heado

README.md

@@ -3,6 +3,7 @@ license: apache-2.0
 base_model: google/vit-base-patch16-224-in21k
 tags:
 - image-classification
+- ViT
 - generated_from_trainer
 metrics:
 - accuracy
@@ -18,8 +19,8 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on the beans dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.0377
-- Accuracy: 0.9925
+- Loss: 0.1842
+- Accuracy: 0.9531
 
 ## Model description
 
@@ -51,8 +52,8 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch  | Step | Validation Loss | Accuracy |
 |:-------------:|:------:|:----:|:---------------:|:--------:|
-| 0.1354        | 1.5385 | 100  | 0.2087          | 0.9098   |
-| 0.0389        | 3.0769 | 200  | 0.0377          | 0.9925   |
+| 0.0316        | 1.5385 | 100  | 0.0977          | 0.9774   |
+| 0.0106        | 3.0769 | 200  | 0.1979          | 0.9624   |
 
 
 ### Framework versions

all_results.json

@@ -1,13 +1,13 @@
 {
     "epoch": 4.0,
-    "eval_accuracy": 0.9924812030075187,
-    "eval_loss": 0.037701886147260666,
-    "eval_runtime": 3.8301,
-    "eval_samples_per_second": 34.725,
-    "eval_steps_per_second": 4.438,
+    "eval_accuracy": 0.953125,
+    "eval_loss": 0.1842205822467804,
+    "eval_runtime": 1.4634,
+    "eval_samples_per_second": 87.465,
+    "eval_steps_per_second": 10.933,
     "total_flos": 3.205097416476426e+17,
-    "train_loss": 0.11890983564349321,
-    "train_runtime": 97.8669,
-    "train_samples_per_second": 42.261,
-    "train_steps_per_second": 2.657
+    "train_loss": 0.02204044573009014,
+    "train_runtime": 91.5843,
+    "train_samples_per_second": 45.161,
+    "train_steps_per_second": 2.839
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:576384d6255228702fc2b93d17baff22c0691f39524086047fb19b3ad7f106c6
+oid sha256:e5443676f9f75c55203d1d8085f08d7be4c158d489cce154442d42d958a547b8
 size 343227052

runs/Jul04_01-50-15_20f4ee5cd7bf/events.out.tfevents.1720059209.20f4ee5cd7bf.3737.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:55bbc6310d9e39a5bfb276ebb8f5860e978a2fb95c7a2e5df3e5924aae948aea
+size 11323

runs/Jul04_01-50-15_20f4ee5cd7bf/events.out.tfevents.1720059562.20f4ee5cd7bf.3737.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:33a88e3ebf177fc7eefb2d9efff7375ea1dd9aaf0ea3e5e6027998b31a70e722
+size 11417

runs/Jul04_01-50-15_20f4ee5cd7bf/events.out.tfevents.1720059997.20f4ee5cd7bf.3737.2

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bbfd11b9b4d8abe85c452faa5493baea3757dd55b9d8a8039dfa97bd0ef7da44
+size 411

test_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 4.0,
+    "eval_accuracy": 0.953125,
+    "eval_loss": 0.1842205822467804,
+    "eval_runtime": 1.4634,
+    "eval_samples_per_second": 87.465,
+    "eval_steps_per_second": 10.933
+}

train_results.json

@@ -1,8 +1,8 @@
 {
     "epoch": 4.0,
     "total_flos": 3.205097416476426e+17,
-    "train_loss": 0.11890983564349321,
-    "train_runtime": 97.8669,
-    "train_samples_per_second": 42.261,
-    "train_steps_per_second": 2.657
+    "train_loss": 0.02204044573009014,
+    "train_runtime": 91.5843,
+    "train_samples_per_second": 45.161,
+    "train_steps_per_second": 2.839
 }

