Edit model card

Chess_Images

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the imagefolder dataset. It achieves the following results on the evaluation set:

  • Loss: 0.5284
  • Accuracy: 0.9

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: 5e-05
  • train_batch_size: 16
  • eval_batch_size: 16
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 64
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 20

Training results

Training Loss Epoch Step Validation Loss Accuracy
No log 1.0 2 1.0120 0.7
No log 2.0 4 0.9958 0.8
No log 3.0 6 0.9576 0.8333
No log 4.0 8 0.8673 0.8333
0.8292 5.0 10 0.8140 0.8667
0.8292 6.0 12 0.7034 0.9
0.8292 7.0 14 0.7036 0.9
0.8292 8.0 16 0.6949 0.9333
0.8292 9.0 18 0.5620 0.9667
0.6112 10.0 20 0.5829 0.9333
0.6112 11.0 22 0.6530 0.9
0.6112 12.0 24 0.5664 0.9333
0.6112 13.0 26 0.5084 1.0
0.6112 14.0 28 0.6490 0.8333
0.4805 15.0 30 0.4700 1.0
0.4805 16.0 32 0.5473 0.9333
0.4805 17.0 34 0.4928 0.9667
0.4805 18.0 36 0.5023 0.9667
0.4805 19.0 38 0.4885 0.9333
0.4145 20.0 40 0.5284 0.9

Framework versions

  • Transformers 4.38.2
  • Pytorch 2.2.1+cu121
  • Datasets 2.18.0
  • Tokenizers 0.15.2
Downloads last month
14
Safetensors
Model size
85.8M params
Tensor type
F32
·

Finetuned from

Evaluation results