05,
+      "loss": 1.0017,
       "step": 230
     },
     {
       "epoch": 0.22,
-      "learning_rate": 1.7792576921184374e-05,
-      "loss": 0.983,
+      "learning_rate": 1.9196895195764363e-05,
+      "loss": 0.9957,
       "step": 235
     },
     {
       "epoch": 0.22,
-      "learning_rate": 1.7701456787229805e-05,
-      "loss": 1.0082,
+      "learning_rate": 1.913284760423745e-05,
+      "loss": 1.0214,
       "step": 240
     },
     {
       "epoch": 0.22,
-      "learning_rate": 1.7608737271566004e-05,
-      "loss": 1.0129,
+      "learning_rate": 1.9066458488400586e-05,
+      "loss": 1.0259,
       "step": 245
     },
     {
       "epoch": 0.23,
-      "learning_rate": 1.751443762949772e-05,
-      "loss": 0.9906,
+      "learning_rate": 1.8997744869421248e-05,
+      "loss": 1.0034,
       "step": 250
     },
     {
       "epoch": 0.23,
-      "learning_rate": 1.741857744447869e-05,
-      "loss": 0.9831,
+      "learning_rate": 1.8926724364434447e-05,
+      "loss": 0.9952,
       "step": 255
     },
     {
       "epoch": 0.24,
-      "learning_rate": 1.732117662404469e-05,
-      "loss": 0.9964,
+      "learning_rate": 1.8853415182025953e-05,
+      "loss": 1.0086,
       "step": 260
     },
     {
       "epoch": 0.24,
-      "learning_rate": 1.7222255395679298e-05,
-      "loss": 1.0063,
+      "learning_rate": 1.8777836117563894e-05,
+      "loss": 1.0196,
       "step": 265
     },
     {
       "epoch": 0.25,
-      "learning_rate": 1.712183430261319e-05,
-      "loss": 0.9783,
+      "learning_rate": 1.8700006548379898e-05,
+      "loss": 0.9918,
       "step": 270
     },
     {
       "epoch": 0.25,
-      "learning_rate": 1.7019934199557868e-05,
-      "loss": 0.9893,
+      "learning_rate": 1.861994642880105e-05,
+      "loss": 1.0037,
       "step": 275
     },
     {
       "epoch": 0.26,
-      "learning_rate": 1.691657624837472e-05,
-      "loss": 0.9931,
+      "learning_rate": 1.8537676285033886e-05,
+      "loss": 1.0062,
       "step": 280
     },
     {
       "epoch": 0.26,
-      "learning_rate": 1.6811781913680273e-05,
-      "loss": 0.9965,
+      "learning_rate": 1.845321720990181e-05,
+      "loss": 1.0095,
       "step": 285
     },
     {
       "epoch": 0.27,
-      "learning_rate": 1.6705572958388576e-05,
-      "loss": 1.0013,
+      "learning_rate": 1.8366590857437182e-05,
+      "loss": 1.0152,
       "step": 290
     },
     {
       "epoch": 0.27,
-      "learning_rate": 1.659797143919165e-05,
-      "loss": 0.9802,
+      "learning_rate": 1.8277819437329577e-05,
+      "loss": 0.993,
       "step": 295
     },
     {
       "epoch": 0.28,
-      "learning_rate": 1.6488999701978905e-05,
-      "loss": 0.9741,
+      "learning_rate": 1.8186925709231534e-05,
+      "loss": 0.9875,
       "step": 300
     },
     {
       "epoch": 0.28,
-      "learning_rate": 1.6378680377196526e-05,
-      "loss": 0.9751,
+      "learning_rate": 1.809393297692334e-05,
+      "loss": 0.989,
       "step": 305
     },
     {
       "epoch": 0.28,
-      "learning_rate": 1.6267036375147728e-05,
-      "loss": 0.9728,
+      "learning_rate": 1.799886508233829e-05,
+      "loss": 0.9866,
       "step": 310
     },
     {
       "epoch": 0.29,
-      "learning_rate": 1.615409088123493e-05,
-      "loss": 0.9754,
+      "learning_rate": 1.790174639944997e-05,
+      "loss": 0.9894,
       "step": 315
     },
     {
       "epoch": 0.29,
-      "learning_rate": 1.6039867351144778e-05,
-      "loss": 0.984,
+      "learning_rate": 1.780260182802314e-05,
+      "loss": 0.9981,
       "step": 320
     },
     {
       "epoch": 0.3,
-      "learning_rate": 1.5924389505977038e-05,
-      "loss": 0.9737,
+      "learning_rate": 1.7701456787229805e-05,
+      "loss": 0.9877,
       "step": 325
     },
     {
       "epoch": 0.3,
-      "learning_rate": 1.5807681327318372e-05,
-      "loss": 0.996,
+      "learning_rate": 1.7598337209132142e-05,
+      "loss": 1.0103,
       "step": 330
     },
     {
       "epoch": 0.31,
-      "learning_rate": 1.5689767052262028e-05,
-      "loss": 0.9795,
+      "learning_rate": 1.7493269532033882e-05,
+      "loss": 0.9938,
       "step": 335
     },
     {
       "epoch": 0.31,
-      "learning_rate": 1.557067116837444e-05,
-      "loss": 0.984,
+      "learning_rate": 1.738628069370195e-05,
