Model save

README.md

@@ -18,8 +18,8 @@ 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: 0.5612
-- Accuracy: 0.7935
+- Loss: 0.3988
+- Accuracy: 0.8714
 
 ## Model description
 
@@ -52,10 +52,10 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch  | Step | Validation Loss | Accuracy |
 |:-------------:|:------:|:----:|:---------------:|:--------:|
-| 2.1719        | 0.2568 | 38   | 2.0526          | 0.1714   |
-| 1.4069        | 1.2568 | 76   | 1.2280          | 0.4714   |
-| 0.6534        | 2.2568 | 114  | 0.7264          | 0.7      |
-| 0.3677        | 3.2297 | 148  | 0.5529          | 0.8      |
+| 2.1392        | 0.2568 | 38   | 1.8769          | 0.4143   |
+| 1.0079        | 1.2568 | 76   | 0.9862          | 0.7286   |
+| 0.4814        | 2.2568 | 114  | 0.5106          | 0.8857   |
+| 0.3286        | 3.2297 | 148  | 0.3988          | 0.8714   |
 
 
 ### Framework versions

all_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 3.22972972972973,
-    "eval_accuracy": 0.7935483870967742,
-    "eval_loss": 0.5611749291419983,
-    "eval_runtime": 7.9378,
-    "eval_samples_per_second": 19.527,
-    "eval_steps_per_second": 2.52
+    "eval_accuracy": 0.8064516129032258,
+    "eval_loss": 0.5119282603263855,
+    "eval_runtime": 8.1588,
+    "eval_samples_per_second": 18.998,
+    "eval_steps_per_second": 2.451
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:37a1911d5bea29d14fd90e9de8f41ae225fd93992cce7669f3dfcabdb0d9f7c2
+oid sha256:7e9782a92fe2b92d055c33d710abfcc464ea7d62fbe46fdb190fd0425137b6ec
 size 344961984

runs/Dec09_21-13-29_psroka/events.out.tfevents.1733776568.psroka.814400.4

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:76522f4b5ba1d595c1a2886a016d42b55605e664ef5c9005a887f2eb26f131fe
-size 8974
+oid sha256:2bd07efd3e0816d797b9333d4833b4946f392cdcaebe2e5440cfa6f851c7646a
+size 10280

runs/Dec09_21-13-29_psroka/events.out.tfevents.1733776831.psroka.814400.5

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4e92dc18f22572168499bd1ee5ef4188ba27b95050f95fb3880bfd9d2eba77d7
+size 411

runs/Dec09_21-53-53_psroka/events.out.tfevents.1733777647.psroka.827004.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a8b9549d01cb49190bff93e568409bc7bdd1485d0503193ffa9f7000927a3f78
+size 10280

test_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 3.22972972972973,
-    "eval_accuracy": 0.7935483870967742,
-    "eval_loss": 0.5611749291419983,
-    "eval_runtime": 7.9378,
-    "eval_samples_per_second": 19.527,
-    "eval_steps_per_second": 2.52
+    "eval_accuracy": 0.8064516129032258,
+    "eval_loss": 0.5119282603263855,
+    "eval_runtime": 8.1588,
+    "eval_samples_per_second": 18.998,
+    "eval_steps_per_second": 2.451
 }

