🍻 cheers

all_results.json

@@ -1,12 +1,12 @@
 {
-    "epoch": 0.97,
-    "eval_accuracy": 0.24808104557084573,
-    "eval_loss": 2.6692957878112793,
-    "eval_runtime": 153.5066,
-    "eval_samples_per_second": 188.409,
-    "eval_steps_per_second": 23.556,
-    "train_loss": 1.2184505571637836,
-    "train_runtime": 3232.5785,
-    "train_samples_per_second": 357.883,
-    "train_steps_per_second": 11.185
+    "epoch": 3.6,
+    "eval_accuracy": 0.7120531083604177,
+    "eval_loss": 1.0266574621200562,
+    "eval_runtime": 155.4593,
+    "eval_samples_per_second": 186.042,
+    "eval_steps_per_second": 23.26,
+    "train_loss": 1.0817440608097957,
+    "train_runtime": 12426.2916,
+    "train_samples_per_second": 93.1,
+    "train_steps_per_second": 2.91
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 0.97,
-    "eval_accuracy": 0.24808104557084573,
-    "eval_loss": 2.6692957878112793,
-    "eval_runtime": 153.5066,
-    "eval_samples_per_second": 188.409,
-    "eval_steps_per_second": 23.556
+    "epoch": 3.6,
+    "eval_accuracy": 0.7120531083604177,
+    "eval_loss": 1.0266574621200562,
+    "eval_runtime": 155.4593,
+    "eval_samples_per_second": 186.042,
+    "eval_steps_per_second": 23.26
 }

runs/May30_14-50-58_adrian-development/events.out.tfevents.1685463710.adrian-development.97399.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4fd99b8df63d4946d5622301444a8e7c90db53de0ef0b3d1586916ef048e3d08
+size 369

train_results.json

@@ -1,7 +1,7 @@
 {
-    "epoch": 0.97,
-    "train_loss": 1.2184505571637836,
-    "train_runtime": 3232.5785,
-    "train_samples_per_second": 357.883,
-    "train_steps_per_second": 11.185
+    "epoch": 3.6,
+    "train_loss": 1.0817440608097957,
+    "train_runtime": 12426.2916,
+    "train_samples_per_second": 93.1,
+    "train_steps_per_second": 2.91
 }

