End of training

README.md

@@ -2,6 +2,8 @@
 license: apache-2.0
 base_model: google/vit-base-patch16-224-in21k
 tags:
+- image-classification
+- vision
 - generated_from_trainer
 metrics:
 - accuracy
@@ -15,7 +17,7 @@ should probably proofread and complete it, then remove this comment. -->
 
 # vit-base-beans
 
-This model is a fine-tuned version of [google/vit-base-patch16-224-in21k](https://huggingface.co/google/vit-base-patch16-224-in21k) on an unknown dataset.
+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: 0.5648
 - Accuracy: 0.9699

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 7.0,
+    "epoch": 3.0,
     "eval_accuracy": 0.9699248120300752,
-    "eval_loss": 0.22669997811317444,
-    "eval_runtime": 5.4349,
-    "eval_samples_per_second": 24.472,
-    "eval_steps_per_second": 0.552,
-    "total_flos": 5.608920478833746e+17,
-    "train_loss": 0.4668571989075476,
-    "train_runtime": 403.193,
-    "train_samples_per_second": 17.952,
-    "train_steps_per_second": 0.295
+    "eval_loss": 0.5648306012153625,
+    "eval_runtime": 5.8619,
+    "eval_samples_per_second": 22.689,
+    "eval_steps_per_second": 0.512,
+    "total_flos": 2.4038230623573197e+17,
+    "train_loss": 0.7542892180237115,
+    "train_runtime": 180.8324,
+    "train_samples_per_second": 17.154,
+    "train_steps_per_second": 0.282
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 7.0,
+    "epoch": 3.0,
     "eval_accuracy": 0.9699248120300752,
-    "eval_loss": 0.22669997811317444,
-    "eval_runtime": 5.4349,
-    "eval_samples_per_second": 24.472,
-    "eval_steps_per_second": 0.552
+    "eval_loss": 0.5648306012153625,
+    "eval_runtime": 5.8619,
+    "eval_samples_per_second": 22.689,
+    "eval_steps_per_second": 0.512
 }

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 7.0,
-    "total_flos": 5.608920478833746e+17,
-    "train_loss": 0.4668571989075476,
-    "train_runtime": 403.193,
-    "train_samples_per_second": 17.952,
-    "train_steps_per_second": 0.295
+    "epoch": 3.0,
+    "total_flos": 2.4038230623573197e+17,
+    "train_loss": 0.7542892180237115,
+    "train_runtime": 180.8324,
+    "train_samples_per_second": 17.154,
+    "train_steps_per_second": 0.282
 }

