Model save

README.md

@@ -17,8 +17,8 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [MCG-NJU/videomae-base](https://huggingface.co/MCG-NJU/videomae-base) on an unknown dataset.
 It achieves the following results on the evaluation set:
-- Loss: 2.1163
-- Accuracy: 0.3594
+- Loss: 1.7371
+- Accuracy: 0.3704
 
 ## Model description
 

all_results.json

@@ -1,12 +1,7 @@
 {
-    "epoch": 8.09,
-    "eval_accuracy": 0.25925925925925924,
-    "eval_loss": 1.6813513040542603,
-    "eval_runtime": 64.4906,
-    "eval_samples_per_second": 0.419,
-    "eval_steps_per_second": 0.109,
-    "train_loss": 1.5179050036839077,
-    "train_runtime": 4008.8343,
-    "train_samples_per_second": 0.14,
-    "train_steps_per_second": 0.009
+    "epoch": 17.03,
+    "train_loss": 1.4362963676452636,
+    "train_runtime": 8809.684,
+    "train_samples_per_second": 0.127,
+    "train_steps_per_second": 0.004
 }

train_results.json

@@ -1,7 +1,7 @@
 {
-    "epoch": 8.09,
-    "train_loss": 1.5179050036839077,
-    "train_runtime": 4008.8343,
-    "train_samples_per_second": 0.14,
-    "train_steps_per_second": 0.009
+    "epoch": 17.03,
+    "train_loss": 1.4362963676452636,
+    "train_runtime": 8809.684,
+    "train_samples_per_second": 0.127,
+    "train_steps_per_second": 0.004
 }

trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 0.34375,
-  "best_model_checkpoint": "finetuned-Accident-MultipleLabels-Video-subset-v2-checkpointing/checkpoint-24",
-  "epoch": 8.085714285714285,
+  "best_metric": 0.390625,
+  "best_model_checkpoint": "finetuned-Accident-MultipleLabels-Video-subset-v2-checkpointing/checkpoint-4",
+  "epoch": 17.02857142857143,
   "eval_steps": 500,
   "global_step": 35,
   "is_hyper_param_search": false,
@@ -9,137 +9,200 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.11,
-      "eval_accuracy": 0.265625,
-      "eval_loss": 1.856452226638794,
-      "eval_runtime": 208.3919,
-      "eval_samples_per_second": 0.307,
-      "eval_steps_per_second": 0.077,
+      "epoch": 0.06,
+      "eval_accuracy": 0.359375,
+      "eval_loss": 1.7265217304229736,
+      "eval_runtime": 228.6142,
+      "eval_samples_per_second": 0.28,
+      "eval_steps_per_second": 0.035,
+      "step": 2
+    },
+    {
+      "epoch": 1.06,
+      "eval_accuracy": 0.390625,
+      "eval_loss": 1.6976414918899536,
+      "eval_runtime": 225.5421,
+      "eval_samples_per_second": 0.284,
+      "eval_steps_per_second": 0.035,
       "step": 4
     },
     {
-      "epoch": 1.11,
-      "eval_accuracy": 0.265625,
-      "eval_loss": 1.867037296295166,
-      "eval_runtime": 204.858,
-      "eval_samples_per_second": 0.312,
-      "eval_steps_per_second": 0.078,
+      "epoch": 2.06,
+      "eval_accuracy": 0.359375,
+      "eval_loss": 1.7503018379211426,
+      "eval_runtime": 176.7414,
+      "eval_samples_per_second": 0.362,
+      "eval_steps_per_second": 0.045,
+      "step": 6
+    },
+    {
+      "epoch": 3.06,
+      "eval_accuracy": 0.3125,
+      "eval_loss": 1.8830795288085938,
+      "eval_runtime": 175.1171,
+      "eval_samples_per_second": 0.365,
+      "eval_steps_per_second": 0.046,
       "step": 8
     },
     {
-      "epoch": 2.06,
+      "epoch": 4.06,
       "learning_rate": 4.032258064516129e-05,
-      "loss": 1.8022,
+      "loss": 1.7254,
+      "step": 10
+    },
+    {
+      "epoch": 4.06,
+      "eval_accuracy": 0.171875,
+      "eval_loss": 2.0285067558288574,
+      "eval_runtime": 184.7182,
+      "eval_samples_per_second": 0.346,
+      "eval_steps_per_second": 0.043,
       "step": 10
     },
     {
-      "epoch": 2.11,
-      "eval_accuracy": 0.21875,
-      "eval_loss": 1.9593775272369385,
-      "eval_runtime": 158.8606,
-      "eval_samples_per_second": 0.403,
-      "eval_steps_per_second": 0.101,
+      "epoch": 5.06,
+      "eval_accuracy": 0.28125,
+      "eval_loss": 2.039059638977051,
+      "eval_runtime": 178.8587,
+      "eval_samples_per_second": 0.358,
+      "eval_steps_per_second": 0.045,
       "step": 12
     },
     {
-      "epoch": 3.11,
-      "eval_accuracy": 0.296875,
-      "eval_loss": 1.9278614521026611,
-      "eval_runtime": 163.5324,
-      "eval_samples_per_second": 0.391,
-      "eval_steps_per_second": 0.098,
+      "epoch": 6.06,
+      "eval_accuracy": 0.328125,
+      "eval_loss": 1.9736883640289307,
+      "eval_runtime": 179.8546,
+      "eval_samples_per_second": 0.356,
+      "eval_steps_per_second": 0.044,
+      "step": 14
+    },
+    {
+      "epoch": 7.06,
+      "eval_accuracy": 0.375,
+      "eval_loss": 1.8998150825500488,
+      "eval_runtime": 172.5948,
+      "eval_samples_per_second": 0.371,
+      "eval_steps_per_second": 0.046,
       "step": 16
     },
     {
-      "epoch": 4.11,
+      "epoch": 8.06,
+      "eval_accuracy": 0.375,
+      "eval_loss": 1.8785591125488281,
+      "eval_runtime": 169.9356,
+      "eval_samples_per_second": 0.377,
+      "eval_steps_per_second": 0.047,
+      "step": 18
+    },
+    {
+      "epoch": 9.06,
       "learning_rate": 2.4193548387096777e-05,
-      "loss": 1.4448,
+      "loss": 1.394,
       "step": 20
     },
     {
-      "epoch": 4.11,
-      "eval_accuracy": 0.296875,
-      "eval_loss": 1.910567283630371,
-      "eval_runtime": 161.4632,
-      "eval_samples_per_second": 0.396,
-      "eval_steps_per_second": 0.099,
+      "epoch": 9.06,
+      "eval_accuracy": 0.34375,
+      "eval_loss": 1.9054185152053833,
+      "eval_runtime": 179.2066,
+      "eval_samples_per_second": 0.357,