trainer_state.json

@@ -1,5 +1,5 @@
 {
-  "best_metric": 0.8,
+  "best_metric": 0.8714285714285714,
   "best_model_checkpoint": "videomae-base-finetuned-ucf101-subset\\checkpoint-148",
   "epoch": 3.22972972972973,
   "eval_steps": 500,
@@ -10,172 +10,154 @@
   "log_history": [
     {
       "epoch": 0.06756756756756757,
-      "grad_norm": 16.5500545501709,
-      "learning_rate": 2.6666666666666667e-05,
-      "loss": 2.3071,
+      "grad_norm": 11.614557266235352,
+      "learning_rate": 2.3333333333333336e-05,
+      "loss": 2.3407,
       "step": 10
     },
     {
       "epoch": 0.13513513513513514,
-      "grad_norm": 7.210005760192871,
+      "grad_norm": 7.542444229125977,
       "learning_rate": 4.9248120300751884e-05,
-      "loss": 2.2612,
+      "loss": 2.2367,
       "step": 20
     },
     {
       "epoch": 0.20270270270270271,
-      "grad_norm": 7.997417449951172,
+      "grad_norm": 8.507399559020996,
       "learning_rate": 4.548872180451128e-05,
-      "loss": 2.1719,
+      "loss": 2.201,
       "step": 30
     },
     {
       "epoch": 0.25675675675675674,
-      "eval_accuracy": 0.17142857142857143,
-      "eval_loss": 2.0525670051574707,
-      "eval_runtime": 3.6401,
-      "eval_samples_per_second": 19.23,
-      "eval_steps_per_second": 2.472,
+      "eval_accuracy": 0.5428571428571428,
+      "eval_loss": 1.803983211517334,
+      "eval_runtime": 3.6229,
+      "eval_samples_per_second": 19.322,
+      "eval_steps_per_second": 2.484,
       "step": 38
     },
     {
       "epoch": 1.0135135135135136,
-      "grad_norm": 7.0121283531188965,
+      "grad_norm": 9.280707359313965,
       "learning_rate": 4.172932330827068e-05,
-      "loss": 1.9451,
+      "loss": 2.0,
       "step": 40
     },
     {
       "epoch": 1.0810810810810811,
-      "grad_norm": 11.407858848571777,
+      "grad_norm": 10.3395357131958,
       "learning_rate": 3.796992481203008e-05,
-      "loss": 1.7221,
+      "loss": 1.6049,
       "step": 50
     },
     {
       "epoch": 1.1486486486486487,
-      "grad_norm": 14.631987571716309,
+      "grad_norm": 9.636208534240723,
       "learning_rate": 3.421052631578947e-05,
-      "loss": 1.4956,
+      "loss": 1.2228,
       "step": 60
     },
     {
       "epoch": 1.2162162162162162,
-      "grad_norm": 9.013343811035156,
+      "grad_norm": 7.421963691711426,
       "learning_rate": 3.0451127819548874e-05,
-      "loss": 1.4069,
+      "loss": 0.9633,
       "step": 70
     },
     {
       "epoch": 1.2567567567567568,
-      "eval_accuracy": 0.4714285714285714,
-      "eval_loss": 1.2280380725860596,
-      "eval_runtime": 3.6911,
-      "eval_samples_per_second": 18.964,
-      "eval_steps_per_second": 2.438,
+      "eval_accuracy": 0.6285714285714286,
+      "eval_loss": 0.8886963129043579,
+      "eval_runtime": 3.7412,
+      "eval_samples_per_second": 18.711,
+      "eval_steps_per_second": 2.406,
       "step": 76
     },
     {
       "epoch": 2.027027027027027,
-      "grad_norm": 9.636581420898438,
+      "grad_norm": 10.210707664489746,
       "learning_rate": 2.6691729323308275e-05,
-      "loss": 0.995,
+      "loss": 0.794,
       "step": 80
     },
     {
       "epoch": 2.0945945945945947,
-      "grad_norm": 9.935404777526855,
+      "grad_norm": 11.060843467712402,
       "learning_rate": 2.293233082706767e-05,
-      "loss": 0.8051,
+      "loss": 0.6296,
       "step": 90
     },
     {
       "epoch": 2.1621621621621623,
-      "grad_norm": 9.97338581085205,
+      "grad_norm": 10.571476936340332,
       "learning_rate": 1.9172932330827066e-05,
