🍻 cheers

all_results.json

@@ -1,12 +1,12 @@
 {
-    "epoch": 3.0,
-    "eval_accuracy": 0.7101168660535233,
-    "eval_loss": 0.9965542554855347,
-    "eval_runtime": 155.1678,
-    "eval_samples_per_second": 186.392,
-    "eval_steps_per_second": 23.304,
-    "train_loss": 0.9789225330756357,
-    "train_runtime": 10316.8833,
-    "train_samples_per_second": 67.281,
-    "train_steps_per_second": 2.103
+    "epoch": 4.0,
+    "eval_accuracy": 0.7118110780720559,
+    "eval_loss": 0.9865913391113281,
+    "eval_runtime": 155.9763,
+    "eval_samples_per_second": 185.426,
+    "eval_steps_per_second": 23.183,
+    "train_loss": 0.8423061020030403,
+    "train_runtime": 13978.3951,
+    "train_samples_per_second": 66.21,
+    "train_steps_per_second": 2.069
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 3.0,
-    "eval_accuracy": 0.7101168660535233,
-    "eval_loss": 0.9965542554855347,
-    "eval_runtime": 155.1678,
-    "eval_samples_per_second": 186.392,
-    "eval_steps_per_second": 23.304
+    "epoch": 4.0,
+    "eval_accuracy": 0.7118110780720559,
+    "eval_loss": 0.9865913391113281,
+    "eval_runtime": 155.9763,
+    "eval_samples_per_second": 185.426,
+    "eval_steps_per_second": 23.183
 }

runs/Jun05_18-03-44_adrian-development/events.out.tfevents.1685995184.adrian-development.8520.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1deafe540a75ee7e34f0e9c0e1c90089c13a43a21a37012718471e2c213f3c3c
+size 369

train_results.json

@@ -1,7 +1,7 @@
 {
-    "epoch": 3.0,
-    "train_loss": 0.9789225330756357,
-    "train_runtime": 10316.8833,
-    "train_samples_per_second": 67.281,
-    "train_steps_per_second": 2.103
+    "epoch": 4.0,
+    "train_loss": 0.8423061020030403,
+    "train_runtime": 13978.3951,
+    "train_samples_per_second": 66.21,
+    "train_steps_per_second": 2.069
 }

trainer_state.json

@@ -1,340 +1,445 @@
 {
-  "best_metric": 0.9965542554855347,
-  "best_model_checkpoint": "Transformers/vit-base-clothing-leafs-example-full-simple/checkpoint-21000",
-  "epoch": 3.0,
-  "global_step": 21693,
+  "best_metric": 0.9865913391113281,
+  "best_model_checkpoint": "Transformers/vit-base-clothing-leafs-example-full-simple/checkpoint-12000",
+  "epoch": 4.0,
+  "global_step": 28924,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.14,
-      "learning_rate": 0.00019078965564928778,
-      "loss": 1.5419,
+      "learning_rate": 4.8273060434241463e-05,
+      "loss": 1.7384,
       "step": 1000
     },
     {
       "epoch": 0.14,
-      "eval_accuracy": 0.6223981744001107,
-      "eval_loss": 1.3017297983169556,
-      "eval_runtime": 154.5699,
-      "eval_samples_per_second": 187.113,
-      "eval_steps_per_second": 23.394,
+      "eval_accuracy": 0.6473272941013761,
+      "eval_loss": 1.328142523765564,
+      "eval_runtime": 162.5228,
+      "eval_samples_per_second": 177.957,
+      "eval_steps_per_second": 22.249,
       "step": 1000
     },
     {
       "epoch": 0.28,
-      "learning_rate": 0.00018157931129857559,
-      "loss": 1.2619,
+      "learning_rate": 4.654439220024893e-05,
+      "loss": 1.2367,
       "step": 2000
     },
     {
       "epoch": 0.28,
-      "eval_accuracy": 0.6317336283797801,
-      "eval_loss": 1.2609293460845947,