+      "loss": 0.9979,
       "step": 340
     },
     {
       "epoch": 0.32,
-      "learning_rate": 1.545041840860986e-05,
-      "loss": 0.9932,
+      "learning_rate": 1.7277398124460022e-05,
+      "loss": 1.007,
       "step": 345
     },
     {
       "epoch": 0.32,
-      "learning_rate": 1.5329033746173975e-05,
-      "loss": 0.9787,
+      "learning_rate": 1.71666497401558e-05,
+      "loss": 0.9926,
       "step": 350
     },
     {
       "epoch": 0.33,
-      "learning_rate": 1.520654238933767e-05,
-      "loss": 0.9825,
+      "learning_rate": 1.7054063935003813e-05,
+      "loss": 0.9971,
       "step": 355
     },
     {
       "epoch": 0.33,
-      "learning_rate": 1.5082969776201948e-05,
-      "loss": 0.9931,
+      "learning_rate": 1.6939669574305565e-05,
+      "loss": 1.0069,
       "step": 360
     },
     {
       "epoch": 0.33,
-      "learning_rate": 1.4958341569415149e-05,
-      "loss": 0.9568,
+      "learning_rate": 1.6823495987048922e-05,
+      "loss": 0.9705,
       "step": 365
     },
     {
       "epoch": 0.34,
-      "learning_rate": 1.483268365084351e-05,
-      "loss": 0.9846,
+      "learning_rate": 1.6705572958388576e-05,
+      "loss": 0.999,
       "step": 370
     },
     {
       "epoch": 0.34,
-      "learning_rate": 1.4706022116196208e-05,
-      "loss": 0.9818,
+      "learning_rate": 1.6585930722009602e-05,
+      "loss": 0.9952,
       "step": 375
     },
     {
       "epoch": 0.35,
-      "learning_rate": 1.4578383269606004e-05,
-      "loss": 0.984,
+      "learning_rate": 1.6464599952375998e-05,
+      "loss": 0.9972,
       "step": 380
     },
     {
       "epoch": 0.35,
-      "learning_rate": 1.4449793618166594e-05,
-      "loss": 0.9678,
+      "learning_rate": 1.63416117568662e-05,
+      "loss": 0.9813,
       "step": 385
     },
     {
       "epoch": 0.36,
-      "learning_rate": 1.4320279866427798e-05,
-      "loss": 0.9944,
+      "learning_rate": 1.621699766779763e-05,
+      "loss": 1.0089,
       "step": 390
     },
     {
       "epoch": 0.36,
-      "learning_rate": 1.4189868910849779e-05,
-      "loss": 0.9879,
+      "learning_rate": 1.6090789634342278e-05,
+      "loss": 1.0026,
       "step": 395
     },
     {
       "epoch": 0.37,
-      "learning_rate": 1.4058587834217356e-05,
-      "loss": 0.9791,
+      "learning_rate": 1.5963020014335437e-05,
+      "loss": 0.9933,
       "step": 400
     },
     {
       "epoch": 0.37,
-      "learning_rate": 1.392646390001569e-05,
-      "loss": 0.9614,
+      "learning_rate": 1.583372156597961e-05,
+      "loss": 0.9754,
       "step": 405
     },
     {
       "epoch": 0.38,
-      "learning_rate": 1.3793524546768358e-05,
-      "loss": 0.996,
+      "learning_rate": 1.570292743944583e-05,
+      "loss": 1.0099,
       "step": 410
     },
     {
       "epoch": 0.38,
-      "learning_rate": 1.3659797382339162e-05,
-      "loss": 0.9549,
+      "learning_rate": 1.557067116837444e-05,
+      "loss": 0.9682,
       "step": 415
     },
     {
       "epoch": 0.39,
-      "learning_rate": 1.3525310178198707e-05,
-      "loss": 0.9819,
+      "learning_rate": 1.5436986661277578e-05,
+      "loss": 0.9963,
       "step": 420
     },
     {
       "epoch": 0.39,
-      "learning_rate": 1.3390090863657048e-05,
-      "loss": 0.9993,
+      "learning_rate": 1.530190819284555e-05,
+      "loss": 1.0133,
       "step": 425
     },
     {
       "epoch": 0.39,
-      "learning_rate": 1.325416752006351e-05,
-      "loss": 0.9746,
+      "learning_rate": 1.5165470395159314e-05,
+      "loss": 0.9886,
       "step": 430
     },
     {
       "epoch": 0.4,
-      "learning_rate": 1.311756837497499e-05,
-      "loss": 0.9756,
+      "learning_rate": 1.5027708248811331e-05,
+      "loss": 0.9891,
       "step": 435
     },
     {
       "epoch": 0.4,
-      "learning_rate": 1.2980321796293838e-05,
-      "loss": 0.9647,
+      "learning_rate": 1.4888657073937077e-05,
+      "loss": 0.9792,
       "step": 440
     },
     {
       "epoch": 0.41,
-      "learning_rate": 1.284245628637665e-05,
-      "loss": 0.963,
+      "learning_rate": 1.4748352521159492e-05,
+      "loss": 0.9769,
       "step": 445
     },
     {
       "epoch": 0.41,
-      "learning_rate": 1.2704000476115079e-05,
-      "loss": 0.974,
+      "learning_rate": 1.4606830562448692e-05,
