test-hasy-4

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

• Loss: 0.7998
• Accuracy: 0.8004

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: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 1787
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	271	3.8386	0.3139
4.0735	2.0	542	3.3305	0.4241
4.0735	3.0	813	2.9722	0.4990
3.3508	4.0	1084	2.6703	0.5426
3.3508	5.0	1355	2.4288	0.5821
2.906	6.0	1626	2.2306	0.6216
2.906	7.0	1897	2.0742	0.6341
2.6285	8.0	2168	1.9164	0.6632
2.6285	9.0	2439	1.8033	0.6840
2.3956	10.0	2710	1.6762	0.6840
2.3956	11.0	2981	1.5895	0.6944
2.2082	12.0	3252	1.5358	0.6861
2.0728	13.0	3523	1.4479	0.7193
2.0728	14.0	3794	1.3645	0.7256
1.9407	15.0	4065	1.3169	0.7235
1.9407	16.0	4336	1.2634	0.7422
1.8696	17.0	4607	1.2372	0.7360
1.8696	18.0	4878	1.1951	0.7505
1.7775	19.0	5149	1.1390	0.7609
1.7775	20.0	5420	1.1082	0.7588
1.7085	21.0	5691	1.0813	0.7630
1.7085	22.0	5962	1.0518	0.7692
1.6762	23.0	6233	1.0408	0.7630
1.6285	24.0	6504	1.0030	0.7651
1.6285	25.0	6775	1.0110	0.7609
1.6043	26.0	7046	0.9640	0.7651
1.6043	27.0	7317	0.9965	0.7588
1.5504	28.0	7588	0.9415	0.7692
1.5504	29.0	7859	0.9091	0.7713
1.4772	30.0	8130	0.9386	0.7588
1.4772	31.0	8401	0.9123	0.7755
1.5073	32.0	8672	0.8747	0.7817
1.5073	33.0	8943	0.8821	0.7734
1.4772	34.0	9214	0.8465	0.7942
1.4772	35.0	9485	0.8403	0.7963
1.4324	36.0	9756	0.8517	0.7775
1.4379	37.0	10027	0.8311	0.7879
1.4379	38.0	10298	0.8262	0.7900
1.4033	39.0	10569	0.8166	0.7755
1.4033	40.0	10840	0.8189	0.7817
1.3963	41.0	11111	0.8170	0.7879
1.3963	42.0	11382	0.8115	0.7900
1.3769	43.0	11653	0.8076	0.7838
1.3769	44.0	11924	0.7998	0.8004
1.3529	45.0	12195	0.8049	0.7900
1.3529	46.0	12466	0.7937	0.7879
1.366	47.0	12737	0.7955	0.7859
1.3569	48.0	13008	0.7853	0.7921
1.3569	49.0	13279	0.7877	0.7921
1.3223	50.0	13550	0.7875	0.7942

Framework versions

• Transformers 4.38.2
• Pytorch 2.1.0+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2