Training in progress, epoch 0

all_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 9.49,
-    "total_flos": 4.447730655960883e+17,
-    "train_loss": 0.31461275049618315,
-    "train_runtime": 310.8231,
-    "train_samples_per_second": 60.645,
-    "train_steps_per_second": 0.45
+    "epoch": 28.57,
+    "total_flos": 1.4797115229312e+18,
+    "train_loss": 0.175064886448284,
+    "train_runtime": 1076.0914,
+    "train_samples_per_second": 18.205,
+    "train_steps_per_second": 0.139
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:906024750d8d4fc61fb267c5ef650bafa6d3a7f15b568fc2f56be4112895decc
+oid sha256:a4bf131fb2e50399ddd95fe34bae231d6b74ce7241f4f8f97c5f209d228d5ece
 size 110348984

runs/Apr08_03-52-04_pytorch-1-10-gpu-py-ml-g4dn-xlarge-bd81d1f288c5d63d0571c57fb0b0/events.out.tfevents.1712548468.pytorch-1-10-gpu-py-ml-g4dn-xlarge-bd81d1f288c5d63d0571c57fb0b0.746.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:167405626ed62bd5ecc4744f235518f54fd21eb515afd7b6f9ecfcf5a774666e
+size 5450

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 9.49,
-    "total_flos": 4.447730655960883e+17,
-    "train_loss": 0.31461275049618315,
-    "train_runtime": 310.8231,
-    "train_samples_per_second": 60.645,
-    "train_steps_per_second": 0.45
+    "epoch": 28.57,
+    "total_flos": 1.4797115229312e+18,
+    "train_loss": 0.175064886448284,
+    "train_runtime": 1076.0914,
+    "train_samples_per_second": 18.205,
+    "train_steps_per_second": 0.139
 }

trainer_state.json

@@ -1,217 +1,395 @@
 {
-  "best_metric": 0.921760391198044,
-  "best_model_checkpoint": "swin-tiny-patch4-window7-224-finetuned-sealv1/checkpoint-118",
-  "epoch": 9.491525423728813,
+  "best_metric": 0.9862068965517241,
+  "best_model_checkpoint": "swin-tiny-patch4-window7-224-finetuned-sealv1/checkpoint-94",
+  "epoch": 28.571428571428573,
   "eval_steps": 500,
-  "global_step": 140,
+  "global_step": 150,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.68,
-      "grad_norm": 8.243694305419922,
-      "learning_rate": 3.571428571428572e-05,
-      "loss": 1.1068,
+      "epoch": 0.95,
+      "eval_accuracy": 0.496551724137931,
+      "eval_loss": 1.2919867038726807,
+      "eval_runtime": 3.1223,
+      "eval_samples_per_second": 46.44,
+      "eval_steps_per_second": 1.601,
+      "step": 5
+    },
+    {
+      "epoch": 1.9,
+      "grad_norm": 29.579647064208984,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 1.1379,
       "step": 10
     },
     {
-      "epoch": 0.95,
-      "eval_accuracy": 0.706601466992665,
-      "eval_loss": 0.6517727375030518,
-      "eval_runtime": 2.518,
-      "eval_samples_per_second": 162.428,
-      "eval_steps_per_second": 5.163,
-      "step": 14
-    },
-    {
-      "epoch": 1.36,
-      "grad_norm": 11.443496704101562,
-      "learning_rate": 4.761904761904762e-05,
-      "loss": 0.4912,
+      "epoch": 1.9,
+      "eval_accuracy": 0.496551724137931,
+      "eval_loss": 1.017675518989563,
+      "eval_runtime": 3.0482,
+      "eval_samples_per_second": 47.569,
+      "eval_steps_per_second": 1.64,
+      "step": 10
+    },
+    {
+      "epoch": 2.86,
+      "eval_accuracy": 0.8758620689655172,
+      "eval_loss": 0.7626163959503174,
+      "eval_runtime": 3.0592,
+      "eval_samples_per_second": 47.398,
+      "eval_steps_per_second": 1.634,
+      "step": 15
+    },
+    {
+      "epoch": 3.81,