+      "loss": 0.9878,
       "step": 450
     },
     {
       "epoch": 0.42,
-      "learning_rate": 1.256498311899001e-05,
-      "loss": 0.9874,
+      "learning_rate": 1.4464127481899312e-05,
+      "loss": 1.0,
       "step": 455
     },
     {
       "epoch": 0.42,
-      "learning_rate": 1.2425433085100224e-05,
-      "loss": 0.9654,
+      "learning_rate": 1.4320279866427798e-05,
+      "loss": 0.9789,
       "step": 460
     },
     {
       "epoch": 0.43,
-      "learning_rate": 1.2285379355166893e-05,
-      "loss": 0.9704,
+      "learning_rate": 1.4175324596392075e-05,
+      "loss": 0.9832,
       "step": 465
     },
     {
       "epoch": 0.43,
-      "learning_rate": 1.2144851014515055e-05,
-      "loss": 0.9639,
+      "learning_rate": 1.402929883613599e-05,
+      "loss": 0.9766,
       "step": 470
     },
     {
       "epoch": 0.44,
-      "learning_rate": 1.2003877247033411e-05,
-      "loss": 0.9734,
+      "learning_rate": 1.3882240024460928e-05,
+      "loss": 0.9861,
       "step": 475
     },
     {
       "epoch": 0.44,
-      "learning_rate": 1.1862487329113606e-05,
-      "loss": 0.971,
+      "learning_rate": 1.3734185865027061e-05,
+      "loss": 0.9832,
       "step": 480
     },
     {
       "epoch": 0.44,
-      "learning_rate": 1.172071062357035e-05,
-      "loss": 0.9701,
+      "learning_rate": 1.358517431668672e-05,
+      "loss": 0.9832,
       "step": 485
     },
     {
       "epoch": 0.45,
-      "learning_rate": 1.1578576573543541e-05,
-      "loss": 0.9624,
+      "learning_rate": 1.3435243583752294e-05,
+      "loss": 0.9752,
       "step": 490
     },
     {
       "epoch": 0.45,
-      "learning_rate": 1.1436114696383749e-05,
-      "loss": 0.9673,
+      "learning_rate": 1.3284432106201233e-05,
+      "loss": 0.9799,
       "step": 495
     },
     {
       "epoch": 0.46,
-      "learning_rate": 1.1293354577522264e-05,
-      "loss": 0.9452,
+      "learning_rate": 1.313277854982062e-05,
+      "loss": 0.9579,
       "step": 500
     },
     {
       "epoch": 0.46,
-      "learning_rate": 1.1150325864327003e-05,
-      "loss": 0.977,
+      "learning_rate": 1.2980321796293838e-05,
+      "loss": 0.9886,
       "step": 505
     },
     {
       "epoch": 0.47,
-      "learning_rate": 1.1007058259945584e-05,
-      "loss": 0.9575,
+      "learning_rate": 1.2827100933231904e-05,
+      "loss": 0.9696,
       "step": 510
     },
     {
       "epoch": 0.47,
-      "learning_rate": 1.0863581517136776e-05,
-      "loss": 0.9677,
+      "learning_rate": 1.2673155244151985e-05,
+      "loss": 0.9798,
       "step": 515
     },
     {
       "epoch": 0.48,
-      "learning_rate": 1.0719925432091671e-05,
-      "loss": 0.9542,
+      "learning_rate": 1.2518524198405699e-05,
+      "loss": 0.9663,
       "step": 520
     },
     {
       "epoch": 0.48,
-      "learning_rate": 1.0576119838245843e-05,
-      "loss": 0.9599,
+      "learning_rate": 1.2363247441059775e-05,
+      "loss": 0.9711,
       "step": 525
     },
     {
       "epoch": 0.49,
-      "learning_rate": 1.043219460008374e-05,
-      "loss": 0.969,
+      "learning_rate": 1.2207364782731657e-05,
+      "loss": 0.98,
       "step": 530
     },
     {
       "epoch": 0.49,
-      "learning_rate": 1.0288179606936666e-05,
-      "loss": 0.9527,
+      "learning_rate": 1.2050916189382646e-05,
+      "loss": 0.9636,
       "step": 535
     },
     {
       "epoch": 0.5,
-      "learning_rate": 1.0144104766775574e-05,
-      "loss": 0.972,
+      "learning_rate": 1.189394177207125e-05,
+      "loss": 0.9826,
       "step": 540
     },
     {
       "epoch": 0.5,
-      "learning_rate": 1e-05,
-      "loss": 0.9699,
+      "learning_rate": 1.1736481776669307e-05,
+      "loss": 0.9807,
       "step": 545
     },
     {
       "epoch": 0.5,
-      "learning_rate": 9.855895233224431e-06,
-      "loss": 0.9501,
+      "learning_rate": 1.1578576573543541e-05,
+      "loss": 0.9614,
       "step": 550
     },
     {
       "epoch": 0.51,
-      "learning_rate": 9.711820393063337e-06,
-      "loss": 0.9573,
+      "learning_rate": 1.1420266647205232e-05,
+      "loss": 0.9692,
       "step": 555
     },
     {
       "epoch": 0.51,
-      "learning_rate": 9.56780539991626e-06,
-      "loss": 0.9764,
+      "learning_rate": 1.1261592585930576e-05,
