🍻 cheers

README.md

@@ -2,6 +2,7 @@
 license: apache-2.0
 base_model: google/vit-base-patch16-224-in21k
 tags:
+- image-classification
 - generated_from_trainer
 model-index:
 - name: ryan_model3272024
@@ -13,12 +14,12 @@ should probably proofread and complete it, then remove this comment. -->
 
 # ryan_model3272024
 
-This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the None dataset.
+This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the beans dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.2868
+- Loss: 0.2552
 - Na Accuracy: 0.95
-- Ordinal Accuracy: 0.6711
-- Ordinal Mae: 1.1626
+- Ordinal Accuracy: 0.6267
+- Ordinal Mae: 1.1586
 
 ## Model description
 

all_results.json

@@ -1,14 +1,14 @@
 {
     "epoch": 3.83,
-    "eval_loss": 1.0327022075653076,
-    "eval_na_accuracy": 0.956,
-    "eval_ordinal_accuracy": 0.568,
-    "eval_ordinal_mae": 56.34840202213265,
-    "eval_runtime": 20.4149,
-    "eval_samples_per_second": 24.492,
-    "eval_steps_per_second": 3.086,
-    "train_loss": 0.6305676238735517,
-    "train_runtime": 1636.5913,
-    "train_samples_per_second": 12.221,
-    "train_steps_per_second": 0.765
+    "eval_loss": 0.2552177309989929,
+    "eval_na_accuracy": 0.95,
+    "eval_ordinal_accuracy": 0.6266666666666667,
+    "eval_ordinal_mae": 1.158560517811113,
+    "eval_runtime": 19.6745,
+    "eval_samples_per_second": 25.414,
+    "eval_steps_per_second": 3.202,
+    "train_loss": 0.17935538868109385,
+    "train_runtime": 1702.5744,
+    "train_samples_per_second": 11.747,
+    "train_steps_per_second": 0.735
 }

eval_results.json

@@ -1,10 +1,10 @@
 {
     "epoch": 3.83,
-    "eval_loss": 1.0327022075653076,
-    "eval_na_accuracy": 0.956,
-    "eval_ordinal_accuracy": 0.568,
-    "eval_ordinal_mae": 56.34840202213265,
-    "eval_runtime": 20.4149,
-    "eval_samples_per_second": 24.492,
-    "eval_steps_per_second": 3.086
+    "eval_loss": 0.2552177309989929,
+    "eval_na_accuracy": 0.95,
+    "eval_ordinal_accuracy": 0.6266666666666667,
+    "eval_ordinal_mae": 1.158560517811113,
+    "eval_runtime": 19.6745,
+    "eval_samples_per_second": 25.414,
+    "eval_steps_per_second": 3.202
 }

runs/Mar27_20-06-26_ryanserver/events.out.tfevents.1711586196.ryanserver.24064.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7608e0bba49dd02a3edba076af19d5783dcf52171b4e0113a42f37cdb474c484
+size 529

train_results.json

@@ -1,7 +1,7 @@
 {
     "epoch": 3.83,
-    "train_loss": 0.6305676238735517,
-    "train_runtime": 1636.5913,
-    "train_samples_per_second": 12.221,
-    "train_steps_per_second": 0.765
+    "train_loss": 0.17935538868109385,
+    "train_runtime": 1702.5744,
+    "train_samples_per_second": 11.747,
+    "train_steps_per_second": 0.735
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": 1.0327022075653076,
-  "best_model_checkpoint": "./ryan_model3272024/checkpoint-500",
+  "best_metric": 0.2552177309989929,
+  "best_model_checkpoint": "./ryan_model3272024/checkpoint-600",
   "epoch": 3.8338658146964857,
   "eval_steps": 100,
   "global_step": 1200,
@@ -10,984 +10,984 @@
   "log_history": [
     {
       "epoch": 0.03,
-      "grad_norm": 2.4977924823760986,
+      "grad_norm": 1.4023665189743042,
       "learning_rate": 0.00019840255591054313,
-      "loss": 1.9988,
+      "loss": 0.5486,
       "step": 10
     },
     {
       "epoch": 0.06,
-      "grad_norm": 2.3356950283050537,
+      "grad_norm": 1.2863692045211792,
       "learning_rate": 0.00019680511182108628,
-      "loss": 1.6981,
+      "loss": 0.4543,
       "step": 20
     },
     {
       "epoch": 0.1,
-      "grad_norm": 2.1436116695404053,
+      "grad_norm": 0.8842328190803528,
       "learning_rate": 0.0001952076677316294,
-      "loss": 1.6585,
+      "loss": 0.4222,
       "step": 30
     },
     {
       "epoch": 0.13,
-      "grad_norm": 1.9430339336395264,
+      "grad_norm": 0.8728455901145935,
       "learning_rate": 0.00019361022364217253,
-      "loss": 1.6929,
+      "loss": 0.3764,
       "step": 40
     },
     {
       "epoch": 0.16,
-      "grad_norm": 1.7540597915649414,
+      "grad_norm": 0.6641435027122498,
       "learning_rate": 0.00019201277955271565,
-      "loss": 1.4822,
+      "loss": 0.3214,
       "step": 50
     },
     {
       "epoch": 0.19,
-      "grad_norm": 2.9073445796966553,
+      "grad_norm": 1.4344050884246826,
       "learning_rate": 0.0001904153354632588,
-      "loss": 1.5968,
+      "loss": 0.3286,
       "step": 60
     },
     {
       "epoch": 0.22,
-      "grad_norm": 1.6934313774108887,
+      "grad_norm": 0.8919397592544556,
       "learning_rate": 0.00018881789137380192,
-      "loss": 1.4684,
+      "loss": 0.33,
       "step": 70
     },
     {
       "epoch": 0.26,
-      "grad_norm": 3.451613426208496,
+      "grad_norm": 1.7052876949310303,
       "learning_rate": 0.00018722044728434505,
-      "loss": 1.4425,
+      "loss": 0.3337,
       "step": 80
     },
     {
       "epoch": 0.29,
-      "grad_norm": 1.747863531112671,
+      "grad_norm": 0.4728272259235382,
       "learning_rate": 0.0001856230031948882,
-      "loss": 1.3641,
+      "loss": 0.3784,
       "step": 90
     },
     {
       "epoch": 0.32,
-      "grad_norm": 2.263026714324951,
+      "grad_norm": 1.1663854122161865,
       "learning_rate": 0.00018402555910543132,