-      "loss": 0.8465,
+      "loss": 0.5047,
       "step": 100
     },
     {
       "epoch": 2.22972972972973,
-      "grad_norm": 8.2127103805542,
+      "grad_norm": 5.044342041015625,
       "learning_rate": 1.5413533834586467e-05,
-      "loss": 0.6534,
+      "loss": 0.4839,
       "step": 110
     },
     {
       "epoch": 2.2567567567567566,
-      "eval_accuracy": 0.7,
-      "eval_loss": 0.7263602018356323,
-      "eval_runtime": 3.6476,
-      "eval_samples_per_second": 19.191,
-      "eval_steps_per_second": 2.467,
+      "eval_accuracy": 0.7714285714285715,
+      "eval_loss": 0.599827229976654,
+      "eval_runtime": 3.9209,
+      "eval_samples_per_second": 17.853,
+      "eval_steps_per_second": 2.295,
       "step": 114
     },
     {
       "epoch": 3.0405405405405403,
-      "grad_norm": 8.435063362121582,
+      "grad_norm": 8.765264511108398,
       "learning_rate": 1.1654135338345864e-05,
-      "loss": 0.5543,
+      "loss": 0.3181,
       "step": 120
     },
     {
       "epoch": 3.108108108108108,
-      "grad_norm": 3.56019926071167,
+      "grad_norm": 3.8785786628723145,
       "learning_rate": 7.894736842105263e-06,
-      "loss": 0.4172,
+      "loss": 0.2858,
       "step": 130
     },
     {
       "epoch": 3.175675675675676,
-      "grad_norm": 7.430360317230225,
+      "grad_norm": 2.8601646423339844,
       "learning_rate": 4.135338345864662e-06,
-      "loss": 0.3677,
+      "loss": 0.2545,
       "step": 140
     },
     {
       "epoch": 3.22972972972973,
-      "eval_accuracy": 0.8,
-      "eval_loss": 0.552931547164917,
-      "eval_runtime": 3.5081,
-      "eval_samples_per_second": 19.954,
-      "eval_steps_per_second": 2.565,
+      "eval_accuracy": 0.8714285714285714,
+      "eval_loss": 0.40526869893074036,
+      "eval_runtime": 3.6308,
+      "eval_samples_per_second": 19.28,
+      "eval_steps_per_second": 2.479,
       "step": 148
     },
     {
       "epoch": 3.22972972972973,
       "step": 148,
       "total_flos": 1.460491890402263e+18,
-      "train_loss": 1.2348162676836993,
-      "train_runtime": 139.5019,
-      "train_samples_per_second": 8.487,
-      "train_steps_per_second": 1.061
+      "train_loss": 1.0872333371961438,
+      "train_runtime": 198.8948,
+      "train_samples_per_second": 5.953,
+      "train_steps_per_second": 0.744
     },
     {
       "epoch": 3.22972972972973,
-      "eval_accuracy": 0.7935483870967742,
-      "eval_loss": 0.561172604560852,
-      "eval_runtime": 7.6883,
-      "eval_samples_per_second": 20.161,
-      "eval_steps_per_second": 2.601,
-      "step": 148
-    },
-    {
-      "epoch": 3.22972972972973,
-      "eval_accuracy": 0.7935483870967742,
-      "eval_loss": 0.5611749291419983,
-      "eval_runtime": 7.9378,
-      "eval_samples_per_second": 19.527,
-      "eval_steps_per_second": 2.52,
-      "step": 148
-    },
-    {
-      "epoch": 3.22972972972973,
-      "eval_accuracy": 0.7935483870967742,
-      "eval_loss": 0.5611773133277893,
-      "eval_runtime": 7.4937,
-      "eval_samples_per_second": 20.684,
-      "eval_steps_per_second": 2.669,
+      "eval_accuracy": 0.8064516129032258,
+      "eval_loss": 0.5119282603263855,
+      "eval_runtime": 8.1588,
+      "eval_samples_per_second": 18.998,
+      "eval_steps_per_second": 2.451,
       "step": 148
     }
   ],

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ea7cab49c4e93ed45d49ec0933d4da4a7a51594614614d7d065fefc9650ee781
+oid sha256:acc02801356b15f5ed557a809a6daa80fc3b7ab0f45609a901a8689ac4560959
 size 5304