+      "loss": 0.9877,
       "step": 560
     },
     {
       "epoch": 0.52,
-      "learning_rate": 9.423880161754158e-06,
-      "loss": 0.9346,
+      "learning_rate": 1.1102595071354471e-05,
+      "loss": 0.9449,
       "step": 565
     },
     {
       "epoch": 0.52,
-      "learning_rate": 9.28007456790833e-06,
-      "loss": 0.9478,
+      "learning_rate": 1.0943314868040365e-05,
+      "loss": 0.9583,
       "step": 570
     },
     {
       "epoch": 0.53,
-      "learning_rate": 9.13641848286323e-06,
-      "loss": 0.9674,
+      "learning_rate": 1.0783792813028828e-05,
+      "loss": 0.9779,
       "step": 575
     },
     {
       "epoch": 0.53,
-      "learning_rate": 8.992941740054418e-06,
-      "loss": 0.9302,
+      "learning_rate": 1.0624069805367558e-05,
+      "loss": 0.9404,
       "step": 580
     },
     {
       "epoch": 0.54,
-      "learning_rate": 8.849674135672999e-06,
-      "loss": 0.9536,
+      "learning_rate": 1.0464186795625481e-05,
+      "loss": 0.9636,
       "step": 585
     },
     {
       "epoch": 0.54,
-      "learning_rate": 8.706645422477739e-06,
-      "loss": 0.9476,
+      "learning_rate": 1.0304184775393642e-05,
+      "loss": 0.9574,
       "step": 590
     },
     {
       "epoch": 0.55,
-      "learning_rate": 8.563885303616253e-06,
-      "loss": 0.9606,
+      "learning_rate": 1.0144104766775574e-05,
+      "loss": 0.9702,
       "step": 595
     },
     {
       "epoch": 0.55,
-      "learning_rate": 8.42142342645646e-06,
-      "loss": 0.9574,
+      "learning_rate": 9.983987811869863e-06,
+      "loss": 0.9671,
       "step": 600
     },
     {
       "epoch": 0.56,
-      "learning_rate": 8.279289376429653e-06,
-      "loss": 0.9551,
+      "learning_rate": 9.823874962247565e-06,
+      "loss": 0.9655,
       "step": 605
     },
     {
       "epoch": 0.56,
-      "learning_rate": 8.137512670886397e-06,
-      "loss": 0.9431,
+      "learning_rate": 9.663807268427197e-06,
+      "loss": 0.9518,
       "step": 610
     },
     {
       "epoch": 0.56,
-      "learning_rate": 7.996122752966596e-06,
-      "loss": 0.9636,
+      "learning_rate": 9.503825769350016e-06,
+      "loss": 0.9739,
       "step": 615
     },
     {
       "epoch": 0.57,
-      "learning_rate": 7.855148985484946e-06,
-      "loss": 0.9409,
+      "learning_rate": 9.343971481858246e-06,
+      "loss": 0.949,
       "step": 620
     },
     {
       "epoch": 0.57,
-      "learning_rate": 7.71462064483311e-06,
-      "loss": 0.955,
+      "learning_rate": 9.184285390178978e-06,
+      "loss": 0.9641,
       "step": 625
     },
     {
       "epoch": 0.58,
-      "learning_rate": 7.574566914899779e-06,
-      "loss": 0.9336,
+      "learning_rate": 9.024808435416435e-06,
+      "loss": 0.9431,
       "step": 630
     },
     {
       "epoch": 0.58,
-      "learning_rate": 7.4350168810099955e-06,
-      "loss": 0.9423,
+      "learning_rate": 8.865581505055292e-06,
+      "loss": 0.9504,
       "step": 635
     },
     {
       "epoch": 0.59,
-      "learning_rate": 7.295999523884921e-06,
-      "loss": 0.9544,
+      "learning_rate": 8.706645422477739e-06,
+      "loss": 0.962,
       "step": 640
     },
     {
       "epoch": 0.59,
-      "learning_rate": 7.157543713623353e-06,
-      "loss": 0.9384,
+      "learning_rate": 8.548040936496989e-06,
+      "loss": 0.9456,
       "step": 645
     },
     {
       "epoch": 0.6,
-      "learning_rate": 7.019678203706164e-06,
-      "loss": 0.9662,
+      "learning_rate": 8.389808710909881e-06,
+      "loss": 0.9738,
       "step": 650
     },
     {
       "epoch": 0.6,
-      "learning_rate": 6.882431625025016e-06,
-      "loss": 0.9527,
+      "learning_rate": 8.231989314071318e-06,
+      "loss": 0.9611,
       "step": 655
     },
     {
       "epoch": 0.61,
-      "learning_rate": 6.745832479936492e-06,
-      "loss": 0.952,
+      "learning_rate": 8.07462320849313e-06,
+      "loss": 0.9601,
       "step": 660
     },
     {
       "epoch": 0.61,
-      "learning_rate": 6.609909136342956e-06,
-      "loss": 0.9614,
+      "learning_rate": 7.917750740470116e-06,
+      "loss": 0.9699,
       "step": 665
     },
     {
       "epoch": 0.61,
-      "learning_rate": 6.474689821801295e-06,
-      "loss": 0.9555,
+      "learning_rate": 7.761412129735853e-06,