-      "loss": 1.3578,
+      "loss": 0.3853,
       "step": 100
     },
     {
       "epoch": 0.32,
-      "eval_loss": 1.2057933807373047,
-      "eval_na_accuracy": 0.944,
-      "eval_ordinal_accuracy": 0.54,
-      "eval_ordinal_mae": 90.9845891143661,
-      "eval_runtime": 46.445,
-      "eval_samples_per_second": 10.765,
-      "eval_steps_per_second": 1.356,
+      "eval_loss": 0.3272034823894501,
+      "eval_na_accuracy": 0.924,
+      "eval_ordinal_accuracy": 0.52,
+      "eval_ordinal_mae": 1.210578082634343,
+      "eval_runtime": 52.9914,
+      "eval_samples_per_second": 9.435,
+      "eval_steps_per_second": 1.189,
       "step": 100
     },
     {
       "epoch": 0.35,
-      "grad_norm": 2.7085518836975098,
+      "grad_norm": 0.8579528331756592,
       "learning_rate": 0.00018242811501597444,
-      "loss": 1.3463,
+      "loss": 0.3585,
       "step": 110
     },
     {
       "epoch": 0.38,
-      "grad_norm": 2.2408878803253174,
+      "grad_norm": 1.02351975440979,
       "learning_rate": 0.00018083067092651756,
-      "loss": 1.1704,
+      "loss": 0.3621,
       "step": 120
     },
     {
       "epoch": 0.42,
-      "grad_norm": 2.1588001251220703,
+      "grad_norm": 1.3286011219024658,
       "learning_rate": 0.00017923322683706071,
-      "loss": 1.24,
+      "loss": 0.3714,
       "step": 130
     },
     {
       "epoch": 0.45,
-      "grad_norm": 2.582127094268799,
+      "grad_norm": 0.6290095448493958,
       "learning_rate": 0.00017763578274760384,
-      "loss": 1.2735,
+      "loss": 0.3275,
       "step": 140
     },
     {
       "epoch": 0.48,
-      "grad_norm": 1.9934537410736084,
+      "grad_norm": 1.269338846206665,
       "learning_rate": 0.000176038338658147,
-      "loss": 1.344,
+      "loss": 0.4287,
       "step": 150
     },
     {
       "epoch": 0.51,
-      "grad_norm": 2.1636180877685547,
+      "grad_norm": 0.6244733333587646,
       "learning_rate": 0.0001744408945686901,
-      "loss": 1.1532,
+      "loss": 0.3067,
       "step": 160
     },
     {
       "epoch": 0.54,
-      "grad_norm": 3.9455337524414062,
+      "grad_norm": 1.1287596225738525,
       "learning_rate": 0.00017284345047923323,
-      "loss": 1.1231,
+      "loss": 0.2982,
       "step": 170
     },
     {
       "epoch": 0.58,
-      "grad_norm": 3.46232271194458,
+      "grad_norm": 1.436303734779358,
       "learning_rate": 0.00017124600638977638,
-      "loss": 1.0945,
+      "loss": 0.2946,
       "step": 180
     },
     {
       "epoch": 0.61,
-      "grad_norm": 2.913996696472168,
+      "grad_norm": 0.8159350752830505,
       "learning_rate": 0.00016964856230031948,
-      "loss": 1.3168,
+      "loss": 0.3514,
       "step": 190
     },
     {
       "epoch": 0.64,
-      "grad_norm": 2.16892671585083,
+      "grad_norm": 0.7363901138305664,
       "learning_rate": 0.00016805111821086263,
-      "loss": 1.089,
+      "loss": 0.3396,
       "step": 200
     },
     {
       "epoch": 0.64,
-      "eval_loss": 1.0986785888671875,
-      "eval_na_accuracy": 0.95,
-      "eval_ordinal_accuracy": 0.548,
-      "eval_ordinal_mae": 120.70968111835792,
-      "eval_runtime": 20.2841,
-      "eval_samples_per_second": 24.65,
-      "eval_steps_per_second": 3.106,
+      "eval_loss": 0.27412503957748413,
+      "eval_na_accuracy": 0.94,
+      "eval_ordinal_accuracy": 0.5644444444444444,
+      "eval_ordinal_mae": 1.1640199238227473,
+      "eval_runtime": 21.3186,
+      "eval_samples_per_second": 23.454,
+      "eval_steps_per_second": 2.955,
       "step": 200
     },
     {
       "epoch": 0.67,
-      "grad_norm": 2.9793784618377686,
+      "grad_norm": 0.6321592330932617,
       "learning_rate": 0.00016645367412140575,
-      "loss": 1.1788,
+      "loss": 0.3952,
       "step": 210
     },
     {
       "epoch": 0.7,
-      "grad_norm": 2.662822723388672,
+      "grad_norm": 0.6153714656829834,
       "learning_rate": 0.0001648562300319489,
-      "loss": 1.0067,
+      "loss": 0.2947,
       "step": 220
     },
     {
       "epoch": 0.73,
-      "grad_norm": 2.3921396732330322,
+      "grad_norm": 1.3031296730041504,
       "learning_rate": 0.00016325878594249202,
-      "loss": 1.2542,
+      "loss": 0.3556,
       "step": 230
     },
     {
       "epoch": 0.77,
-      "grad_norm": 2.670137882232666,
+      "grad_norm": 1.058060646057129,
       "learning_rate": 0.00016166134185303515,
-      "loss": 1.1284,
+      "loss": 0.3432,
       "step": 240
     },
     {
       "epoch": 0.8,
-      "grad_norm": 3.212069511413574,
+      "grad_norm": 0.957135796546936,
       "learning_rate": 0.0001600638977635783,
-      "loss": 1.3407,
+      "loss": 0.3675,
       "step": 250
     },
     {
       "epoch": 0.83,
-      "grad_norm": 2.284740447998047,
+      "grad_norm": 1.6347941160202026,
       "learning_rate": 0.00015846645367412142,
-      "loss": 1.1876,
+      "loss": 0.3008,
       "step": 260
     },
     {
       "epoch": 0.86,
-      "grad_norm": 2.98988938331604,
+      "grad_norm": 1.1190528869628906,
       "learning_rate": 0.00015686900958466454,
-      "loss": 1.0993,
+      "loss": 0.2944,
       "step": 270
     },
     {
       "epoch": 0.89,
-      "grad_norm": 3.687721014022827,
+      "grad_norm": 0.8016924858093262,
       "learning_rate": 0.00015527156549520767,
