🍻 cheers

README.md

@@ -2,6 +2,7 @@
 license: apache-2.0
 base_model: google/vit-base-patch16-224
 tags:
+- image-classification
 - generated_from_trainer
 metrics:
 - accuracy
@@ -18,13 +19,13 @@ should probably proofread and complete it, then remove this comment. -->
 
 # vit-epsilon-1e-9
 
-This model is a fine-tuned version of [google/vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224) on an unknown dataset.
+This model is a fine-tuned version of [google/vit-base-patch16-224](https://huggingface.co/google/vit-base-patch16-224) on the skin-cancer dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.6021
-- Accuracy: 0.8627
-- Precision: 0.8567
-- Recall: 0.8627
-- F1: 0.8572
+- Loss: 0.5143
+- Accuracy: 0.8131
+- Precision: 0.8359
+- Recall: 0.8131
+- F1: 0.8206
 
 ## Model description
 

all_results.json

@@ -0,0 +1,16 @@
+{
+    "epoch": 21.0,
+    "eval_accuracy": 0.8131067961165048,
+    "eval_f1": 0.8206122662707227,
+    "eval_loss": 0.5143479704856873,
+    "eval_precision": 0.8358668720104463,
+    "eval_recall": 0.8131067961165048,
+    "eval_runtime": 22.7414,
+    "eval_samples_per_second": 126.817,
+    "eval_steps_per_second": 15.874,
+    "total_flos": 8.345331343683404e+18,
+    "train_loss": 0.6620033840241126,
+    "train_runtime": 2520.6758,
+    "train_samples_per_second": 203.438,
+    "train_steps_per_second": 12.735
+}

eval_results.json

@@ -0,0 +1,11 @@
+{
+    "epoch": 21.0,
+    "eval_accuracy": 0.8131067961165048,
+    "eval_f1": 0.8206122662707227,
+    "eval_loss": 0.5143479704856873,
+    "eval_precision": 0.8358668720104463,
+    "eval_recall": 0.8131067961165048,
+    "eval_runtime": 22.7414,
+    "eval_samples_per_second": 126.817,
+    "eval_steps_per_second": 15.874
+}

runs/Mar22_09-25-07_80a90583e6d6/events.out.tfevents.1711102086.80a90583e6d6.2864.7

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e1df46c5213de5a3cc61e2f5bf2cd507ea848039d103020ed0ad0d661e6e9065
+size 560

train_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 21.0,
+    "total_flos": 8.345331343683404e+18,
+    "train_loss": 0.6620033840241126,
+    "train_runtime": 2520.6758,
+    "train_samples_per_second": 203.438,
+    "train_steps_per_second": 12.735
+}

