Edit model card

hushem_1x_deit_tiny_rms_lr001_fold1

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: 1.6888
  • Accuracy: 0.3778

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: 0.001
  • train_batch_size: 32
  • eval_batch_size: 32
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 50

Training results

Training Loss Epoch Step Validation Loss Accuracy
No log 1.0 6 5.6136 0.2667
3.3911 2.0 12 1.9975 0.2444
3.3911 3.0 18 1.6108 0.2444
1.8768 4.0 24 1.5122 0.2667
1.5956 5.0 30 1.5150 0.2444
1.5956 6.0 36 1.8016 0.2444
1.4916 7.0 42 1.5690 0.4444
1.4916 8.0 48 1.4991 0.2667
1.4756 9.0 54 1.4574 0.2444
1.4478 10.0 60 1.4022 0.2444
1.4478 11.0 66 1.4406 0.2667
1.421 12.0 72 1.3666 0.2444
1.421 13.0 78 1.3200 0.2667
1.4101 14.0 84 1.4311 0.2444
1.366 15.0 90 1.5240 0.4
1.366 16.0 96 1.1533 0.5333
1.3311 17.0 102 1.1480 0.4667
1.3311 18.0 108 1.5207 0.2444
1.1912 19.0 114 1.6588 0.2889
1.1923 20.0 120 1.4947 0.4667
1.1923 21.0 126 1.3281 0.2444
1.1796 22.0 132 1.3569 0.3333
1.1796 23.0 138 1.7298 0.2444
1.1031 24.0 144 1.6401 0.3556
1.2056 25.0 150 1.3732 0.2889
1.2056 26.0 156 1.8651 0.3333
1.1039 27.0 162 1.2494 0.4667
1.1039 28.0 168 1.4459 0.2889
1.0659 29.0 174 1.4875 0.3333
1.0534 30.0 180 1.4599 0.3556
1.0534 31.0 186 1.3781 0.3556
1.0466 32.0 192 1.7266 0.4
1.0466 33.0 198 1.5340 0.4222
0.973 34.0 204 1.5429 0.3111
1.0226 35.0 210 1.6233 0.3111
1.0226 36.0 216 1.7204 0.3111
0.9676 37.0 222 1.7918 0.4
0.9676 38.0 228 1.6933 0.3333
0.8379 39.0 234 1.6708 0.4222
0.8938 40.0 240 1.6748 0.4
0.8938 41.0 246 1.6963 0.3778
0.8462 42.0 252 1.6888 0.3778
0.8462 43.0 258 1.6888 0.3778
0.8764 44.0 264 1.6888 0.3778
0.8676 45.0 270 1.6888 0.3778
0.8676 46.0 276 1.6888 0.3778
0.8418 47.0 282 1.6888 0.3778
0.8418 48.0 288 1.6888 0.3778
0.8647 49.0 294 1.6888 0.3778
0.872 50.0 300 1.6888 0.3778

Framework versions

  • Transformers 4.35.0
  • Pytorch 2.1.0+cu118
  • Datasets 2.14.6
  • Tokenizers 0.14.1
Downloads last month
14
Safetensors
Model size
5.53M 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 hkivancoral/hushem_1x_deit_tiny_rms_lr001_fold1

Finetuned
(363)
this model

Evaluation results