🍻 cheers

all_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 4.0,
-    "eval_accuracy": 1.0,
-    "eval_loss": 0.38402849435806274,
-    "eval_runtime": 1.7644,
-    "eval_samples_per_second": 117.888,
-    "eval_steps_per_second": 7.368,
-    "total_flos": 3.22436019290112e+17,
-    "train_loss": 0.9762612545129025,
-    "train_runtime": 74.7912,
-    "train_samples_per_second": 55.622,
-    "train_steps_per_second": 1.765
+    "eval_accuracy": 0.9572649572649573,
+    "eval_loss": 0.3159617781639099,
+    "eval_runtime": 1.9143,
+    "eval_samples_per_second": 122.237,
+    "eval_steps_per_second": 7.836,
+    "total_flos": 7.25481043402752e+17,
+    "train_loss": 0.6950366432602341,
+    "train_runtime": 183.7888,
+    "train_samples_per_second": 50.928,
+    "train_steps_per_second": 1.611
 }

runs/Jan12_00-14-49_7ecde4810381/events.out.tfevents.1705018829.7ecde4810381.24.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b1abb09f82b29639033ec2d8840fe5f97bcc0dc901e6545208ddcab894e7d997
+size 363

test_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 4.0,
-    "eval_accuracy": 1.0,
-    "eval_loss": 0.38402849435806274,
-    "eval_runtime": 1.7644,
-    "eval_samples_per_second": 117.888,
-    "eval_steps_per_second": 7.368
+    "eval_accuracy": 0.9572649572649573,
+    "eval_loss": 0.3159617781639099,
+    "eval_runtime": 1.9143,
+    "eval_samples_per_second": 122.237,
+    "eval_steps_per_second": 7.836
 }

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 4.0,
-    "total_flos": 3.22436019290112e+17,
-    "train_loss": 0.9762612545129025,
-    "train_runtime": 74.7912,
-    "train_samples_per_second": 55.622,
-    "train_steps_per_second": 1.765
+    "total_flos": 7.25481043402752e+17,
+    "train_loss": 0.6950366432602341,
+    "train_runtime": 183.7888,
+    "train_samples_per_second": 50.928,
+    "train_steps_per_second": 1.611
 }