trainer_state.json

@@ -1,130 +1,415 @@
 {
-  "best_metric": 2.6692957878112793,
-  "best_model_checkpoint": "Transformers/vit-base-clothing-leafs-example-full-simple/checkpoint-2000",
-  "epoch": 0.968054211035818,
-  "global_step": 7000,
+  "best_metric": 1.0266574621200562,
+  "best_model_checkpoint": "Transformers/vit-base-clothing-leafs-example-full-simple/checkpoint-21000",
+  "epoch": 3.5956299267044667,
+  "global_step": 26000,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.14,
-      "learning_rate": 0.0019446826165122391,
-      "loss": 2.7222,
+      "learning_rate": 9.723413082561196e-06,
+      "loss": 2.4732,
       "step": 1000
     },
     {
       "epoch": 0.14,
-      "eval_accuracy": 0.25281792407164094,
-      "eval_loss": 2.676572322845459,
-      "eval_runtime": 156.0521,
-      "eval_samples_per_second": 185.336,
-      "eval_steps_per_second": 23.172,
+      "eval_accuracy": 0.5465735426319065,
+      "eval_loss": 1.9312118291854858,
+      "eval_runtime": 155.4998,
+      "eval_samples_per_second": 185.994,
+      "eval_steps_per_second": 23.254,
       "step": 1000
     },
     {
       "epoch": 0.28,
-      "learning_rate": 0.001889586502558429,
-      "loss": 2.6768,
+      "learning_rate": 9.44682616512239e-06,
+      "loss": 1.7297,
       "step": 2000
     },
     {
       "epoch": 0.28,
-      "eval_accuracy": 0.24808104557084573,
-      "eval_loss": 2.6692957878112793,
-      "eval_runtime": 154.1979,
-      "eval_samples_per_second": 187.564,
-      "eval_steps_per_second": 23.45,
+      "eval_accuracy": 0.6410345066039693,
+      "eval_loss": 1.5575428009033203,
+      "eval_runtime": 154.3161,
+      "eval_samples_per_second": 187.42,
+      "eval_steps_per_second": 23.432,
       "step": 2000
     },
     {
       "epoch": 0.41,
-      "learning_rate": 0.0018360392753422764,
-      "loss": 2.6689,
+      "learning_rate": 9.170515834601023e-06,
+      "loss": 1.4684,
       "step": 3000
     },
     {
       "epoch": 0.41,
-      "eval_accuracy": 0.22304819860313949,
-      "eval_loss": NaN,
-      "eval_runtime": 151.9095,
-      "eval_samples_per_second": 190.39,
-      "eval_steps_per_second": 23.804,
+      "eval_accuracy": 0.6578383237673743,
+      "eval_loss": 1.3884613513946533,
+      "eval_runtime": 156.4025,
+      "eval_samples_per_second": 184.92,
+      "eval_steps_per_second": 23.12,
       "step": 3000
     },
     {
       "epoch": 0.55,
-      "learning_rate": 0.0017879131517079244,
-      "loss": 0.4613,
+      "learning_rate": 8.89392891716222e-06,
+      "loss": 1.3161,
       "step": 4000
     },
     {
       "epoch": 0.55,
-      "eval_accuracy": 0.051241269621741234,
-      "eval_loss": NaN,
-      "eval_runtime": 152.778,
-      "eval_samples_per_second": 189.307,
-      "eval_steps_per_second": 23.668,
+      "eval_accuracy": 0.6701472927183459,
+      "eval_loss": 1.2846978902816772,
+      "eval_runtime": 154.8377,
+      "eval_samples_per_second": 186.789,
+      "eval_steps_per_second": 23.353,
       "step": 4000
     },
     {
       "epoch": 0.69,
-      "learning_rate": 0.0017325957682201632,
-      "loss": 0.0,
+      "learning_rate": 8.617618586640853e-06,
+      "loss": 1.2652,
       "step": 5000
     },
     {
       "epoch": 0.69,
-      "eval_accuracy": 0.051241269621741234,
-      "eval_loss": NaN,
-      "eval_runtime": 152.584,
-      "eval_samples_per_second": 189.548,
-      "eval_steps_per_second": 23.698,
+      "eval_accuracy": 0.680381716340502,
+      "eval_loss": 1.2144489288330078,
+      "eval_runtime": 153.0458,
+      "eval_samples_per_second": 188.976,
+      "eval_steps_per_second": 23.627,
       "step": 5000
     },
     {
       "epoch": 0.83,
-      "learning_rate": 0.0016772783847324023,
-      "loss": 0.0,
+      "learning_rate": 8.341031669202048e-06,
+      "loss": 1.1962,
       "step": 6000
     },
     {
       "epoch": 0.83,
-      "eval_accuracy": 0.051241269621741234,
-      "eval_loss": NaN,
-      "eval_runtime": 152.8491,
-      "eval_samples_per_second": 189.219,
-      "eval_steps_per_second": 23.657,
+      "eval_accuracy": 0.6856718069289814,
+      "eval_loss": 1.1739206314086914,
+      "eval_runtime": 153.7562,
+      "eval_samples_per_second": 188.103,
+      "eval_steps_per_second": 23.518,
       "step": 6000