trainer_state.json

@@ -0,0 +1,429 @@
+{
+  "best_metric": 0.5143479704856873,
+  "best_model_checkpoint": "./vit-epsilon-1e-9/checkpoint-3531",
+  "epoch": 21.0,
+  "eval_steps": 500,
+  "global_step": 6741,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 1.0,
+      "grad_norm": 31.074954986572266,
+      "learning_rate": 1.8291979226774382e-05,
+      "loss": 1.7127,
+      "step": 321
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.6986823855755895,
+      "eval_f1": 0.6331209029522241,
+      "eval_loss": 0.9578500390052795,
+      "eval_precision": 0.6581233216765519,
+      "eval_recall": 0.6986823855755895,
+      "eval_runtime": 22.9329,
+      "eval_samples_per_second": 125.758,
+      "eval_steps_per_second": 15.742,
+      "step": 321
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 32.117530822753906,
+      "learning_rate": 3.675706866705136e-05,
+      "loss": 1.1596,
+      "step": 642
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.7298890429958391,
+      "eval_f1": 0.7299352756165304,
+      "eval_loss": 0.7026094794273376,
+      "eval_precision": 0.7441954806441219,
+      "eval_recall": 0.7298890429958391,
+      "eval_runtime": 22.7527,
+      "eval_samples_per_second": 126.754,
+      "eval_steps_per_second": 15.866,
+      "step": 642
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 7.492339134216309,
+      "learning_rate": 5.52798615118292e-05,
+      "loss": 1.0337,
+      "step": 963
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.7548543689320388,
+      "eval_f1": 0.7356331445906263,
+      "eval_loss": 0.6560060977935791,
+      "eval_precision": 0.7647340906949731,
+      "eval_recall": 0.7548543689320388,
+      "eval_runtime": 22.6575,
+      "eval_samples_per_second": 127.287,
+      "eval_steps_per_second": 15.933,
+      "step": 963
+    },
+    {
+      "epoch": 4.0,
+      "grad_norm": 26.88030433654785,
+      "learning_rate": 7.380265435660705e-05,
+      "loss": 0.9695,
+      "step": 1284
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.7656033287101248,
+      "eval_f1": 0.7758337241261131,
+      "eval_loss": 0.5708493590354919,
+      "eval_precision": 0.7963463250386764,
+      "eval_recall": 0.7656033287101248,
+      "eval_runtime": 23.0619,
+      "eval_samples_per_second": 125.055,
+      "eval_steps_per_second": 15.654,
+      "step": 1284
+    },
+    {
+      "epoch": 5.0,
+      "grad_norm": 5.717092514038086,
+      "learning_rate": 9.232544720138489e-05,
+      "loss": 0.9137,
+      "step": 1605
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.7610957004160888,
+      "eval_f1": 0.7625168248891897,
+      "eval_loss": 0.646024227142334,
+      "eval_precision": 0.7828031789425305,
+      "eval_recall": 0.7610957004160888,
+      "eval_runtime": 22.9345,
+      "eval_samples_per_second": 125.75,
+      "eval_steps_per_second": 15.74,
+      "step": 1605
+    },
+    {
+      "epoch": 6.0,
+      "grad_norm": 8.619205474853516,
+      "learning_rate": 9.984876489938473e-05,
+      "loss": 1.0053,
+      "step": 1926
+    },
+    {
+      "epoch": 6.0,
+      "eval_accuracy": 0.7673370319001387,
+      "eval_f1": 0.7796592423510157,
+      "eval_loss": 0.6020165681838989,
+      "eval_precision": 0.8048902741712624,
+      "eval_recall": 0.7673370319001387,
+      "eval_runtime": 22.5129,
+      "eval_samples_per_second": 128.104,
+      "eval_steps_per_second": 16.035,
+      "step": 1926
+    },
+    {
+      "epoch": 7.0,
+      "grad_norm": 7.978899955749512,
+      "learning_rate": 9.889494151200358e-05,
+      "loss": 0.896,
+      "step": 2247
+    },
+    {
+      "epoch": 7.0,
+      "eval_accuracy": 0.7271151178918169,
+      "eval_f1": 0.747671205921563,
+      "eval_loss": 0.7087484002113342,
+      "eval_precision": 0.8055197232423271,
+      "eval_recall": 0.7271151178918169,
+      "eval_runtime": 22.9629,
+      "eval_samples_per_second": 125.594,
+      "eval_steps_per_second": 15.721,
+      "step": 2247
+    },
+    {
+      "epoch": 8.0,
