End of training

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 3.0,
-    "eval_f1": 0.6808059384941676,
-    "eval_loss": 0.6882463097572327,
-    "eval_runtime": 15.1393,
-    "eval_samples_per_second": 42.274,
-    "eval_steps_per_second": 1.321,
-    "total_flos": 1.3494104908807864e+18,
-    "train_loss": 0.7046969166508428,
-    "train_runtime": 848.1758,
-    "train_samples_per_second": 20.345,
-    "train_steps_per_second": 0.159
+    "epoch": 2.9881422924901186,
+    "eval_accuracy": 0.7536231884057971,
+    "eval_loss": 0.5222525000572205,
+    "eval_runtime": 21.9256,
+    "eval_samples_per_second": 40.911,
+    "eval_steps_per_second": 1.323,
+    "total_flos": 1.8869538215666074e+18,
+    "train_loss": 0.48228894844257014,
+    "train_runtime": 1217.9669,
+    "train_samples_per_second": 19.865,
+    "train_steps_per_second": 0.155
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 3.0,
-    "eval_f1": 0.6808059384941676,
-    "eval_loss": 0.6882463097572327,
-    "eval_runtime": 15.1393,
-    "eval_samples_per_second": 42.274,
-    "eval_steps_per_second": 1.321
+    "epoch": 2.9881422924901186,
+    "eval_accuracy": 0.7536231884057971,
+    "eval_loss": 0.5222525000572205,
+    "eval_runtime": 21.9256,
+    "eval_samples_per_second": 40.911,
+    "eval_steps_per_second": 1.323
 }

runs/Apr29_09-49-38_ecfd88f9ba4a/events.out.tfevents.1714386232.ecfd88f9ba4a.1021.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e30c23c7adb0dff8f698e718c2405866b075d2cf78698734314c74cfdf1fc22
+size 411

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 3.0,
-    "total_flos": 1.3494104908807864e+18,
-    "train_loss": 0.7046969166508428,
-    "train_runtime": 848.1758,
-    "train_samples_per_second": 20.345,
-    "train_steps_per_second": 0.159
+    "epoch": 2.9881422924901186,
+    "total_flos": 1.8869538215666074e+18,
+    "train_loss": 0.48228894844257014,
+    "train_runtime": 1217.9669,
+    "train_samples_per_second": 19.865,
+    "train_steps_per_second": 0.155
 }