-      "loss": 1.1273,
+      "loss": 0.2361,
       "step": 280
     },
     {
       "epoch": 0.93,
-      "grad_norm": 2.0068211555480957,
+      "grad_norm": 1.3622130155563354,
       "learning_rate": 0.00015367412140575082,
-      "loss": 1.0204,
+      "loss": 0.3569,
       "step": 290
     },
     {
       "epoch": 0.96,
-      "grad_norm": 2.719209909439087,
+      "grad_norm": 0.6603774428367615,
       "learning_rate": 0.00015207667731629394,
-      "loss": 0.924,
+      "loss": 0.2075,
       "step": 300
     },
     {
       "epoch": 0.96,
-      "eval_loss": 1.083807110786438,
+      "eval_loss": 0.2772314250469208,
       "eval_na_accuracy": 0.946,
-      "eval_ordinal_accuracy": 0.568,
-      "eval_ordinal_mae": 76.79816162186209,
-      "eval_runtime": 20.3196,
-      "eval_samples_per_second": 24.607,
-      "eval_steps_per_second": 3.1,
+      "eval_ordinal_accuracy": 0.5933333333333334,
+      "eval_ordinal_mae": 1.194209214001894,
+      "eval_runtime": 20.7347,
+      "eval_samples_per_second": 24.114,
+      "eval_steps_per_second": 3.038,
       "step": 300
     },
     {
       "epoch": 0.99,
-      "grad_norm": 6.174163341522217,
-      "learning_rate": 0.00015063897763578277,
-      "loss": 1.0079,
+      "grad_norm": 1.3968242406845093,
+      "learning_rate": 0.00015047923322683706,
+      "loss": 0.2232,
       "step": 310
     },
     {
       "epoch": 1.02,
-      "grad_norm": 1.8171716928482056,
-      "learning_rate": 0.0001490415335463259,
-      "loss": 0.9573,
+      "grad_norm": 0.7815521359443665,
+      "learning_rate": 0.0001488817891373802,
+      "loss": 0.3132,
       "step": 320
     },
     {
       "epoch": 1.05,
-      "grad_norm": 1.7882392406463623,
-      "learning_rate": 0.00014744408945686902,
-      "loss": 0.8008,
+      "grad_norm": 1.1288195848464966,
+      "learning_rate": 0.00014728434504792333,
+      "loss": 0.255,
       "step": 330
     },
     {
       "epoch": 1.09,
-      "grad_norm": 2.8711063861846924,
-      "learning_rate": 0.00014584664536741214,
-      "loss": 0.7867,
+      "grad_norm": 0.7704196572303772,
+      "learning_rate": 0.00014568690095846646,
+      "loss": 0.2415,
       "step": 340
     },
     {
       "epoch": 1.12,
-      "grad_norm": 3.8461153507232666,
-      "learning_rate": 0.00014424920127795526,
-      "loss": 0.7375,
+      "grad_norm": 1.9226877689361572,
+      "learning_rate": 0.00014408945686900958,
+      "loss": 0.1975,
       "step": 350
     },
     {
       "epoch": 1.15,
-      "grad_norm": 3.599330186843872,
-      "learning_rate": 0.0001426517571884984,
-      "loss": 0.8057,
+      "grad_norm": 0.5694310069084167,
+      "learning_rate": 0.00014249201277955273,
+      "loss": 0.1722,
       "step": 360
     },
     {
       "epoch": 1.18,
-      "grad_norm": 1.9864487648010254,
-      "learning_rate": 0.00014105431309904153,
-      "loss": 0.7876,
+      "grad_norm": 1.719147801399231,
+      "learning_rate": 0.00014089456869009585,
+      "loss": 0.2175,
       "step": 370
     },
     {
       "epoch": 1.21,
-      "grad_norm": 3.377739191055298,
-      "learning_rate": 0.00013945686900958468,
-      "loss": 0.7835,
+      "grad_norm": 0.9247463941574097,
+      "learning_rate": 0.000139297124600639,
+      "loss": 0.2088,
       "step": 380
     },
     {
       "epoch": 1.25,
-      "grad_norm": 3.5138072967529297,
-      "learning_rate": 0.0001378594249201278,
-      "loss": 0.8265,
+      "grad_norm": 1.0941154956817627,
+      "learning_rate": 0.00013769968051118212,
+      "loss": 0.2854,
       "step": 390
     },
     {
       "epoch": 1.28,
-      "grad_norm": 3.5262341499328613,
-      "learning_rate": 0.00013626198083067093,
-      "loss": 0.694,
+      "grad_norm": 1.0274015665054321,
+      "learning_rate": 0.00013610223642172525,
+      "loss": 0.196,
       "step": 400
     },
     {
       "epoch": 1.28,
-      "eval_loss": 1.068002700805664,
-      "eval_na_accuracy": 0.942,
-      "eval_ordinal_accuracy": 0.556,
-      "eval_ordinal_mae": 105.63124973691814,
-      "eval_runtime": 19.9749,
-      "eval_samples_per_second": 25.031,
-      "eval_steps_per_second": 3.154,
+      "eval_loss": 0.273777574300766,
+      "eval_na_accuracy": 0.95,
+      "eval_ordinal_accuracy": 0.6133333333333333,
+      "eval_ordinal_mae": 1.198390154937903,
+      "eval_runtime": 20.9145,
+      "eval_samples_per_second": 23.907,
+      "eval_steps_per_second": 3.012,
       "step": 400
     },
     {
       "epoch": 1.31,
-      "grad_norm": 2.814100980758667,
-      "learning_rate": 0.00013466453674121405,
-      "loss": 0.7246,
+      "grad_norm": 2.912687063217163,
+      "learning_rate": 0.00013450479233226837,
+      "loss": 0.2156,
       "step": 410
     },
     {
       "epoch": 1.34,
-      "grad_norm": 1.9095064401626587,
-      "learning_rate": 0.00013306709265175718,
-      "loss": 0.768,
+      "grad_norm": 0.6906268000602722,
+      "learning_rate": 0.0001329073482428115,
+      "loss": 0.1366,
       "step": 420
     },
     {
       "epoch": 1.37,
-      "grad_norm": 2.849485158920288,
-      "learning_rate": 0.00013146964856230033,
-      "loss": 0.6669,
+      "grad_norm": 0.43070048093795776,
+      "learning_rate": 0.00013130990415335464,
+      "loss": 0.2174,
