README.md

metadata

license: apache-2.0
base_model: google/vit-base-patch16-224-in21k
tags:
  - generated_from_trainer
datasets:
  - imagefolder
metrics:
  - accuracy
model-index:
  - name: emotion_classifier
    results:
      - task:
          name: Emotion Classifier
          type: emotion-classifier
        dataset:
          name: imagefolder
          type: imagefolder
          config: default
          split: train
          args: default
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.5520833333333334

emotion_classifier

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: 1.2783
• Accuracy: 0.5521

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: 1e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant_with_warmup
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	35	2.0697	0.2014
No log	2.0	70	2.0539	0.1875
No log	3.0	105	2.0278	0.2014
No log	4.0	140	1.9869	0.2639
No log	5.0	175	1.9248	0.2986
No log	6.0	210	1.8172	0.3403
No log	7.0	245	1.7661	0.375
No log	8.0	280	1.6933	0.4306
No log	9.0	315	1.6493	0.4514
No log	10.0	350	1.6028	0.4514
No log	11.0	385	1.5580	0.4444
No log	12.0	420	1.5267	0.5
No log	13.0	455	1.4934	0.4861
No log	14.0	490	1.4605	0.5208
1.6139	15.0	525	1.4499	0.5278
1.6139	16.0	560	1.4228	0.5347
1.6139	17.0	595	1.4109	0.5208
1.6139	18.0	630	1.3872	0.5139
1.6139	19.0	665	1.3640	0.5556
1.6139	20.0	700	1.3787	0.5208
1.6139	21.0	735	1.3820	0.5278
1.6139	22.0	770	1.3649	0.5069
1.6139	23.0	805	1.3508	0.5347
1.6139	24.0	840	1.3322	0.5417
1.6139	25.0	875	1.3577	0.5347
1.6139	26.0	910	1.3337	0.5625
1.6139	27.0	945	1.3578	0.5139
1.6139	28.0	980	1.3256	0.5556
0.8303	29.0	1015	1.3139	0.5833
0.8303	30.0	1050	1.3575	0.5139
0.8303	31.0	1085	1.3214	0.5625
0.8303	32.0	1120	1.3185	0.5486
0.8303	33.0	1155	1.3285	0.5417
0.8303	34.0	1190	1.3259	0.5903
0.8303	35.0	1225	1.3492	0.5556
0.8303	36.0	1260	1.3164	0.5764
0.8303	37.0	1295	1.3645	0.5417
0.8303	38.0	1330	1.3431	0.5347
0.8303	39.0	1365	1.3272	0.5278
0.8303	40.0	1400	1.3326	0.5972
0.8303	41.0	1435	1.3375	0.5486
0.8303	42.0	1470	1.3641	0.5556
0.3516	43.0	1505	1.3633	0.5278
0.3516	44.0	1540	1.3532	0.5278
0.3516	45.0	1575	1.3473	0.5903
0.3516	46.0	1610	1.3413	0.5833
0.3516	47.0	1645	1.4158	0.5556
0.3516	48.0	1680	1.3747	0.5903
0.3516	49.0	1715	1.4364	0.5347
0.3516	50.0	1750	1.4659	0.5417

Framework versions

• Transformers 4.37.2
• Pytorch 2.2.0+cu121
• Datasets 2.16.1
• Tokenizers 0.15.1