-      "eval_runtime": 156.0227,
-      "eval_samples_per_second": 185.37,
-      "eval_steps_per_second": 23.176,
+      "eval_accuracy": 0.6703201714957472,
+      "eval_loss": 1.181534767150879,
+      "eval_runtime": 155.6299,
+      "eval_samples_per_second": 185.838,
+      "eval_steps_per_second": 23.235,
       "step": 2000
     },
     {
       "epoch": 0.41,
-      "learning_rate": 0.00017235974738394875,
-      "loss": 1.2244,
+      "learning_rate": 4.481745263449039e-05,
+      "loss": 1.1348,
       "step": 3000
     },
     {
       "epoch": 0.41,
-      "eval_accuracy": 0.6370928704792199,
-      "eval_loss": 1.2237269878387451,
-      "eval_runtime": 154.8118,
-      "eval_samples_per_second": 186.82,
-      "eval_steps_per_second": 23.357,
+      "eval_accuracy": 0.6793790194315745,
+      "eval_loss": 1.128984808921814,
+      "eval_runtime": 155.0985,
+      "eval_samples_per_second": 186.475,
+      "eval_steps_per_second": 23.314,
       "step": 3000
     },
     {
       "epoch": 0.55,
-      "learning_rate": 0.00016314018346932192,
-      "loss": 1.1924,
+      "learning_rate": 4.3088784400497855e-05,
+      "loss": 1.1003,
       "step": 4000
     },
     {
       "epoch": 0.55,
-      "eval_accuracy": 0.6339119009750364,
-      "eval_loss": 1.2324565649032593,
-      "eval_runtime": 151.2093,
-      "eval_samples_per_second": 191.271,
-      "eval_steps_per_second": 23.914,
+      "eval_accuracy": 0.6882995643454809,
+      "eval_loss": 1.0926539897918701,
+      "eval_runtime": 154.2349,
+      "eval_samples_per_second": 187.519,
+      "eval_steps_per_second": 23.445,
       "step": 4000
     },
     {
       "epoch": 0.69,
-      "learning_rate": 0.0001539298391186097,
-      "loss": 1.1572,
+      "learning_rate": 4.136357350297331e-05,
+      "loss": 1.0695,
       "step": 5000
     },
     {
       "epoch": 0.69,
-      "eval_accuracy": 0.6550031118179932,
-      "eval_loss": 1.1636905670166016,
-      "eval_runtime": 151.8183,
-      "eval_samples_per_second": 190.504,
-      "eval_steps_per_second": 23.818,
+      "eval_accuracy": 0.6911347762948621,
+      "eval_loss": 1.0641425848007202,
+      "eval_runtime": 156.9018,
+      "eval_samples_per_second": 184.332,
+      "eval_steps_per_second": 23.046,
       "step": 5000
     },
     {
       "epoch": 0.83,
-      "learning_rate": 0.00014471027520398285,
-      "loss": 1.1382,
+      "learning_rate": 3.963490526898078e-05,
+      "loss": 1.0426,
       "step": 6000
     },
     {
       "epoch": 0.83,
-      "eval_accuracy": 0.6618491114030841,
-      "eval_loss": 1.1443654298782349,
-      "eval_runtime": 150.6223,
-      "eval_samples_per_second": 192.017,
-      "eval_steps_per_second": 24.007,
+      "eval_accuracy": 0.6957679275292166,
+      "eval_loss": 1.0410244464874268,
+      "eval_runtime": 154.2668,
+      "eval_samples_per_second": 187.48,
+      "eval_steps_per_second": 23.44,
       "step": 6000
     },
     {
       "epoch": 0.97,
-      "learning_rate": 0.00013549993085327063,
-      "loss": 1.1403,
+      "learning_rate": 3.790623703498825e-05,
+      "loss": 1.0247,
       "step": 7000
     },
     {
       "epoch": 0.97,
-      "eval_accuracy": 0.6624023234907682,