       "step": 430
     },
     {
       "epoch": 1.41,
-      "grad_norm": 2.2935235500335693,
-      "learning_rate": 0.00012987220447284345,
-      "loss": 0.6708,
+      "grad_norm": 0.5173763632774353,
+      "learning_rate": 0.00012971246006389777,
+      "loss": 0.2016,
       "step": 440
     },
     {
       "epoch": 1.44,
-      "grad_norm": 2.6063146591186523,
-      "learning_rate": 0.0001282747603833866,
-      "loss": 0.7481,
+      "grad_norm": 1.04314386844635,
+      "learning_rate": 0.00012811501597444092,
+      "loss": 0.2233,
       "step": 450
     },
     {
       "epoch": 1.47,
-      "grad_norm": 2.329845428466797,
-      "learning_rate": 0.00012667731629392972,
-      "loss": 0.7341,
+      "grad_norm": 0.523073673248291,
+      "learning_rate": 0.00012651757188498404,
+      "loss": 0.2231,
       "step": 460
     },
     {
       "epoch": 1.5,
-      "grad_norm": 4.184568881988525,
-      "learning_rate": 0.00012507987220447287,
-      "loss": 0.785,
+      "grad_norm": 3.259795904159546,
+      "learning_rate": 0.00012492012779552716,
+      "loss": 0.2366,
       "step": 470
     },
     {
       "epoch": 1.53,
-      "grad_norm": 2.142719268798828,
-      "learning_rate": 0.00012348242811501597,
-      "loss": 0.702,
+      "grad_norm": 0.6846562027931213,
+      "learning_rate": 0.00012332268370607028,
+      "loss": 0.2144,
       "step": 480
     },
     {
       "epoch": 1.57,
-      "grad_norm": 2.7038168907165527,
-      "learning_rate": 0.0001218849840255591,
-      "loss": 0.8301,
+      "grad_norm": 1.2122007608413696,
+      "learning_rate": 0.00012172523961661342,
+      "loss": 0.2938,
       "step": 490
     },
     {
       "epoch": 1.6,
-      "grad_norm": 3.093721389770508,
-      "learning_rate": 0.00012028753993610224,
-      "loss": 0.7739,
+      "grad_norm": 1.3790067434310913,
+      "learning_rate": 0.00012012779552715656,
+      "loss": 0.2228,
       "step": 500
     },
     {
       "epoch": 1.6,
-      "eval_loss": 1.0327022075653076,
+      "eval_loss": 0.26852139830589294,
       "eval_na_accuracy": 0.956,
-      "eval_ordinal_accuracy": 0.568,
-      "eval_ordinal_mae": 56.34840202213265,
-      "eval_runtime": 20.2051,
-      "eval_samples_per_second": 24.746,
-      "eval_steps_per_second": 3.118,
+      "eval_ordinal_accuracy": 0.62,
+      "eval_ordinal_mae": 1.1989026491012837,
+      "eval_runtime": 20.0158,
+      "eval_samples_per_second": 24.98,
+      "eval_steps_per_second": 3.148,
       "step": 500
     },
     {
       "epoch": 1.63,
-      "grad_norm": 3.727419853210449,
-      "learning_rate": 0.00011869009584664536,
-      "loss": 0.7536,
+      "grad_norm": 0.7108421921730042,
+      "learning_rate": 0.00011853035143769968,
+      "loss": 0.1916,
       "step": 510
     },
     {
       "epoch": 1.66,
-      "grad_norm": 3.3148770332336426,
-      "learning_rate": 0.00011709265175718851,
-      "loss": 0.7675,
+      "grad_norm": 0.42910462617874146,
+      "learning_rate": 0.00011693290734824283,
+      "loss": 0.2478,
       "step": 520
     },
     {
       "epoch": 1.69,
-      "grad_norm": 3.7952725887298584,
-      "learning_rate": 0.00011549520766773163,
-      "loss": 0.7543,
+      "grad_norm": 0.9730465412139893,
+      "learning_rate": 0.00011533546325878595,
+      "loss": 0.189,
       "step": 530
     },
     {
       "epoch": 1.73,
-      "grad_norm": 3.9252896308898926,
-      "learning_rate": 0.00011389776357827477,
-      "loss": 0.7754,
+      "grad_norm": 0.9566612243652344,
+      "learning_rate": 0.00011373801916932908,
+      "loss": 0.1768,
       "step": 540
     },
     {
       "epoch": 1.76,
-      "grad_norm": 3.045405387878418,
-      "learning_rate": 0.0001123003194888179,
-      "loss": 0.5713,
+      "grad_norm": 0.5167070627212524,
+      "learning_rate": 0.00011214057507987221,
+      "loss": 0.1385,
       "step": 550
     },
     {
       "epoch": 1.79,
-      "grad_norm": 1.8544007539749146,
-      "learning_rate": 0.00011070287539936102,
-      "loss": 0.5592,
+      "grad_norm": 0.5880122780799866,
+      "learning_rate": 0.00011054313099041533,
+      "loss": 0.1262,
       "step": 560
     },
     {
       "epoch": 1.82,
-      "grad_norm": 3.626079797744751,
-      "learning_rate": 0.00010910543130990417,
-      "loss": 0.6631,
+      "grad_norm": 1.202286720275879,
+      "learning_rate": 0.00010894568690095847,
+      "loss": 0.1721,
       "step": 570
     },
     {
       "epoch": 1.85,
-      "grad_norm": 4.355627536773682,
-      "learning_rate": 0.00010750798722044728,
-      "loss": 0.5981,
+      "grad_norm": 2.6997601985931396,
+      "learning_rate": 0.0001073482428115016,
+      "loss": 0.2128,
       "step": 580
     },
     {
       "epoch": 1.88,
-      "grad_norm": 4.025274276733398,
-      "learning_rate": 0.00010591054313099043,
-      "loss": 0.7375,
+      "grad_norm": 1.1591830253601074,
+      "learning_rate": 0.00010575079872204474,
+      "loss": 0.2402,
       "step": 590
     },
     {
       "epoch": 1.92,
-      "grad_norm": 2.506023645401001,
-      "learning_rate": 0.00010431309904153355,
-      "loss": 0.5935,
+      "grad_norm": 0.5840221643447876,
+      "learning_rate": 0.00010415335463258787,
+      "loss": 0.1816,
       "step": 600
     },
     {
