End of training

README.md

@@ -2,6 +2,8 @@
 license: apache-2.0
 base_model: google/vit-large-patch16-224-in21k
 tags:
+- image-classification
+- vision
 - generated_from_trainer
 datasets:
 - imagefolder
@@ -14,7 +16,7 @@ model-index:
       name: Image Classification
       type: image-classification
     dataset:
-      name: imagefolder
+      name: touchtech/fashion-images-perspectives
       type: imagefolder
       config: default
       split: train
@@ -22,7 +24,7 @@ model-index:
     metrics:
     - name: Accuracy
       type: accuracy
-      value: 0.9356252884171666
+      value: 0.92455006922012
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -30,10 +32,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # fashion-images-perspectives-vit-large-patch16-224-in21k-v3
 
-This model is a fine-tuned version of [google/vit-large-patch16-224-in21k](https://huggingface.co/google/vit-large-patch16-224-in21k) on the imagefolder dataset.
+This model is a fine-tuned version of [google/vit-large-patch16-224-in21k](https://huggingface.co/google/vit-large-patch16-224-in21k) on the touchtech/fashion-images-perspectives dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.2444
-- Accuracy: 0.9356
+- Loss: 0.2419
+- Accuracy: 0.9246
 
 ## Model description
 

all_results.json

@@ -0,0 +1,12 @@
+{
+    "epoch": 5.0,
+    "eval_accuracy": 0.92455006922012,
+    "eval_loss": 0.24191519618034363,
+    "eval_runtime": 317.7384,
+    "eval_samples_per_second": 13.64,
+    "eval_steps_per_second": 1.706,
+    "train_loss": 0.3319820684874097,
+    "train_runtime": 13677.2564,
+    "train_samples_per_second": 8.977,
+    "train_steps_per_second": 1.122
+}

eval_results.json

@@ -0,0 +1,8 @@
+{
+    "epoch": 5.0,
+    "eval_accuracy": 0.92455006922012,
+    "eval_loss": 0.24191519618034363,
+    "eval_runtime": 317.7384,
+    "eval_samples_per_second": 13.64,
+    "eval_steps_per_second": 1.706
+}

train_results.json

@@ -0,0 +1,7 @@
+{
+    "epoch": 5.0,
+    "train_loss": 0.3319820684874097,
+    "train_runtime": 13677.2564,
+    "train_samples_per_second": 8.977,
+    "train_steps_per_second": 1.122
+}