-      "eval_loss": 1.1293965578079224,
-      "eval_runtime": 151.5259,
-      "eval_samples_per_second": 190.872,
-      "eval_steps_per_second": 23.864,
+      "eval_accuracy": 0.6936933822004011,
+      "eval_loss": 1.0401992797851562,
+      "eval_runtime": 155.38,
+      "eval_samples_per_second": 186.137,
+      "eval_steps_per_second": 23.272,
       "step": 7000
     },
     {
       "epoch": 1.11,
-      "learning_rate": 0.0001262803669386438,
-      "loss": 1.0204,
+      "learning_rate": 3.61792974692297e-05,
+      "loss": 0.9406,
       "step": 8000
     },
     {
       "epoch": 1.11,
-      "eval_accuracy": 0.6720835350252403,
-      "eval_loss": 1.1081748008728027,
-      "eval_runtime": 154.3075,
-      "eval_samples_per_second": 187.431,
-      "eval_steps_per_second": 23.434,
+      "eval_accuracy": 0.7003665030080908,
+      "eval_loss": 1.024382472038269,
+      "eval_runtime": 154.8686,
+      "eval_samples_per_second": 186.752,
+      "eval_steps_per_second": 23.349,
       "step": 8000
     },
     {
       "epoch": 1.24,
-      "learning_rate": 0.00011706080302401696,
-      "loss": 0.9853,
+      "learning_rate": 3.4450629235237175e-05,
+      "loss": 0.8824,
       "step": 9000
     },
     {
       "epoch": 1.24,
-      "eval_accuracy": 0.6693174745868197,
-      "eval_loss": 1.109432339668274,
-      "eval_runtime": 151.4973,
-      "eval_samples_per_second": 190.908,
-      "eval_steps_per_second": 23.868,
+      "eval_accuracy": 0.699329230343683,
+      "eval_loss": 1.0364962816238403,
+      "eval_runtime": 156.6606,
+      "eval_samples_per_second": 184.616,
+      "eval_steps_per_second": 23.082,
       "step": 9000
     },
     {
       "epoch": 1.38,
-      "learning_rate": 0.00010784123910939012,
-      "loss": 0.9767,
+      "learning_rate": 3.2721961001244647e-05,
+      "loss": 0.8979,
       "step": 10000
     },
     {
       "epoch": 1.38,
-      "eval_accuracy": 0.6742272318650162,
-      "eval_loss": 1.0898501873016357,
-      "eval_runtime": 152.2804,
-      "eval_samples_per_second": 189.926,
-      "eval_steps_per_second": 23.746,
+      "eval_accuracy": 0.7066938662609779,
+      "eval_loss": 1.0050914287567139,
+      "eval_runtime": 154.5086,
+      "eval_samples_per_second": 187.187,
+      "eval_steps_per_second": 23.403,
       "step": 10000
     },
     {
       "epoch": 1.52,
-      "learning_rate": 9.86216751947633e-05,
-      "loss": 0.9815,
+      "learning_rate": 3.099502143548611e-05,
+      "loss": 0.8947,
       "step": 11000
     },
     {
       "epoch": 1.52,
-      "eval_accuracy": 0.6871931401701127,
-      "eval_loss": 1.0531729459762573,
-      "eval_runtime": 152.4524,
-      "eval_samples_per_second": 189.712,
-      "eval_steps_per_second": 23.719,
+      "eval_accuracy": 0.7089067146117143,
+      "eval_loss": 0.9985986948013306,
+      "eval_runtime": 156.682,
+      "eval_samples_per_second": 184.59,
+      "eval_steps_per_second": 23.079,
       "step": 11000
     },
     {
       "epoch": 1.66,
-      "learning_rate": 8.941133084405108e-05,
-      "loss": 0.9672,
+      "learning_rate": 2.926635320149357e-05,
+      "loss": 0.8785,
       "step": 12000
     },
     {
       "epoch": 1.66,
-      "eval_accuracy": 0.6864324735495471,
