README.md

metadata

license: apache-2.0
base_model: google/vit-base-patch16-224-in21k
tags:
  - generated_from_trainer
datasets:
  - imagefolder
metrics:
  - accuracy
  - precision
  - f1
model-index:
  - name: emotion_classification
    results:
      - task:
          name: Image Classification
          type: image-classification
        dataset:
          name: imagefolder
          type: imagefolder
          config: default
          split: train
          args: default
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.63125
          - name: Precision
            type: precision
            value: 0.6580684399341683
          - name: F1
            type: f1
            value: 0.6375321878900636

emotion_classification

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: 1.1145
• Accuracy: 0.6312
• Precision: 0.6581
• F1: 0.6375

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: 3e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine_with_restarts
• lr_scheduler_warmup_steps: 150
• num_epochs: 300

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	F1
2.0848	1.0	10	2.0806	0.1625	0.1527	0.1483
2.0824	2.0	20	2.0784	0.1688	0.1556	0.1538
2.0785	3.0	30	2.0748	0.175	0.1612	0.1606
2.0709	4.0	40	2.0698	0.1812	0.1684	0.1661
2.067	5.0	50	2.0635	0.1812	0.1787	0.1697
2.0554	6.0	60	2.0553	0.2	0.1958	0.1893
2.0461	7.0	70	2.0438	0.2313	0.2434	0.2272
2.0263	8.0	80	2.0260	0.2437	0.2763	0.2472
1.9963	9.0	90	1.9959	0.275	0.3073	0.2780
1.9512	10.0	100	1.9435	0.3312	0.3481	0.3307
1.8885	11.0	110	1.8610	0.4313	0.4426	0.4138
1.7908	12.0	120	1.7604	0.4688	0.4485	0.4243
1.6944	13.0	130	1.6677	0.4813	0.4369	0.4349
1.6245	14.0	140	1.6105	0.4625	0.4071	0.4124
1.5745	15.0	150	1.5671	0.5062	0.4551	0.4690
1.5132	16.0	160	1.5169	0.4688	0.4481	0.4201
1.471	17.0	170	1.4772	0.4813	0.4203	0.4404
1.4272	18.0	180	1.4426	0.4938	0.4453	0.4496
1.3896	19.0	190	1.4153	0.4813	0.4409	0.4370
1.3347	20.0	200	1.3976	0.5062	0.4694	0.4662
1.3145	21.0	210	1.3840	0.4813	0.4459	0.4366
1.3319	22.0	220	1.3511	0.5062	0.4867	0.4655
1.2438	23.0	230	1.3186	0.5312	0.5804	0.4945
1.2202	24.0	240	1.3012	0.5375	0.5342	0.5023
1.1838	25.0	250	1.2879	0.5563	0.6162	0.5295
1.1448	26.0	260	1.2534	0.5687	0.5631	0.5456
1.113	27.0	270	1.2398	0.55	0.5645	0.5359
1.0862	28.0	280	1.2357	0.5437	0.6075	0.5143
1.0837	29.0	290	1.2095	0.5687	0.5653	0.5471
1.0609	30.0	300	1.2095	0.5437	0.5729	0.5393
1.0112	31.0	310	1.1859	0.575	0.5989	0.5490
0.9584	32.0	320	1.1683	0.5875	0.6019	0.5777
0.941	33.0	330	1.1649	0.5938	0.6083	0.5875
0.904	34.0	340	1.1896	0.5875	0.6078	0.5720
0.921	35.0	350	1.1662	0.6062	0.6352	0.5975
0.9026	36.0	360	1.1441	0.5875	0.5981	0.5841
0.8217	37.0	370	1.1602	0.5813	0.6098	0.5779
0.8292	38.0	380	1.2140	0.5437	0.5588	0.5258
0.8017	39.0	390	1.1545	0.5563	0.5459	0.5294
0.7787	40.0	400	1.1358	0.6062	0.6300	0.5948
0.7473	41.0	410	1.1285	0.5813	0.5996	0.5779
0.6941	42.0	420	1.1311	0.575	0.5982	0.5757
0.7009	43.0	430	1.1296	0.6125	0.6371	0.6076
0.6537	44.0	440	1.0996	0.5813	0.5866	0.5684
0.6524	45.0	450	1.1477	0.5875	0.6077	0.5813
0.674	46.0	460	1.1063	0.6188	0.6322	0.6127
0.5999	47.0	470	1.1077	0.6	0.6035	0.5951
0.6194	48.0	480	1.1249	0.5813	0.5936	0.5805
0.595	49.0	490	1.1331	0.6	0.5955	0.5876
0.5403	50.0	500	1.1577	0.5875	0.6010	0.5781
0.5932	51.0	510	1.1352	0.5938	0.6214	0.5851
0.621	52.0	520	1.0893	0.6062	0.6044	0.6007
0.5157	53.0	530	1.1382	0.6125	0.6173	0.6075
0.5318	54.0	540	1.1402	0.6	0.6158	0.5970
0.4757	55.0	550	1.1668	0.5938	0.6096	0.5930
0.4826	56.0	560	1.1506	0.6062	0.6367	0.6051
0.5058	57.0	570	1.1857	0.5875	0.5873	0.5767
0.4791	58.0	580	1.1618	0.5813	0.5670	0.5587
0.4322	59.0	590	1.2007	0.5625	0.5628	0.5532
0.442	60.0	600	1.1862	0.5875	0.5681	0.5560
0.431	61.0	610	1.1145	0.6312	0.6581	0.6375
0.4131	62.0	620	1.2081	0.575	0.5912	0.5705
0.3911	63.0	630	1.1380	0.6062	0.6043	0.5988
0.4281	64.0	640	1.1189	0.6188	0.6157	0.6138
0.385	65.0	650	1.2177	0.5625	0.5888	0.5615
0.398	66.0	660	1.2204	0.6	0.6321	0.6008
0.4821	67.0	670	1.2037	0.5938	0.6065	0.5804
0.4127	68.0	680	1.1473	0.6	0.6193	0.5996
0.4062	69.0	690	1.2160	0.5938	0.5950	0.5806
0.3906	70.0	700	1.1763	0.5938	0.6421	0.6034
0.352	71.0	710	1.2355	0.5687	0.5836	0.5613
0.3801	72.0	720	1.1623	0.5813	0.5800	0.5789
0.333	73.0	730	1.1770	0.5875	0.5920	0.5851
0.3562	74.0	740	1.2140	0.5875	0.6367	0.5917
0.3403	75.0	750	1.1679	0.6	0.6209	0.6044
0.3456	76.0	760	1.2496	0.5625	0.5465	0.5409
0.3331	77.0	770	1.1975	0.575	0.6042	0.5759
0.3408	78.0	780	1.2381	0.575	0.5606	0.5565
0.2964	79.0	790	1.1792	0.6	0.6204	0.6009
0.2833	80.0	800	1.1840	0.6	0.6059	0.5933
0.2875	81.0	810	1.2024	0.5875	0.5920	0.5841
0.327	82.0	820	1.2190	0.5813	0.5799	0.5728
0.3027	83.0	830	1.2520	0.5813	0.5682	0.5704
0.2731	84.0	840	1.2167	0.5875	0.6021	0.5847
0.2821	85.0	850	1.2805	0.575	0.5659	0.5527
0.3192	86.0	860	1.2453	0.5625	0.5585	0.5575

Framework versions

• Transformers 4.33.1
• Pytorch 2.0.0
• Datasets 2.14.5
• Tokenizers 0.13.3