End of training

README.md

@@ -1,6 +1,11 @@
 ---
 library_name: transformers
+license: apache-2.0
+base_model: facebook/detr-resnet-50
 tags:
+- image-regression
+- human-movement
+- vision
 - generated_from_trainer
 model-index:
 - name: target_hold
@@ -12,7 +17,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # target_hold
 
-This model is a fine-tuned version of [](https://huggingface.co/) on an unknown dataset.
+This model is a fine-tuned version of [facebook/detr-resnet-50](https://huggingface.co/facebook/detr-resnet-50) on the c14kevincardenas/beta_caller_284_target_hold dataset.
 It achieves the following results on the evaluation set:
 - Loss: 0.8720
 - Iou: 0.0008

all_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 20.0,
-    "eval_IoU": 0.0,
-    "eval_loss": 1.3494025468826294,
-    "eval_runtime": 10.1301,
-    "eval_samples_per_second": 110.858,
-    "eval_steps_per_second": 1.777,
+    "eval_IoU": 0.0008468033398795426,
+    "eval_loss": 0.871971845626831,
+    "eval_runtime": 15.8906,
+    "eval_samples_per_second": 70.671,
+    "eval_steps_per_second": 1.133,
     "total_flos": 0.0,
-    "train_loss": 1.3746022605895996,
-    "train_runtime": 3619.0694,
-    "train_samples_per_second": 35.142,
-    "train_steps_per_second": 0.553
+    "train_loss": 0.9022787961959838,
+    "train_runtime": 6945.9457,
+    "train_samples_per_second": 18.31,
+    "train_steps_per_second": 0.288
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 20.0,
-    "eval_IoU": 0.0,
-    "eval_loss": 1.3494025468826294,
-    "eval_runtime": 10.1301,
-    "eval_samples_per_second": 110.858,
-    "eval_steps_per_second": 1.777
+    "eval_IoU": 0.0008468033398795426,
+    "eval_loss": 0.871971845626831,
+    "eval_runtime": 15.8906,
+    "eval_samples_per_second": 70.671,
+    "eval_steps_per_second": 1.133
 }

runs/Oct25_18-04-31_galactica.ad.cirange.net/events.out.tfevents.1729886459.galactica.ad.cirange.net.184438.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6680e58d5b569fa6fc518905ce2f201bb5bea5ca4b335027e1eac3782c745948
+size 406

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 20.0,
     "total_flos": 0.0,
-    "train_loss": 1.3746022605895996,
-    "train_runtime": 3619.0694,
-    "train_samples_per_second": 35.142,
-    "train_steps_per_second": 0.553
+    "train_loss": 0.9022787961959838,
+    "train_runtime": 6945.9457,
+    "train_samples_per_second": 18.31,
+    "train_steps_per_second": 0.288
 }

trainer_state.json

@@ -1,5 +1,5 @@
 {
-  "best_metric": 1.3494025468826294,
+  "best_metric": 0.871971845626831,
   "best_model_checkpoint": "/mnt/ml_drive/kcardenas/target_hold/checkpoint-2000",
   "epoch": 20.0,
   "eval_steps": 500,
@@ -10,752 +10,752 @@
   "log_history": [
     {
       "epoch": 0.25,
-      "grad_norm": 771.6892700195312,
+      "grad_norm": 671978.625,
       "learning_rate": 5e-06,
-      "loss": 1.4609,
+      "loss": 1.3666,
       "step": 25
     },
     {
       "epoch": 0.5,
-      "grad_norm": 742042.25,
+      "grad_norm": 686243.6875,
       "learning_rate": 1e-05,
-      "loss": 1.4537,
+      "loss": 1.3413,
       "step": 50
     },
     {
       "epoch": 0.75,
-      "grad_norm": 778067.125,
+      "grad_norm": 512857.65625,
       "learning_rate": 1.5e-05,
-      "loss": 1.4627,
+      "loss": 1.2866,
       "step": 75
     },
     {
       "epoch": 1.0,
-      "grad_norm": 913205.0625,
+      "grad_norm": 536989.5625,
       "learning_rate": 2e-05,
-      "loss": 1.4703,
+      "loss": 1.2348,
       "step": 100
     },
     {
       "epoch": 1.0,
-      "eval_IoU": 0.0,