+      "grad_norm": 11.914603233337402,
+      "learning_rate": 4.814814814814815e-05,
+      "loss": 0.6784,
       "step": 20
     },
     {
-      "epoch": 1.97,
-      "eval_accuracy": 0.843520782396088,
-      "eval_loss": 0.46679040789604187,
-      "eval_runtime": 2.5032,
-      "eval_samples_per_second": 163.39,
-      "eval_steps_per_second": 5.193,
-      "step": 29
+      "epoch": 4.0,
+      "eval_accuracy": 0.9310344827586207,
+      "eval_loss": 0.5387857556343079,
+      "eval_runtime": 3.0489,
+      "eval_samples_per_second": 47.558,
+      "eval_steps_per_second": 1.64,
+      "step": 21
+    },
+    {
+      "epoch": 4.95,
+      "eval_accuracy": 0.9103448275862069,
+      "eval_loss": 0.4190843105316162,
+      "eval_runtime": 3.0202,
+      "eval_samples_per_second": 48.01,
+      "eval_steps_per_second": 1.656,
+      "step": 26
     },
     {
-      "epoch": 2.03,
-      "grad_norm": 9.79934024810791,
-      "learning_rate": 4.3650793650793655e-05,
-      "loss": 0.3366,
+      "epoch": 5.71,
+      "grad_norm": 7.094911575317383,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 0.3269,
       "step": 30
     },
     {
-      "epoch": 2.71,
-      "grad_norm": 8.352068901062012,
-      "learning_rate": 3.968253968253968e-05,
-      "loss": 0.2749,
+      "epoch": 5.9,
+      "eval_accuracy": 0.8896551724137931,
+      "eval_loss": 0.39897656440734863,
+      "eval_runtime": 3.0245,
+      "eval_samples_per_second": 47.942,
+      "eval_steps_per_second": 1.653,
+      "step": 31
+    },
+    {
+      "epoch": 6.86,
+      "eval_accuracy": 0.9517241379310345,
+      "eval_loss": 0.20898482203483582,
+      "eval_runtime": 3.0505,
+      "eval_samples_per_second": 47.534,
+      "eval_steps_per_second": 1.639,
+      "step": 36
+    },
+    {
+      "epoch": 7.62,
+      "grad_norm": 6.531408786773682,
+      "learning_rate": 4.074074074074074e-05,
+      "loss": 0.2068,
       "step": 40
     },
     {
-      "epoch": 2.98,
-      "eval_accuracy": 0.8704156479217604,
-      "eval_loss": 0.41268301010131836,
-      "eval_runtime": 2.4728,
-      "eval_samples_per_second": 165.399,
-      "eval_steps_per_second": 5.257,
-      "step": 44
+      "epoch": 8.0,
+      "eval_accuracy": 0.9586206896551724,
+      "eval_loss": 0.18190069496631622,
+      "eval_runtime": 3.053,
+      "eval_samples_per_second": 47.494,
+      "eval_steps_per_second": 1.638,
+      "step": 42
+    },
+    {
+      "epoch": 8.95,
+      "eval_accuracy": 0.9655172413793104,
+      "eval_loss": 0.11921437084674835,
+      "eval_runtime": 3.0392,
+      "eval_samples_per_second": 47.71,
+      "eval_steps_per_second": 1.645,
+      "step": 47
     },
     {
-      "epoch": 3.39,
-      "grad_norm": 7.004426002502441,
-      "learning_rate": 3.571428571428572e-05,
-      "loss": 0.3189,
+      "epoch": 9.52,
+      "grad_norm": 5.500025272369385,
+      "learning_rate": 3.7037037037037037e-05,
+      "loss": 0.1104,
       "step": 50
     },
     {
-      "epoch": 4.0,
-      "eval_accuracy": 0.8875305623471883,
-      "eval_loss": 0.36258456110954285,
-      "eval_runtime": 2.4887,
-      "eval_samples_per_second": 164.344,
-      "eval_steps_per_second": 5.224,
-      "step": 59
-    },
-    {
-      "epoch": 4.07,
-      "grad_norm": 4.621389865875244,
-      "learning_rate": 3.1746031746031745e-05,
-      "loss": 0.2412,