-      "eval_loss": 1.056949496269226,
-      "eval_runtime": 153.3882,
-      "eval_samples_per_second": 188.554,
-      "eval_steps_per_second": 23.574,
+      "eval_accuracy": 0.7118110780720559,
+      "eval_loss": 0.9865913391113281,
+      "eval_runtime": 154.6399,
+      "eval_samples_per_second": 187.028,
+      "eval_steps_per_second": 23.383,
       "step": 12000
     },
     {
       "epoch": 1.8,
-      "learning_rate": 8.019176692942424e-05,
-      "loss": 0.9439,
+      "learning_rate": 2.7539413635735027e-05,
+      "loss": 0.8881,
       "step": 13000
     },
     {
       "epoch": 1.8,
-      "eval_accuracy": 0.6933822004010788,
-      "eval_loss": 1.0358134508132935,
-      "eval_runtime": 154.0974,
-      "eval_samples_per_second": 187.687,
-      "eval_steps_per_second": 23.466,
+      "eval_accuracy": 0.7112232902288915,
+      "eval_loss": 0.9892340898513794,
+      "eval_runtime": 155.1615,
+      "eval_samples_per_second": 186.399,
+      "eval_steps_per_second": 23.305,
       "step": 13000
     },
     {
       "epoch": 1.94,
-      "learning_rate": 7.09722030147974e-05,
-      "loss": 0.9228,
+      "learning_rate": 2.58107454017425e-05,
+      "loss": 0.8652,
       "step": 14000
     },
     {
       "epoch": 1.94,
-      "eval_accuracy": 0.6884032916119217,
-      "eval_loss": 1.0430152416229248,
-      "eval_runtime": 152.7147,
-      "eval_samples_per_second": 189.386,
-      "eval_steps_per_second": 23.678,
+      "eval_accuracy": 0.7111887144734113,
+      "eval_loss": 0.987538754940033,
+      "eval_runtime": 154.34,
+      "eval_samples_per_second": 187.392,
+      "eval_steps_per_second": 23.429,
       "step": 14000
     },
     {
       "epoch": 2.07,
-      "learning_rate": 6.17618586640852e-05,
-      "loss": 0.8511,
+      "learning_rate": 2.4082077167749967e-05,
+      "loss": 0.7969,
       "step": 15000
     },
     {
       "epoch": 2.07,
-      "eval_accuracy": 0.695837079040177,
-      "eval_loss": 1.0438477993011475,
-      "eval_runtime": 155.2102,
-      "eval_samples_per_second": 186.341,
-      "eval_steps_per_second": 23.297,
+      "eval_accuracy": 0.7082843510130696,
+      "eval_loss": 1.003035068511963,
+      "eval_runtime": 154.584,
+      "eval_samples_per_second": 187.096,
+      "eval_steps_per_second": 23.392,
       "step": 15000
     },
     {
       "epoch": 2.21,
-      "learning_rate": 5.2542294749458354e-05,
-      "loss": 0.7619,
+      "learning_rate": 2.2353408933757432e-05,
+      "loss": 0.7153,
       "step": 16000
     },
     {
       "epoch": 2.21,
-      "eval_accuracy": 0.697980775879953,
-      "eval_loss": 1.0432237386703491,
-      "eval_runtime": 155.9321,
-      "eval_samples_per_second": 185.478,
-      "eval_steps_per_second": 23.19,
+      "eval_accuracy": 0.7085263813014314,
+      "eval_loss": 1.0069255828857422,
+      "eval_runtime": 154.1761,
+      "eval_samples_per_second": 187.591,
+      "eval_steps_per_second": 23.454,
       "step": 16000
     },
     {
       "epoch": 2.35,
-      "learning_rate": 4.333195039874614e-05,
-      "loss": 0.7672,
+      "learning_rate": 2.0626469367998893e-05,
+      "loss": 0.7158,
       "step": 17000
     },
     {
       "epoch": 2.35,
