Training in progress, step 10

all_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 56.666666666666664,
     "total_flos": 2.576884875160781e+16,
-    "train_loss": 0.6197609144098618,
-    "train_runtime": 31.4964,
-    "train_samples_per_second": 121.443,
-    "train_steps_per_second": 2.699
+    "train_loss": 0.17622246672125425,
+    "train_runtime": 32.2918,
+    "train_samples_per_second": 118.451,
+    "train_steps_per_second": 2.632
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:23824b97c77ae2fd25220117b1334df8700e8a8b0646b55d46866d4ff625b997
+oid sha256:c0f28d90eba1f59db494c41e3d5586f383f85c53de367bf8c854f3f593aa52ef
 size 44774596

runs/Nov11_11-43-44_9dbf826d7e63/events.out.tfevents.1731325602.9dbf826d7e63.838.3

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:21848382f7839e28accec5243991cc6f3f36584205454041c17c8651c5e6faf8
+size 5798

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 56.666666666666664,
     "total_flos": 2.576884875160781e+16,
-    "train_loss": 0.6197609144098618,
-    "train_runtime": 31.4964,
-    "train_samples_per_second": 121.443,
-    "train_steps_per_second": 2.699
+    "train_loss": 0.17622246672125425,
+    "train_runtime": 32.2918,
+    "train_samples_per_second": 118.451,
+    "train_steps_per_second": 2.632
 }