trainer_state.json

@@ -1,167 +1,89 @@
 {
-  "best_metric": 0.22669997811317444,
-  "best_model_checkpoint": "./beans_outputs/checkpoint-119",
-  "epoch": 7.0,
+  "best_metric": 0.5648306012153625,
+  "best_model_checkpoint": "./beans_outputs/checkpoint-51",
+  "epoch": 3.0,
   "eval_steps": 500,
-  "global_step": 119,
+  "global_step": 51,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
     {
       "epoch": 0.5882352941176471,
-      "grad_norm": 1.3734956979751587,
-      "learning_rate": 1.831932773109244e-05,
-      "loss": 0.9859,
+      "grad_norm": 1.3434114456176758,
+      "learning_rate": 1.607843137254902e-05,
+      "loss": 0.9886,
       "step": 10
     },
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.9323308270676691,
-      "eval_loss": 0.7495129704475403,
-      "eval_runtime": 5.9299,
-      "eval_samples_per_second": 22.429,
-      "eval_steps_per_second": 0.506,
+      "eval_accuracy": 0.924812030075188,
+      "eval_loss": 0.7761334776878357,
+      "eval_runtime": 5.4724,
+      "eval_samples_per_second": 24.304,
+      "eval_steps_per_second": 0.548,
       "step": 17
     },
     {
       "epoch": 1.1764705882352942,
-      "grad_norm": 1.1843513250350952,
-      "learning_rate": 1.6638655462184877e-05,
-      "loss": 0.808,
+      "grad_norm": 1.1961711645126343,
+      "learning_rate": 1.215686274509804e-05,
+      "loss": 0.8267,
       "step": 20
     },
     {
       "epoch": 1.7647058823529411,
-      "grad_norm": 1.209150791168213,
-      "learning_rate": 1.4957983193277313e-05,
-      "loss": 0.6767,
+      "grad_norm": 1.177196979522705,
+      "learning_rate": 8.23529411764706e-06,
+      "loss": 0.7227,
       "step": 30
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.9624060150375939,
-      "eval_loss": 0.5292961001396179,
-      "eval_runtime": 5.7583,
-      "eval_samples_per_second": 23.097,
-      "eval_steps_per_second": 0.521,
+      "eval_accuracy": 0.9548872180451128,
+      "eval_loss": 0.6194689869880676,
+      "eval_runtime": 5.9342,
+      "eval_samples_per_second": 22.413,
+      "eval_steps_per_second": 0.506,
       "step": 34
     },
     {
       "epoch": 2.3529411764705883,
-      "grad_norm": 1.226790428161621,
-      "learning_rate": 1.3277310924369749e-05,
-      "loss": 0.5504,
+      "grad_norm": 1.2419898509979248,
+      "learning_rate": 4.313725490196079e-06,
+      "loss": 0.6422,
       "step": 40
     },
     {
       "epoch": 2.9411764705882355,
-      "grad_norm": 1.0332392454147339,
-      "learning_rate": 1.1596638655462186e-05,
-      "loss": 0.4611,
+      "grad_norm": 1.1358920335769653,
+      "learning_rate": 3.921568627450981e-07,
+      "loss": 0.6041,
       "step": 50
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.9624060150375939,
-      "eval_loss": 0.3726067841053009,
-      "eval_runtime": 5.9181,
-      "eval_samples_per_second": 22.473,
-      "eval_steps_per_second": 0.507,
-      "step": 51
-    },
-    {
-      "epoch": 3.5294117647058822,
-      "grad_norm": 1.1689759492874146,
-      "learning_rate": 9.915966386554622e-06,
-      "loss": 0.4046,
-      "step": 60
-    },
-    {
-      "epoch": 4.0,
       "eval_accuracy": 0.9699248120300752,
-      "eval_loss": 0.2977140247821808,
-      "eval_runtime": 5.7869,
-      "eval_samples_per_second": 22.983,
-      "eval_steps_per_second": 0.518,
-      "step": 68
-    },
-    {
-      "epoch": 4.117647058823529,
-      "grad_norm": 1.0469905138015747,
-      "learning_rate": 8.23529411764706e-06,
-      "loss": 0.3266,
-      "step": 70
-    },
-    {
-      "epoch": 4.705882352941177,
-      "grad_norm": 1.3861141204833984,
-      "learning_rate": 6.5546218487394966e-06,
-      "loss": 0.3173,
-      "step": 80
-    },
-    {
-      "epoch": 5.0,
-      "eval_accuracy": 0.9699248120300752,
-      "eval_loss": 0.2532085180282593,
-      "eval_runtime": 5.8905,
-      "eval_samples_per_second": 22.579,
-      "eval_steps_per_second": 0.509,
-      "step": 85
-    },
-    {
-      "epoch": 5.294117647058823,
-      "grad_norm": 1.1056371927261353,
-      "learning_rate": 4.873949579831933e-06,
-      "loss": 0.2773,
-      "step": 90
-    },
-    {
-      "epoch": 5.882352941176471,
-      "grad_norm": 0.9023756980895996,
-      "learning_rate": 3.1932773109243696e-06,
-      "loss": 0.2547,
-      "step": 100
-    },
-    {
-      "epoch": 6.0,
-      "eval_accuracy": 0.9774436090225563,
-      "eval_loss": 0.22800950706005096,
-      "eval_runtime": 5.8041,
-      "eval_samples_per_second": 22.915,
-      "eval_steps_per_second": 0.517,
-      "step": 102
-    },
-    {
-      "epoch": 6.470588235294118,
-      "grad_norm": 1.1000926494598389,
-      "learning_rate": 1.5126050420168068e-06,
-      "loss": 0.2629,
-      "step": 110
-    },
-    {
-      "epoch": 7.0,
-      "eval_accuracy": 0.9699248120300752,
-      "eval_loss": 0.22669997811317444,
-      "eval_runtime": 6.0113,
-      "eval_samples_per_second": 22.125,
-      "eval_steps_per_second": 0.499,
-      "step": 119
+      "eval_loss": 0.5648306012153625,
+      "eval_runtime": 5.9544,
+      "eval_samples_per_second": 22.336,
+      "eval_steps_per_second": 0.504,
+      "step": 51
     },
     {
-      "epoch": 7.0,
-      "step": 119,
-      "total_flos": 5.608920478833746e+17,
-      "train_loss": 0.4668571989075476,
-      "train_runtime": 403.193,
-      "train_samples_per_second": 17.952,
-      "train_steps_per_second": 0.295
+      "epoch": 3.0,
+      "step": 51,
+      "total_flos": 2.4038230623573197e+17,
+      "train_loss": 0.7542892180237115,
+      "train_runtime": 180.8324,
+      "train_samples_per_second": 17.154,
+      "train_steps_per_second": 0.282
     }
   ],
   "logging_steps": 10,
-  "max_steps": 119,
+  "max_steps": 51,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 7,
+  "num_train_epochs": 3,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -175,7 +97,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 5.608920478833746e+17,
+  "total_flos": 2.4038230623573197e+17,
   "train_batch_size": 64,
   "trial_name": null,
   "trial_params": null