-      "eval_loss": 1.4399977922439575,
-      "eval_runtime": 10.5794,
-      "eval_samples_per_second": 106.15,
-      "eval_steps_per_second": 1.701,
+      "eval_IoU": 6.266252117587144e-05,
+      "eval_loss": 1.1666094064712524,
+      "eval_runtime": 16.7443,
+      "eval_samples_per_second": 67.068,
+      "eval_steps_per_second": 1.075,
       "step": 100
     },
     {
       "epoch": 1.25,
-      "grad_norm": 805371.25,
+      "grad_norm": 261045.0625,
       "learning_rate": 2.5e-05,
-      "loss": 1.4476,
+      "loss": 1.1566,
       "step": 125
     },
     {
       "epoch": 1.5,
-      "grad_norm": 179319.0625,
+      "grad_norm": 215879.234375,
       "learning_rate": 3e-05,
-      "loss": 1.4673,
+      "loss": 1.094,
       "step": 150
     },
     {
       "epoch": 1.75,
-      "grad_norm": 132580.140625,
+      "grad_norm": 149507.1875,
       "learning_rate": 3.5e-05,
-      "loss": 1.4393,
+      "loss": 1.037,
       "step": 175
     },
     {
       "epoch": 2.0,
-      "grad_norm": 4426164.5,
+      "grad_norm": 179843.1875,
       "learning_rate": 4e-05,
-      "loss": 1.4686,
+      "loss": 1.0043,
       "step": 200
     },
     {
       "epoch": 2.0,
-      "eval_IoU": 0.0,
-      "eval_loss": 1.4274886846542358,
-      "eval_runtime": 10.333,
-      "eval_samples_per_second": 108.681,
-      "eval_steps_per_second": 1.742,
+      "eval_IoU": 0.002259200195430286,
+      "eval_loss": 0.9816299080848694,
+      "eval_runtime": 16.1893,
+      "eval_samples_per_second": 69.367,
+      "eval_steps_per_second": 1.112,
       "step": 200
     },
     {
       "epoch": 2.25,
-      "grad_norm": 3633516.5,
+      "grad_norm": 79894.609375,
       "learning_rate": 4.5e-05,
-      "loss": 1.4373,
+      "loss": 0.9636,
       "step": 225
     },
     {
       "epoch": 2.5,
-      "grad_norm": 602330.6875,
+      "grad_norm": 58901.21484375,
       "learning_rate": 5e-05,
-      "loss": 1.4544,
+      "loss": 0.9395,
       "step": 250
     },
     {
       "epoch": 2.75,
-      "grad_norm": 3835361.5,
+      "grad_norm": 46244.05859375,
       "learning_rate": 4.928571428571429e-05,
-      "loss": 1.4161,
+      "loss": 0.9225,
       "step": 275
     },
     {
       "epoch": 3.0,
-      "grad_norm": 4627148.5,
+      "grad_norm": 31114.857421875,
       "learning_rate": 4.8571428571428576e-05,
-      "loss": 1.463,
+      "loss": 0.9101,
       "step": 300
     },
     {
       "epoch": 3.0,
-      "eval_IoU": 0.0,
-      "eval_loss": 1.5115046501159668,
-      "eval_runtime": 10.026,
-      "eval_samples_per_second": 112.009,
-      "eval_steps_per_second": 1.795,
+      "eval_IoU": 0.002016003165160829,
+      "eval_loss": 0.9057585597038269,
+      "eval_runtime": 16.1324,
+      "eval_samples_per_second": 69.611,
+      "eval_steps_per_second": 1.116,
       "step": 300
     },
     {
       "epoch": 3.25,
-      "grad_norm": 5056040.5,
+      "grad_norm": 25909.61328125,
       "learning_rate": 4.785714285714286e-05,
-      "loss": 1.4849,
+      "loss": 0.9011,
       "step": 325
     },
     {
       "epoch": 3.5,
-      "grad_norm": 1441977.25,
+      "grad_norm": 25156.486328125,
       "learning_rate": 4.714285714285714e-05,
-      "loss": 1.4628,
+      "loss": 0.8919,
       "step": 350
     },
     {
       "epoch": 3.75,
-      "grad_norm": 123561.8515625,
+      "grad_norm": 31189.755859375,
       "learning_rate": 4.642857142857143e-05,
-      "loss": 1.4627,
+      "loss": 0.8876,
       "step": 375
     },
     {
       "epoch": 4.0,
-      "grad_norm": 1065073.75,
+      "grad_norm": 23475.142578125,
