mobileViTV2-128-2

This model is a fine-tuned version of apple/mobilevitv2-1.0-imagenet1k-256 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.1494
• Accuracy: 0.9480
• F1: 0.9484
• Precision: 0.9498
• Recall: 0.9480

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
1.628	1.0	93	1.6205	0.1697	0.1607	0.1720	0.1697
1.6155	2.0	186	1.6120	0.2242	0.2119	0.2293	0.2242
1.6031	3.0	279	1.5941	0.2667	0.2650	0.2872	0.2667
1.5639	4.0	372	1.5756	0.3091	0.3022	0.3163	0.3091
1.525	5.0	465	1.5384	0.4	0.3949	0.4097	0.4
1.4892	6.0	558	1.4831	0.5091	0.4924	0.4945	0.5091
1.3605	7.0	651	1.3807	0.6	0.5801	0.5833	0.6
1.1423	8.0	744	1.2166	0.6242	0.5967	0.6139	0.6242
1.0801	9.0	837	1.0558	0.7030	0.6663	0.7190	0.7030
0.8504	10.0	930	0.8856	0.7697	0.7487	0.8081	0.7697
0.7659	11.0	1023	0.7253	0.8121	0.7954	0.8362	0.8121
0.547	12.0	1116	0.5812	0.8485	0.8418	0.8581	0.8485
0.5522	13.0	1209	0.4633	0.8970	0.8926	0.9078	0.8970
0.3583	14.0	1302	0.3797	0.9152	0.9119	0.9257	0.9152
0.3421	15.0	1395	0.3431	0.9273	0.9263	0.9379	0.9273
0.3624	16.0	1488	0.3010	0.9273	0.9265	0.9389	0.9273
0.2069	17.0	1581	0.2989	0.9152	0.9146	0.9260	0.9152
0.1639	18.0	1674	0.2797	0.9212	0.9205	0.9300	0.9212
0.2428	19.0	1767	0.2815	0.9273	0.9263	0.9379	0.9273
0.304	20.0	1860	0.2587	0.9333	0.9325	0.9432	0.9333
0.1349	21.0	1953	0.2617	0.9273	0.9266	0.9378	0.9273
0.2299	22.0	2046	0.2552	0.9333	0.9325	0.9432	0.9333
0.0894	23.0	2139	0.2560	0.9152	0.9149	0.9222	0.9152
0.1049	24.0	2232	0.2689	0.9152	0.9146	0.9226	0.9152
0.1201	25.0	2325	0.2921	0.9152	0.9144	0.9223	0.9152
0.1162	26.0	2418	0.3317	0.9212	0.9206	0.9283	0.9212
0.049	27.0	2511	0.2916	0.9273	0.9266	0.9346	0.9273
0.107	28.0	2604	0.2921	0.9273	0.9266	0.9346	0.9273
0.0521	29.0	2697	0.3267	0.9212	0.9207	0.9264	0.9212
0.1911	30.0	2790	0.3661	0.9091	0.9089	0.9147	0.9091
0.1636	31.0	2883	0.3444	0.9152	0.9147	0.9201	0.9152
0.0615	32.0	2976	0.3879	0.9212	0.9208	0.9277	0.9212
0.0581	33.0	3069	0.3606	0.9212	0.9207	0.9264	0.9212
0.1042	34.0	3162	0.3910	0.9333	0.9327	0.9404	0.9333
0.1468	35.0	3255	0.4503	0.9273	0.9266	0.9346	0.9273
0.0303	36.0	3348	0.4035	0.9152	0.9146	0.9207	0.9152
0.0512	37.0	3441	0.4157	0.9212	0.9207	0.9264	0.9212
0.0627	38.0	3534	0.4399	0.9273	0.9268	0.9323	0.9273
0.091	39.0	3627	0.4023	0.9273	0.9268	0.9323	0.9273
0.1877	40.0	3720	0.4463	0.9152	0.9148	0.9187	0.9152
0.072	41.0	3813	0.4729	0.9212	0.9207	0.9264	0.9212
0.0611	42.0	3906	0.5000	0.9152	0.9147	0.9201	0.9152
0.0308	43.0	3999	0.5051	0.9091	0.9087	0.9130	0.9091
0.0801	44.0	4092	0.5044	0.9273	0.9268	0.9323	0.9273
0.0548	45.0	4185	0.5312	0.9212	0.9211	0.9259	0.9212
0.054	46.0	4278	0.5439	0.8970	0.8965	0.9031	0.8970
0.036	47.0	4371	0.5276	0.8970	0.8963	0.9027	0.8970
0.0172	48.0	4464	0.5379	0.8970	0.8966	0.9003	0.8970
0.0573	49.0	4557	0.5380	0.9152	0.9146	0.9207	0.9152