trainer_state.json

@@ -1,6 +1,6 @@
 {
-  "best_metric": 0.037701886147260666,
-  "best_model_checkpoint": "./vit-base-beans-demo-v5/checkpoint-200",
+  "best_metric": 0.09767168015241623,
+  "best_model_checkpoint": "./vit-base-beans-demo-v5/checkpoint-100",
   "epoch": 4.0,
   "eval_steps": 100,
   "global_step": 260,
@@ -10,212 +10,212 @@
   "log_history": [
     {
       "epoch": 0.15384615384615385,
-      "grad_norm": 1.473002552986145,
+      "grad_norm": 0.0549064539372921,
       "learning_rate": 0.00019230769230769233,
-      "loss": 0.8228,
+      "loss": 0.011,
       "step": 10
     },
     {
       "epoch": 0.3076923076923077,
-      "grad_norm": 1.797827124595642,
+      "grad_norm": 0.053657062351703644,
       "learning_rate": 0.00018461538461538463,
-      "loss": 0.357,
+      "loss": 0.0099,
       "step": 20
     },
     {
       "epoch": 0.46153846153846156,
-      "grad_norm": 2.562692880630493,
+      "grad_norm": 0.04586618021130562,
       "learning_rate": 0.00017692307692307693,
-      "loss": 0.2274,
+      "loss": 0.0085,
       "step": 30
     },
     {
       "epoch": 0.6153846153846154,
-      "grad_norm": 1.5373003482818604,
+      "grad_norm": 0.03834086284041405,
       "learning_rate": 0.00016923076923076923,
-      "loss": 0.1787,
+      "loss": 0.0074,
       "step": 40
     },
     {
       "epoch": 0.7692307692307693,
-      "grad_norm": 2.497523307800293,
+      "grad_norm": 0.0455704927444458,
       "learning_rate": 0.00016153846153846155,
-      "loss": 0.2233,
+      "loss": 0.0081,
       "step": 50
     },
     {
       "epoch": 0.9230769230769231,
-      "grad_norm": 2.7328505516052246,
+      "grad_norm": 4.641096591949463,
       "learning_rate": 0.00015384615384615385,
-      "loss": 0.26,
+      "loss": 0.0547,
       "step": 60
     },
     {
       "epoch": 1.0769230769230769,
-      "grad_norm": 1.830623984336853,
+      "grad_norm": 10.7581205368042,
       "learning_rate": 0.00014615384615384615,
-      "loss": 0.1442,
+      "loss": 0.028,
       "step": 70
     },
     {
       "epoch": 1.2307692307692308,
-      "grad_norm": 0.41552042961120605,
+      "grad_norm": 0.04066062346100807,
       "learning_rate": 0.00013846153846153847,
-      "loss": 0.0919,
+      "loss": 0.0387,
       "step": 80
     },
     {
       "epoch": 1.3846153846153846,
-      "grad_norm": 0.18147416412830353,
+      "grad_norm": 0.042889270931482315,
       "learning_rate": 0.00013076923076923077,
-      "loss": 0.1297,
+      "loss": 0.0808,
       "step": 90
     },
     {
       "epoch": 1.5384615384615383,
-      "grad_norm": 1.4338873624801636,
+      "grad_norm": 0.039594896137714386,
       "learning_rate": 0.0001230769230769231,
-      "loss": 0.1354,
+      "loss": 0.0316,
       "step": 100
     },
     {
       "epoch": 1.5384615384615383,
-      "eval_accuracy": 0.9097744360902256,
-      "eval_loss": 0.20873208343982697,
-      "eval_runtime": 1.9944,
-      "eval_samples_per_second": 66.687,
-      "eval_steps_per_second": 8.524,
+      "eval_accuracy": 0.9774436090225563,
+      "eval_loss": 0.09767168015241623,
+      "eval_runtime": 1.2787,
+      "eval_samples_per_second": 104.015,
+      "eval_steps_per_second": 13.295,
       "step": 100
     },
     {
       "epoch": 1.6923076923076923,
-      "grad_norm": 0.26245740056037903,
+      "grad_norm": 0.4228585660457611,
       "learning_rate": 0.00011538461538461538,
-      "loss": 0.106,
+      "loss": 0.0357,
       "step": 110
     },
     {
       "epoch": 1.8461538461538463,
-      "grad_norm": 0.13711966574192047,
+      "grad_norm": 0.03368207439780235,
       "learning_rate": 0.0001076923076923077,
-      "loss": 0.1161,
+      "loss": 0.0721,
       "step": 120
     },
     {
       "epoch": 2.0,
-      "grad_norm": 0.6398559212684631,
+      "grad_norm": 0.041465017944574356,
       "learning_rate": 0.0001,
-      "loss": 0.0433,
+      "loss": 0.0286,
       "step": 130
     },
     {
       "epoch": 2.1538461538461537,
-      "grad_norm": 0.13836634159088135,
+      "grad_norm": 4.102177619934082,
       "learning_rate": 9.230769230769232e-05,