       "learning_rate": 4.5714285714285716e-05,
-      "loss": 1.4399,
+      "loss": 0.8846,
       "step": 400
     },
     {
       "epoch": 4.0,
-      "eval_IoU": 0.0,
-      "eval_loss": 1.4285613298416138,
-      "eval_runtime": 10.0206,
-      "eval_samples_per_second": 112.069,
-      "eval_steps_per_second": 1.796,
+      "eval_IoU": 0.001328349647772117,
+      "eval_loss": 0.8883054256439209,
+      "eval_runtime": 16.0469,
+      "eval_samples_per_second": 69.982,
+      "eval_steps_per_second": 1.122,
       "step": 400
     },
     {
       "epoch": 4.25,
-      "grad_norm": 1083903.625,
+      "grad_norm": 17651.619140625,
       "learning_rate": 4.5e-05,
-      "loss": 1.4181,
+      "loss": 0.8822,
       "step": 425
     },
     {
       "epoch": 4.5,
-      "grad_norm": 989387.875,
+      "grad_norm": 12211.6630859375,
       "learning_rate": 4.428571428571428e-05,
-      "loss": 1.3905,
+      "loss": 0.8777,
       "step": 450
     },
     {
       "epoch": 4.75,
-      "grad_norm": 504863.90625,
+      "grad_norm": 28110.140625,
       "learning_rate": 4.3571428571428576e-05,
-      "loss": 1.4001,
+      "loss": 0.8786,
       "step": 475
     },
     {
       "epoch": 5.0,
-      "grad_norm": 1111330.75,
+      "grad_norm": 10401.6005859375,
       "learning_rate": 4.2857142857142856e-05,
-      "loss": 1.4252,
+      "loss": 0.8755,
       "step": 500
     },
     {
       "epoch": 5.0,
-      "eval_IoU": 0.0,
-      "eval_loss": 1.4052964448928833,
-      "eval_runtime": 10.1806,
-      "eval_samples_per_second": 110.307,
-      "eval_steps_per_second": 1.768,
+      "eval_IoU": 0.0011379811073136117,
+      "eval_loss": 0.8818888664245605,
+      "eval_runtime": 15.912,
+      "eval_samples_per_second": 70.575,
+      "eval_steps_per_second": 1.131,
       "step": 500
     },
     {
       "epoch": 5.25,
-      "grad_norm": 20536042.0,
+      "grad_norm": 14044.361328125,
       "learning_rate": 4.214285714285714e-05,
-      "loss": 1.4065,
+      "loss": 0.8734,
       "step": 525
     },
     {
       "epoch": 5.5,
-      "grad_norm": 1043716.0625,
+      "grad_norm": 8820.140625,
       "learning_rate": 4.1428571428571437e-05,
-      "loss": 1.4236,
+      "loss": 0.875,
       "step": 550
     },
     {
       "epoch": 5.75,
-      "grad_norm": 6279096.0,
+      "grad_norm": 7751.67822265625,
       "learning_rate": 4.0714285714285717e-05,
-      "loss": 1.4061,
+      "loss": 0.8758,
       "step": 575
     },
     {
       "epoch": 6.0,
-      "grad_norm": 1782540.375,
+      "grad_norm": 29417.427734375,
       "learning_rate": 4e-05,
-      "loss": 1.3961,
+      "loss": 0.8714,
       "step": 600
     },
     {
       "epoch": 6.0,
-      "eval_IoU": 0.0,
-      "eval_loss": 1.3607368469238281,
-      "eval_runtime": 10.205,
-      "eval_samples_per_second": 110.045,
-      "eval_steps_per_second": 1.764,
+      "eval_IoU": 0.0010282730649858857,
+      "eval_loss": 0.8788518905639648,
+      "eval_runtime": 16.1196,
+      "eval_samples_per_second": 69.667,
+      "eval_steps_per_second": 1.117,
       "step": 600
     },
     {
       "epoch": 6.25,
-      "grad_norm": 2240105.75,
+      "grad_norm": 13193.9521484375,
       "learning_rate": 3.928571428571429e-05,
-      "loss": 1.45,
+      "loss": 0.87,
       "step": 625
     },
     {
       "epoch": 6.5,
-      "grad_norm": 778888.9375,
+      "grad_norm": 9639.7734375,
       "learning_rate": 3.857142857142858e-05,
-      "loss": 1.4902,
+      "loss": 0.8709,
       "step": 650
     },
     {
       "epoch": 6.75,
-      "grad_norm": 871022.9375,
