vit-msn-small-wbc-classifier-100

This model is a fine-tuned version of Melo1512/vit-msn-small-wbc-blur-detector on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.2005
• Accuracy: 0.9272

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: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 256
• 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
0.2356	1.0	208	0.2005	0.9272
0.2305	2.0	416	0.2259	0.9195
0.246	3.0	624	0.2097	0.9210
0.2585	4.0	832	0.2184	0.9180
0.2593	5.0	1040	0.2331	0.9171
0.2483	6.0	1248	0.2170	0.9198
0.268	7.0	1456	0.2228	0.9181
0.3112	8.0	1664	0.2361	0.9171
0.2679	9.0	1872	0.2273	0.9185
0.3099	10.0	2080	0.2303	0.9144
0.2749	11.0	2288	0.2658	0.9125
0.2475	12.0	2496	0.2247	0.9179
0.2338	13.0	2704	0.2333	0.9139
0.2731	14.0	2912	0.2295	0.9153
0.229	15.0	3120	0.2505	0.9138
0.2462	16.0	3328	0.2534	0.9137
0.2274	17.0	3536	0.2652	0.9079
0.2339	18.0	3744	0.2550	0.9153
0.2053	19.0	3952	0.2819	0.9106
0.2063	20.0	4160	0.2747	0.9129
0.1964	21.0	4368	0.2975	0.9118
0.1953	22.0	4576	0.2799	0.9145
0.1938	23.0	4784	0.3197	0.9100
0.1851	24.0	4992	0.3143	0.9138
0.1931	25.0	5200	0.3331	0.9125
0.1877	26.0	5408	0.3044	0.9110
0.177	27.0	5616	0.3271	0.9109
0.1529	28.0	5824	0.3382	0.9094
0.1684	29.0	6032	0.3415	0.9128
0.176	30.0	6240	0.3463	0.9095
0.1496	31.0	6448	0.3952	0.9136
0.1509	32.0	6656	0.3690	0.9121
0.1463	33.0	6864	0.3999	0.9094
0.1354	34.0	7072	0.3996	0.9135
0.1546	35.0	7280	0.3810	0.9116
0.1513	36.0	7488	0.3992	0.9121
0.115	37.0	7696	0.4295	0.9132
0.1479	38.0	7904	0.4363	0.9123
0.1455	39.0	8112	0.4220	0.9140
0.1353	40.0	8320	0.4112	0.9127
0.141	41.0	8528	0.4322	0.9139
0.1272	42.0	8736	0.4176	0.9119
0.1402	43.0	8944	0.4041	0.9108
0.1236	44.0	9152	0.4478	0.9095
0.1349	45.0	9360	0.4211	0.9112
0.1472	46.0	9568	0.4510	0.9113
0.1115	47.0	9776	0.4373	0.9119
0.1122	48.0	9984	0.4689	0.9129
0.1297	49.0	10192	0.4569	0.9140
0.1337	50.0	10400	0.4622	0.9111
0.1194	51.0	10608	0.4579	0.9151
0.1322	52.0	10816	0.4728	0.9104
0.1179	53.0	11024	0.4729	0.9125
0.1216	54.0	11232	0.5199	0.9114
0.1234	55.0	11440	0.4769	0.9135
0.1125	56.0	11648	0.4871	0.9118
0.1234	57.0	11856	0.4667	0.9146
0.1103	58.0	12064	0.4741	0.9119
0.1103	59.0	12272	0.4864	0.9129
0.1222	60.0	12480	0.4550	0.9143
0.127	61.0	12688	0.4919	0.9135
0.1117	62.0	12896	0.4946	0.9139
0.1078	63.0	13104	0.5040	0.9133
0.1127	64.0	13312	0.4804	0.9126
0.1122	65.0	13520	0.4997	0.9136
0.1089	66.0	13728	0.5134	0.9139
0.1179	67.0	13936	0.5246	0.9155
0.0934	68.0	14144	0.5158	0.9126
0.1011	69.0	14352	0.5361	0.9140
0.1063	70.0	14560	0.5326	0.9135
0.1021	71.0	14768	0.5151	0.9143
0.1007	72.0	14976	0.5390	0.9143
0.0946	73.0	15184	0.5256	0.9114
0.097	74.0	15392	0.5247	0.9135
0.0967	75.0	15600	0.5154	0.9144
0.0985	76.0	15808	0.5412	0.9154
0.0856	77.0	16016	0.5335	0.9148
0.103	78.0	16224	0.5210	0.9162
0.1033	79.0	16432	0.5165	0.9156
0.109	80.0	16640	0.5303	0.9150
0.0999	81.0	16848	0.5299	0.9158
0.0966	82.0	17056	0.5324	0.9167
0.0952	83.0	17264	0.5229	0.9168
0.1071	84.0	17472	0.5303	0.9176
0.0899	85.0	17680	0.5228	0.9160
0.0868	86.0	17888	0.5297	0.9149
0.1011	87.0	18096	0.5370	0.9156
0.0867	88.0	18304	0.5430	0.9158
0.0936	89.0	18512	0.5346	0.9165
0.0929	90.0	18720	0.5387	0.9163
0.0792	91.0	18928	0.5459	0.9150
0.0918	92.0	19136	0.5257	0.9165
0.0853	93.0	19344	0.5426	0.9155
0.0908	94.0	19552	0.5429	0.9153
0.0981	95.0	19760	0.5394	0.9155
0.0825	96.0	19968	0.5345	0.9168
0.0849	97.0	20176	0.5388	0.9164
0.0992	98.0	20384	0.5357	0.9168
0.0909	99.0	20592	0.5375	0.9167
0.0861	100.0	20800	0.5372	0.9166

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.2.0
• Tokenizers 0.19.1