+      "grad_norm": 3.2379393577575684,
+      "learning_rate": 9.707962612088379e-05,
+      "loss": 0.7646,
+      "step": 2568
+    },
+    {
+      "epoch": 8.0,
+      "eval_accuracy": 0.7441054091539528,
+      "eval_f1": 0.7622121595512239,
+      "eval_loss": 0.6615051627159119,
+      "eval_precision": 0.8133816759254769,
+      "eval_recall": 0.7441054091539528,
+      "eval_runtime": 22.527,
+      "eval_samples_per_second": 128.024,
+      "eval_steps_per_second": 16.025,
+      "step": 2568
+    },
+    {
+      "epoch": 9.0,
+      "grad_norm": 1.0829046964645386,
+      "learning_rate": 9.443480321450928e-05,
+      "loss": 0.7262,
+      "step": 2889
+    },
+    {
+      "epoch": 9.0,
+      "eval_accuracy": 0.79750346740638,
+      "eval_f1": 0.7985217217665854,
+      "eval_loss": 0.5611338019371033,
+      "eval_precision": 0.8189863473770944,
+      "eval_recall": 0.79750346740638,
+      "eval_runtime": 22.5299,
+      "eval_samples_per_second": 128.008,
+      "eval_steps_per_second": 16.023,
+      "step": 2889
+    },
+    {
+      "epoch": 10.0,
+      "grad_norm": 12.064545631408691,
+      "learning_rate": 9.100707257835249e-05,
+      "loss": 0.7025,
+      "step": 3210
+    },
+    {
+      "epoch": 10.0,
+      "eval_accuracy": 0.79750346740638,
+      "eval_f1": 0.805841955667511,
+      "eval_loss": 0.5337752103805542,
+      "eval_precision": 0.8277828822554938,
+      "eval_recall": 0.79750346740638,
+      "eval_runtime": 22.769,
+      "eval_samples_per_second": 126.663,
+      "eval_steps_per_second": 15.855,
+      "step": 3210
+    },
+    {
+      "epoch": 11.0,
+      "grad_norm": 22.526174545288086,
+      "learning_rate": 8.685682824178951e-05,
+      "loss": 0.6138,
+      "step": 3531
+    },
+    {
+      "epoch": 11.0,
+      "eval_accuracy": 0.8131067961165048,
+      "eval_f1": 0.8206122662707227,
+      "eval_loss": 0.5143479704856873,
+      "eval_precision": 0.8358668720104463,
+      "eval_recall": 0.8131067961165048,
+      "eval_runtime": 22.3689,
+      "eval_samples_per_second": 128.929,
+      "eval_steps_per_second": 16.138,
+      "step": 3531
+    },
+    {
+      "epoch": 12.0,
+      "grad_norm": 4.508119583129883,
+      "learning_rate": 8.205719438083829e-05,
+      "loss": 0.5582,
+      "step": 3852
+    },
+    {
+      "epoch": 12.0,
+      "eval_accuracy": 0.7864077669902912,
+      "eval_f1": 0.7992483326872344,
+      "eval_loss": 0.6157103180885315,
+      "eval_precision": 0.8253081034753366,
+      "eval_recall": 0.7864077669902912,
+      "eval_runtime": 22.2433,
+      "eval_samples_per_second": 129.657,
+      "eval_steps_per_second": 16.23,
+      "step": 3852
+    },
+    {
+      "epoch": 13.0,
+      "grad_norm": 37.043174743652344,
+      "learning_rate": 7.669273692531118e-05,
+      "loss": 0.4736,
+      "step": 4173
+    },
+    {
+      "epoch": 13.0,
+      "eval_accuracy": 0.8117198335644937,
+      "eval_f1": 0.8007107317782642,
+      "eval_loss": 0.6899198293685913,
+      "eval_precision": 0.8252732818495465,
+      "eval_recall": 0.8117198335644937,
+      "eval_runtime": 23.2981,
+      "eval_samples_per_second": 123.787,
+      "eval_steps_per_second": 15.495,
+      "step": 4173
+    },
+    {
+      "epoch": 14.0,
+      "grad_norm": 2.777528762817383,
+      "learning_rate": 7.085797357089247e-05,
+      "loss": 0.4581,
+      "step": 4494
+    },
+    {
+      "epoch": 14.0,
+      "eval_accuracy": 0.812760055478502,
+      "eval_f1": 0.8198980869707609,
+      "eval_loss": 0.6062142252922058,
+      "eval_precision": 0.8446721645536698,
+      "eval_recall": 0.812760055478502,
+      "eval_runtime": 22.8029,
+      "eval_samples_per_second": 126.475,
+      "eval_steps_per_second": 15.831,
+      "step": 4494
+    },
+    {
+      "epoch": 15.0,
+      "grad_norm": 1.7661919593811035,
+      "learning_rate": 6.46557084486047e-05,
+      "loss": 0.407,
+      "step": 4815
+    },
+    {
+      "epoch": 15.0,
+      "eval_accuracy": 0.8307905686546463,
+      "eval_f1": 0.8338660350277718,