+      "loss": 0.9627,
       "step": 670
     },
     {
       "epoch": 0.62,
-      "learning_rate": 6.340202617660842e-06,
-      "loss": 0.9416,
+      "learning_rate": 7.605647459150961e-06,
+      "loss": 0.9486,
       "step": 675
     },
     {
       "epoch": 0.62,
-      "learning_rate": 6.206475453231644e-06,
-      "loss": 0.9413,
+      "learning_rate": 7.4504966644264775e-06,
+      "loss": 0.9484,
       "step": 680
     },
     {
       "epoch": 0.63,
-      "learning_rate": 6.073536099984314e-06,
-      "loss": 0.9361,
+      "learning_rate": 7.295999523884921e-06,
+      "loss": 0.9425,
       "step": 685
     },
     {
       "epoch": 0.63,
-      "learning_rate": 5.941412165782645e-06,
-      "loss": 0.9594,
+      "learning_rate": 7.142195648261747e-06,
+      "loss": 0.9648,
       "step": 690
     },
     {
       "epoch": 0.64,
-      "learning_rate": 5.810131089150228e-06,
-      "loss": 0.9547,
+      "learning_rate": 6.989124470549746e-06,
+      "loss": 0.9604,
       "step": 695
     },
     {
       "epoch": 0.64,
-      "learning_rate": 5.6797201335722064e-06,
-      "loss": 0.9376,
+      "learning_rate": 6.83682523588902e-06,
+      "loss": 0.9428,
       "step": 700
     },
     {
       "epoch": 0.65,
-      "learning_rate": 5.550206381833409e-06,
-      "loss": 0.9512,
+      "learning_rate": 6.685336991505122e-06,
+      "loss": 0.9575,
       "step": 705
     },
     {
       "epoch": 0.65,
-      "learning_rate": 5.421616730394e-06,
-      "loss": 0.9445,
+      "learning_rate": 6.5346985766979384e-06,
+      "loss": 0.9501,
       "step": 710
     },
     {
       "epoch": 0.66,
-      "learning_rate": 5.293977883803797e-06,
-      "loss": 0.9419,
+      "learning_rate": 6.384948612883872e-06,
+      "loss": 0.9475,
       "step": 715
     },
     {
       "epoch": 0.66,
-      "learning_rate": 5.167316349156495e-06,
-      "loss": 0.9328,
+      "learning_rate": 6.2361254936939e-06,
+      "loss": 0.9386,
       "step": 720
     },
     {
       "epoch": 0.67,
-      "learning_rate": 5.041658430584852e-06,
-      "loss": 0.9453,
+      "learning_rate": 6.0882673751300235e-06,
+      "loss": 0.9501,
       "step": 725
     },
     {
       "epoch": 0.67,
-      "learning_rate": 4.917030223798057e-06,
-      "loss": 0.944,
+      "learning_rate": 5.941412165782645e-06,
+      "loss": 0.9488,
       "step": 730
     },
     {
       "epoch": 0.67,
-      "learning_rate": 4.793457610662334e-06,
-      "loss": 0.9309,
+      "learning_rate": 5.79559751711138e-06,
+      "loss": 0.9351,
       "step": 735
     },
     {
       "epoch": 0.68,
-      "learning_rate": 4.670966253826027e-06,
-      "loss": 0.9394,
+      "learning_rate": 5.650860813791786e-06,
+      "loss": 0.9444,
       "step": 740
     },
     {
       "epoch": 0.68,
-      "learning_rate": 4.549581591390142e-06,
-      "loss": 0.9479,
+      "learning_rate": 5.507239164130501e-06,
+      "loss": 0.9524,
       "step": 745
     },
     {
       "epoch": 0.69,
-      "learning_rate": 4.429328831625565e-06,
-      "loss": 0.9465,
+      "learning_rate": 5.364769390551225e-06,
+      "loss": 0.9511,
       "step": 750
     },
     {
       "epoch": 0.69,
-      "learning_rate": 4.310232947737979e-06,
-      "loss": 0.943,
+      "learning_rate": 5.223488020154028e-06,
+      "loss": 0.9462,
       "step": 755
     },
     {
       "epoch": 0.7,
-      "learning_rate": 4.192318672681631e-06,
-      "loss": 0.9366,
+      "learning_rate": 5.083431275350312e-06,
+      "loss": 0.9397,
       "step": 760
     },
     {
       "epoch": 0.7,
-      "learning_rate": 4.0756104940229645e-06,
-      "loss": 0.9438,
+      "learning_rate": 4.9446350645759885e-06,
+      "loss": 0.9469,
       "step": 765
     },
     {
       "epoch": 0.71,
-      "learning_rate": 3.960132648855226e-06,
-      "loss": 0.9395,
+      "learning_rate": 4.807134973085036e-06,
+      "loss": 0.9439,
       "step": 770
     },
     {
       "epoch": 0.71,
-      "learning_rate": 3.845909118765073e-06,
-      "loss": 0.9472,
+      "learning_rate": 4.670966253826027e-06,
+      "loss": 0.9502,
       "step": 775
     },
     {
       "epoch": 0.72,
-      "learning_rate": 3.732963624852275e-06,
-      "loss": 0.9317,
+      "learning_rate": 4.53616381840377e-06,
+      "loss": 0.9336,