+      "grad_norm": 42599.96484375,
       "learning_rate": 3.785714285714286e-05,
-      "loss": 1.3991,
+      "loss": 0.8693,
       "step": 675
     },
     {
       "epoch": 7.0,
-      "grad_norm": 2324823.5,
+      "grad_norm": 22185.576171875,
       "learning_rate": 3.7142857142857143e-05,
-      "loss": 1.3791,
+      "loss": 0.8684,
       "step": 700
     },
     {
       "epoch": 7.0,
-      "eval_IoU": 0.0,
-      "eval_loss": 1.3694654703140259,
-      "eval_runtime": 9.9921,
-      "eval_samples_per_second": 112.389,
-      "eval_steps_per_second": 1.801,
+      "eval_IoU": 0.0009165315499091501,
+      "eval_loss": 0.8773335814476013,
+      "eval_runtime": 15.6201,
+      "eval_samples_per_second": 71.894,
+      "eval_steps_per_second": 1.152,
       "step": 700
     },
     {
       "epoch": 7.25,
-      "grad_norm": 842470.1875,
+      "grad_norm": 8237.5849609375,
       "learning_rate": 3.642857142857143e-05,
-      "loss": 1.3698,
+      "loss": 0.8692,
       "step": 725
     },
     {
       "epoch": 7.5,
-      "grad_norm": 1414357.5,
+      "grad_norm": 9722.630859375,
       "learning_rate": 3.571428571428572e-05,
-      "loss": 1.3673,
+      "loss": 0.8709,
       "step": 750
     },
     {
       "epoch": 7.75,
-      "grad_norm": 1465819.625,
+      "grad_norm": 11062.8916015625,
       "learning_rate": 3.5e-05,
-      "loss": 1.3714,
+      "loss": 0.8652,
       "step": 775
     },
     {
       "epoch": 8.0,
-      "grad_norm": 5227108.0,
+      "grad_norm": 34961.71484375,
       "learning_rate": 3.428571428571429e-05,
-      "loss": 1.3555,
+      "loss": 0.8664,
       "step": 800
     },
     {
       "epoch": 8.0,
-      "eval_IoU": 0.0,
-      "eval_loss": 1.369844675064087,
-      "eval_runtime": 10.1046,
-      "eval_samples_per_second": 111.138,
-      "eval_steps_per_second": 1.781,
+      "eval_IoU": 0.0008482180748956496,
+      "eval_loss": 0.8764283061027527,
+      "eval_runtime": 15.6883,
+      "eval_samples_per_second": 71.582,
+      "eval_steps_per_second": 1.147,
       "step": 800
     },
     {
       "epoch": 8.25,
-      "grad_norm": 1675220.125,
+      "grad_norm": 11363.548828125,
       "learning_rate": 3.357142857142857e-05,
-      "loss": 1.4224,
+      "loss": 0.8658,
       "step": 825
     },
     {
       "epoch": 8.5,
-      "grad_norm": 2374479.0,
+      "grad_norm": 21425.958984375,
       "learning_rate": 3.285714285714286e-05,
-      "loss": 1.3854,
+      "loss": 0.8673,
       "step": 850
     },
     {
       "epoch": 8.75,
-      "grad_norm": 741768.375,
+      "grad_norm": 20737.205078125,
       "learning_rate": 3.2142857142857144e-05,
-      "loss": 1.3449,
+      "loss": 0.8637,
       "step": 875
     },
     {
       "epoch": 9.0,
-      "grad_norm": 4549860.0,
+      "grad_norm": 15000.978515625,
       "learning_rate": 3.142857142857143e-05,
-      "loss": 1.3402,
+      "loss": 0.8677,
       "step": 900
     },
     {
       "epoch": 9.0,
-      "eval_IoU": 0.0,
-      "eval_loss": 1.3592411279678345,
-      "eval_runtime": 9.9765,
-      "eval_samples_per_second": 112.564,
-      "eval_steps_per_second": 1.804,
+      "eval_IoU": 0.0008826124813546499,
+      "eval_loss": 0.875173807144165,
+      "eval_runtime": 15.8392,
+      "eval_samples_per_second": 70.9,
+      "eval_steps_per_second": 1.136,
       "step": 900
     },
     {
       "epoch": 9.25,
-      "grad_norm": 29592694.0,
+      "grad_norm": 20330.591796875,
       "learning_rate": 3.071428571428572e-05,
-      "loss": 1.3638,
+      "loss": 0.8675,
       "step": 925
     },
     {
       "epoch": 9.5,
