🍻 cheers

all_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 4.0,
-    "total_flos": 3.233737255326843e+17,
-    "train_loss": 0.21213595934140567,
-    "train_runtime": 75.7831,
-    "train_samples_per_second": 54.577,
-    "train_steps_per_second": 3.431
+    "total_flos": 3.205097416476426e+17,
+    "train_loss": 0.13431097194552422,
+    "train_runtime": 68.8344,
+    "train_samples_per_second": 60.086,
+    "train_steps_per_second": 3.777
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 4.0,
-    "total_flos": 3.233737255326843e+17,
-    "train_loss": 0.21213595934140567,
-    "train_runtime": 75.7831,
-    "train_samples_per_second": 54.577,
-    "train_steps_per_second": 3.431
+    "total_flos": 3.205097416476426e+17,
+    "train_loss": 0.13431097194552422,
+    "train_runtime": 68.8344,
+    "train_samples_per_second": 60.086,
+    "train_steps_per_second": 3.777
 }

trainer_state.json

@@ -1,5 +1,5 @@
 {
-  "best_metric": 0.006127053406089544,
+  "best_metric": 0.013820217922329903,
   "best_model_checkpoint": "./vit-base-beans/checkpoint-200",
   "epoch": 4.0,
   "global_step": 260,
@@ -9,191 +9,191 @@
   "log_history": [
     {
       "epoch": 0.15,
-      "learning_rate": 0.0001930769230769231,
-      "loss": 2.9271,
+      "learning_rate": 0.00019230769230769233,
+      "loss": 0.8816,
       "step": 10
     },
     {
       "epoch": 0.31,
-      "learning_rate": 0.0001853846153846154,
-      "loss": 0.4676,
+      "learning_rate": 0.00018461538461538463,
+      "loss": 0.4369,
       "step": 20
     },
     {
       "epoch": 0.46,
-      "learning_rate": 0.0001776923076923077,
-      "loss": 0.3661,
+      "learning_rate": 0.00017692307692307693,
+      "loss": 0.2885,
       "step": 30
     },
     {
       "epoch": 0.62,
-      "learning_rate": 0.00017,
-      "loss": 0.2549,
+      "learning_rate": 0.00016923076923076923,
+      "loss": 0.1892,
       "step": 40
     },
     {
       "epoch": 0.77,
       "learning_rate": 0.0001623076923076923,
-      "loss": 0.229,
+      "loss": 0.2412,
       "step": 50
     },
     {
       "epoch": 0.92,
       "learning_rate": 0.00015461538461538464,
-      "loss": 0.2858,
+      "loss": 0.265,
       "step": 60
     },
     {
       "epoch": 1.08,
       "learning_rate": 0.00014692307692307693,
-      "loss": 0.1084,
+      "loss": 0.1797,
       "step": 70
     },
     {
       "epoch": 1.23,
       "learning_rate": 0.00013923076923076923,
-      "loss": 0.1062,
+      "loss": 0.0923,
       "step": 80
     },
     {
       "epoch": 1.38,
       "learning_rate": 0.00013153846153846156,
-      "loss": 0.1174,
+      "loss": 0.0401,
       "step": 90
     },
     {
       "epoch": 1.54,
       "learning_rate": 0.00012384615384615385,
-      "loss": 0.0994,
+      "loss": 0.1595,
       "step": 100
     },
     {
       "epoch": 1.54,
-      "eval_accuracy": 0.9849624060150376,
-      "eval_loss": 0.05971372127532959,
-      "eval_runtime": 1.0813,
-      "eval_samples_per_second": 122.997,
-      "eval_steps_per_second": 15.721,
+      "eval_accuracy": 0.9699248120300752,
+      "eval_loss": 0.1211724504828453,
+      "eval_runtime": 1.0092,
+      "eval_samples_per_second": 131.786,
+      "eval_steps_per_second": 16.845,
       "step": 100
     },
     {
       "epoch": 1.69,
       "learning_rate": 0.00011615384615384617,