-      "eval_accuracy": 0.7022681695595049,
-      "eval_loss": 1.0282564163208008,
-      "eval_runtime": 156.0109,
-      "eval_samples_per_second": 185.385,
-      "eval_steps_per_second": 23.178,
+      "eval_accuracy": 0.7080077449692276,
+      "eval_loss": 1.0076011419296265,
+      "eval_runtime": 156.3358,
+      "eval_samples_per_second": 184.999,
+      "eval_steps_per_second": 23.13,
       "step": 17000
     },
     {
       "epoch": 2.49,
-      "learning_rate": 3.41123864841193e-05,
-      "loss": 0.7378,
+      "learning_rate": 1.8897801134006362e-05,
+      "loss": 0.7248,
       "step": 18000
     },
     {
       "epoch": 2.49,
-      "eval_accuracy": 0.7029942604245902,
-      "eval_loss": 1.017476201057434,
-      "eval_runtime": 153.8464,
-      "eval_samples_per_second": 187.993,
-      "eval_steps_per_second": 23.504,
+      "eval_accuracy": 0.7108083811631284,
+      "eval_loss": 1.0020238161087036,
+      "eval_runtime": 154.0953,
+      "eval_samples_per_second": 187.689,
+      "eval_steps_per_second": 23.466,
       "step": 18000
     },
     {
       "epoch": 2.63,
-      "learning_rate": 2.4892822569492465e-05,
-      "loss": 0.7217,
+      "learning_rate": 1.716913290001383e-05,
+      "loss": 0.7204,
       "step": 19000
     },
     {
       "epoch": 2.63,
-      "eval_accuracy": 0.7042044118663993,
-      "eval_loss": 1.018778681755066,
-      "eval_runtime": 156.8745,
-      "eval_samples_per_second": 184.364,
-      "eval_steps_per_second": 23.05,
+      "eval_accuracy": 0.7130903810248254,
+      "eval_loss": 0.992910623550415,
+      "eval_runtime": 154.166,
+      "eval_samples_per_second": 187.603,
+      "eval_steps_per_second": 23.455,
       "step": 19000
     },
     {
       "epoch": 2.77,
-      "learning_rate": 1.5682478218780253e-05,
-      "loss": 0.7285,
+      "learning_rate": 1.5442193334255288e-05,
+      "loss": 0.7127,
       "step": 20000
     },
     {
       "epoch": 2.77,
-      "eval_accuracy": 0.7103588963418851,
-      "eval_loss": 0.9978100061416626,
-      "eval_runtime": 153.4823,
-      "eval_samples_per_second": 188.439,
-      "eval_steps_per_second": 23.56,
+      "eval_accuracy": 0.7138510476453911,
+      "eval_loss": 0.992859423160553,
+      "eval_runtime": 155.0252,
+      "eval_samples_per_second": 186.563,
+      "eval_steps_per_second": 23.325,
       "step": 20000
     },
     {
       "epoch": 2.9,
-      "learning_rate": 6.462914304153414e-06,
-      "loss": 0.7206,
+      "learning_rate": 1.3713525100262758e-05,
+      "loss": 0.7274,
       "step": 21000
     },
     {
       "epoch": 2.9,
-      "eval_accuracy": 0.7101168660535233,
-      "eval_loss": 0.9965542554855347,
-      "eval_runtime": 156.6977,
-      "eval_samples_per_second": 184.572,
-      "eval_steps_per_second": 23.076,
+      "eval_accuracy": 0.7104626236083258,
+      "eval_loss": 0.9929330945014954,
+      "eval_runtime": 155.2366,
+      "eval_samples_per_second": 186.309,
+      "eval_steps_per_second": 23.293,
       "step": 21000
     },
     {
-      "epoch": 3.0,
-      "step": 21693,
-      "total_flos": 5.381080460363188e+19,
-      "train_loss": 0.9789225330756357,
-      "train_runtime": 10316.8833,
-      "train_samples_per_second": 67.281,
