Edit model card

Adam_ViTL-32_224-2e-4-batch_16_epoch_4_classes_24

This model is a fine-tuned version of google/vit-large-patch32-224-in21k on the imagefolder dataset. It achieves the following results on the evaluation set:

  • Loss: 0.2083
  • Accuracy: 0.9483

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 0.0002
  • train_batch_size: 16
  • eval_batch_size: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 3
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Accuracy
0.7984 0.07 100 0.8039 0.8606
0.4352 0.14 200 0.5735 0.8463
0.3651 0.21 300 0.3951 0.8937
0.3133 0.28 400 0.4525 0.8894
0.2641 0.35 500 0.3618 0.9023
0.2104 0.42 600 0.4240 0.8922
0.1787 0.49 700 0.4070 0.8879
0.1412 0.56 800 0.3259 0.9124
0.2121 0.63 900 0.3575 0.8994
0.1491 0.7 1000 0.2769 0.9152
0.268 0.77 1100 0.3432 0.9195
0.2378 0.84 1200 0.3622 0.9109
0.0812 0.91 1300 0.2857 0.9210
0.127 0.97 1400 0.2787 0.9253
0.0256 1.04 1500 0.3116 0.9267
0.027 1.11 1600 0.2889 0.9282
0.0508 1.18 1700 0.3048 0.9310
0.0932 1.25 1800 0.2732 0.9382
0.0745 1.32 1900 0.3275 0.9195
0.0675 1.39 2000 0.2505 0.9440
0.0347 1.46 2100 0.2686 0.9382
0.0121 1.53 2200 0.2888 0.9454
0.1104 1.6 2300 0.2375 0.9440
0.0778 1.67 2400 0.2345 0.9411
0.0029 1.74 2500 0.2924 0.9282
0.0063 1.81 2600 0.2867 0.9353
0.0394 1.88 2700 0.3384 0.9224
0.0043 1.95 2800 0.2855 0.9195
0.025 2.02 2900 0.3218 0.9296
0.0096 2.09 3000 0.2810 0.9368
0.0018 2.16 3100 0.1971 0.9526
0.0102 2.23 3200 0.2175 0.9497
0.0016 2.3 3300 0.2341 0.9454
0.0024 2.37 3400 0.2607 0.9425
0.0024 2.44 3500 0.2380 0.9440
0.0019 2.51 3600 0.2422 0.9382
0.0062 2.58 3700 0.2191 0.9483
0.0416 2.65 3800 0.2491 0.9483
0.002 2.72 3900 0.2201 0.9497
0.0013 2.79 4000 0.2242 0.9468
0.0012 2.86 4100 0.2182 0.9440
0.0011 2.92 4200 0.2079 0.9497
0.001 2.99 4300 0.2083 0.9483

Framework versions

  • Transformers 4.39.3
  • Pytorch 2.1.2
  • Datasets 2.18.0
  • Tokenizers 0.15.2
Downloads last month
0
Safetensors
Model size
306M params
Tensor type
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for ZaneHorrible/Adam_ViTL-32_224-2e-4-batch_16_epoch_4_classes_24

Finetuned
(4)
this model

Evaluation results