-      "loss": 0.1219,
+      "loss": 0.0987,
       "step": 110
     },
     {
       "epoch": 1.85,
       "learning_rate": 0.00010846153846153846,
-      "loss": 0.0793,
+      "loss": 0.1505,
       "step": 120
     },
     {
       "epoch": 2.0,
       "learning_rate": 0.00010076923076923077,
-      "loss": 0.0741,
+      "loss": 0.0768,
       "step": 130
     },
     {
       "epoch": 2.15,
       "learning_rate": 9.307692307692309e-05,
-      "loss": 0.0281,
+      "loss": 0.0241,
       "step": 140
     },
     {
       "epoch": 2.31,
       "learning_rate": 8.538461538461538e-05,
-      "loss": 0.0309,
+      "loss": 0.0415,
       "step": 150
     },
     {
       "epoch": 2.46,
       "learning_rate": 7.76923076923077e-05,
-      "loss": 0.1636,
+      "loss": 0.07,
       "step": 160
     },
     {
       "epoch": 2.62,
       "learning_rate": 7e-05,
-      "loss": 0.0244,
+      "loss": 0.0488,
       "step": 170
     },
     {
       "epoch": 2.77,
       "learning_rate": 6.23076923076923e-05,
-      "loss": 0.0115,
+      "loss": 0.0738,
       "step": 180
     },
     {
       "epoch": 2.92,
       "learning_rate": 5.461538461538461e-05,
-      "loss": 0.0046,
+      "loss": 0.0162,
       "step": 190
     },
     {
       "epoch": 3.08,
       "learning_rate": 4.692307692307693e-05,
-      "loss": 0.001,
+      "loss": 0.014,
       "step": 200
     },
     {
       "epoch": 3.08,
       "eval_accuracy": 1.0,
-      "eval_loss": 0.006127053406089544,
-      "eval_runtime": 1.5958,
-      "eval_samples_per_second": 83.342,
-      "eval_steps_per_second": 10.653,
+      "eval_loss": 0.013820217922329903,
+      "eval_runtime": 1.0516,
+      "eval_samples_per_second": 126.478,
+      "eval_steps_per_second": 16.166,
       "step": 200
     },
     {
       "epoch": 3.23,
       "learning_rate": 3.923076923076923e-05,
-      "loss": 0.0081,
+      "loss": 0.0131,
       "step": 210
     },
     {
       "epoch": 3.38,
       "learning_rate": 3.153846153846154e-05,
-      "loss": 0.0004,
+      "loss": 0.0421,
       "step": 220
     },
     {
       "epoch": 3.54,
       "learning_rate": 2.384615384615385e-05,
-      "loss": 0.0009,
+      "loss": 0.0122,
       "step": 230
     },
     {
       "epoch": 3.69,
       "learning_rate": 1.6153846153846154e-05,
-      "loss": 0.0004,
+      "loss": 0.0123,
       "step": 240
     },
     {
       "epoch": 3.85,
       "learning_rate": 8.461538461538462e-06,
-      "loss": 0.0019,
+      "loss": 0.0121,
       "step": 250
     },
     {
       "epoch": 4.0,
       "learning_rate": 7.692307692307694e-07,
-      "loss": 0.0026,
+      "loss": 0.012,
       "step": 260
     },
     {
       "epoch": 4.0,
       "step": 260,
-      "total_flos": 3.233737255326843e+17,
-      "train_loss": 0.21213595934140567,
-      "train_runtime": 75.7831,
-      "train_samples_per_second": 54.577,
-      "train_steps_per_second": 3.431
+      "total_flos": 3.205097416476426e+17,
+      "train_loss": 0.13431097194552422,
+      "train_runtime": 68.8344,
+      "train_samples_per_second": 60.086,
+      "train_steps_per_second": 3.777
     }
   ],
   "max_steps": 260,
   "num_train_epochs": 4,
-  "total_flos": 3.233737255326843e+17,
+  "total_flos": 3.205097416476426e+17,
   "trial_name": null,
   "trial_params": null
 }