-      "grad_norm": 1362061.625,
+      "grad_norm": 14069.2763671875,
       "learning_rate": 3e-05,
-      "loss": 1.3651,
+      "loss": 0.8662,
       "step": 950
     },
     {
       "epoch": 9.75,
-      "grad_norm": 4175936.75,
+      "grad_norm": 18826.400390625,
       "learning_rate": 2.9285714285714288e-05,
-      "loss": 1.3495,
+      "loss": 0.8625,
       "step": 975
     },
     {
       "epoch": 10.0,
-      "grad_norm": 1573651.75,
+      "grad_norm": 19264.611328125,
       "learning_rate": 2.857142857142857e-05,
-      "loss": 1.3551,
+      "loss": 0.863,
       "step": 1000
     },
     {
       "epoch": 10.0,
-      "eval_IoU": 0.0,
-      "eval_loss": 1.378575086593628,
-      "eval_runtime": 10.1575,
-      "eval_samples_per_second": 110.559,
-      "eval_steps_per_second": 1.772,
+      "eval_IoU": 0.0008654725467694338,
+      "eval_loss": 0.8746740221977234,
+      "eval_runtime": 16.0579,
+      "eval_samples_per_second": 69.935,
+      "eval_steps_per_second": 1.121,
       "step": 1000
     },
     {
       "epoch": 10.25,
-      "grad_norm": 2796916.25,
+      "grad_norm": 8239.904296875,
       "learning_rate": 2.785714285714286e-05,
-      "loss": 1.3547,
+      "loss": 0.8644,
       "step": 1025
     },
     {
       "epoch": 10.5,
-      "grad_norm": 5406704.5,
+      "grad_norm": 11806.5556640625,
       "learning_rate": 2.714285714285714e-05,
-      "loss": 1.3367,
+      "loss": 0.8629,
       "step": 1050
     },
     {
       "epoch": 10.75,
-      "grad_norm": 1702886.625,
+      "grad_norm": 8895.9296875,
       "learning_rate": 2.642857142857143e-05,
-      "loss": 1.3551,
+      "loss": 0.8665,
       "step": 1075
     },
     {
       "epoch": 11.0,
-      "grad_norm": 3208807.0,
+      "grad_norm": 14509.6494140625,
       "learning_rate": 2.5714285714285714e-05,
-      "loss": 1.3517,
+      "loss": 0.8619,
       "step": 1100
     },
     {
       "epoch": 11.0,
-      "eval_IoU": 0.0,
-      "eval_loss": 1.3585208654403687,
-      "eval_runtime": 10.1978,
-      "eval_samples_per_second": 110.122,
-      "eval_steps_per_second": 1.765,
+      "eval_IoU": 0.0008878591602175693,
+      "eval_loss": 0.873702883720398,
+      "eval_runtime": 15.7003,
+      "eval_samples_per_second": 71.527,
+      "eval_steps_per_second": 1.146,
       "step": 1100
     },
     {
       "epoch": 11.25,
-      "grad_norm": 676928.4375,
+      "grad_norm": 16054.029296875,
       "learning_rate": 2.5e-05,
-      "loss": 1.3503,
+      "loss": 0.8632,
       "step": 1125
     },
     {
       "epoch": 11.5,
-      "grad_norm": 2296206.75,
+      "grad_norm": 13585.3857421875,
       "learning_rate": 2.4285714285714288e-05,
-      "loss": 1.344,
+      "loss": 0.8644,
       "step": 1150
     },
     {
       "epoch": 11.75,
-      "grad_norm": 21383138.0,
+      "grad_norm": 13984.9892578125,
       "learning_rate": 2.357142857142857e-05,
-      "loss": 1.3309,
+      "loss": 0.8627,
       "step": 1175
     },
     {
       "epoch": 12.0,
-      "grad_norm": 2441665.5,
+      "grad_norm": 10830.5859375,
       "learning_rate": 2.2857142857142858e-05,
-      "loss": 1.3629,
+      "loss": 0.8637,
       "step": 1200
     },
     {
       "epoch": 12.0,
-      "eval_IoU": 0.0,
-      "eval_loss": 1.3613262176513672,
-      "eval_runtime": 10.2624,
-      "eval_samples_per_second": 109.429,
-      "eval_steps_per_second": 1.754,
+      "eval_IoU": 0.0008766921482666689,
+      "eval_loss": 0.8732270002365112,
+      "eval_runtime": 15.714,
+      "eval_samples_per_second": 71.465,
+      "eval_steps_per_second": 1.145,