       "epoch": 1.92,
-      "eval_loss": 1.0479341745376587,
-      "eval_na_accuracy": 0.932,
-      "eval_ordinal_accuracy": 0.598,
-      "eval_ordinal_mae": 50.45202396943141,
-      "eval_runtime": 20.1021,
-      "eval_samples_per_second": 24.873,
-      "eval_steps_per_second": 3.134,
+      "eval_loss": 0.2552177309989929,
+      "eval_na_accuracy": 0.95,
+      "eval_ordinal_accuracy": 0.6266666666666667,
+      "eval_ordinal_mae": 1.158560517811113,
+      "eval_runtime": 19.5011,
+      "eval_samples_per_second": 25.64,
+      "eval_steps_per_second": 3.231,
       "step": 600
     },
     {
       "epoch": 1.95,
-      "grad_norm": 4.494864463806152,
-      "learning_rate": 0.00010271565495207669,
-      "loss": 0.651,
+      "grad_norm": 0.7560299634933472,
+      "learning_rate": 0.000102555910543131,
+      "loss": 0.2021,
       "step": 610
     },
     {
       "epoch": 1.98,
-      "grad_norm": 4.595503807067871,
-      "learning_rate": 0.00010111821086261981,
-      "loss": 0.6743,
+      "grad_norm": 1.8860361576080322,
+      "learning_rate": 0.00010095846645367413,
+      "loss": 0.2092,
       "step": 620
     },
     {
       "epoch": 2.01,
-      "grad_norm": 3.0126616954803467,
-      "learning_rate": 9.952076677316294e-05,
-      "loss": 0.5441,
+      "grad_norm": 0.7235255837440491,
+      "learning_rate": 9.936102236421726e-05,
+      "loss": 0.1131,
       "step": 630
     },
     {
       "epoch": 2.04,
-      "grad_norm": 2.2599124908447266,
-      "learning_rate": 9.792332268370608e-05,
-      "loss": 0.3816,
+      "grad_norm": 0.3656529486179352,
+      "learning_rate": 9.77635782747604e-05,
+      "loss": 0.0867,
       "step": 640
     },
     {
       "epoch": 2.08,
-      "grad_norm": 1.8868279457092285,
-      "learning_rate": 9.63258785942492e-05,
-      "loss": 0.4175,
+      "grad_norm": 0.3450271785259247,
+      "learning_rate": 9.616613418530351e-05,
+      "loss": 0.0903,
       "step": 650
     },
     {
       "epoch": 2.11,
-      "grad_norm": 2.2594425678253174,
-      "learning_rate": 9.472843450479234e-05,
-      "loss": 0.3208,
+      "grad_norm": 1.0603750944137573,
+      "learning_rate": 9.456869009584664e-05,
+      "loss": 0.1234,
       "step": 660
     },
     {
       "epoch": 2.14,
-      "grad_norm": 1.4436126947402954,
-      "learning_rate": 9.313099041533548e-05,
-      "loss": 0.2459,
+      "grad_norm": 0.6790297031402588,
+      "learning_rate": 9.297124600638978e-05,
+      "loss": 0.0936,
       "step": 670
     },
     {
       "epoch": 2.17,
-      "grad_norm": 2.018686532974243,
-      "learning_rate": 9.15335463258786e-05,
-      "loss": 0.2556,
+      "grad_norm": 0.5596363544464111,
+      "learning_rate": 9.137380191693292e-05,
+      "loss": 0.0651,
       "step": 680
     },
     {
       "epoch": 2.2,
-      "grad_norm": 3.8970091342926025,
-      "learning_rate": 8.993610223642172e-05,
-      "loss": 0.2912,
+      "grad_norm": 0.5989049673080444,
+      "learning_rate": 8.977635782747604e-05,
+      "loss": 0.1218,
       "step": 690
     },
     {
       "epoch": 2.24,
-      "grad_norm": 7.18654727935791,
-      "learning_rate": 8.833865814696486e-05,
-      "loss": 0.3525,
+      "grad_norm": 0.9003208875656128,
+      "learning_rate": 8.817891373801918e-05,
+      "loss": 0.0682,
       "step": 700
     },
     {
       "epoch": 2.24,
-      "eval_loss": 1.1914987564086914,
-      "eval_na_accuracy": 0.94,
-      "eval_ordinal_accuracy": 0.578,
-      "eval_ordinal_mae": 68.50985479798727,
-      "eval_runtime": 20.035,
-      "eval_samples_per_second": 24.956,
-      "eval_steps_per_second": 3.145,
+      "eval_loss": 0.27212005853652954,
+      "eval_na_accuracy": 0.952,
+      "eval_ordinal_accuracy": 0.6577777777777778,
+      "eval_ordinal_mae": 1.1557789803379113,
+      "eval_runtime": 19.5966,
+      "eval_samples_per_second": 25.515,
+      "eval_steps_per_second": 3.215,
       "step": 700
     },
     {
       "epoch": 2.27,
-      "grad_norm": 3.8287854194641113,
-      "learning_rate": 8.6741214057508e-05,
-      "loss": 0.3517,
+      "grad_norm": 0.6663013100624084,
+      "learning_rate": 8.658146964856231e-05,
+      "loss": 0.0714,
       "step": 710
     },
     {
       "epoch": 2.3,
-      "grad_norm": 4.602190971374512,
-      "learning_rate": 8.514376996805112e-05,
-      "loss": 0.3595,
+      "grad_norm": 1.0458776950836182,
+      "learning_rate": 8.498402555910544e-05,
+      "loss": 0.102,
       "step": 720
     },
     {
       "epoch": 2.33,
-      "grad_norm": 4.8052978515625,
-      "learning_rate": 8.354632587859425e-05,
-      "loss": 0.3686,
+      "grad_norm": 0.9246501922607422,
+      "learning_rate": 8.338658146964856e-05,
+      "loss": 0.1623,
       "step": 730
     },
     {
       "epoch": 2.36,
-      "grad_norm": 1.684538722038269,
-      "learning_rate": 8.194888178913739e-05,
-      "loss": 0.3499,
+      "grad_norm": 1.0837684869766235,
+      "learning_rate": 8.17891373801917e-05,
+      "loss": 0.0934,
       "step": 740
     },
     {
       "epoch": 2.4,
-      "grad_norm": 2.9729549884796143,
-      "learning_rate": 8.035143769968051e-05,
-      "loss": 0.3116,
+      "grad_norm": 0.564241349697113,
+      "learning_rate": 8.019169329073483e-05,
+      "loss": 0.0853,