     },
     {
       "epoch": 0.97,
-      "learning_rate": 0.001621961001244641,
-      "loss": 0.0,
+      "learning_rate": 8.064444751763242e-06,
+      "loss": 1.1536,
       "step": 7000
     },
     {
       "epoch": 0.97,
-      "eval_accuracy": 0.051241269621741234,
-      "eval_loss": NaN,
-      "eval_runtime": 151.1266,
-      "eval_samples_per_second": 191.376,
-      "eval_steps_per_second": 23.927,
+      "eval_accuracy": 0.6906852914736187,
+      "eval_loss": 1.1354219913482666,
+      "eval_runtime": 154.4082,
+      "eval_samples_per_second": 187.309,
+      "eval_steps_per_second": 23.418,
       "step": 7000
     },
     {
-      "epoch": 0.97,
-      "step": 7000,
-      "total_flos": 1.736505102237696e+19,
-      "train_loss": 1.2184505571637836,
-      "train_runtime": 3232.5785,
-      "train_samples_per_second": 357.883,
-      "train_steps_per_second": 11.185
+      "epoch": 1.11,
+      "learning_rate": 7.788134421241876e-06,
+      "loss": 1.0931,
+      "step": 8000
+    },
+    {
+      "epoch": 1.11,
+      "eval_accuracy": 0.6934859276675195,
+      "eval_loss": 1.1159623861312866,
+      "eval_runtime": 153.8754,
+      "eval_samples_per_second": 187.957,
+      "eval_steps_per_second": 23.5,
+      "step": 8000
+    },
+    {
+      "epoch": 1.24,
+      "learning_rate": 7.51182409072051e-06,
+      "loss": 1.0486,
+      "step": 9000
+    },
+    {
+      "epoch": 1.24,
+      "eval_accuracy": 0.6964940183943019,
+      "eval_loss": 1.1065187454223633,
+      "eval_runtime": 153.5864,
+      "eval_samples_per_second": 188.311,
+      "eval_steps_per_second": 23.544,
+      "step": 9000
+    },
+    {
+      "epoch": 1.38,
+      "learning_rate": 7.235513760199143e-06,
+      "loss": 1.0242,
+      "step": 10000
+    },
+    {
+      "epoch": 1.38,
+      "eval_accuracy": 0.6990180485443607,
+      "eval_loss": 1.080113410949707,
+      "eval_runtime": 154.1408,
+      "eval_samples_per_second": 187.634,
+      "eval_steps_per_second": 23.459,
+      "step": 10000
+    },
+    {
+      "epoch": 1.52,
+      "learning_rate": 6.9589268427603384e-06,
+      "loss": 1.0203,
+      "step": 11000
+    },
+    {
+      "epoch": 1.52,
+      "eval_accuracy": 0.7030634119355508,
+      "eval_loss": 1.0677887201309204,
+      "eval_runtime": 153.95,
+      "eval_samples_per_second": 187.866,
+      "eval_steps_per_second": 23.488,
+      "step": 11000
+    },
+    {
+      "epoch": 1.66,
+      "learning_rate": 6.682339925321532e-06,
+      "loss": 1.0079,
+      "step": 12000
+    },
+    {
+      "epoch": 1.66,
+      "eval_accuracy": 0.703754927045156,
+      "eval_loss": 1.0623527765274048,
+      "eval_runtime": 155.0978,
+      "eval_samples_per_second": 186.476,
+      "eval_steps_per_second": 23.314,
+      "step": 12000
+    },
+    {
+      "epoch": 1.8,
+      "learning_rate": 6.405753007882728e-06,
+      "loss": 0.9962,
+      "step": 13000
+    },
+    {
+      "epoch": 1.8,
+      "eval_accuracy": 0.7038932300670769,
+      "eval_loss": 1.0550236701965332,
+      "eval_runtime": 154.4093,
+      "eval_samples_per_second": 187.307,
+      "eval_steps_per_second": 23.418,
+      "step": 13000
+    },
+    {
+      "epoch": 1.94,
+      "learning_rate": 6.1291660904439225e-06,
+      "loss": 0.9975,
+      "step": 14000
+    },
+    {
+      "epoch": 1.94,
+      "eval_accuracy": 0.7048613512205242,
+      "eval_loss": 1.0496357679367065,
+      "eval_runtime": 153.3268,
+      "eval_samples_per_second": 188.63,
+      "eval_steps_per_second": 23.584,
+      "step": 14000
+    },
+    {
+      "epoch": 2.07,
+      "learning_rate": 5.852579173005117e-06,
+      "loss": 0.9572,
+      "step": 15000
+    },
+    {
+      "epoch": 2.07,
+      "eval_accuracy": 0.7075928359034644,
+      "eval_loss": 1.0450878143310547,
+      "eval_runtime": 153.3954,
+      "eval_samples_per_second": 188.545,
+      "eval_steps_per_second": 23.573,
+      "step": 15000
+    },
+    {
+      "epoch": 2.21,
+      "learning_rate": 5.575992255566313e-06,
+      "loss": 0.8944,
+      "step": 16000
+    },
+    {
+      "epoch": 2.21,
+      "eval_accuracy": 0.7076274116589447,
+      "eval_loss": 1.0498133897781372,
+      "eval_runtime": 155.6828,
+      "eval_samples_per_second": 185.775,