+      "eval_steps_per_second": 0.045,
       "step": 20
     },
     {
-      "epoch": 5.11,
-      "eval_accuracy": 0.34375,
-      "eval_loss": 1.8732622861862183,
-      "eval_runtime": 158.3884,
-      "eval_samples_per_second": 0.404,
-      "eval_steps_per_second": 0.101,
+      "epoch": 10.06,
+      "eval_accuracy": 0.328125,
+      "eval_loss": 1.9473646879196167,
+      "eval_runtime": 173.6519,
+      "eval_samples_per_second": 0.369,
+      "eval_steps_per_second": 0.046,
+      "step": 22
+    },
+    {
+      "epoch": 11.06,
+      "eval_accuracy": 0.328125,
+      "eval_loss": 2.003166675567627,
+      "eval_runtime": 174.1677,
+      "eval_samples_per_second": 0.367,
+      "eval_steps_per_second": 0.046,
       "step": 24
     },
     {
-      "epoch": 6.11,
-      "eval_accuracy": 0.3125,
-      "eval_loss": 1.8884299993515015,
-      "eval_runtime": 159.9497,
-      "eval_samples_per_second": 0.4,
-      "eval_steps_per_second": 0.1,
+      "epoch": 12.06,
+      "eval_accuracy": 0.328125,
+      "eval_loss": 2.072862386703491,
+      "eval_runtime": 196.1804,
+      "eval_samples_per_second": 0.326,
+      "eval_steps_per_second": 0.041,
+      "step": 26
+    },
+    {
+      "epoch": 13.06,
+      "eval_accuracy": 0.34375,
+      "eval_loss": 2.108105421066284,
+      "eval_runtime": 180.4592,
+      "eval_samples_per_second": 0.355,
+      "eval_steps_per_second": 0.044,
       "step": 28
     },
     {
-      "epoch": 7.06,
+      "epoch": 14.06,
       "learning_rate": 8.064516129032258e-06,
-      "loss": 1.3906,
+      "loss": 1.285,
       "step": 30
     },
     {
-      "epoch": 7.11,
-      "eval_accuracy": 0.28125,
-      "eval_loss": 1.8995435237884521,
-      "eval_runtime": 165.6513,
-      "eval_samples_per_second": 0.386,
-      "eval_steps_per_second": 0.097,
+      "epoch": 14.06,
+      "eval_accuracy": 0.328125,
+      "eval_loss": 2.1189751625061035,
+      "eval_runtime": 180.6898,
+      "eval_samples_per_second": 0.354,
+      "eval_steps_per_second": 0.044,
+      "step": 30
+    },
+    {
+      "epoch": 15.06,
+      "eval_accuracy": 0.34375,
+      "eval_loss": 2.1188015937805176,
+      "eval_runtime": 172.3883,
+      "eval_samples_per_second": 0.371,
+      "eval_steps_per_second": 0.046,
       "step": 32
     },
     {
-      "epoch": 8.09,
-      "eval_accuracy": 0.28125,
-      "eval_loss": 1.9018282890319824,
-      "eval_runtime": 160.8854,
-      "eval_samples_per_second": 0.398,
-      "eval_steps_per_second": 0.099,
-      "step": 35
+      "epoch": 16.06,
+      "eval_accuracy": 0.359375,
+      "eval_loss": 2.1155242919921875,
+      "eval_runtime": 169.8241,
+      "eval_samples_per_second": 0.377,
+      "eval_steps_per_second": 0.047,
+      "step": 34
     },
     {
-      "epoch": 8.09,
-      "step": 35,
-      "total_flos": 6.978272853255782e+17,
-      "train_loss": 1.5179050036839077,
-      "train_runtime": 4008.8343,
-      "train_samples_per_second": 0.14,
-      "train_steps_per_second": 0.009
-    },
-    {
-      "epoch": 8.09,
-      "eval_accuracy": 0.25925925925925924,
-      "eval_loss": 1.6813514232635498,
-      "eval_runtime": 64.3768,
-      "eval_samples_per_second": 0.419,
-      "eval_steps_per_second": 0.109,
+      "epoch": 17.03,
+      "eval_accuracy": 0.359375,
+      "eval_loss": 2.116347312927246,
+      "eval_runtime": 177.6799,
+      "eval_samples_per_second": 0.36,
+      "eval_steps_per_second": 0.045,
       "step": 35
     },
     {
-      "epoch": 8.09,
-      "eval_accuracy": 0.25925925925925924,
-      "eval_loss": 1.6813513040542603,
-      "eval_runtime": 64.4906,
-      "eval_samples_per_second": 0.419,
-      "eval_steps_per_second": 0.109,
-      "step": 35
+      "epoch": 17.03,
+      "step": 35,
+      "total_flos": 1.3956545706511565e+18,
+      "train_loss": 1.4362963676452636,
+      "train_runtime": 8809.684,
+      "train_samples_per_second": 0.127,
+      "train_steps_per_second": 0.004
     }
   ],
   "logging_steps": 10,
   "max_steps": 35,
   "num_train_epochs": 9223372036854775807,
   "save_steps": 500,
-  "total_flos": 6.978272853255782e+17,
+  "total_flos": 1.3956545706511565e+18,
   "trial_name": null,
   "trial_params": null
 }