0.0593	50.0	4650	0.5323	0.9091	0.9089	0.9113	0.9091
0.073	51.0	4743	0.5931	0.9030	0.9029	0.9110	0.9030
0.0959	52.0	4836	0.5285	0.9152	0.9148	0.9187	0.9152
0.0251	53.0	4929	0.5081	0.9152	0.9150	0.9173	0.9152
0.0129	54.0	5022	0.5469	0.9212	0.9207	0.9264	0.9212
0.0073	55.0	5115	0.5533	0.9273	0.9268	0.9323	0.9273
0.0922	56.0	5208	0.5499	0.9273	0.9268	0.9323	0.9273
0.0468	57.0	5301	0.5510	0.9273	0.9268	0.9323	0.9273
0.0217	58.0	5394	0.5798	0.9273	0.9268	0.9323	0.9273
0.0949	59.0	5487	0.5748	0.9273	0.9268	0.9323	0.9273
0.0569	60.0	5580	0.5744	0.9273	0.9268	0.9323	0.9273
0.0187	61.0	5673	0.5989	0.9273	0.9268	0.9323	0.9273
0.0333	62.0	5766	0.6353	0.9273	0.9268	0.9323	0.9273
0.002	63.0	5859	0.6033	0.9273	0.9268	0.9323	0.9273
0.0169	64.0	5952	0.6128	0.9273	0.9268	0.9323	0.9273
0.0303	65.0	6045	0.6143	0.9212	0.9207	0.9264	0.9212
0.0451	66.0	6138	0.6139	0.9273	0.9268	0.9323	0.9273
0.0291	67.0	6231	0.6058	0.9152	0.9148	0.9187	0.9152
0.0163	68.0	6324	0.6154	0.9212	0.9207	0.9264	0.9212
0.0181	69.0	6417	0.5810	0.9091	0.9089	0.9114	0.9091
0.0441	70.0	6510	0.6019	0.9152	0.9147	0.9203	0.9152
0.0395	71.0	6603	0.6018	0.9212	0.9207	0.9264	0.9212
0.0229	72.0	6696	0.6280	0.9091	0.9084	0.9152	0.9091
0.0509	73.0	6789	0.6442	0.9273	0.9268	0.9323	0.9273
0.0178	74.0	6882	0.6510	0.9273	0.9268	0.9323	0.9273
0.0048	75.0	6975	0.6086	0.9273	0.9268	0.9323	0.9273
0.0207	76.0	7068	0.6676	0.9212	0.9207	0.9264	0.9212
0.0354	77.0	7161	0.6055	0.9152	0.9148	0.9187	0.9152
0.0233	78.0	7254	0.6043	0.9152	0.9150	0.9173	0.9152
0.0522	79.0	7347	0.6388	0.9152	0.9148	0.9187	0.9152
0.0519	80.0	7440	0.6531	0.9273	0.9268	0.9323	0.9273
0.0129	81.0	7533	0.6346	0.9212	0.9209	0.9246	0.9212
0.0092	82.0	7626	0.6650	0.9273	0.9268	0.9323	0.9273
0.0289	83.0	7719	0.6390	0.9091	0.9089	0.9114	0.9091
0.0561	84.0	7812	0.6260	0.9091	0.9090	0.9103	0.9091
0.063	85.0	7905	0.6484	0.9212	0.9207	0.9263	0.9212
0.0372	86.0	7998	0.6375	0.9273	0.9268	0.9323	0.9273
0.0042	87.0	8091	0.6384	0.9152	0.9150	0.9173	0.9152
0.0659	88.0	8184	0.6734	0.9273	0.9268	0.9323	0.9273
0.016	89.0	8277	0.6275	0.9091	0.9089	0.9114	0.9091
0.0567	90.0	8370	0.6611	0.9212	0.9207	0.9264	0.9212
0.0467	91.0	8463	0.6528	0.9273	0.9268	0.9323	0.9273
0.0337	92.0	8556	0.6726	0.9273	0.9268	0.9323	0.9273
0.0159	93.0	8649	0.6528	0.9212	0.9207	0.9264	0.9212
0.1206	94.0	8742	0.6997	0.9273	0.9268	0.9323	0.9273
0.051	95.0	8835	0.6729	0.9152	0.9150	0.9173	0.9152
0.0459	96.0	8928	0.6691	0.9212	0.9209	0.9246	0.9212
0.0338	97.0	9021	0.6332	0.9212	0.9207	0.9264	0.9212
0.0823	98.0	9114	0.6550	0.9273	0.9268	0.9323	0.9273
0.0259	99.0	9207	0.6553	0.9212	0.9209	0.9246	0.9212
0.2724	100.0	9300	0.6473	0.9273	0.9268	0.9323	0.9273

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1+cu124
• Datasets 3.3.2
• Tokenizers 0.21.0