swiftformer-xs-ve-U13-b-80b

This model is a fine-tuned version of MBZUAI/swiftformer-xs on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 1.2197
• Accuracy: 0.6522

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: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 80

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.92	6	1.3862	0.2174
1.3862	2.0	13	1.3856	0.3261
1.3862	2.92	19	1.3849	0.2826
1.3848	4.0	26	1.3830	0.2609
1.3806	4.92	32	1.3804	0.1739
1.3806	6.0	39	1.3758	0.1957
1.3662	6.92	45	1.3700	0.1739
1.3261	8.0	52	1.3652	0.1739
1.3261	8.92	58	1.3625	0.1522
1.2588	10.0	65	1.3629	0.1304
1.1972	10.92	71	1.3592	0.1304
1.1972	12.0	78	1.3570	0.2174
1.1578	12.92	84	1.3590	0.1957
1.124	14.0	91	1.3731	0.2174
1.124	14.92	97	1.3718	0.1522
1.1045	16.0	104	1.3736	0.1739
1.0703	16.92	110	1.4983	0.2174
1.0703	18.0	117	1.5455	0.1739
1.0663	18.92	123	1.4473	0.1739
1.01	20.0	130	1.4011	0.2609
1.01	20.92	136	1.4053	0.2826
0.9961	22.0	143	1.4186	0.2174
0.9961	22.92	149	1.5168	0.2609
0.9754	24.0	156	1.3873	0.2826
0.9417	24.92	162	1.4656	0.3261
0.9417	26.0	169	1.3499	0.2609
0.9286	26.92	175	1.3902	0.3043
0.9216	28.0	182	1.4819	0.3261
0.9216	28.92	188	1.4133	0.3043
0.8868	30.0	195	1.4124	0.4130
0.8908	30.92	201	1.4421	0.3478
0.8908	32.0	208	1.5085	0.3043
0.8729	32.92	214	1.3854	0.3478
0.8685	34.0	221	1.3264	0.3043
0.8685	34.92	227	1.3947	0.3043
0.8739	36.0	234	1.3455	0.3913
0.8288	36.92	240	1.3621	0.3913
0.8288	38.0	247	1.3875	0.3913
0.8369	38.92	253	1.4274	0.3696
0.8101	40.0	260	1.3251	0.4565
0.8101	40.92	266	1.3039	0.4783
0.8126	42.0	273	1.2523	0.5435
0.8126	42.92	279	1.3060	0.5217
0.7971	44.0	286	1.2678	0.5217
0.7806	44.92	292	1.3332	0.5
0.7806	46.0	299	1.2550	0.5652
0.7899	46.92	305	1.2517	0.5870
0.7602	48.0	312	1.2627	0.5870
0.7602	48.92	318	1.2620	0.6087
0.7748	50.0	325	1.2286	0.5652
0.7613	50.92	331	1.1997	0.6087
0.7613	52.0	338	1.2353	0.5870
0.7514	52.92	344	1.2466	0.5870
0.7581	54.0	351	1.2161	0.5870
0.7581	54.92	357	1.2396	0.5435
0.7401	56.0	364	1.1859	0.6087
0.7421	56.92	370	1.1757	0.6304
0.7421	58.0	377	1.1754	0.5870
0.7261	58.92	383	1.1630	0.6304
0.709	60.0	390	1.2157	0.5870
0.709	60.92	396	1.2124	0.6087
0.7075	62.0	403	1.2095	0.6087
0.7075	62.92	409	1.2543	0.5652
0.7141	64.0	416	1.2210	0.6087
0.6907	64.92	422	1.3190	0.5435
0.6907	66.0	429	1.2197	0.6522
0.7237	66.92	435	1.2365	0.5652
0.6918	68.0	442	1.1570	0.6304
0.6918	68.92	448	1.1790	0.6087
0.7137	70.0	455	1.1968	0.6087
0.6954	70.92	461	1.1959	0.6304
0.6954	72.0	468	1.1782	0.6304
0.6961	72.92	474	1.1935	0.5652
0.6889	73.85	480	1.1835	0.6087

Framework versions

• Transformers 4.36.2
• Pytorch 2.1.2+cu118
• Datasets 2.16.1
• Tokenizers 0.15.0