       "step": 750
     },
     {
       "epoch": 2.43,
-      "grad_norm": 4.066229343414307,
-      "learning_rate": 7.875399361022364e-05,
-      "loss": 0.301,
+      "grad_norm": 4.335838794708252,
+      "learning_rate": 7.859424920127795e-05,
+      "loss": 0.1246,
       "step": 760
     },
     {
       "epoch": 2.46,
-      "grad_norm": 4.297142505645752,
-      "learning_rate": 7.715654952076677e-05,
-      "loss": 0.3501,
+      "grad_norm": 0.957082211971283,
+      "learning_rate": 7.699680511182109e-05,
+      "loss": 0.1292,
       "step": 770
     },
     {
       "epoch": 2.49,
-      "grad_norm": 3.189189910888672,
-      "learning_rate": 7.555910543130991e-05,
-      "loss": 0.4095,
+      "grad_norm": 0.9633702039718628,
+      "learning_rate": 7.539936102236423e-05,
+      "loss": 0.1916,
       "step": 780
     },
     {
       "epoch": 2.52,
-      "grad_norm": 4.354482173919678,
-      "learning_rate": 7.396166134185304e-05,
-      "loss": 0.2547,
+      "grad_norm": 0.7254676222801208,
+      "learning_rate": 7.380191693290735e-05,
+      "loss": 0.1054,
       "step": 790
     },
     {
       "epoch": 2.56,
-      "grad_norm": 2.2913010120391846,
-      "learning_rate": 7.236421725239617e-05,
-      "loss": 0.2385,
+      "grad_norm": 0.5885197520256042,
+      "learning_rate": 7.220447284345049e-05,
+      "loss": 0.0795,
       "step": 800
     },
     {
       "epoch": 2.56,
-      "eval_loss": 1.1303021907806396,
+      "eval_loss": 0.2753521502017975,
       "eval_na_accuracy": 0.948,
-      "eval_ordinal_accuracy": 0.586,
-      "eval_ordinal_mae": 43.02211729006702,
-      "eval_runtime": 20.2963,
-      "eval_samples_per_second": 24.635,
-      "eval_steps_per_second": 3.104,
+      "eval_ordinal_accuracy": 0.6333333333333333,
+      "eval_ordinal_mae": 1.1599188842872779,
+      "eval_runtime": 20.0506,
+      "eval_samples_per_second": 24.937,
+      "eval_steps_per_second": 3.142,
       "step": 800
     },
     {
       "epoch": 2.59,
-      "grad_norm": 4.907588481903076,
-      "learning_rate": 7.07667731629393e-05,
-      "loss": 0.2578,
+      "grad_norm": 0.5671622157096863,
+      "learning_rate": 7.060702875399361e-05,
+      "loss": 0.0948,
       "step": 810
     },
     {
       "epoch": 2.62,
-      "grad_norm": 5.770906925201416,
-      "learning_rate": 6.916932907348244e-05,
-      "loss": 0.3875,
+      "grad_norm": 0.9914100766181946,
+      "learning_rate": 6.900958466453674e-05,
+      "loss": 0.0715,
       "step": 820
     },
     {
       "epoch": 2.65,
-      "grad_norm": 3.4392714500427246,
-      "learning_rate": 6.757188498402556e-05,
-      "loss": 0.3435,
+      "grad_norm": 0.4819205105304718,
+      "learning_rate": 6.741214057507987e-05,
+      "loss": 0.0839,
       "step": 830
     },
     {
       "epoch": 2.68,
-      "grad_norm": 1.785447359085083,
-      "learning_rate": 6.597444089456869e-05,
-      "loss": 0.2681,
+      "grad_norm": 0.3811684250831604,
+      "learning_rate": 6.5814696485623e-05,
+      "loss": 0.0825,
       "step": 840
     },
     {
       "epoch": 2.72,
-      "grad_norm": 3.527557373046875,
-      "learning_rate": 6.437699680511182e-05,
-      "loss": 0.3525,
+      "grad_norm": 0.9750994443893433,
+      "learning_rate": 6.421725239616614e-05,
+      "loss": 0.0968,
       "step": 850
     },
     {
       "epoch": 2.75,
-      "grad_norm": 0.4556908905506134,
-      "learning_rate": 6.277955271565496e-05,
-      "loss": 0.2462,
+      "grad_norm": 0.35765138268470764,
+      "learning_rate": 6.261980830670928e-05,
+      "loss": 0.1605,
       "step": 860
     },
     {
       "epoch": 2.78,
-      "grad_norm": 2.7756271362304688,
-      "learning_rate": 6.118210862619808e-05,
-      "loss": 0.3107,
+      "grad_norm": 0.3497343361377716,
+      "learning_rate": 6.1022364217252406e-05,
+      "loss": 0.0933,
       "step": 870
     },
     {
       "epoch": 2.81,
-      "grad_norm": 4.53245735168457,
-      "learning_rate": 5.958466453674122e-05,
-      "loss": 0.319,
+      "grad_norm": 0.4838835299015045,
+      "learning_rate": 5.942492012779552e-05,
+      "loss": 0.0859,
       "step": 880
     },
     {
       "epoch": 2.84,
-      "grad_norm": 1.7556302547454834,
-      "learning_rate": 5.7987220447284354e-05,
-      "loss": 0.2754,
+      "grad_norm": 0.7002846002578735,
+      "learning_rate": 5.782747603833866e-05,
+      "loss": 0.1021,
       "step": 890
     },
     {
       "epoch": 2.88,
-      "grad_norm": 4.295087814331055,
-      "learning_rate": 5.6389776357827484e-05,
-      "loss": 0.3423,
+      "grad_norm": 2.312203884124756,
+      "learning_rate": 5.623003194888179e-05,
+      "loss": 0.1367,
       "step": 900
     },
     {
       "epoch": 2.88,
-      "eval_loss": 1.1767209768295288,
-      "eval_na_accuracy": 0.94,
-      "eval_ordinal_accuracy": 0.604,
-      "eval_ordinal_mae": 72.14369884862379,
-      "eval_runtime": 19.1966,
-      "eval_samples_per_second": 26.046,
-      "eval_steps_per_second": 3.282,
+      "eval_loss": 0.29526129364967346,
+      "eval_na_accuracy": 0.946,
+      "eval_ordinal_accuracy": 0.64,
+      "eval_ordinal_mae": 1.166716830432415,
+      "eval_runtime": 20.0091,
+      "eval_samples_per_second": 24.989,
+      "eval_steps_per_second": 3.149,