+      "eval_steps_per_second": 23.227,
+      "step": 16000
+    },
+    {
+      "epoch": 2.35,
+      "learning_rate": 5.2999585119623845e-06,
+      "loss": 0.9027,
+      "step": 17000
+    },
+    {
+      "epoch": 2.35,
+      "eval_accuracy": 0.707938593458267,
+      "eval_loss": 1.0396690368652344,
+      "eval_runtime": 162.118,
+      "eval_samples_per_second": 178.401,
+      "eval_steps_per_second": 22.305,
+      "step": 17000
+    },
+    {
+      "epoch": 2.49,
+      "learning_rate": 5.023371594523579e-06,
+      "loss": 0.8806,
+      "step": 18000
+    },
+    {
+      "epoch": 2.49,
+      "eval_accuracy": 0.7092870479219971,
+      "eval_loss": 1.0353764295578003,
+      "eval_runtime": 157.661,
+      "eval_samples_per_second": 183.444,
+      "eval_steps_per_second": 22.935,
+      "step": 18000
+    },
+    {
+      "epoch": 2.63,
+      "learning_rate": 4.747061264002213e-06,
+      "loss": 0.8968,
+      "step": 19000
+    },
+    {
+      "epoch": 2.63,
+      "eval_accuracy": 0.709010441878155,
+      "eval_loss": 1.0362025499343872,
+      "eval_runtime": 159.5334,
+      "eval_samples_per_second": 181.291,
+      "eval_steps_per_second": 22.666,
+      "step": 19000
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 4.470474346563408e-06,
+      "loss": 0.8895,
+      "step": 20000
+    },
+    {
+      "epoch": 2.77,
+      "eval_accuracy": 0.7106009266302469,
+      "eval_loss": 1.0285927057266235,
+      "eval_runtime": 158.0807,
+      "eval_samples_per_second": 182.957,
+      "eval_steps_per_second": 22.874,
+      "step": 20000
+    },
+    {
+      "epoch": 2.9,
+      "learning_rate": 4.193887429124603e-06,
+      "loss": 0.8764,
+      "step": 21000
+    },
+    {
+      "epoch": 2.9,
+      "eval_accuracy": 0.7120531083604177,
+      "eval_loss": 1.0266574621200562,
+      "eval_runtime": 157.7263,
+      "eval_samples_per_second": 183.368,
+      "eval_steps_per_second": 22.926,
+      "step": 21000
+    },
+    {
+      "epoch": 3.04,
+      "learning_rate": 3.917300511685797e-06,
+      "loss": 0.8495,
+      "step": 22000
+    },
+    {
+      "epoch": 3.04,
+      "eval_accuracy": 0.7089412903671946,
+      "eval_loss": 1.0354034900665283,
+      "eval_runtime": 157.704,
+      "eval_samples_per_second": 183.394,
+      "eval_steps_per_second": 22.929,
+      "step": 22000
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 3.6412667680818703e-06,
+      "loss": 0.7935,
+      "step": 23000
+    },
+    {
+      "epoch": 3.18,
+      "eval_accuracy": 0.7067630177719383,
+      "eval_loss": 1.0451409816741943,
+      "eval_runtime": 158.4061,
+      "eval_samples_per_second": 182.581,
+      "eval_steps_per_second": 22.827,
+      "step": 23000
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 3.364679850643065e-06,
+      "loss": 0.7876,
+      "step": 24000
+    },
+    {
+      "epoch": 3.32,
+      "eval_accuracy": 0.7097019569877602,
+      "eval_loss": 1.0455540418624878,
+      "eval_runtime": 158.0269,
+      "eval_samples_per_second": 183.02,
+      "eval_steps_per_second": 22.882,
+      "step": 24000
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 3.0880929332042597e-06,
+      "loss": 0.8093,
+      "step": 25000
+    },
+    {
+      "epoch": 3.46,
+      "eval_accuracy": 0.7088029873452735,
+      "eval_loss": 1.04353928565979,
+      "eval_runtime": 159.3368,
+      "eval_samples_per_second": 181.515,
+      "eval_steps_per_second": 22.694,
+      "step": 25000
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 2.8117826026828933e-06,
+      "loss": 0.7977,
+      "step": 26000
+    },
+    {
+      "epoch": 3.6,
+      "eval_accuracy": 0.7094599266993984,
+      "eval_loss": 1.0454202890396118,
+      "eval_runtime": 156.4417,
+      "eval_samples_per_second": 184.874,
+      "eval_steps_per_second": 23.114,
+      "step": 26000
+    },
+    {
+      "epoch": 3.6,
+      "step": 26000,
+      "total_flos": 6.449527242554296e+19,
+      "train_loss": 1.0817440608097957,
+      "train_runtime": 12426.2916,
+      "train_samples_per_second": 93.1,
+      "train_steps_per_second": 2.91
     }
   ],
   "max_steps": 36155,
   "num_train_epochs": 5,
-  "total_flos": 1.736505102237696e+19,
+  "total_flos": 6.449527242554296e+19,
   "trial_name": null,
   "trial_params": null
 }