       "step": 1200
     },
     {
       "epoch": 12.25,
-      "grad_norm": 4257276.5,
+      "grad_norm": 9003.2802734375,
       "learning_rate": 2.214285714285714e-05,
-      "loss": 1.3485,
+      "loss": 0.8638,
       "step": 1225
     },
     {
       "epoch": 12.5,
-      "grad_norm": 810549.6875,
+      "grad_norm": 20906.34375,
       "learning_rate": 2.1428571428571428e-05,
-      "loss": 1.3481,
+      "loss": 0.8607,
       "step": 1250
     },
     {
       "epoch": 12.75,
-      "grad_norm": 1533887.125,
+      "grad_norm": 6575.8974609375,
       "learning_rate": 2.0714285714285718e-05,
-      "loss": 1.3258,
+      "loss": 0.8616,
       "step": 1275
     },
     {
       "epoch": 13.0,
-      "grad_norm": 4748758.0,
+      "grad_norm": 12234.8251953125,
       "learning_rate": 2e-05,
-      "loss": 1.3322,
+      "loss": 0.8632,
       "step": 1300
     },
     {
       "epoch": 13.0,
-      "eval_IoU": 0.0,
-      "eval_loss": 1.3555089235305786,
-      "eval_runtime": 10.3606,
-      "eval_samples_per_second": 108.391,
-      "eval_steps_per_second": 1.737,
+      "eval_IoU": 0.0008608651426555169,
+      "eval_loss": 0.8730344772338867,
+      "eval_runtime": 16.3659,
+      "eval_samples_per_second": 68.618,
+      "eval_steps_per_second": 1.1,
       "step": 1300
     },
     {
       "epoch": 13.25,
-      "grad_norm": 2167070.25,
+      "grad_norm": 8454.9755859375,
       "learning_rate": 1.928571428571429e-05,
-      "loss": 1.3428,
+      "loss": 0.864,
       "step": 1325
     },
     {
       "epoch": 13.5,
-      "grad_norm": 1552955.5,
+      "grad_norm": 16713.82421875,
       "learning_rate": 1.8571428571428572e-05,
-      "loss": 1.3287,
+      "loss": 0.8623,
       "step": 1350
     },
     {
       "epoch": 13.75,
-      "grad_norm": 4735588.5,
+      "grad_norm": 17913.958984375,
       "learning_rate": 1.785714285714286e-05,
-      "loss": 1.3376,
+      "loss": 0.8649,
       "step": 1375
     },
     {
       "epoch": 14.0,
-      "grad_norm": 6000364.0,
+      "grad_norm": 33508.1640625,
       "learning_rate": 1.7142857142857145e-05,
-      "loss": 1.3185,
+      "loss": 0.8581,
       "step": 1400
     },
     {
       "epoch": 14.0,
-      "eval_IoU": 0.0,
-      "eval_loss": 1.3536880016326904,
-      "eval_runtime": 10.1961,
-      "eval_samples_per_second": 110.14,
-      "eval_steps_per_second": 1.765,
+      "eval_IoU": 0.0008642116435639467,
+      "eval_loss": 0.8726587295532227,
+      "eval_runtime": 15.9379,
+      "eval_samples_per_second": 70.461,
+      "eval_steps_per_second": 1.129,
       "step": 1400
     },
     {
       "epoch": 14.25,
-      "grad_norm": 1262304.125,
+      "grad_norm": 10902.5810546875,
       "learning_rate": 1.642857142857143e-05,
-      "loss": 1.3381,
+      "loss": 0.8626,
       "step": 1425
     },
     {
       "epoch": 14.5,
-      "grad_norm": 923730.6875,
+      "grad_norm": 5604.90576171875,
       "learning_rate": 1.5714285714285715e-05,
-      "loss": 1.3302,
+      "loss": 0.8581,
       "step": 1450
     },
     {
       "epoch": 14.75,
-      "grad_norm": 654002.0,
+      "grad_norm": 11584.734375,
       "learning_rate": 1.5e-05,
-      "loss": 1.3242,
+      "loss": 0.8644,
       "step": 1475
     },
     {
       "epoch": 15.0,
-      "grad_norm": 1146749.125,
+      "grad_norm": 10246.080078125,
       "learning_rate": 1.4285714285714285e-05,
-      "loss": 1.3254,
+      "loss": 0.8615,
       "step": 1500
     },
     {
       "epoch": 15.0,
-      "eval_IoU": 0.0,
-      "eval_loss": 1.3588523864746094,
-      "eval_runtime": 10.2312,
