End of training

README.md

@@ -23,11 +23,11 @@ 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 medmnist-v2 dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.2690
-- Accuracy: 0.8846
-- Precision: 0.9018
-- Recall: 0.8008
-- F1: 0.8342
+- Loss: 0.3129
+- Accuracy: 0.8782
+- Precision: 0.8971
+- Recall: 0.7888
+- F1: 0.8232
 
 ## Model description
 

all_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 9.142857142857142,
+    "eval_accuracy": 0.8782051282051282,
+    "eval_f1": 0.8232241903739489,
+    "eval_loss": 0.3128599524497986,
+    "eval_precision": 0.8970714900947458,
+    "eval_recall": 0.7888471177944862,
+    "eval_runtime": 0.8256,
+    "eval_samples_per_second": 188.95,
+    "eval_steps_per_second": 12.112,
+    "total_flos": 3.8966228771394355e+17,
+    "train_loss": 0.41416080892086027,
+    "train_runtime": 62.0133,
+    "train_samples_per_second": 88.046,
+    "train_steps_per_second": 1.29
+}

eval_results.json

@@ -0,0 +1,11 @@
+{
+    "epoch": 9.142857142857142,
+    "eval_accuracy": 0.8782051282051282,
+    "eval_f1": 0.8232241903739489,
+    "eval_loss": 0.3128599524497986,
+    "eval_precision": 0.8970714900947458,
+    "eval_recall": 0.7888471177944862,
+    "eval_runtime": 0.8256,
+    "eval_samples_per_second": 188.95,
+    "eval_steps_per_second": 12.112
+}

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 9.142857142857142,
+    "total_flos": 3.8966228771394355e+17,
+    "train_loss": 0.41416080892086027,
+    "train_runtime": 62.0133,
+    "train_samples_per_second": 88.046,
+    "train_steps_per_second": 1.29
+}