trainer_state.json

@@ -1,112 +1,217 @@
 {
-  "best_metric": 0.38402849435806274,
-  "best_model_checkpoint": "./vit-base-uppercase-english-characters/checkpoint-100",
+  "best_metric": 0.3159617781639099,
+  "best_model_checkpoint": "./vit-base-uppercase-english-characters/checkpoint-200",
   "epoch": 4.0,
-  "global_step": 132,
+  "global_step": 296,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.3,
-      "learning_rate": 0.00018484848484848484,
-      "loss": 3.0027,
+      "epoch": 0.14,
+      "learning_rate": 0.00019324324324324326,
+      "loss": 3.1876,
       "step": 10
     },
     {
-      "epoch": 0.61,
-      "learning_rate": 0.00016969696969696972,
-      "loss": 2.2789,
+      "epoch": 0.27,
+      "learning_rate": 0.0001864864864864865,
+      "loss": 2.7336,
       "step": 20
     },
     {
-      "epoch": 0.91,
-      "learning_rate": 0.00015454545454545454,
-      "loss": 1.6762,
+      "epoch": 0.41,
+      "learning_rate": 0.00017972972972972974,
+      "loss": 2.2045,
       "step": 30
     },
     {
-      "epoch": 1.21,
-      "learning_rate": 0.0001393939393939394,
-      "loss": 1.2222,
+      "epoch": 0.54,
+      "learning_rate": 0.000172972972972973,
+      "loss": 1.7453,
       "step": 40
     },
     {
-      "epoch": 1.52,
-      "learning_rate": 0.00012424242424242425,
-      "loss": 0.9241,
+      "epoch": 0.68,
+      "learning_rate": 0.00016621621621621622,
+      "loss": 1.4371,
       "step": 50
     },
     {
-      "epoch": 1.82,
-      "learning_rate": 0.00010909090909090909,
-      "loss": 0.7407,
+      "epoch": 0.81,
+      "learning_rate": 0.00015945945945945947,
+      "loss": 1.1114,
       "step": 60
     },
     {
-      "epoch": 2.12,
-      "learning_rate": 9.393939393939395e-05,
-      "loss": 0.5808,
+      "epoch": 0.95,
+      "learning_rate": 0.0001527027027027027,
+      "loss": 0.9965,
       "step": 70
     },
     {
-      "epoch": 2.42,
-      "learning_rate": 7.878787878787879e-05,
-      "loss": 0.4945,
+      "epoch": 1.08,
+      "learning_rate": 0.00014594594594594595,
+      "loss": 0.7593,
       "step": 80
     },
     {
-      "epoch": 2.73,
-      "learning_rate": 6.363636363636364e-05,
-      "loss": 0.4364,
+      "epoch": 1.22,
+      "learning_rate": 0.0001391891891891892,
+      "loss": 0.6622,
       "step": 90
     },
     {
-      "epoch": 3.03,
-      "learning_rate": 4.848484848484849e-05,
-      "loss": 0.3936,
+      "epoch": 1.35,
+      "learning_rate": 0.00013243243243243243,
+      "loss": 0.5944,
       "step": 100
     },
     {
-      "epoch": 3.03,
-      "eval_accuracy": 1.0,
-      "eval_loss": 0.38402849435806274,
-      "eval_runtime": 1.6842,
-      "eval_samples_per_second": 123.504,
-      "eval_steps_per_second": 7.719,
+      "epoch": 1.35,
+      "eval_accuracy": 0.9487179487179487,
+      "eval_loss": 0.5538277626037598,
+      "eval_runtime": 1.9221,
+      "eval_samples_per_second": 121.742,
+      "eval_steps_per_second": 7.804,
       "step": 100
     },
     {
-      "epoch": 3.33,
-      "learning_rate": 3.3333333333333335e-05,
-      "loss": 0.3722,
+      "epoch": 1.49,
+      "learning_rate": 0.00012567567567567568,
+      "loss": 0.4882,
       "step": 110
     },
     {
-      "epoch": 3.64,
-      "learning_rate": 1.8181818181818182e-05,
-      "loss": 0.3534,
+      "epoch": 1.62,
+      "learning_rate": 0.00011891891891891893,
+      "loss": 0.46,
       "step": 120
     },
     {
-      "epoch": 3.94,
-      "learning_rate": 3.0303030303030305e-06,
-      "loss": 0.3421,
+      "epoch": 1.76,
+      "learning_rate": 0.00011216216216216217,
+      "loss": 0.4453,
       "step": 130
     },
+    {
+      "epoch": 1.89,
+      "learning_rate": 0.0001054054054054054,
+      "loss": 0.4454,
+      "step": 140
+    },
+    {
+      "epoch": 2.03,
+      "learning_rate": 9.864864864864865e-05,
+      "loss": 0.3618,
+      "step": 150
+    },
+    {
+      "epoch": 2.16,
+      "learning_rate": 9.18918918918919e-05,
+      "loss": 0.3138,
+      "step": 160
+    },
+    {
+      "epoch": 2.3,
+      "learning_rate": 8.513513513513514e-05,
+      "loss": 0.2605,
+      "step": 170
+    },
+    {
+      "epoch": 2.43,
+      "learning_rate": 7.837837837837838e-05,
+      "loss": 0.292,
+      "step": 180
+    },
+    {
+      "epoch": 2.57,
+      "learning_rate": 7.162162162162162e-05,
+      "loss": 0.2502,
+      "step": 190
+    },
+    {
+      "epoch": 2.7,
+      "learning_rate": 6.486486486486487e-05,
+      "loss": 0.2241,
+      "step": 200
+    },
+    {
+      "epoch": 2.7,
+      "eval_accuracy": 0.9572649572649573,
+      "eval_loss": 0.3159617781639099,
+      "eval_runtime": 1.8767,
+      "eval_samples_per_second": 124.688,
+      "eval_steps_per_second": 7.993,
+      "step": 200
+    },
+    {
+      "epoch": 2.84,
+      "learning_rate": 5.8108108108108105e-05,
+      "loss": 0.2273,
+      "step": 210
+    },
+    {
+      "epoch": 2.97,
+      "learning_rate": 5.135135135135135e-05,
+      "loss": 0.2002,
+      "step": 220
+    },
+    {
+      "epoch": 3.11,
+      "learning_rate": 4.4594594594594596e-05,
+      "loss": 0.163,
+      "step": 230
+    },
+    {
+      "epoch": 3.24,
+      "learning_rate": 3.783783783783784e-05,
+      "loss": 0.1654,
+      "step": 240
+    },
+    {
+      "epoch": 3.38,
+      "learning_rate": 3.108108108108108e-05,
+      "loss": 0.1569,
+      "step": 250
+    },
+    {
+      "epoch": 3.51,
+      "learning_rate": 2.4324324324324327e-05,
+      "loss": 0.1507,
+      "step": 260
+    },
+    {
+      "epoch": 3.65,
+      "learning_rate": 1.756756756756757e-05,
+      "loss": 0.1543,
+      "step": 270
+    },
+    {
+      "epoch": 3.78,
+      "learning_rate": 1.0810810810810812e-05,
+      "loss": 0.158,
+      "step": 280
+    },
+    {
+      "epoch": 3.92,
+      "learning_rate": 4.0540540540540545e-06,
+      "loss": 0.14,
+      "step": 290
+    },
     {
       "epoch": 4.0,
-      "step": 132,
-      "total_flos": 3.22436019290112e+17,
-      "train_loss": 0.9762612545129025,
-      "train_runtime": 74.7912,
-      "train_samples_per_second": 55.622,
-      "train_steps_per_second": 1.765
+      "step": 296,
+      "total_flos": 7.25481043402752e+17,
+      "train_loss": 0.6950366432602341,
+      "train_runtime": 183.7888,
+      "train_samples_per_second": 50.928,
+      "train_steps_per_second": 1.611
     }
   ],
-  "max_steps": 132,
+  "max_steps": 296,
   "num_train_epochs": 4,
-  "total_flos": 3.22436019290112e+17,
+  "total_flos": 7.25481043402752e+17,
   "trial_name": null,
   "trial_params": null
 }