metadata

license: apache-2.0
base_model: facebook/deit-tiny-patch16-224
tags:
  - generated_from_trainer
datasets:
  - imagefolder
metrics:
  - accuracy
model-index:
  - name: deit-tiny-patch16-224-finetuned-piid
    results:
      - task:
          name: Image Classification
          type: image-classification
        dataset:
          name: imagefolder
          type: imagefolder
          config: default
          split: val
          args: default
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.7899543378995434

deit-tiny-patch16-224-finetuned-piid

This model is a fine-tuned version of facebook/deit-tiny-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

Loss: 0.5695
Accuracy: 0.7900

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: 8
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 32
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
1.2584	0.98	20	1.1962	0.4064
0.7575	2.0	41	0.8537	0.6347
0.6732	2.98	61	0.7349	0.6758
0.5892	4.0	82	0.6902	0.7032
0.5785	4.98	102	0.6303	0.7352
0.4276	6.0	123	0.5948	0.7397
0.3684	6.98	143	0.6197	0.7260
0.3669	8.0	164	0.5451	0.7580
0.3391	8.98	184	0.6707	0.7352
0.3359	10.0	205	0.5079	0.8082
0.3417	10.98	225	0.5678	0.7580
0.2714	12.0	246	0.5774	0.7443
0.3166	12.98	266	0.5747	0.7534
0.2319	14.0	287	0.5598	0.7945
0.227	14.98	307	0.5853	0.7397
0.1801	16.0	328	0.6237	0.7580
0.158	16.98	348	0.5609	0.7854
0.199	18.0	369	0.6128	0.7443
0.1407	18.98	389	0.5727	0.7900
0.1787	19.51	400	0.5695	0.7900

Framework versions

Transformers 4.35.0
Pytorch 2.1.0+cu118
Datasets 2.14.6
Tokenizers 0.14.1