       "step": 900
     },
     {
       "epoch": 2.91,
-      "grad_norm": 3.532813549041748,
-      "learning_rate": 5.479233226837061e-05,
-      "loss": 0.2459,
+      "grad_norm": 0.44126951694488525,
+      "learning_rate": 5.4632587859424925e-05,
+      "loss": 0.0854,
       "step": 910
     },
     {
       "epoch": 2.94,
-      "grad_norm": 2.6231563091278076,
-      "learning_rate": 5.3194888178913736e-05,
-      "loss": 0.2713,
+      "grad_norm": 1.0075191259384155,
+      "learning_rate": 5.3035143769968054e-05,
+      "loss": 0.0823,
       "step": 920
     },
     {
       "epoch": 2.97,
-      "grad_norm": 1.7921823263168335,
-      "learning_rate": 5.159744408945687e-05,
-      "loss": 0.1915,
+      "grad_norm": 0.9991279244422913,
+      "learning_rate": 5.1437699680511184e-05,
+      "loss": 0.1156,
       "step": 930
     },
     {
       "epoch": 3.0,
-      "grad_norm": 0.42470985651016235,
-      "learning_rate": 5e-05,
-      "loss": 0.2758,
+      "grad_norm": 0.8888081312179565,
+      "learning_rate": 4.984025559105431e-05,
+      "loss": 0.0876,
       "step": 940
     },
     {
       "epoch": 3.04,
-      "grad_norm": 2.1311330795288086,
-      "learning_rate": 4.840255591054313e-05,
-      "loss": 0.1447,
+      "grad_norm": 0.3761376738548279,
+      "learning_rate": 4.824281150159744e-05,
+      "loss": 0.0452,
       "step": 950
     },
     {
       "epoch": 3.07,
-      "grad_norm": 2.371814489364624,
-      "learning_rate": 4.680511182108626e-05,
-      "loss": 0.0985,
+      "grad_norm": 0.365622341632843,
+      "learning_rate": 4.664536741214058e-05,
+      "loss": 0.0428,
       "step": 960
     },
     {
       "epoch": 3.1,
-      "grad_norm": 0.3315708637237549,
-      "learning_rate": 4.520766773162939e-05,
-      "loss": 0.0887,
+      "grad_norm": 0.35657036304473877,
+      "learning_rate": 4.504792332268371e-05,
+      "loss": 0.033,
       "step": 970
     },
     {
       "epoch": 3.13,
-      "grad_norm": 5.331906795501709,
-      "learning_rate": 4.361022364217253e-05,
-      "loss": 0.1125,
+      "grad_norm": 0.5636401176452637,
+      "learning_rate": 4.345047923322684e-05,
+      "loss": 0.0356,
       "step": 980
     },
     {
       "epoch": 3.16,
-      "grad_norm": 1.143639326095581,
-      "learning_rate": 4.201277955271566e-05,
-      "loss": 0.1029,
+      "grad_norm": 0.431383341550827,
+      "learning_rate": 4.185303514376997e-05,
+      "loss": 0.0463,
       "step": 990
     },
     {
       "epoch": 3.19,
-      "grad_norm": 0.28226566314697266,
-      "learning_rate": 4.041533546325879e-05,
-      "loss": 0.0674,
+      "grad_norm": 0.583328127861023,
+      "learning_rate": 4.0255591054313104e-05,
+      "loss": 0.0387,
       "step": 1000
     },
     {
       "epoch": 3.19,
-      "eval_loss": 1.2294036149978638,
-      "eval_na_accuracy": 0.938,
-      "eval_ordinal_accuracy": 0.606,
-      "eval_ordinal_mae": 28.07016432112176,
-      "eval_runtime": 20.2957,
-      "eval_samples_per_second": 24.636,
-      "eval_steps_per_second": 3.104,
+      "eval_loss": 0.2923290431499481,
+      "eval_na_accuracy": 0.944,
+      "eval_ordinal_accuracy": 0.6377777777777778,
+      "eval_ordinal_mae": 1.2024743282463815,
+      "eval_runtime": 19.3226,
+      "eval_samples_per_second": 25.876,
+      "eval_steps_per_second": 3.26,
       "step": 1000
     },
     {
       "epoch": 3.23,
-      "grad_norm": 0.1891467422246933,
-      "learning_rate": 3.8817891373801916e-05,
-      "loss": 0.0911,
+      "grad_norm": 2.440162420272827,
+      "learning_rate": 3.8658146964856234e-05,
+      "loss": 0.0607,
       "step": 1010
     },
     {
       "epoch": 3.26,
-      "grad_norm": 0.14374223351478577,
-      "learning_rate": 3.722044728434505e-05,
-      "loss": 0.1007,
+      "grad_norm": 0.29546236991882324,
+      "learning_rate": 3.7060702875399364e-05,
+      "loss": 0.0515,
       "step": 1020
     },
     {
       "epoch": 3.29,
-      "grad_norm": 0.22984538972377777,
-      "learning_rate": 3.562300319488818e-05,
-      "loss": 0.0686,
+      "grad_norm": 0.44689303636550903,
+      "learning_rate": 3.546325878594249e-05,
+      "loss": 0.0273,
       "step": 1030
     },
     {
       "epoch": 3.32,
-      "grad_norm": 6.988112926483154,
-      "learning_rate": 3.402555910543131e-05,
-      "loss": 0.0882,
+      "grad_norm": 0.3288978040218353,
+      "learning_rate": 3.386581469648562e-05,
+      "loss": 0.0352,
       "step": 1040
     },
     {
       "epoch": 3.35,
-      "grad_norm": 0.12949563562870026,
-      "learning_rate": 3.242811501597444e-05,
-      "loss": 0.1098,
+      "grad_norm": 0.41706767678260803,
+      "learning_rate": 3.226837060702875e-05,
+      "loss": 0.0345,
       "step": 1050
     },
     {
       "epoch": 3.39,
-      "grad_norm": 0.40354716777801514,
-      "learning_rate": 3.083067092651757e-05,
-      "loss": 0.0809,
+      "grad_norm": 0.31060507893562317,
+      "learning_rate": 3.067092651757188e-05,
+      "loss": 0.0294,
       "step": 1060
     },
     {
       "epoch": 3.42,
-      "grad_norm": 0.6469861268997192,
-      "learning_rate": 2.9233226837060707e-05,
-      "loss": 0.0998,
+      "grad_norm": 0.2541821599006653,
+      "learning_rate": 2.907348242811502e-05,
+      "loss": 0.0354,