trainer_state.json

@@ -0,0 +1,218 @@
+{
+  "best_metric": 0.8846153846153846,
+  "best_model_checkpoint": "vit-base-patch16-224-in21k-finetuned-lora-medmnistv2/checkpoint-61",
+  "epoch": 9.142857142857142,
+  "eval_steps": 500,
+  "global_step": 80,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.9142857142857143,
+      "eval_accuracy": 0.782051282051282,
+      "eval_f1": 0.595114503816794,
+      "eval_loss": 0.4751149117946625,
+      "eval_precision": 0.8851351351351351,
+      "eval_recall": 0.5952380952380952,
+      "eval_runtime": 0.4075,
+      "eval_samples_per_second": 191.417,
+      "eval_steps_per_second": 12.27,
+      "step": 8
+    },
+    {
+      "epoch": 1.1428571428571428,
+      "grad_norm": 0.6409297585487366,
+      "learning_rate": 0.004375,
+      "loss": 0.5516,
+      "step": 10
+    },
+    {
+      "epoch": 1.9428571428571428,
+      "eval_accuracy": 0.8461538461538461,
+      "eval_f1": 0.7982758620689654,
+      "eval_loss": 0.4165884256362915,
+      "eval_precision": 0.8090990187332738,
+      "eval_recall": 0.7894736842105263,
+      "eval_runtime": 0.3611,
+      "eval_samples_per_second": 215.989,
+      "eval_steps_per_second": 13.845,
+      "step": 17
+    },
+    {
+      "epoch": 2.2857142857142856,
+      "grad_norm": 0.398532509803772,
+      "learning_rate": 0.00375,
+      "loss": 0.478,
+      "step": 20
+    },
+    {
+      "epoch": 2.9714285714285715,
+      "eval_accuracy": 0.8205128205128205,
+      "eval_f1": 0.7564674397859055,
+      "eval_loss": 0.3675924241542816,
+      "eval_precision": 0.7791706846673095,
+      "eval_recall": 0.7418546365914787,
+      "eval_runtime": 0.3868,
+      "eval_samples_per_second": 201.635,
+      "eval_steps_per_second": 12.925,
+      "step": 26
+    },
+    {
+      "epoch": 3.4285714285714284,
+      "grad_norm": 0.3681485056877136,
+      "learning_rate": 0.003125,
+      "loss": 0.4617,
+      "step": 30
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.8717948717948718,
+      "eval_f1": 0.8194444444444444,
+      "eval_loss": 0.31802448630332947,
+      "eval_precision": 0.8698412698412699,
+      "eval_recall": 0.7919799498746867,
+      "eval_runtime": 0.3831,
+      "eval_samples_per_second": 203.577,
+      "eval_steps_per_second": 13.05,
+      "step": 35
+    },
+    {
+      "epoch": 4.571428571428571,
+      "grad_norm": 0.6881831288337708,
+      "learning_rate": 0.0025,
+      "loss": 0.4208,
+      "step": 40
+    },
+    {
+      "epoch": 4.914285714285715,
+      "eval_accuracy": 0.8589743589743589,
+      "eval_f1": 0.8325200078079251,
+      "eval_loss": 0.4562018811702728,
+      "eval_precision": 0.8173076923076923,
+      "eval_recall": 0.8583959899749374,
+      "eval_runtime": 0.382,
+      "eval_samples_per_second": 204.197,
+      "eval_steps_per_second": 13.09,
+      "step": 43
+    },
+    {
+      "epoch": 5.714285714285714,
+      "grad_norm": 0.34260129928588867,
+      "learning_rate": 0.001875,
+      "loss": 0.3759,
+      "step": 50
+    },
+    {
+      "epoch": 5.942857142857143,
+      "eval_accuracy": 0.8717948717948718,
+      "eval_f1": 0.8417207792207791,
+      "eval_loss": 0.3780059218406677,
+      "eval_precision": 0.833201581027668,
+      "eval_recall": 0.8521303258145363,
+      "eval_runtime": 0.3775,
+      "eval_samples_per_second": 206.647,
+      "eval_steps_per_second": 13.247,
+      "step": 52
+    },
+    {
+      "epoch": 6.857142857142857,
+      "grad_norm": 0.5428153872489929,
+      "learning_rate": 0.00125,
+      "loss": 0.3689,
+      "step": 60
+    },
+    {
+      "epoch": 6.9714285714285715,
+      "eval_accuracy": 0.8846153846153846,
+      "eval_f1": 0.8342384887839434,
+      "eval_loss": 0.29926371574401855,
+      "eval_precision": 0.9017857142857143,
+      "eval_recall": 0.8007518796992481,
+      "eval_runtime": 0.4642,
+      "eval_samples_per_second": 168.033,
+      "eval_steps_per_second": 10.771,
+      "step": 61
+    },
+    {
+      "epoch": 8.0,
+      "grad_norm": 0.7398069500923157,
+      "learning_rate": 0.000625,
+      "loss": 0.3322,
+      "step": 70
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.8717948717948718,
+      "eval_f1": 0.8194444444444444,
+      "eval_loss": 0.2784874737262726,
+      "eval_precision": 0.8698412698412699,
+      "eval_recall": 0.7919799498746867,
+      "eval_runtime": 0.3832,
+      "eval_samples_per_second": 203.524,
+      "eval_steps_per_second": 13.046,
+      "step": 70
+    },
+    {
+      "epoch": 8.914285714285715,
+      "eval_accuracy": 0.8846153846153846,
+      "eval_f1": 0.8342384887839434,
+      "eval_loss": 0.2699558734893799,
+      "eval_precision": 0.9017857142857143,
+      "eval_recall": 0.8007518796992481,
+      "eval_runtime": 0.456,
+      "eval_samples_per_second": 171.058,
+      "eval_steps_per_second": 10.965,
+      "step": 78
+    },
+    {
+      "epoch": 9.142857142857142,
+      "grad_norm": 0.5633471012115479,
+      "learning_rate": 0.0,
+      "loss": 0.3242,
+      "step": 80
+    },
+    {
+      "epoch": 9.142857142857142,
+      "eval_accuracy": 0.8846153846153846,
+      "eval_f1": 0.8342384887839434,
+      "eval_loss": 0.2689835727214813,
+      "eval_precision": 0.9017857142857143,
+      "eval_recall": 0.8007518796992481,
+      "eval_runtime": 0.4354,
+      "eval_samples_per_second": 179.156,
+      "eval_steps_per_second": 11.484,
+      "step": 80
+    },
+    {
+      "epoch": 9.142857142857142,
+      "step": 80,
+      "total_flos": 3.8966228771394355e+17,
+      "train_loss": 0.41416080892086027,
+      "train_runtime": 62.0133,
+      "train_samples_per_second": 88.046,
+      "train_steps_per_second": 1.29
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 80,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 3.8966228771394355e+17,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}