-      "eval_samples_per_second": 109.762,
-      "eval_steps_per_second": 1.759,
+      "eval_IoU": 0.0008597120680148628,
+      "eval_loss": 0.8724328279495239,
+      "eval_runtime": 15.6896,
+      "eval_samples_per_second": 71.576,
+      "eval_steps_per_second": 1.147,
       "step": 1500
     },
     {
       "epoch": 15.25,
-      "grad_norm": 1684872.625,
+      "grad_norm": 8243.01953125,
       "learning_rate": 1.357142857142857e-05,
-      "loss": 1.3347,
+      "loss": 0.8617,
       "step": 1525
     },
     {
       "epoch": 15.5,
-      "grad_norm": 1137702.0,
+      "grad_norm": 9738.013671875,
       "learning_rate": 1.2857142857142857e-05,
-      "loss": 1.3213,
+      "loss": 0.8618,
       "step": 1550
     },
     {
       "epoch": 15.75,
-      "grad_norm": 1545808.0,
+      "grad_norm": 10498.345703125,
       "learning_rate": 1.2142857142857144e-05,
-      "loss": 1.3084,
+      "loss": 0.8603,
       "step": 1575
     },
     {
       "epoch": 16.0,
-      "grad_norm": 3496195.75,
+      "grad_norm": 31001.509765625,
       "learning_rate": 1.1428571428571429e-05,
-      "loss": 1.3343,
+      "loss": 0.8604,
       "step": 1600
     },
     {
       "epoch": 16.0,
-      "eval_IoU": 0.0,
-      "eval_loss": 1.3545335531234741,
-      "eval_runtime": 10.4337,
-      "eval_samples_per_second": 107.632,
-      "eval_steps_per_second": 1.725,
+      "eval_IoU": 0.0008376356411078918,
+      "eval_loss": 0.8724232912063599,
+      "eval_runtime": 16.5134,
+      "eval_samples_per_second": 68.005,
+      "eval_steps_per_second": 1.09,
       "step": 1600
     },
     {
       "epoch": 16.25,
-      "grad_norm": 3201461.5,
+      "grad_norm": 23480.4375,
       "learning_rate": 1.0714285714285714e-05,
-      "loss": 1.3219,
+      "loss": 0.8622,
       "step": 1625
     },
     {
       "epoch": 16.5,
-      "grad_norm": 1067260.875,
+      "grad_norm": 10336.7646484375,
       "learning_rate": 1e-05,
-      "loss": 1.3258,
+      "loss": 0.8593,
       "step": 1650
     },
     {
       "epoch": 16.75,
-      "grad_norm": 2405265.25,
+      "grad_norm": 13781.5810546875,
       "learning_rate": 9.285714285714286e-06,
-      "loss": 1.3412,
+      "loss": 0.8627,
       "step": 1675
     },
     {
       "epoch": 17.0,
-      "grad_norm": 1497889.125,
+      "grad_norm": 8223.326171875,
       "learning_rate": 8.571428571428573e-06,
-      "loss": 1.3282,
+      "loss": 0.8606,
       "step": 1700
     },
     {
       "epoch": 17.0,
-      "eval_IoU": 0.0,
-      "eval_loss": 1.3675432205200195,
-      "eval_runtime": 10.0896,
-      "eval_samples_per_second": 111.302,
-      "eval_steps_per_second": 1.784,
+      "eval_IoU": 0.0008638472124057635,
+      "eval_loss": 0.8720372319221497,
+      "eval_runtime": 16.0284,
+      "eval_samples_per_second": 70.063,
+      "eval_steps_per_second": 1.123,
       "step": 1700
     },
     {
       "epoch": 17.25,
-      "grad_norm": 2305034.25,
+      "grad_norm": 6056.0810546875,
       "learning_rate": 7.857142857142858e-06,
-      "loss": 1.3297,
+      "loss": 0.8616,
       "step": 1725
     },
     {
       "epoch": 17.5,
-      "grad_norm": 1990793.25,
+      "grad_norm": 13163.3916015625,
       "learning_rate": 7.142857142857143e-06,
-      "loss": 1.3318,
+      "loss": 0.8626,
       "step": 1750
     },
     {
       "epoch": 17.75,
-      "grad_norm": 832195.9375,
+      "grad_norm": 7198.291015625,
       "learning_rate": 6.428571428571429e-06,
-      "loss": 1.3334,
+      "loss": 0.862,
       "step": 1775
     },
     {
       "epoch": 18.0,