       "step": 780
     },
     {
       "epoch": 0.72,
-      "learning_rate": 3.6213196228034796e-06,
-      "loss": 0.9377,
+      "learning_rate": 4.402762228128531e-06,
+      "loss": 0.9408,
       "step": 785
     },
     {
       "epoch": 0.72,
-      "learning_rate": 3.511000298021098e-06,
-      "loss": 0.9465,
+      "learning_rate": 4.270795685155001e-06,
+      "loss": 0.9483,
       "step": 790
     },
     {
       "epoch": 0.73,
-      "learning_rate": 3.402028560808357e-06,
-      "loss": 0.9264,
+      "learning_rate": 4.140298023713416e-06,
+      "loss": 0.9289,
       "step": 795
     },
     {
       "epoch": 0.73,
-      "learning_rate": 3.2944270416114256e-06,
-      "loss": 0.9412,
+      "learning_rate": 4.0113027014349374e-06,
+      "loss": 0.9439,
       "step": 800
     },
     {
       "epoch": 0.74,
-      "learning_rate": 3.1882180863197308e-06,
-      "loss": 0.9301,
+      "learning_rate": 3.883842790773647e-06,
+      "loss": 0.9325,
       "step": 805
     },
     {
       "epoch": 0.74,
-      "learning_rate": 3.0834237516252817e-06,
-      "loss": 0.9333,
+      "learning_rate": 3.757950970527249e-06,
+      "loss": 0.9364,
       "step": 810
     },
     {
       "epoch": 0.75,
-      "learning_rate": 2.980065800442137e-06,
-      "loss": 0.9302,
+      "learning_rate": 3.633659517458736e-06,
+      "loss": 0.9311,
       "step": 815
     },
     {
       "epoch": 0.75,
-      "learning_rate": 2.878165697386812e-06,
-      "loss": 0.9481,
+      "learning_rate": 3.511000298021098e-06,
+      "loss": 0.9496,
       "step": 820
     },
     {
       "epoch": 0.76,
-      "learning_rate": 2.777744604320706e-06,
-      "loss": 0.9429,
+      "learning_rate": 3.39000476018726e-06,
+      "loss": 0.9441,
       "step": 825
     },
     {
       "epoch": 0.76,
-      "learning_rate": 2.678823375955314e-06,
-      "loss": 0.9343,
+      "learning_rate": 3.2707039253872796e-06,
+      "loss": 0.9352,
       "step": 830
     },
     {
       "epoch": 0.77,
-      "learning_rate": 2.581422555521316e-06,
-      "loss": 0.9414,
+      "learning_rate": 3.153128380554941e-06,
+      "loss": 0.9418,
       "step": 835
     },
     {
       "epoch": 0.77,
-      "learning_rate": 2.485562370502279e-06,
-      "loss": 0.919,
+      "learning_rate": 3.037308270285709e-06,
+      "loss": 0.9204,
       "step": 840
     },
     {
       "epoch": 0.78,
-      "learning_rate": 2.391262728433995e-06,
-      "loss": 0.946,
+      "learning_rate": 2.923273289108115e-06,
+      "loss": 0.9462,
       "step": 845
     },
     {
       "epoch": 0.78,
-      "learning_rate": 2.2985432127701945e-06,
-      "loss": 0.9358,
+      "learning_rate": 2.8110526738705345e-06,
+      "loss": 0.9368,
       "step": 850
     },
     {
       "epoch": 0.78,
-      "learning_rate": 2.2074230788156325e-06,
-      "loss": 0.946,
+      "learning_rate": 2.700675196245288e-06,
+      "loss": 0.9467,
       "step": 855
     },
     {
       "epoch": 0.79,
-      "learning_rate": 2.1179212497272582e-06,
-      "loss": 0.9148,
+      "learning_rate": 2.592169155352031e-06,
+      "loss": 0.9157,
       "step": 860
     },
     {
       "epoch": 0.79,
-      "learning_rate": 2.030056312584424e-06,
-      "loss": 0.9282,
+      "learning_rate": 2.485562370502279e-06,
+      "loss": 0.9283,
       "step": 865
     },
     {
       "epoch": 0.8,
-      "learning_rate": 1.9438465145288377e-06,
-      "loss": 0.9295,
+      "learning_rate": 2.3808821740669608e-06,
+      "loss": 0.9294,
       "step": 870
     },
     {
       "epoch": 0.8,
-      "learning_rate": 1.8593097589751318e-06,
-      "loss": 0.9405,
+      "learning_rate": 2.2781554044688015e-06,
+      "loss": 0.9404,
       "step": 875
     },
     {
       "epoch": 0.81,
-      "learning_rate": 1.7764636018928249e-06,
-      "loss": 0.9349,
+      "learning_rate": 2.1774083993013715e-06,
+      "loss": 0.9363,
       "step": 880
     },
     {
       "epoch": 0.81,
-      "learning_rate": 1.6953252481604198e-06,
-      "loss": 0.935,
+      "learning_rate": 2.0786669885765044e-06,
+      "loss": 0.9344,
       "step": 885
     },
     {
       "epoch": 0.82,
-      "learning_rate": 1.6159115479924259e-06,
-      "loss": 0.9241,
+      "learning_rate": 1.981956488101898e-06,
+      "loss": 0.9245,