trainer_state.json

@@ -1,147 +1,182 @@
 {
-  "best_metric": 0.6808059384941676,
-  "best_model_checkpoint": "dinov2-base-finetuned-eurosat/checkpoint-90",
-  "epoch": 3.0,
+  "best_metric": 0.7536231884057971,
+  "best_model_checkpoint": "dinov2-base-finetuned-eurosat/checkpoint-189",
+  "epoch": 2.9881422924901186,
   "eval_steps": 500,
-  "global_step": 135,
+  "global_step": 189,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.2222222222222222,
-      "grad_norm": 8.703512191772461,
-      "learning_rate": 3.571428571428572e-05,
-      "loss": 0.7234,
+      "epoch": 0.15810276679841898,
+      "grad_norm": 17.875991821289062,
+      "learning_rate": 2.6315789473684212e-05,
+      "loss": 0.5618,
       "step": 10
     },
     {
-      "epoch": 0.4444444444444444,
-      "grad_norm": 5.468322277069092,
-      "learning_rate": 4.75206611570248e-05,
-      "loss": 0.7517,
+      "epoch": 0.31620553359683795,
+      "grad_norm": 8.501226425170898,
+      "learning_rate": 4.970588235294118e-05,
+      "loss": 0.5333,
       "step": 20
     },
     {
-      "epoch": 0.6666666666666666,
-      "grad_norm": 6.569424629211426,
-      "learning_rate": 4.338842975206612e-05,
-      "loss": 0.7212,
+      "epoch": 0.4743083003952569,
+      "grad_norm": 22.501171112060547,
+      "learning_rate": 4.6764705882352944e-05,
+      "loss": 0.6441,
       "step": 30
     },
     {
-      "epoch": 0.8888888888888888,
-      "grad_norm": 3.0393030643463135,
-      "learning_rate": 3.925619834710744e-05,
-      "loss": 0.6972,
+      "epoch": 0.6324110671936759,
+      "grad_norm": 11.3024320602417,
+      "learning_rate": 4.382352941176471e-05,
+      "loss": 0.5519,
       "step": 40
     },
     {
-      "epoch": 1.0,
-      "eval_f1": 0.5401929260450161,
-      "eval_loss": 0.6850488781929016,
-      "eval_runtime": 14.9955,
-      "eval_samples_per_second": 42.679,
-      "eval_steps_per_second": 1.334,
-      "step": 45
-    },
-    {
-      "epoch": 1.1111111111111112,
-      "grad_norm": 19.1541748046875,
-      "learning_rate": 3.512396694214876e-05,
-      "loss": 0.7277,
+      "epoch": 0.7905138339920948,
+      "grad_norm": 7.681583404541016,
+      "learning_rate": 4.0882352941176474e-05,
+      "loss": 0.5391,
       "step": 50
     },
     {
-      "epoch": 1.3333333333333333,
-      "grad_norm": 16.108083724975586,
-      "learning_rate": 3.099173553719008e-05,
-      "loss": 0.7382,
+      "epoch": 0.9486166007905138,
+      "grad_norm": 6.365460395812988,
+      "learning_rate": 3.794117647058824e-05,
+      "loss": 0.5271,
       "step": 60
     },
     {
-      "epoch": 1.5555555555555556,
-      "grad_norm": 1.285812258720398,
-      "learning_rate": 2.6859504132231405e-05,
-      "loss": 0.7041,
-      "step": 70
+      "epoch": 0.9960474308300395,
+      "eval_accuracy": 0.7134894091415831,
+      "eval_loss": 0.5548548698425293,
+      "eval_runtime": 21.298,
+      "eval_samples_per_second": 42.117,
+      "eval_steps_per_second": 1.362,
+      "step": 63
     },
     {
-      "epoch": 1.7777777777777777,
-      "grad_norm": 3.1954760551452637,
-      "learning_rate": 2.272727272727273e-05,
-      "loss": 0.6853,
-      "step": 80
+      "epoch": 1.1067193675889329,
+      "grad_norm": 9.361493110656738,
+      "learning_rate": 3.5e-05,
+      "loss": 0.5572,
+      "step": 70
     },
     {
-      "epoch": 2.0,
-      "grad_norm": 1.845994472503662,
-      "learning_rate": 1.859504132231405e-05,
-      "loss": 0.6839,
-      "step": 90
+      "epoch": 1.2648221343873518,
+      "grad_norm": 7.021822452545166,
+      "learning_rate": 3.205882352941177e-05,
+      "loss": 0.5003,
+      "step": 80
     },
     {
-      "epoch": 2.0,
-      "eval_f1": 0.6808059384941676,
-      "eval_loss": 0.6882463097572327,
-      "eval_runtime": 15.0867,
-      "eval_samples_per_second": 42.421,
-      "eval_steps_per_second": 1.326,
+      "epoch": 1.4229249011857708,