+      "epoch": 9.9,
+      "eval_accuracy": 0.9724137931034482,
+      "eval_loss": 0.06815902888774872,
+      "eval_runtime": 3.0143,
+      "eval_samples_per_second": 48.105,
+      "eval_steps_per_second": 1.659,
+      "step": 52
+    },
+    {
+      "epoch": 10.86,
+      "eval_accuracy": 0.9724137931034482,
+      "eval_loss": 0.08536545187234879,
+      "eval_runtime": 3.0425,
+      "eval_samples_per_second": 47.658,
+      "eval_steps_per_second": 1.643,
+      "step": 57
+    },
+    {
+      "epoch": 11.43,
+      "grad_norm": 2.2279155254364014,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 0.0571,
       "step": 60
     },
     {
-      "epoch": 4.75,
-      "grad_norm": 6.152388572692871,
-      "learning_rate": 2.777777777777778e-05,
-      "loss": 0.2226,
+      "epoch": 12.0,
+      "eval_accuracy": 0.9655172413793104,
+      "eval_loss": 0.08157803863286972,
+      "eval_runtime": 3.0775,
+      "eval_samples_per_second": 47.116,
+      "eval_steps_per_second": 1.625,
+      "step": 63
+    },
+    {
+      "epoch": 12.95,
+      "eval_accuracy": 0.9793103448275862,
+      "eval_loss": 0.053488027304410934,
+      "eval_runtime": 3.0426,
+      "eval_samples_per_second": 47.656,
+      "eval_steps_per_second": 1.643,
+      "step": 68
+    },
+    {
+      "epoch": 13.33,
+      "grad_norm": 3.4508020877838135,
+      "learning_rate": 2.962962962962963e-05,
+      "loss": 0.0382,
       "step": 70
     },
     {
-      "epoch": 4.95,
-      "eval_accuracy": 0.9046454767726161,
-      "eval_loss": 0.26376911997795105,
-      "eval_runtime": 2.4808,
-      "eval_samples_per_second": 164.864,
-      "eval_steps_per_second": 5.24,
+      "epoch": 13.9,
+      "eval_accuracy": 0.9793103448275862,
+      "eval_loss": 0.04910058155655861,
+      "eval_runtime": 3.0539,
+      "eval_samples_per_second": 47.48,
+      "eval_steps_per_second": 1.637,
       "step": 73
     },
     {
-      "epoch": 5.42,
-      "grad_norm": 6.929192066192627,
-      "learning_rate": 2.380952380952381e-05,
-      "loss": 0.2394,
+      "epoch": 14.86,
+      "eval_accuracy": 0.9793103448275862,
+      "eval_loss": 0.05336721986532211,
+      "eval_runtime": 3.0215,
+      "eval_samples_per_second": 47.989,
+      "eval_steps_per_second": 1.655,
+      "step": 78
+    },
+    {
+      "epoch": 15.24,
+      "grad_norm": 0.396097868680954,
+      "learning_rate": 2.5925925925925925e-05,
+      "loss": 0.0158,
       "step": 80
     },
     {
-      "epoch": 5.97,
-      "eval_accuracy": 0.8801955990220048,
-      "eval_loss": 0.358397513628006,
-      "eval_runtime": 2.4802,
-      "eval_samples_per_second": 164.908,
-      "eval_steps_per_second": 5.242,
-      "step": 88
+      "epoch": 16.0,
+      "eval_accuracy": 0.9793103448275862,
+      "eval_loss": 0.03685032203793526,
+      "eval_runtime": 3.0204,
+      "eval_samples_per_second": 48.006,
+      "eval_steps_per_second": 1.655,
+      "step": 84
+    },
+    {
+      "epoch": 16.95,
+      "eval_accuracy": 0.9724137931034482,
+      "eval_loss": 0.11105308681726456,
+      "eval_runtime": 3.0328,
+      "eval_samples_per_second": 47.81,
+      "eval_steps_per_second": 1.649,
+      "step": 89
     },
     {
-      "epoch": 6.1,
-      "grad_norm": 6.410078525543213,
-      "learning_rate": 1.984126984126984e-05,
-      "loss": 0.2156,
+      "epoch": 17.14,
+      "grad_norm": 1.3435662984848022,