-      "loss": 0.0412,
+      "loss": 0.0239,
       "step": 140
     },
     {
       "epoch": 2.3076923076923075,
-      "grad_norm": 2.5153419971466064,
+      "grad_norm": 1.726514458656311,
       "learning_rate": 8.461538461538461e-05,
-      "loss": 0.0231,
+      "loss": 0.0085,
       "step": 150
     },
     {
       "epoch": 2.4615384615384617,
-      "grad_norm": 0.08133295178413391,
+      "grad_norm": 0.025889277458190918,
       "learning_rate": 7.692307692307693e-05,
-      "loss": 0.0166,
+      "loss": 0.0052,
       "step": 160
     },
     {
       "epoch": 2.6153846153846154,
-      "grad_norm": 0.07401230931282043,
+      "grad_norm": 0.025063877925276756,
       "learning_rate": 6.923076923076924e-05,
-      "loss": 0.0167,
+      "loss": 0.0046,
       "step": 170
     },
     {
       "epoch": 2.769230769230769,
-      "grad_norm": 0.8673214316368103,
+      "grad_norm": 0.024265997111797333,
       "learning_rate": 6.153846153846155e-05,
-      "loss": 0.0149,
+      "loss": 0.0046,
       "step": 180
     },
     {
       "epoch": 2.9230769230769234,
-      "grad_norm": 0.06832073628902435,
+      "grad_norm": 15.569941520690918,
       "learning_rate": 5.384615384615385e-05,
-      "loss": 0.0337,
+      "loss": 0.0232,
       "step": 190
     },
     {
       "epoch": 3.076923076923077,
-      "grad_norm": 0.06998351216316223,
+      "grad_norm": 0.023741118609905243,
       "learning_rate": 4.615384615384616e-05,
-      "loss": 0.0389,
+      "loss": 0.0106,
       "step": 200
     },
     {
       "epoch": 3.076923076923077,
-      "eval_accuracy": 0.9924812030075187,
-      "eval_loss": 0.037701886147260666,
-      "eval_runtime": 1.2077,
-      "eval_samples_per_second": 110.122,
-      "eval_steps_per_second": 14.076,
+      "eval_accuracy": 0.9624060150375939,
+      "eval_loss": 0.19794000685214996,
+      "eval_runtime": 1.6975,
+      "eval_samples_per_second": 78.351,
+      "eval_steps_per_second": 10.015,
       "step": 200
     },
     {
       "epoch": 3.230769230769231,
-      "grad_norm": 0.061544470489025116,
+      "grad_norm": 0.03866928815841675,
       "learning_rate": 3.846153846153846e-05,
-      "loss": 0.0122,
+      "loss": 0.0377,
       "step": 210
     },
     {
       "epoch": 3.3846153846153846,
-      "grad_norm": 0.06420169025659561,
+      "grad_norm": 0.023310711607336998,
       "learning_rate": 3.0769230769230774e-05,
-      "loss": 0.0125,
+      "loss": 0.0221,
       "step": 220
     },
     {
       "epoch": 3.5384615384615383,
-      "grad_norm": 0.06388475000858307,
+      "grad_norm": 0.02209044247865677,
       "learning_rate": 2.307692307692308e-05,
-      "loss": 0.0117,
+      "loss": 0.0057,
       "step": 230
     },
     {
       "epoch": 3.6923076923076925,
-      "grad_norm": 0.06183198094367981,
+      "grad_norm": 0.020965544506907463,
       "learning_rate": 1.5384615384615387e-05,
-      "loss": 0.0113,
+      "loss": 0.004,
       "step": 240
     },
     {
       "epoch": 3.8461538461538463,
-      "grad_norm": 0.05793595686554909,
+      "grad_norm": 0.02124304138123989,
       "learning_rate": 7.692307692307694e-06,
-      "loss": 0.0114,
+      "loss": 0.0039,
       "step": 250
     },
     {
       "epoch": 4.0,
-      "grad_norm": 0.056076530367136,
+      "grad_norm": 0.0217234306037426,
       "learning_rate": 0.0,
-      "loss": 0.0116,
+      "loss": 0.0039,
       "step": 260
     },
     {
       "epoch": 4.0,
       "step": 260,
       "total_flos": 3.205097416476426e+17,
-      "train_loss": 0.11890983564349321,
-      "train_runtime": 97.8669,
-      "train_samples_per_second": 42.261,
-      "train_steps_per_second": 2.657
+      "train_loss": 0.02204044573009014,
+      "train_runtime": 91.5843,
+      "train_samples_per_second": 45.161,
+      "train_steps_per_second": 2.839
     }
   ],
   "logging_steps": 10,

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5020becdec1b9528f546f5b3874db748b6fb45822f70f644c839d27e503bab49
+oid sha256:44164b4d4762f13debd0722ab6612948375146c034c87f96580824bc81c4f4c8
 size 5112