mobilevit-xx-small-finetuned-eurosat

This model is a fine-tuned version of apple/mobilevit-xx-small on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 7.3961
• Accuracy: 0.09

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: 0.003
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 4
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.2991	1.0	100	2.2896	0.16
2.3041	2.0	200	2.4578	0.12
2.2833	3.0	300	2.3022	0.12
2.2755	4.0	400	2.4039	0.17
2.3063	5.0	500	2.5689	0.1
2.3247	6.0	600	2.5307	0.05
2.2867	7.0	700	4.1296	0.08
2.2696	8.0	800	3.0869	0.07
2.2688	9.0	900	3.6086	0.08
2.2616	10.0	1000	6.5422	0.13
2.3896	11.0	1100	3.2715	0.11
2.3264	12.0	1200	2.6975	0.08
2.2603	13.0	1300	2.4012	0.17
2.2845	14.0	1400	3.0856	0.19
2.2813	15.0	1500	3.2556	0.17
2.2232	16.0	1600	3.5357	0.18
2.2332	17.0	1700	3.8758	0.11
2.3568	18.0	1800	3.0675	0.13
2.2627	19.0	1900	3.1308	0.16
2.2528	20.0	2000	2.7741	0.1
2.2039	21.0	2100	2.7257	0.14
2.389	22.0	2200	2.6245	0.08
2.31	23.0	2300	3.1870	0.1
2.1471	24.0	2400	2.8313	0.02
2.1658	25.0	2500	2.9323	0.11
2.0946	26.0	2600	2.8372	0.14
2.0924	27.0	2700	2.7403	0.16
2.2634	28.0	2800	2.8991	0.14
2.1897	29.0	2900	2.8778	0.13
2.144	30.0	3000	2.6043	0.15
2.108	31.0	3100	2.9231	0.1
2.0792	32.0	3200	2.8421	0.12
2.1552	33.0	3300	2.8106	0.12
1.9701	34.0	3400	2.8279	0.11
1.9291	35.0	3500	3.0954	0.2
2.0341	36.0	3600	3.8294	0.14
1.9165	37.0	3700	4.5289	0.11
1.9736	38.0	3800	3.0090	0.14
1.9811	39.0	3900	5.3900	0.14
1.9522	40.0	4000	3.5710	0.08
2.047	41.0	4100	3.4724	0.13
1.9999	42.0	4200	7.2604	0.11
1.9869	43.0	4300	7.9946	0.06
1.9428	44.0	4400	6.1566	0.08
1.7922	45.0	4500	4.9919	0.03
1.9047	46.0	4600	7.1934	0.13
1.9419	47.0	4700	4.3265	0.08
1.7765	48.0	4800	4.6136	0.12
1.7962	49.0	4900	13.4765	0.14
2.0226	50.0	5000	8.1225	0.08
2.1393	51.0	5100	7.7941	0.17
1.8256	52.0	5200	5.4134	0.12
1.9116	53.0	5300	6.1129	0.08
2.1156	54.0	5400	4.1454	0.14
1.7501	55.0	5500	6.2134	0.09
1.8722	56.0	5600	6.4985	0.12
1.9432	57.0	5700	5.2718	0.12
1.7713	58.0	5800	12.3311	0.08
1.6786	59.0	5900	7.1599	0.07
1.5969	60.0	6000	6.0869	0.08
1.8203	61.0	6100	8.8250	0.14
1.7148	62.0	6200	19.0942	0.11
1.6627	63.0	6300	12.4329	0.16
1.7134	64.0	6400	5.5367	0.11
1.8841	65.0	6500	9.1239	0.11
1.6822	66.0	6600	9.4719	0.11
1.8892	67.0	6700	5.6084	0.09
1.72	68.0	6800	8.7854	0.12
1.8751	69.0	6900	7.5571	0.11
1.3783	70.0	7000	11.6321	0.12
1.6403	71.0	7100	7.5354	0.15
2.087	72.0	7200	13.7248	0.11
1.6402	73.0	7300	5.4883	0.12
1.8016	74.0	7400	7.8351	0.13
1.4308	75.0	7500	4.6966	0.13
1.6833	76.0	7600	5.9138	0.12
1.5684	77.0	7700	11.9864	0.15
1.6765	78.0	7800	12.2146	0.1
1.7482	79.0	7900	4.6041	0.12
1.7836	80.0	8000	9.7217	0.13
1.5195	81.0	8100	7.5132	0.12
1.4384	82.0	8200	6.6091	0.13
1.5538	83.0	8300	7.0786	0.13
1.5705	84.0	8400	12.5851	0.14
1.7255	85.0	8500	9.9331	0.11
1.6063	86.0	8600	11.3630	0.14
1.5201	87.0	8700	20.8011	0.08
1.3734	88.0	8800	5.2354	0.09
1.5931	89.0	8900	6.5090	0.1
1.5562	90.0	9000	11.8341	0.1
1.576	91.0	9100	6.9521	0.11
1.542	92.0	9200	5.4470	0.11
1.4968	93.0	9300	11.3896	0.08
1.5031	94.0	9400	11.9717	0.09
1.797	95.0	9500	5.6596	0.15
1.5389	96.0	9600	5.3947	0.15
1.6494	97.0	9700	12.2707	0.09
1.73	98.0	9800	7.7482	0.09
1.6781	99.0	9900	8.2178	0.09
1.6353	100.0	10000	7.3961	0.09

Framework versions

• Transformers 4.44.0
• Pytorch 2.4.0+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1