       "step": 890
     },
     {
       "epoch": 0.82,
-      "learning_rate": 1.5382389934400199e-06,
-      "loss": 0.9334,
+      "learning_rate": 1.8873016929904942e-06,
+      "loss": 0.9335,
       "step": 895
     },
     {
       "epoch": 0.83,
-      "learning_rate": 1.462323714966114e-06,
-      "loss": 0.9485,
+      "learning_rate": 1.7947268713034128e-06,
+      "loss": 0.9481,
       "step": 900
     },
     {
       "epoch": 0.83,
-      "learning_rate": 1.3881814780954916e-06,
-      "loss": 0.9365,
+      "learning_rate": 1.704255757827963e-06,
+      "loss": 0.9352,
       "step": 905
     },
     {
       "epoch": 0.83,
-      "learning_rate": 1.3158276801407432e-06,
-      "loss": 0.9249,
+      "learning_rate": 1.6159115479924259e-06,
+      "loss": 0.9238,
       "step": 910
     },
     {
       "epoch": 0.84,
-      "learning_rate": 1.2452773470046543e-06,
-      "loss": 0.9212,
+      "learning_rate": 1.529716891919074e-06,
+      "loss": 0.9213,
       "step": 915
     },
     {
       "epoch": 0.84,
-      "learning_rate": 1.1765451300597574e-06,
-      "loss": 0.9377,
+      "learning_rate": 1.4456938886170413e-06,
+      "loss": 0.938,
       "step": 920
     },
     {
       "epoch": 0.85,
-      "learning_rate": 1.1096453031056265e-06,
-      "loss": 0.926,
+      "learning_rate": 1.3638640803164516e-06,
+      "loss": 0.9248,
       "step": 925
     },
     {
       "epoch": 0.85,
-      "learning_rate": 1.0445917594046073e-06,
-      "loss": 0.9171,
+      "learning_rate": 1.2842484469453365e-06,
+      "loss": 0.9157,
       "step": 930
     },
     {
       "epoch": 0.86,
-      "learning_rate": 9.813980087965625e-07,
-      "loss": 0.9296,
+      "learning_rate": 1.2068674007506787e-06,
+      "loss": 0.9289,
       "step": 935
     },
     {
       "epoch": 0.86,
-      "learning_rate": 9.200771748932513e-07,
-      "loss": 0.9218,
+      "learning_rate": 1.1317407810650372e-06,
+      "loss": 0.9215,
       "step": 940
     },
     {
       "epoch": 0.87,
-      "learning_rate": 8.606419923529175e-07,
-      "loss": 0.9351,
+      "learning_rate": 1.0588878492200261e-06,
+      "loss": 0.9343,
       "step": 945
     },
     {
       "epoch": 0.87,
-      "learning_rate": 8.031048042356393e-07,
-      "loss": 0.9317,
+      "learning_rate": 9.883272836080116e-07,
+      "loss": 0.9294,
       "step": 950
     },
     {
       "epoch": 0.88,
-      "learning_rate": 7.474775594400252e-07,
-      "loss": 0.9417,
+      "learning_rate": 9.200771748932513e-07,
+      "loss": 0.9419,
       "step": 955
     },
     {
       "epoch": 0.88,
-      "learning_rate": 6.937718102217461e-07,
-      "loss": 0.9272,
+      "learning_rate": 8.541550213737171e-07,
+      "loss": 0.924,
       "step": 960
     },
     {
       "epoch": 0.89,
-      "learning_rate": 6.41998709794458e-07,
-      "loss": 0.9199,
+      "learning_rate": 7.905777244947954e-07,
+      "loss": 0.9194,
       "step": 965
     },
     {
       "epoch": 0.89,
-      "learning_rate": 5.921690100135713e-07,
-      "loss": 0.9145,
+      "learning_rate": 7.293615845160196e-07,
+      "loss": 0.9124,
       "step": 970
     },
     {
       "epoch": 0.89,
-      "learning_rate": 5.442930591433992e-07,
-      "loss": 0.9287,
+      "learning_rate": 6.705222963319191e-07,
+      "loss": 0.9282,
       "step": 975
     },
     {
       "epoch": 0.9,
-      "learning_rate": 4.983807997080925e-07,
-      "loss": 0.94,
+      "learning_rate": 6.140749454480932e-07,
+      "loss": 0.9384,
       "step": 980
     },
     {
       "epoch": 0.9,
-      "learning_rate": 4.544417664268652e-07,
-      "loss": 0.9284,
+      "learning_rate": 5.600340041135133e-07,
+      "loss": 0.9259,
       "step": 985
     },
     {
       "epoch": 0.91,
-      "learning_rate": 4.124850842338779e-07,
-      "loss": 0.9295,
+      "learning_rate": 5.0841332761005e-07,
+      "loss": 0.9277,
       "step": 990
     },
     {
       "epoch": 0.91,
-      "learning_rate": 3.725194663832521e-07,
-      "loss": 0.9277,
+      "learning_rate": 4.592261507001994e-07,
+      "loss": 0.925,
       "step": 995
     },
     {
       "epoch": 0.92,
-      "learning_rate": 3.345532126395579e-07,
-      "loss": 0.9265,
+      "learning_rate": 4.124850842338779e-07,
+      "loss": 0.9253,
