End of training

README.md

@@ -17,7 +17,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [microsoft/swin-base-patch4-window7-224-in22k](https://huggingface.co/microsoft/swin-base-patch4-window7-224-in22k) on the imagefolder dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.0115
+- Loss: 0.0104
 
 ## Model description
 

all_results.json

@@ -1,8 +1,12 @@
 {
-    "epoch": 4.8,
-    "total_flos": 4.4500275596722176e+17,
-    "train_loss": 0.1861822952826818,
-    "train_runtime": 258.3418,
-    "train_samples_per_second": 22.838,
-    "train_steps_per_second": 0.116
+    "epoch": 9.6,
+    "eval_loss": 0.010416708886623383,
+    "eval_runtime": 2.6618,
+    "eval_samples_per_second": 49.59,
+    "eval_steps_per_second": 1.127,
+    "total_flos": 8.884386008218829e+17,
+    "train_loss": 0.13027287572622298,
+    "train_runtime": 399.2998,
+    "train_samples_per_second": 29.552,
+    "train_steps_per_second": 0.15
 }

eval_results.json

@@ -0,0 +1,7 @@
+{
+    "epoch": 9.6,
+    "eval_loss": 0.010416708886623383,
+    "eval_runtime": 2.6618,
+    "eval_samples_per_second": 49.59,
+    "eval_steps_per_second": 1.127
+}

runs/Apr18_19-54-13_c91955ade82b/events.out.tfevents.1713470470.c91955ade82b.34.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4a8d2a0e4cde4075077ebcc8494a12c95a938f6653107da1438c81b9db11f4c8
+size 354

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 4.8,
-    "total_flos": 4.4500275596722176e+17,
-    "train_loss": 0.1861822952826818,
-    "train_runtime": 258.3418,
-    "train_samples_per_second": 22.838,
-    "train_steps_per_second": 0.116
+    "epoch": 9.6,
+    "total_flos": 8.884386008218829e+17,
+    "train_loss": 0.13027287572622298,
+    "train_runtime": 399.2998,
+    "train_samples_per_second": 29.552,
+    "train_steps_per_second": 0.15
 }