trainer_state.json

@@ -0,0 +1,253 @@
+{
+  "best_metric": 0.24191519618034363,
+  "best_model_checkpoint": "/workspace/training_output/perspectives-vit-large-patch16-224-in21k-v3/checkpoint-9210",
+  "epoch": 5.0,
+  "eval_steps": 500,
+  "global_step": 15350,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.16,
+      "learning_rate": 1.93485342019544e-05,
+      "loss": 0.9017,
+      "step": 500
+    },
+    {
+      "epoch": 0.33,
+      "learning_rate": 1.8697068403908796e-05,
+      "loss": 0.5698,
+      "step": 1000
+    },
+    {
+      "epoch": 0.49,
+      "learning_rate": 1.8045602605863193e-05,
+      "loss": 0.4974,
+      "step": 1500
+    },
+    {
+      "epoch": 0.65,
+      "learning_rate": 1.7394136807817594e-05,
+      "loss": 0.4646,
+      "step": 2000
+    },
+    {
+      "epoch": 0.81,
+      "learning_rate": 1.6742671009771988e-05,
+      "loss": 0.4538,
+      "step": 2500
+    },
+    {
+      "epoch": 0.98,
+      "learning_rate": 1.6091205211726385e-05,
+      "loss": 0.4353,
+      "step": 3000
+    },
+    {
+      "epoch": 1.0,
+      "eval_accuracy": 0.8811721273650208,
+      "eval_loss": 0.33953866362571716,
+      "eval_runtime": 349.0715,
+      "eval_samples_per_second": 12.416,
+      "eval_steps_per_second": 1.553,
+      "step": 3070
+    },
+    {
+      "epoch": 1.14,
+      "learning_rate": 1.5439739413680782e-05,
+      "loss": 0.3613,
+      "step": 3500
+    },
+    {
+      "epoch": 1.3,
+      "learning_rate": 1.478827361563518e-05,
+      "loss": 0.3571,
+      "step": 4000
+    },
+    {
+      "epoch": 1.47,
+      "learning_rate": 1.4136807817589578e-05,
+      "loss": 0.3861,
+      "step": 4500
+    },
+    {
+      "epoch": 1.63,
+      "learning_rate": 1.3485342019543975e-05,
+      "loss": 0.3657,
+      "step": 5000
+    },
+    {
+      "epoch": 1.79,
+      "learning_rate": 1.2833876221498372e-05,
+      "loss": 0.3615,
+      "step": 5500
+    },
+    {
+      "epoch": 1.95,
+      "learning_rate": 1.2182410423452771e-05,
+      "loss": 0.3415,
+      "step": 6000
+    },
+    {
+      "epoch": 2.0,
+      "eval_accuracy": 0.9192431933548685,
+      "eval_loss": 0.25435712933540344,
+      "eval_runtime": 357.646,
+      "eval_samples_per_second": 12.118,
+      "eval_steps_per_second": 1.515,
+      "step": 6140
+    },
+    {
+      "epoch": 2.12,
+      "learning_rate": 1.1530944625407168e-05,
+      "loss": 0.3178,
+      "step": 6500
+    },
+    {
+      "epoch": 2.28,
+      "learning_rate": 1.0879478827361565e-05,
+      "loss": 0.2954,
+      "step": 7000
+    },
+    {
+      "epoch": 2.44,
+      "learning_rate": 1.0228013029315963e-05,
+      "loss": 0.3127,
+      "step": 7500
+    },
+    {
+      "epoch": 2.61,
+      "learning_rate": 9.576547231270358e-06,
+      "loss": 0.3116,
+      "step": 8000
+    },
+    {
+      "epoch": 2.77,
+      "learning_rate": 8.925081433224755e-06,
+      "loss": 0.2747,
+      "step": 8500
+    },
+    {
+      "epoch": 2.93,
+      "learning_rate": 8.273615635179154e-06,
+      "loss": 0.2689,
+      "step": 9000
+    },
+    {
+      "epoch": 3.0,
+      "eval_accuracy": 0.92455006922012,
+      "eval_loss": 0.24191519618034363,
+      "eval_runtime": 319.492,
+      "eval_samples_per_second": 13.565,
+      "eval_steps_per_second": 1.696,
+      "step": 9210
+    },
+    {
+      "epoch": 3.09,
+      "learning_rate": 7.622149837133551e-06,
+      "loss": 0.2727,
+      "step": 9500
+    },
+    {
+      "epoch": 3.26,
+      "learning_rate": 6.9706840390879485e-06,
+      "loss": 0.2362,
+      "step": 10000
+    },
+    {
+      "epoch": 3.42,
+      "learning_rate": 6.319218241042345e-06,
+      "loss": 0.2617,
+      "step": 10500
+    },
+    {
+      "epoch": 3.58,
+      "learning_rate": 5.667752442996744e-06,
+      "loss": 0.2437,
+      "step": 11000
+    },
+    {
+      "epoch": 3.75,
+      "learning_rate": 5.016286644951141e-06,
+      "loss": 0.2494,
+      "step": 11500
+    },
+    {
+      "epoch": 3.91,
+      "learning_rate": 4.364820846905538e-06,
+      "loss": 0.2525,
+      "step": 12000
+    },
+    {
+      "epoch": 4.0,
+      "eval_accuracy": 0.9192431933548685,
+      "eval_loss": 0.29525890946388245,
+      "eval_runtime": 322.9797,
+      "eval_samples_per_second": 13.419,
+      "eval_steps_per_second": 1.678,
+      "step": 12280
+    },
+    {
+      "epoch": 4.07,
+      "learning_rate": 3.7133550488599353e-06,
+      "loss": 0.2265,
+      "step": 12500
+    },
+    {
+      "epoch": 4.23,
+      "learning_rate": 3.061889250814333e-06,
+      "loss": 0.204,
+      "step": 13000
+    },
+    {
+      "epoch": 4.4,
+      "learning_rate": 2.4104234527687296e-06,
+      "loss": 0.2261,
+      "step": 13500
+    },
+    {
+      "epoch": 4.56,
+      "learning_rate": 1.7589576547231272e-06,
+      "loss": 0.2152,
+      "step": 14000
+    },
+    {
+      "epoch": 4.72,
+      "learning_rate": 1.1074918566775244e-06,
+      "loss": 0.1766,
+      "step": 14500
+    },
+    {
+      "epoch": 4.89,
+      "learning_rate": 4.5602605863192187e-07,
+      "loss": 0.1977,
+      "step": 15000
+    },
+    {
+      "epoch": 5.0,
+      "eval_accuracy": 0.9356252884171666,
+      "eval_loss": 0.24443966150283813,
+      "eval_runtime": 321.057,
+      "eval_samples_per_second": 13.499,
+      "eval_steps_per_second": 1.688,
+      "step": 15350
+    },
+    {
+      "epoch": 5.0,
+      "step": 15350,
+      "total_flos": 3.363363283622308e+19,
+      "train_loss": 0.3319820684874097,
+      "train_runtime": 13677.2564,
+      "train_samples_per_second": 8.977,
+      "train_steps_per_second": 1.122
+    }
+  ],
+  "logging_steps": 500,
+  "max_steps": 15350,
+  "num_train_epochs": 5,
+  "save_steps": 500,
+  "total_flos": 3.363363283622308e+19,
+  "trial_name": null,
+  "trial_params": null
+}