       "step": 1000
     },
     {
       "epoch": 0.92,
-      "learning_rate": 2.985942075541848e-07,
-      "loss": 0.9266,
+      "learning_rate": 3.6820211191520127e-07,
+      "loss": 0.9235,
       "step": 1005
     },
     {
       "epoch": 0.93,
-      "learning_rate": 2.646499188279328e-07,
-      "loss": 0.912,
+      "learning_rate": 3.263885872300343e-07,
+      "loss": 0.9086,
       "step": 1010
     },
     {
       "epoch": 0.93,
-      "learning_rate": 2.3272739576017945e-07,
-      "loss": 0.9308,
+      "learning_rate": 2.870552305351382e-07,
+      "loss": 0.9292,
       "step": 1015
     },
     {
       "epoch": 0.94,
-      "learning_rate": 2.028332677849254e-07,
-      "loss": 0.9358,
+      "learning_rate": 2.5021212630962246e-07,
+      "loss": 0.9329,
       "step": 1020
     },
     {
       "epoch": 0.94,
-      "learning_rate": 1.7497374309405346e-07,
-      "loss": 0.9223,
+      "learning_rate": 2.158687205694443e-07,
+      "loss": 0.9214,
       "step": 1025
     },
     {
       "epoch": 0.94,
-      "learning_rate": 1.49154607348051e-07,
-      "loss": 0.9168,
+      "learning_rate": 1.840338184455881e-07,
+      "loss": 0.9144,
       "step": 1030
     },
     {
       "epoch": 0.95,
-      "learning_rate": 1.2538122247448325e-07,
-      "loss": 0.9243,
+      "learning_rate": 1.5471558192656776e-07,
+      "loss": 0.923,
       "step": 1035
     },
     {
       "epoch": 0.95,
-      "learning_rate": 1.0365852555447642e-07,
-      "loss": 0.9139,
+      "learning_rate": 1.279215277658097e-07,
+      "loss": 0.9123,
       "step": 1040
     },
     {
       "epoch": 0.96,
-      "learning_rate": 8.39910277974132e-08,
-      "loss": 0.9425,
+      "learning_rate": 1.0365852555447642e-07,
+      "loss": 0.9406,
       "step": 1045
     },
     {
       "epoch": 0.96,
-      "learning_rate": 6.638281360408339e-08,
-      "loss": 0.9361,
+      "learning_rate": 8.19327959602012e-08,
+      "loss": 0.9364,
       "step": 1050
     },
     {
       "epoch": 0.97,
-      "learning_rate": 5.083753971846239e-08,
-      "loss": 0.9223,
+      "learning_rate": 6.274990913221035e-08,
+      "loss": 0.9204,
       "step": 1055
     },
     {
       "epoch": 0.97,
-      "learning_rate": 3.735843446830867e-08,
-      "loss": 0.9227,
+      "learning_rate": 4.6114783273213395e-08,
+      "loss": 0.9214,
       "step": 1060
     },
     {
       "epoch": 0.98,
-      "learning_rate": 2.5948297094724463e-08,
-      "loss": 0.9145,
+      "learning_rate": 3.203168337845508e-08,
+      "loss": 0.9126,
       "step": 1065
     },
     {
       "epoch": 0.98,
-      "learning_rate": 1.6609497170834154e-08,
-      "loss": 0.929,
+      "learning_rate": 2.05042201422323e-08,
+      "loss": 0.9267,
       "step": 1070
     },
     {
       "epoch": 0.99,
-      "learning_rate": 9.343974109685684e-09,
-      "loss": 0.9277,
+      "learning_rate": 1.1535349032167908e-08,
+      "loss": 0.9244,
       "step": 1075
     },
     {
       "epoch": 0.99,
-      "learning_rate": 4.153236761488266e-09,
-      "loss": 0.9477,
+      "learning_rate": 5.127369531473525e-09,
+      "loss": 0.9455,
       "step": 1080
     },
     {
       "epoch": 1.0,
-      "learning_rate": 1.0383631002686133e-09,
-      "loss": 0.9181,
+      "learning_rate": 1.2819245493955746e-09,
+      "loss": 0.9161,
       "step": 1085
     },
     {
       "epoch": 1.0,
       "learning_rate": 0.0,
-      "loss": 0.8975,
+      "loss": 0.9077,
       "step": 1090
     },
     {
       "epoch": 1.0,
-      "eval_loss": 0.9360852241516113,
-      "eval_runtime": 143.326,
-      "eval_samples_per_second": 107.664,
-      "eval_steps_per_second": 1.688,
+      "eval_loss": 0.9354520440101624,
+      "eval_runtime": 142.5652,
+      "eval_samples_per_second": 108.238,
+      "eval_steps_per_second": 1.697,
       "step": 1090
     },
     {
       "epoch": 1.0,
       "step": 1090,
       "total_flos": 456447649382400.0,
-      "train_loss": 0.07633398782222643,
-      "train_runtime": 563.604,
-      "train_samples_per_second": 247.383,
-      "train_steps_per_second": 1.934
+      "train_loss": 0.977718904259008,
+      "train_runtime": 5540.1569,
+      "train_samples_per_second": 25.166,
+      "train_steps_per_second": 0.197
     }
   ],
   "logging_steps": 5,