+      "learning_rate": 2.2222222222222223e-05,
+      "loss": 0.0082,
       "step": 90
     },
     {
-      "epoch": 6.78,
-      "grad_norm": 6.870254039764404,
-      "learning_rate": 1.5873015873015872e-05,
-      "loss": 0.2241,
+      "epoch": 17.9,
+      "eval_accuracy": 0.9862068965517241,
+      "eval_loss": 0.0514516718685627,
+      "eval_runtime": 3.0168,
+      "eval_samples_per_second": 48.065,
+      "eval_steps_per_second": 1.657,
+      "step": 94
+    },
+    {
+      "epoch": 18.86,
+      "eval_accuracy": 0.9793103448275862,
+      "eval_loss": 0.07132605463266373,
+      "eval_runtime": 3.0424,
+      "eval_samples_per_second": 47.66,
+      "eval_steps_per_second": 1.643,
+      "step": 99
+    },
+    {
+      "epoch": 19.05,
+      "grad_norm": 1.5692905187606812,
+      "learning_rate": 1.8518518518518518e-05,
+      "loss": 0.0105,
       "step": 100
     },
     {
-      "epoch": 6.98,
-      "eval_accuracy": 0.9046454767726161,
-      "eval_loss": 0.2820737659931183,
-      "eval_runtime": 2.4944,
-      "eval_samples_per_second": 163.967,
-      "eval_steps_per_second": 5.212,
-      "step": 103
+      "epoch": 20.0,
+      "eval_accuracy": 0.9793103448275862,
+      "eval_loss": 0.059764981269836426,
+      "eval_runtime": 3.0213,
+      "eval_samples_per_second": 47.993,
+      "eval_steps_per_second": 1.655,
+      "step": 105
     },
     {
-      "epoch": 7.46,
-      "grad_norm": 20.99500274658203,
-      "learning_rate": 1.1904761904761905e-05,
-      "loss": 0.1815,
+      "epoch": 20.95,
+      "grad_norm": 0.6745745539665222,
+      "learning_rate": 1.4814814814814815e-05,
+      "loss": 0.009,
       "step": 110
     },
     {
-      "epoch": 8.0,
-      "eval_accuracy": 0.921760391198044,
-      "eval_loss": 0.21383894979953766,
-      "eval_runtime": 2.4781,
-      "eval_samples_per_second": 165.045,
-      "eval_steps_per_second": 5.246,
-      "step": 118
-    },
-    {
-      "epoch": 8.14,
-      "grad_norm": 10.073310852050781,
-      "learning_rate": 7.936507936507936e-06,
-      "loss": 0.1715,
+      "epoch": 20.95,
+      "eval_accuracy": 0.9724137931034482,
+      "eval_loss": 0.07587717473506927,
+      "eval_runtime": 3.0482,
+      "eval_samples_per_second": 47.569,
+      "eval_steps_per_second": 1.64,
+      "step": 110
+    },
+    {
+      "epoch": 21.9,
+      "eval_accuracy": 0.9793103448275862,
+      "eval_loss": 0.07691250741481781,
+      "eval_runtime": 3.0353,
+      "eval_samples_per_second": 47.771,
+      "eval_steps_per_second": 1.647,
+      "step": 115
+    },
+    {
+      "epoch": 22.86,
+      "grad_norm": 5.058462142944336,
+      "learning_rate": 1.1111111111111112e-05,
+      "loss": 0.0134,
       "step": 120
     },
     {
-      "epoch": 8.81,
-      "grad_norm": 5.79032564163208,
-      "learning_rate": 3.968253968253968e-06,
-      "loss": 0.1862,
+      "epoch": 22.86,
+      "eval_accuracy": 0.9793103448275862,
+      "eval_loss": 0.070177361369133,
+      "eval_runtime": 3.0247,
+      "eval_samples_per_second": 47.938,
+      "eval_steps_per_second": 1.653,
+      "step": 120
+    },
+    {
+      "epoch": 24.0,
+      "eval_accuracy": 0.9793103448275862,
+      "eval_loss": 0.060489553958177567,
+      "eval_runtime": 3.0525,
+      "eval_samples_per_second": 47.501,
+      "eval_steps_per_second": 1.638,
+      "step": 126