-      "grad_norm": 3059447.75,
+      "grad_norm": 30992.525390625,
       "learning_rate": 5.7142857142857145e-06,
-      "loss": 1.335,
+      "loss": 0.8592,
       "step": 1800
     },
     {
       "epoch": 18.0,
-      "eval_IoU": 0.0,
-      "eval_loss": 1.3514825105667114,
-      "eval_runtime": 10.1648,
-      "eval_samples_per_second": 110.48,
-      "eval_steps_per_second": 1.771,
+      "eval_IoU": 0.000850173645477854,
+      "eval_loss": 0.8720242381095886,
+      "eval_runtime": 15.5216,
+      "eval_samples_per_second": 72.351,
+      "eval_steps_per_second": 1.16,
       "step": 1800
     },
     {
       "epoch": 18.25,
-      "grad_norm": 2467115.5,
+      "grad_norm": 8230.2451171875,
       "learning_rate": 5e-06,
-      "loss": 1.3281,
+      "loss": 0.859,
       "step": 1825
     },
     {
       "epoch": 18.5,
-      "grad_norm": 1106492.0,
+      "grad_norm": 9012.4931640625,
       "learning_rate": 4.285714285714286e-06,
-      "loss": 1.3091,
+      "loss": 0.8601,
       "step": 1850
     },
     {
       "epoch": 18.75,
-      "grad_norm": 2224053.0,
+      "grad_norm": 12221.1279296875,
       "learning_rate": 3.5714285714285714e-06,
-      "loss": 1.3299,
+      "loss": 0.8608,
       "step": 1875
     },
     {
       "epoch": 19.0,
-      "grad_norm": 2320941.0,
+      "grad_norm": 26469.41015625,
       "learning_rate": 2.8571428571428573e-06,
-      "loss": 1.3259,
+      "loss": 0.8621,
       "step": 1900
     },
     {
       "epoch": 19.0,
-      "eval_IoU": 0.0,
-      "eval_loss": 1.3499101400375366,
-      "eval_runtime": 9.933,
-      "eval_samples_per_second": 113.057,
-      "eval_steps_per_second": 1.812,
+      "eval_IoU": 0.0008449985251271719,
+      "eval_loss": 0.8720095753669739,
+      "eval_runtime": 15.4454,
+      "eval_samples_per_second": 72.708,
+      "eval_steps_per_second": 1.165,
       "step": 1900
     },
     {
       "epoch": 19.25,
-      "grad_norm": 1948301.25,
+      "grad_norm": 12594.4345703125,
       "learning_rate": 2.142857142857143e-06,
-      "loss": 1.3186,
+      "loss": 0.8623,
       "step": 1925
     },
     {
       "epoch": 19.5,
-      "grad_norm": 497170.53125,
+      "grad_norm": 7997.478515625,
       "learning_rate": 1.4285714285714286e-06,
-      "loss": 1.3113,
+      "loss": 0.8597,
       "step": 1950
     },
     {
       "epoch": 19.75,
-      "grad_norm": 257136.8125,
+      "grad_norm": 8224.8974609375,
       "learning_rate": 7.142857142857143e-07,
-      "loss": 1.3326,
+      "loss": 0.8605,
       "step": 1975
     },
     {
       "epoch": 20.0,
-      "grad_norm": 734737.9375,
+      "grad_norm": 18461.4453125,
       "learning_rate": 0.0,
-      "loss": 1.3144,
+      "loss": 0.8629,
       "step": 2000
     },
     {
       "epoch": 20.0,
-      "eval_IoU": 0.0,
-      "eval_loss": 1.3494025468826294,
-      "eval_runtime": 10.2833,
-      "eval_samples_per_second": 109.206,
-      "eval_steps_per_second": 1.75,
+      "eval_IoU": 0.0008468033398795426,
+      "eval_loss": 0.871971845626831,
+      "eval_runtime": 15.8784,
+      "eval_samples_per_second": 70.725,
+      "eval_steps_per_second": 1.134,
       "step": 2000
     },
     {
       "epoch": 20.0,
       "step": 2000,
       "total_flos": 0.0,
-      "train_loss": 1.3746022605895996,
-      "train_runtime": 3619.0694,
-      "train_samples_per_second": 35.142,
-      "train_steps_per_second": 0.553
+      "train_loss": 0.9022787961959838,
+      "train_runtime": 6945.9457,
+      "train_samples_per_second": 18.31,
+      "train_steps_per_second": 0.288
     }
   ],
   "logging_steps": 25,