+      "grad_norm": 8.887531280517578,
+      "learning_rate": 2.9117647058823534e-05,
+      "loss": 0.4903,
       "step": 90
     },
     {
-      "epoch": 2.2222222222222223,
-      "grad_norm": 6.143222808837891,
-      "learning_rate": 1.4462809917355372e-05,
-      "loss": 0.6805,
+      "epoch": 1.5810276679841897,
+      "grad_norm": 7.857639789581299,
+      "learning_rate": 2.6176470588235295e-05,
+      "loss": 0.5096,
       "step": 100
     },
     {
-      "epoch": 2.4444444444444446,
-      "grad_norm": 1.5121873617172241,
-      "learning_rate": 1.0330578512396695e-05,
-      "loss": 0.6844,
+      "epoch": 1.7391304347826086,
+      "grad_norm": 6.486521244049072,
+      "learning_rate": 2.323529411764706e-05,
+      "loss": 0.4911,
       "step": 110
     },
     {
-      "epoch": 2.6666666666666665,
-      "grad_norm": 0.6525880694389343,
-      "learning_rate": 6.198347107438017e-06,
-      "loss": 0.6964,
+      "epoch": 1.8972332015810278,
+      "grad_norm": 12.094149589538574,
+      "learning_rate": 2.0294117647058825e-05,
+      "loss": 0.4804,
       "step": 120
     },
     {
-      "epoch": 2.888888888888889,
-      "grad_norm": 4.430673599243164,
-      "learning_rate": 2.066115702479339e-06,
-      "loss": 0.6829,
+      "epoch": 1.992094861660079,
+      "eval_accuracy": 0.738015607580825,
+      "eval_loss": 0.5335468053817749,
+      "eval_runtime": 21.4566,
+      "eval_samples_per_second": 41.805,
+      "eval_steps_per_second": 1.352,
+      "step": 126
+    },
+    {
+      "epoch": 2.0553359683794468,
+      "grad_norm": 10.58962345123291,
+      "learning_rate": 1.735294117647059e-05,
+      "loss": 0.4347,
       "step": 130
     },
     {
-      "epoch": 3.0,
-      "eval_f1": 0.6529126213592233,
-      "eval_loss": 0.6814367175102234,
-      "eval_runtime": 15.0867,
-      "eval_samples_per_second": 42.421,
-      "eval_steps_per_second": 1.326,
-      "step": 135
+      "epoch": 2.2134387351778657,
+      "grad_norm": 12.747217178344727,
+      "learning_rate": 1.4411764705882352e-05,
+      "loss": 0.3991,
+      "step": 140
+    },
+    {
+      "epoch": 2.3715415019762847,
+      "grad_norm": 9.04605770111084,
+      "learning_rate": 1.1470588235294118e-05,
+      "loss": 0.3978,
+      "step": 150
+    },
+    {
+      "epoch": 2.5296442687747036,
+      "grad_norm": 11.612887382507324,
+      "learning_rate": 8.529411764705883e-06,
+      "loss": 0.3989,
+      "step": 160
+    },
+    {
+      "epoch": 2.6877470355731226,
+      "grad_norm": 11.906518936157227,
+      "learning_rate": 5.588235294117647e-06,
+      "loss": 0.3797,
+      "step": 170
+    },
+    {
+      "epoch": 2.8458498023715415,
+      "grad_norm": 11.579667091369629,
+      "learning_rate": 2.647058823529412e-06,
+      "loss": 0.3901,
+      "step": 180
+    },
+    {
+      "epoch": 2.9881422924901186,
+      "eval_accuracy": 0.7536231884057971,
+      "eval_loss": 0.5222525000572205,
+      "eval_runtime": 21.3284,
+      "eval_samples_per_second": 42.057,
+      "eval_steps_per_second": 1.36,
+      "step": 189
     },
     {
-      "epoch": 3.0,
-      "step": 135,
-      "total_flos": 1.3494104908807864e+18,
-      "train_loss": 0.7046969166508428,
-      "train_runtime": 848.1758,
-      "train_samples_per_second": 20.345,
-      "train_steps_per_second": 0.159
+      "epoch": 2.9881422924901186,
+      "step": 189,
+      "total_flos": 1.8869538215666074e+18,
+      "train_loss": 0.48228894844257014,
+      "train_runtime": 1217.9669,
+      "train_samples_per_second": 19.865,
+      "train_steps_per_second": 0.155
     }
   ],
   "logging_steps": 10,
-  "max_steps": 135,
+  "max_steps": 189,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 3,
   "save_steps": 500,
-  "total_flos": 1.3494104908807864e+18,
+  "total_flos": 1.8869538215666074e+18,
   "train_batch_size": 32,
   "trial_name": null,
   "trial_params": null