+    },
+    {
+      "epoch": 24.76,
+      "grad_norm": 1.4174549579620361,
+      "learning_rate": 7.4074074074074075e-06,
+      "loss": 0.0042,
       "step": 130
     },
     {
-      "epoch": 8.95,
-      "eval_accuracy": 0.9046454767726161,
-      "eval_loss": 0.27377307415008545,
-      "eval_runtime": 2.4752,
-      "eval_samples_per_second": 165.24,
-      "eval_steps_per_second": 5.252,
-      "step": 132
+      "epoch": 24.95,
+      "eval_accuracy": 0.9793103448275862,
+      "eval_loss": 0.06210276484489441,
+      "eval_runtime": 3.0244,
+      "eval_samples_per_second": 47.944,
+      "eval_steps_per_second": 1.653,
+      "step": 131
     },
     {
-      "epoch": 9.49,
-      "grad_norm": 9.280794143676758,
-      "learning_rate": 0.0,
-      "loss": 0.1942,
-      "step": 140
+      "epoch": 25.9,
+      "eval_accuracy": 0.9793103448275862,
+      "eval_loss": 0.06535915285348892,
+      "eval_runtime": 3.0417,
+      "eval_samples_per_second": 47.671,
+      "eval_steps_per_second": 1.644,
+      "step": 136
     },
     {
-      "epoch": 9.49,
-      "eval_accuracy": 0.9119804400977995,
-      "eval_loss": 0.25534436106681824,
-      "eval_runtime": 2.4962,
-      "eval_samples_per_second": 163.851,
-      "eval_steps_per_second": 5.208,
+      "epoch": 26.67,
+      "grad_norm": 0.7513312101364136,
+      "learning_rate": 3.7037037037037037e-06,
+      "loss": 0.0027,
       "step": 140
     },
     {
-      "epoch": 9.49,
-      "step": 140,
-      "total_flos": 4.447730655960883e+17,
-      "train_loss": 0.31461275049618315,
-      "train_runtime": 310.8231,
-      "train_samples_per_second": 60.645,
-      "train_steps_per_second": 0.45
+      "epoch": 26.86,
+      "eval_accuracy": 0.9724137931034482,
+      "eval_loss": 0.06658191233873367,
+      "eval_runtime": 3.0308,
+      "eval_samples_per_second": 47.843,
+      "eval_steps_per_second": 1.65,
+      "step": 141
+    },
+    {
+      "epoch": 28.0,
+      "eval_accuracy": 0.9793103448275862,
+      "eval_loss": 0.06654076278209686,
+      "eval_runtime": 3.0747,
+      "eval_samples_per_second": 47.159,
+      "eval_steps_per_second": 1.626,
+      "step": 147
+    },
+    {
+      "epoch": 28.57,
+      "grad_norm": 2.4569530487060547,
+      "learning_rate": 0.0,
+      "loss": 0.0065,
+      "step": 150
+    },
+    {
+      "epoch": 28.57,
+      "eval_accuracy": 0.9793103448275862,
+      "eval_loss": 0.06503264605998993,
+      "eval_runtime": 3.0856,
+      "eval_samples_per_second": 46.992,
+      "eval_steps_per_second": 1.62,
+      "step": 150
+    },
+    {
+      "epoch": 28.57,
+      "step": 150,
+      "total_flos": 1.4797115229312e+18,
+      "train_loss": 0.175064886448284,
+      "train_runtime": 1076.0914,
+      "train_samples_per_second": 18.205,
+      "train_steps_per_second": 0.139
     }
   ],
   "logging_steps": 10,
-  "max_steps": 140,
+  "max_steps": 150,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 10,
+  "num_train_epochs": 30,
   "save_steps": 500,
-  "total_flos": 4.447730655960883e+17,
+  "total_flos": 1.4797115229312e+18,
   "train_batch_size": 32,
   "trial_name": null,
   "trial_params": null

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:aa7f6a4d27b095d4d9346004b0f7772126cf0335c77cd7dcbfa4d65e7f0a3ecc
+oid sha256:218eb8bf534d4d2f1bf19a89b978665296be37a95beb535a7cade1f692029704
 size 4527