+      "eval_loss": 0.531688928604126,
+      "eval_precision": 0.8410181561922312,
+      "eval_recall": 0.8307905686546463,
+      "eval_runtime": 22.886,
+      "eval_samples_per_second": 126.016,
+      "eval_steps_per_second": 15.774,
+      "step": 4815
+    },
+    {
+      "epoch": 16.0,
+      "grad_norm": 45.85359573364258,
+      "learning_rate": 5.8195220793532045e-05,
+      "loss": 0.369,
+      "step": 5136
+    },
+    {
+      "epoch": 16.0,
+      "eval_accuracy": 0.8196948682385575,
+      "eval_f1": 0.8270482179777584,
+      "eval_loss": 0.6474519371986389,
+      "eval_precision": 0.8414288606484934,
+      "eval_recall": 0.8196948682385575,
+      "eval_runtime": 22.3448,
+      "eval_samples_per_second": 129.068,
+      "eval_steps_per_second": 16.156,
+      "step": 5136
+    },
+    {
+      "epoch": 17.0,
+      "grad_norm": 32.817665100097656,
+      "learning_rate": 5.15903395270923e-05,
+      "loss": 0.2855,
+      "step": 5457
+    },
+    {
+      "epoch": 17.0,
+      "eval_accuracy": 0.8616504854368932,
+      "eval_f1": 0.8578019708384981,
+      "eval_loss": 0.5152972340583801,
+      "eval_precision": 0.85715321702538,
+      "eval_recall": 0.8616504854368932,
+      "eval_runtime": 22.8719,
+      "eval_samples_per_second": 126.094,
+      "eval_steps_per_second": 15.784,
+      "step": 5457
+    },
+    {
+      "epoch": 18.0,
+      "grad_norm": 13.112251281738281,
+      "learning_rate": 4.495743767726598e-05,
+      "loss": 0.2545,
+      "step": 5778
+    },
+    {
+      "epoch": 18.0,
+      "eval_accuracy": 0.8436199722607489,
+      "eval_f1": 0.847268091121353,
+      "eval_loss": 0.5454699993133545,
+      "eval_precision": 0.8555190582823499,
+      "eval_recall": 0.8436199722607489,
+      "eval_runtime": 22.9113,
+      "eval_samples_per_second": 125.877,
+      "eval_steps_per_second": 15.756,
+      "step": 5778
+    },
+    {
+      "epoch": 19.0,
+      "grad_norm": 11.9920015335083,
+      "learning_rate": 3.843351058509752e-05,
+      "loss": 0.2221,
+      "step": 6099
+    },
+    {
+      "epoch": 19.0,
+      "eval_accuracy": 0.8470873786407767,
+      "eval_f1": 0.8516125159546437,
+      "eval_loss": 0.5955303311347961,
+      "eval_precision": 0.8624218059544237,
+      "eval_recall": 0.8470873786407767,
+      "eval_runtime": 22.6514,
+      "eval_samples_per_second": 127.321,
+      "eval_steps_per_second": 15.937,
+      "step": 6099
+    },
+    {
+      "epoch": 20.0,
+      "grad_norm": 0.16968537867069244,
+      "learning_rate": 3.209279054309716e-05,
+      "loss": 0.2143,
+      "step": 6420
+    },
+    {
+      "epoch": 20.0,
+      "eval_accuracy": 0.8574895977808599,
+      "eval_f1": 0.8581580614443969,
+      "eval_loss": 0.5771576762199402,
+      "eval_precision": 0.86041215849512,
+      "eval_recall": 0.8574895977808599,
+      "eval_runtime": 22.9384,
+      "eval_samples_per_second": 125.728,
+      "eval_steps_per_second": 15.738,
+      "step": 6420
+    },
+    {
+      "epoch": 21.0,
+      "grad_norm": 4.404163360595703,
+      "learning_rate": 2.6067582030088143e-05,
+      "loss": 0.1619,
+      "step": 6741
+    },
+    {
+      "epoch": 21.0,
+      "eval_accuracy": 0.8626907073509015,
+      "eval_f1": 0.8572038210510016,
+      "eval_loss": 0.6020533442497253,
+      "eval_precision": 0.8566861930842659,
+      "eval_recall": 0.8626907073509015,
+      "eval_runtime": 23.4679,
+      "eval_samples_per_second": 122.891,
+      "eval_steps_per_second": 15.383,
+      "step": 6741
+    },
+    {
+      "epoch": 21.0,
+      "step": 6741,
+      "total_flos": 8.345331343683404e+18,
+      "train_loss": 0.6620033840241126,
+      "train_runtime": 2520.6758,
+      "train_samples_per_second": 203.438,
+      "train_steps_per_second": 12.735
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 32100,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 100,
+  "save_steps": 500,
+  "total_flos": 8.345331343683404e+18,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": null
+}