-      "train_steps_per_second": 2.103
+      "epoch": 3.04,
+      "learning_rate": 1.198658553450422e-05,
+      "loss": 0.6769,
+      "step": 22000
+    },
+    {
+      "epoch": 3.04,
+      "eval_accuracy": 0.7118110780720559,
+      "eval_loss": 1.0151628255844116,
+      "eval_runtime": 154.6606,
+      "eval_samples_per_second": 187.003,
+      "eval_steps_per_second": 23.38,
+      "step": 22000
+    },
+    {
+      "epoch": 3.18,
+      "learning_rate": 1.0259645968745679e-05,
+      "loss": 0.5859,
+      "step": 23000
+    },
+    {
+      "epoch": 3.18,
+      "eval_accuracy": 0.708872138856234,
+      "eval_loss": 1.0313763618469238,
+      "eval_runtime": 154.3702,
+      "eval_samples_per_second": 187.355,
+      "eval_steps_per_second": 23.424,
+      "step": 23000
+    },
+    {
+      "epoch": 3.32,
+      "learning_rate": 8.530977734753147e-06,
+      "loss": 0.5811,
+      "step": 24000
+    },
+    {
+      "epoch": 3.32,
+      "eval_accuracy": 0.7105663508747666,
+      "eval_loss": 1.0339767932891846,
+      "eval_runtime": 154.2207,
+      "eval_samples_per_second": 187.536,
+      "eval_steps_per_second": 23.447,
+      "step": 24000
+    },
+    {
+      "epoch": 3.46,
+      "learning_rate": 6.802309500760615e-06,
+      "loss": 0.5863,
+      "step": 25000
+    },
+    {
+      "epoch": 3.46,
+      "eval_accuracy": 0.7104971993638061,
+      "eval_loss": 1.0252958536148071,
+      "eval_runtime": 154.7743,
+      "eval_samples_per_second": 186.866,
+      "eval_steps_per_second": 23.363,
+      "step": 25000
+    },
+    {
+      "epoch": 3.6,
+      "learning_rate": 5.073641266768082e-06,
+      "loss": 0.5656,
+      "step": 26000
+    },
+    {
+      "epoch": 3.6,
+      "eval_accuracy": 0.7103934720973654,
+      "eval_loss": 1.027882695198059,
+      "eval_runtime": 154.9221,
+      "eval_samples_per_second": 186.687,
+      "eval_steps_per_second": 23.341,
+      "step": 26000
+    },
+    {
+      "epoch": 3.73,
+      "learning_rate": 3.346701701009542e-06,
+      "loss": 0.5753,
+      "step": 27000
+    },
+    {
+      "epoch": 3.73,
+      "eval_accuracy": 0.7107738054076481,
+      "eval_loss": 1.0284228324890137,
+      "eval_runtime": 155.4908,
+      "eval_samples_per_second": 186.005,
+      "eval_steps_per_second": 23.255,
+      "step": 27000
+    },
+    {
+      "epoch": 3.87,
+      "learning_rate": 1.61803346701701e-06,
+      "loss": 0.5681,
+      "step": 28000
+    },
+    {
+      "epoch": 3.87,
+      "eval_accuracy": 0.7111887144734113,
+      "eval_loss": 1.0259647369384766,
+      "eval_runtime": 155.1653,
+      "eval_samples_per_second": 186.395,
+      "eval_steps_per_second": 23.304,
+      "step": 28000
+    },
+    {
+      "epoch": 4.0,
+      "step": 28924,
+      "total_flos": 7.174773947150918e+19,
+      "train_loss": 0.8423061020030403,
+      "train_runtime": 13978.3951,
+      "train_samples_per_second": 66.21,
+      "train_steps_per_second": 2.069
     }
   ],
-  "max_steps": 21693,
-  "num_train_epochs": 3,
-  "total_flos": 5.381080460363188e+19,
+  "max_steps": 28924,
+  "num_train_epochs": 4,
+  "total_flos": 7.174773947150918e+19,
   "trial_name": null,
   "trial_params": null
 }