trainer_state.json

@@ -1,5 +1,5 @@
 {
-  "best_metric": 1.5808919668197632,
+  "best_metric": 1.2994384765625,
   "best_model_checkpoint": "resnet-18-dungeons-001/checkpoint-80",
   "epoch": 56.666666666666664,
   "eval_steps": 10,
@@ -10,140 +10,140 @@
   "log_history": [
     {
       "epoch": 6.666666666666667,
-      "grad_norm": 8.576932907104492,
+      "grad_norm": 3.9049274921417236,
       "learning_rate": 1.0000000000000002e-06,
-      "loss": 0.8642,
+      "loss": 0.2236,
       "step": 10
     },
     {
       "epoch": 6.666666666666667,
-      "eval_accuracy": 0.3333333333333333,
-      "eval_loss": 1.9274088144302368,
-      "eval_runtime": 0.1032,
-      "eval_samples_per_second": 116.247,
-      "eval_steps_per_second": 19.374,
+      "eval_accuracy": 0.5833333333333334,
+      "eval_loss": 1.567138671875,
+      "eval_runtime": 0.1015,
+      "eval_samples_per_second": 118.186,
+      "eval_steps_per_second": 19.698,
       "step": 10
     },
     {
       "epoch": 13.333333333333334,
-      "grad_norm": 9.152162551879883,
+      "grad_norm": 4.2215895652771,
       "learning_rate": 2.0000000000000003e-06,
-      "loss": 0.914,
+      "loss": 0.2873,
       "step": 20
     },
     {
       "epoch": 13.333333333333334,
-      "eval_accuracy": 0.3333333333333333,
-      "eval_loss": 1.9344075918197632,
-      "eval_runtime": 0.167,
-      "eval_samples_per_second": 71.854,
-      "eval_steps_per_second": 11.976,
+      "eval_accuracy": 0.5833333333333334,
+      "eval_loss": 1.5972086191177368,
+      "eval_runtime": 0.102,
+      "eval_samples_per_second": 117.645,
+      "eval_steps_per_second": 19.608,
       "step": 20
     },
     {
       "epoch": 20.0,
-      "grad_norm": 5.871454238891602,
+      "grad_norm": 1.4720317125320435,
       "learning_rate": 3e-06,
-      "loss": 0.7867,
+      "loss": 0.2203,
       "step": 30
     },
     {
       "epoch": 20.0,
-      "eval_accuracy": 0.3333333333333333,
-      "eval_loss": 1.8524576425552368,
-      "eval_runtime": 0.1045,
-      "eval_samples_per_second": 114.786,
-      "eval_steps_per_second": 19.131,
+      "eval_accuracy": 0.6666666666666666,
+      "eval_loss": 1.4836832284927368,
+      "eval_runtime": 0.1006,
+      "eval_samples_per_second": 119.334,
+      "eval_steps_per_second": 19.889,
       "step": 30
     },
     {
       "epoch": 26.666666666666668,
-      "grad_norm": 5.765829563140869,
+      "grad_norm": 1.490944266319275,
       "learning_rate": 4.000000000000001e-06,
-      "loss": 0.7229,
+      "loss": 0.2063,
       "step": 40
     },
     {
       "epoch": 26.666666666666668,
-      "eval_accuracy": 0.4166666666666667,
-      "eval_loss": 1.813232421875,
-      "eval_runtime": 0.1046,
-      "eval_samples_per_second": 114.681,
-      "eval_steps_per_second": 19.113,
+      "eval_accuracy": 0.6666666666666666,
+      "eval_loss": 1.4542642831802368,
+      "eval_runtime": 0.1527,
+      "eval_samples_per_second": 78.602,
+      "eval_steps_per_second": 13.1,
       "step": 40
     },
     {
       "epoch": 33.333333333333336,
-      "grad_norm": 8.122588157653809,
+      "grad_norm": 2.723828077316284,
       "learning_rate": 5e-06,
-      "loss": 0.6161,
+      "loss": 0.1631,
       "step": 50
     },
     {
       "epoch": 33.333333333333336,
-      "eval_accuracy": 0.4166666666666667,
-      "eval_loss": 1.8011881113052368,
-      "eval_runtime": 0.1032,
-      "eval_samples_per_second": 116.279,
-      "eval_steps_per_second": 19.38,
+      "eval_accuracy": 0.6666666666666666,
+      "eval_loss": 1.4692789316177368,
+      "eval_runtime": 0.1017,
+      "eval_samples_per_second": 117.95,
+      "eval_steps_per_second": 19.658,
       "step": 50
     },
     {
       "epoch": 40.0,
-      "grad_norm": 6.240235805511475,
+      "grad_norm": 3.7070419788360596,
       "learning_rate": 6e-06,
-      "loss": 0.5369,
+      "loss": 0.1546,
       "step": 60
     },
     {
       "epoch": 40.0,
-      "eval_accuracy": 0.4166666666666667,
-      "eval_loss": 1.7578125,
-      "eval_runtime": 0.1028,
-      "eval_samples_per_second": 116.714,
-      "eval_steps_per_second": 19.452,
+      "eval_accuracy": 0.5833333333333334,
+      "eval_loss": 1.4473876953125,
+      "eval_runtime": 0.1192,
+      "eval_samples_per_second": 100.651,
+      "eval_steps_per_second": 16.775,
       "step": 60
     },
     {
       "epoch": 46.666666666666664,
-      "grad_norm": 4.085389614105225,
+      "grad_norm": 1.3349393606185913,
       "learning_rate": 7e-06,
-      "loss": 0.3855,
+      "loss": 0.1032,
       "step": 70
     },
     {
       "epoch": 46.666666666666664,
-      "eval_accuracy": 0.5,
-      "eval_loss": 1.6616617441177368,
-      "eval_runtime": 0.1047,
-      "eval_samples_per_second": 114.609,
-      "eval_steps_per_second": 19.102,
+      "eval_accuracy": 0.5833333333333334,
+      "eval_loss": 1.3604329824447632,
+      "eval_runtime": 0.1048,
+      "eval_samples_per_second": 114.494,
+      "eval_steps_per_second": 19.082,
       "step": 70
     },
     {
       "epoch": 53.333333333333336,
-      "grad_norm": 3.3966498374938965,
+      "grad_norm": 1.0077532529830933,
       "learning_rate": 8.000000000000001e-06,
-      "loss": 0.3339,
+      "loss": 0.1101,
       "step": 80
     },
     {
       "epoch": 53.333333333333336,
-      "eval_accuracy": 0.5833333333333334,
-      "eval_loss": 1.5808919668197632,
-      "eval_runtime": 0.1125,
-      "eval_samples_per_second": 106.626,
-      "eval_steps_per_second": 17.771,
+      "eval_accuracy": 0.5,
+      "eval_loss": 1.2994384765625,
+      "eval_runtime": 0.1073,
+      "eval_samples_per_second": 111.828,
+      "eval_steps_per_second": 18.638,
       "step": 80
     },
     {
       "epoch": 56.666666666666664,
       "step": 85,
       "total_flos": 2.576884875160781e+16,
-      "train_loss": 0.6197609144098618,
-      "train_runtime": 31.4964,
-      "train_samples_per_second": 121.443,
-      "train_steps_per_second": 2.699
+      "train_loss": 0.17622246672125425,
+      "train_runtime": 32.2918,
+      "train_samples_per_second": 118.451,
+      "train_steps_per_second": 2.632
     }
   ],
   "logging_steps": 10,