trainer_state.json

@@ -1,90 +1,151 @@
 {
-  "best_metric": 0.030113982036709785,
-  "best_model_checkpoint": "swin-base-patch4-window7-224-in22k-Kontur-competition-1.3K/checkpoint-18",
-  "epoch": 4.8,
+  "best_metric": 0.010416708886623383,
+  "best_model_checkpoint": "swin-base-patch4-window7-224-in22k-Kontur-competition-1.3K/checkpoint-56",
+  "epoch": 9.6,
   "eval_steps": 500,
-  "global_step": 30,
+  "global_step": 60,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.96,
-      "eval_loss": 0.3283247947692871,
-      "eval_runtime": 3.7851,
-      "eval_samples_per_second": 34.874,
-      "eval_steps_per_second": 0.793,
+      "eval_loss": 0.4818643033504486,
+      "eval_runtime": 2.5684,
+      "eval_samples_per_second": 51.393,
+      "eval_steps_per_second": 1.168,
       "step": 6
     },
     {
       "epoch": 1.6,
-      "grad_norm": 3.094001293182373,
-      "learning_rate": 3.7037037037037037e-05,
-      "loss": 0.4531,
+      "grad_norm": 3.978776454925537,
+      "learning_rate": 4.62962962962963e-05,
+      "loss": 0.5639,
       "step": 10
     },
     {
       "epoch": 1.92,
-      "eval_loss": 0.08094710856676102,
-      "eval_runtime": 2.5203,
-      "eval_samples_per_second": 52.376,
-      "eval_steps_per_second": 1.19,
+      "eval_loss": 0.22424142062664032,
+      "eval_runtime": 2.618,
+      "eval_samples_per_second": 50.419,
+      "eval_steps_per_second": 1.146,
       "step": 12
     },
     {
       "epoch": 2.88,
-      "eval_loss": 0.030113982036709785,
-      "eval_runtime": 2.6775,
-      "eval_samples_per_second": 49.299,
-      "eval_steps_per_second": 1.12,
+      "eval_loss": 0.0641486644744873,
+      "eval_runtime": 2.5792,
+      "eval_samples_per_second": 51.18,
+      "eval_steps_per_second": 1.163,
       "step": 18
     },
     {
       "epoch": 3.2,
-      "grad_norm": 5.238066673278809,
-      "learning_rate": 1.8518518518518518e-05,
-      "loss": 0.069,
+      "grad_norm": 2.1514062881469727,
+      "learning_rate": 3.7037037037037037e-05,
+      "loss": 0.1359,
       "step": 20
     },
     {
       "epoch": 4.0,
-      "eval_loss": 0.12024853378534317,
-      "eval_runtime": 2.6542,
-      "eval_samples_per_second": 49.733,
-      "eval_steps_per_second": 1.13,
+      "eval_loss": 0.07821954041719437,
+      "eval_runtime": 2.6087,
+      "eval_samples_per_second": 50.601,
+      "eval_steps_per_second": 1.15,
       "step": 25
     },
     {
       "epoch": 4.8,
-      "grad_norm": 3.4171395301818848,
-      "learning_rate": 0.0,
-      "loss": 0.0364,
+      "grad_norm": 1.6489207744598389,
+      "learning_rate": 2.777777777777778e-05,
+      "loss": 0.031,
       "step": 30
     },
     {
-      "epoch": 4.8,
-      "eval_loss": 0.0865083634853363,
-      "eval_runtime": 2.5761,
-      "eval_samples_per_second": 51.24,
-      "eval_steps_per_second": 1.165,
-      "step": 30
+      "epoch": 4.96,
+      "eval_loss": 0.025011500343680382,
+      "eval_runtime": 2.632,
+      "eval_samples_per_second": 50.152,
+      "eval_steps_per_second": 1.14,
+      "step": 31
     },
     {
-      "epoch": 4.8,
-      "step": 30,
-      "total_flos": 4.4500275596722176e+17,
-      "train_loss": 0.1861822952826818,
-      "train_runtime": 258.3418,
-      "train_samples_per_second": 22.838,
-      "train_steps_per_second": 0.116
+      "epoch": 5.92,
+      "eval_loss": 0.017290038987994194,
+      "eval_runtime": 2.6102,
+      "eval_samples_per_second": 50.571,
+      "eval_steps_per_second": 1.149,
+      "step": 37
+    },
+    {
+      "epoch": 6.4,
+      "grad_norm": 7.398428440093994,
+      "learning_rate": 1.8518518518518518e-05,
+      "loss": 0.0221,
+      "step": 40
+    },
+    {
+      "epoch": 6.88,
+      "eval_loss": 0.01186411827802658,
+      "eval_runtime": 2.603,
+      "eval_samples_per_second": 50.712,
+      "eval_steps_per_second": 1.153,
+      "step": 43
+    },
+    {
+      "epoch": 8.0,
+      "grad_norm": 0.057548072189092636,
+      "learning_rate": 9.259259259259259e-06,
+      "loss": 0.0144,
+      "step": 50
+    },
+    {
+      "epoch": 8.0,
+      "eval_loss": 0.01627057045698166,
+      "eval_runtime": 2.5904,
+      "eval_samples_per_second": 50.957,
+      "eval_steps_per_second": 1.158,
+      "step": 50
+    },
+    {
+      "epoch": 8.96,
+      "eval_loss": 0.010416708886623383,
+      "eval_runtime": 2.5905,
+      "eval_samples_per_second": 50.956,
+      "eval_steps_per_second": 1.158,
+      "step": 56
+    },
+    {
+      "epoch": 9.6,
+      "grad_norm": 0.8250289559364319,
+      "learning_rate": 0.0,
+      "loss": 0.0144,
+      "step": 60
+    },
+    {
+      "epoch": 9.6,
+      "eval_loss": 0.011465705931186676,
+      "eval_runtime": 2.9818,
+      "eval_samples_per_second": 44.269,
+      "eval_steps_per_second": 1.006,
+      "step": 60
+    },
+    {
+      "epoch": 9.6,
+      "step": 60,
+      "total_flos": 8.884386008218829e+17,
+      "train_loss": 0.13027287572622298,
+      "train_runtime": 399.2998,
+      "train_samples_per_second": 29.552,
+      "train_steps_per_second": 0.15
     }
   ],
   "logging_steps": 10,
-  "max_steps": 30,
+  "max_steps": 60,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 5,
+  "num_train_epochs": 10,
   "save_steps": 500,
-  "total_flos": 4.4500275596722176e+17,
+  "total_flos": 8.884386008218829e+17,
   "train_batch_size": 48,
   "trial_name": null,
   "trial_params": null