       "step": 1070
     },
     {
       "epoch": 3.45,
-      "grad_norm": 2.1838839054107666,
-      "learning_rate": 2.7635782747603834e-05,
-      "loss": 0.0923,
+      "grad_norm": 0.574343740940094,
+      "learning_rate": 2.747603833865815e-05,
+      "loss": 0.0443,
       "step": 1080
     },
     {
       "epoch": 3.48,
-      "grad_norm": 1.0453554391860962,
-      "learning_rate": 2.6038338658146967e-05,
-      "loss": 0.0669,
+      "grad_norm": 0.47532182931900024,
+      "learning_rate": 2.5878594249201278e-05,
+      "loss": 0.0605,
       "step": 1090
     },
     {
       "epoch": 3.51,
-      "grad_norm": 0.6662182807922363,
-      "learning_rate": 2.44408945686901e-05,
-      "loss": 0.1206,
+      "grad_norm": 0.45276594161987305,
+      "learning_rate": 2.428115015974441e-05,
+      "loss": 0.0293,
       "step": 1100
     },
     {
       "epoch": 3.51,
-      "eval_loss": 1.2335716485977173,
-      "eval_na_accuracy": 0.938,
-      "eval_ordinal_accuracy": 0.616,
-      "eval_ordinal_mae": 65.07940023336745,
-      "eval_runtime": 20.0687,
-      "eval_samples_per_second": 24.914,
-      "eval_steps_per_second": 3.139,
+      "eval_loss": 0.2884800434112549,
+      "eval_na_accuracy": 0.948,
+      "eval_ordinal_accuracy": 0.6644444444444444,
+      "eval_ordinal_mae": 1.1666180535654227,
+      "eval_runtime": 19.9365,
+      "eval_samples_per_second": 25.08,
+      "eval_steps_per_second": 3.16,
       "step": 1100
     },
     {
       "epoch": 3.55,
-      "grad_norm": 3.52840256690979,
-      "learning_rate": 2.284345047923323e-05,
-      "loss": 0.1402,
+      "grad_norm": 0.655549168586731,
+      "learning_rate": 2.268370607028754e-05,
+      "loss": 0.034,
       "step": 1110
     },
     {
       "epoch": 3.58,
-      "grad_norm": 0.1548507809638977,
-      "learning_rate": 2.124600638977636e-05,
-      "loss": 0.1038,
+      "grad_norm": 0.16610193252563477,
+      "learning_rate": 2.108626198083067e-05,
+      "loss": 0.0319,
       "step": 1120
     },
     {
       "epoch": 3.61,
-      "grad_norm": 3.753204584121704,
-      "learning_rate": 1.964856230031949e-05,
-      "loss": 0.1237,
+      "grad_norm": 0.26889652013778687,
+      "learning_rate": 1.9488817891373803e-05,
+      "loss": 0.0479,
       "step": 1130
     },
     {
       "epoch": 3.64,
-      "grad_norm": 2.5928821563720703,
-      "learning_rate": 1.805111821086262e-05,
-      "loss": 0.1653,
+      "grad_norm": 0.2418793886899948,
+      "learning_rate": 1.7891373801916932e-05,
+      "loss": 0.0322,
       "step": 1140
     },
     {
       "epoch": 3.67,
-      "grad_norm": 0.26777195930480957,
-      "learning_rate": 1.645367412140575e-05,
-      "loss": 0.0924,
+      "grad_norm": 0.5379694104194641,
+      "learning_rate": 1.6293929712460065e-05,
+      "loss": 0.0393,
       "step": 1150
     },
     {
       "epoch": 3.71,
-      "grad_norm": 0.11366520822048187,
-      "learning_rate": 1.485623003194888e-05,
-      "loss": 0.0376,
+      "grad_norm": 0.19815516471862793,
+      "learning_rate": 1.4696485623003195e-05,
+      "loss": 0.0217,
       "step": 1160
     },
     {
       "epoch": 3.74,
-      "grad_norm": 2.3529770374298096,
-      "learning_rate": 1.3258785942492014e-05,
-      "loss": 0.1034,
+      "grad_norm": 0.889312207698822,
+      "learning_rate": 1.3099041533546328e-05,
+      "loss": 0.0332,
       "step": 1170
     },
     {
       "epoch": 3.77,
-      "grad_norm": 0.16830460727214813,
-      "learning_rate": 1.1661341853035145e-05,
-      "loss": 0.107,
+      "grad_norm": 0.2865816652774811,
+      "learning_rate": 1.1501597444089457e-05,
+      "loss": 0.0313,
       "step": 1180
     },
     {
       "epoch": 3.8,
-      "grad_norm": 0.9607967734336853,
-      "learning_rate": 1.0063897763578276e-05,
-      "loss": 0.0751,
+      "grad_norm": 0.5947129726409912,
+      "learning_rate": 9.904153354632589e-06,
+      "loss": 0.034,
       "step": 1190
     },
     {
       "epoch": 3.83,
-      "grad_norm": 7.122812271118164,
-      "learning_rate": 8.466453674121406e-06,
-      "loss": 0.1261,
+      "grad_norm": 0.44885268807411194,
+      "learning_rate": 8.306709265175718e-06,
+      "loss": 0.0286,
       "step": 1200
     },
     {
       "epoch": 3.83,
-      "eval_loss": 1.2907226085662842,
-      "eval_na_accuracy": 0.938,
-      "eval_ordinal_accuracy": 0.604,
-      "eval_ordinal_mae": 45.83343084012577,
-      "eval_runtime": 20.016,
-      "eval_samples_per_second": 24.98,
-      "eval_steps_per_second": 3.147,
+      "eval_loss": 0.28681233525276184,
+      "eval_na_accuracy": 0.95,
+      "eval_ordinal_accuracy": 0.6711111111111111,
+      "eval_ordinal_mae": 1.1625636271304554,
+      "eval_runtime": 19.7259,
+      "eval_samples_per_second": 25.347,
+      "eval_steps_per_second": 3.194,
       "step": 1200
     },
     {
       "epoch": 3.83,
       "step": 1200,
       "total_flos": 1.4860396665534874e+18,
-      "train_loss": 0.6305676238735517,
-      "train_runtime": 1636.5913,
-      "train_samples_per_second": 12.221,
-      "train_steps_per_second": 0.765
+      "train_loss": 0.17935538868109385,
+      "train_runtime": 1702.5744,
+      "train_samples_per_second": 11.747,
+      "train_steps_per_second": 0.735
     }
   ],
   "logging_steps": 10,