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

Loss: 1.1135
Accuracy: 0.8627

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: 100
eval_batch_size: 102
seed: 1337
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 120.0

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	6	3.2331	0.0882
3.2363	2.0	12	3.2025	0.1373
3.2363	3.0	18	3.1761	0.1863
3.1622	4.0	24	3.1238	0.2255
3.0918	5.0	30	3.0789	0.3137
3.0918	6.0	36	3.0280	0.3235
3.0081	7.0	42	2.9878	0.3431
3.0081	8.0	48	2.9316	0.3824
2.9118	9.0	54	2.8864	0.4314
2.8231	10.0	60	2.8314	0.4510
2.8231	11.0	66	2.7817	0.5196
2.7149	12.0	72	2.7278	0.5196
2.7149	13.0	78	2.6796	0.5588
2.6202	14.0	84	2.6203	0.5882
2.5243	15.0	90	2.5674	0.5882
2.5243	16.0	96	2.5170	0.6078
2.4279	17.0	102	2.4672	0.6176
2.4279	18.0	108	2.4285	0.5980
2.3404	19.0	114	2.3784	0.6569
2.2633	20.0	120	2.3348	0.6471
2.2633	21.0	126	2.2872	0.6667
2.1838	22.0	132	2.2539	0.6569
2.1838	23.0	138	2.2232	0.6765
2.1022	24.0	144	2.1867	0.6471
2.0364	25.0	150	2.1489	0.6863
2.0364	26.0	156	2.1099	0.7255
1.96	27.0	162	2.0767	0.7157
1.96	28.0	168	2.0417	0.7157
1.9235	29.0	174	2.0162	0.7353
1.8484	30.0	180	1.9787	0.7451
1.8484	31.0	186	1.9548	0.7451
1.7971	32.0	192	1.9329	0.7549
1.7971	33.0	198	1.9052	0.7647
1.7409	34.0	204	1.8827	0.7549
1.7006	35.0	210	1.8589	0.7745
1.7006	36.0	216	1.8294	0.7843
1.6426	37.0	222	1.8098	0.7843
1.6426	38.0	228	1.7809	0.7647
1.6102	39.0	234	1.7643	0.7843
1.5704	40.0	240	1.7399	0.8039
1.5704	41.0	246	1.7193	0.8137
1.5264	42.0	252	1.6980	0.8333
1.5264	43.0	258	1.6840	0.8039
1.4821	44.0	264	1.6644	0.8235
1.4506	45.0	270	1.6467	0.8235
1.4506	46.0	276	1.6333	0.8235
1.4358	47.0	282	1.6095	0.8235
1.4358	48.0	288	1.5906	0.8235
1.3695	49.0	294	1.5720	0.8431
1.367	50.0	300	1.5610	0.8333
1.367	51.0	306	1.5440	0.8529
1.3299	52.0	312	1.5359	0.8333
1.3299	53.0	318	1.5129	0.8333
1.2765	54.0	324	1.5057	0.8235
1.2785	55.0	330	1.4867	0.8235
1.2785	56.0	336	1.4751	0.8333
1.2355	57.0	342	1.4553	0.8235
1.2355	58.0	348	1.4491	0.8235
1.2418	59.0	354	1.4289	0.8431
1.2058	60.0	360	1.4185	0.8235
1.2058	61.0	366	1.4104	0.8333
1.164	62.0	372	1.3968	0.8333
1.164	63.0	378	1.3846	0.8431
1.1529	64.0	384	1.3697	0.8431
1.1408	65.0	390	1.3633	0.8431
1.1408	66.0	396	1.3505	0.8431
1.1102	67.0	402	1.3371	0.8529
1.1102	68.0	408	1.3282	0.8529
1.0906	69.0	414	1.3240	0.8431
1.0759	70.0	420	1.3163	0.8431
1.0759	71.0	426	1.3044	0.8529
1.0651	72.0	432	1.2924	0.8431
1.0651	73.0	438	1.2867	0.8529
1.0501	74.0	444	1.2749	0.8529
1.0238	75.0	450	1.2688	0.8431
1.0238	76.0	456	1.2568	0.8529
1.0046	77.0	462	1.2502	0.8529
1.0046	78.0	468	1.2460	0.8529
0.9946	79.0	474	1.2455	0.8431
0.9998	80.0	480	1.2343	0.8529
0.9998	81.0	486	1.2286	0.8529
0.9709	82.0	492	1.2195	0.8431
0.9709	83.0	498	1.2126	0.8529
0.963	84.0	504	1.2102	0.8431
0.9499	85.0	510	1.2024	0.8431
0.9499	86.0	516	1.1980	0.8529
0.937	87.0	522	1.1912	0.8529
0.937	88.0	528	1.1883	0.8431
0.9389	89.0	534	1.1845	0.8529
0.9181	90.0	540	1.1811	0.8529
0.9181	91.0	546	1.1777	0.8431
0.9219	92.0	552	1.1743	0.8627
0.9219	93.0	558	1.1675	0.8627
0.9067	94.0	564	1.1598	0.8627
0.9009	95.0	570	1.1601	0.8627
0.9009	96.0	576	1.1564	0.8529
0.8914	97.0	582	1.1505	0.8529
0.8914	98.0	588	1.1487	0.8529
0.8739	99.0	594	1.1480	0.8627
0.8742	100.0	600	1.1413	0.8529
0.8742	101.0	606	1.1368	0.8627
0.8679	102.0	612	1.1361	0.8627
0.8679	103.0	618	1.1317	0.8627
0.8516	104.0	624	1.1296	0.8529
0.876	105.0	630	1.1288	0.8627
0.876	106.0	636	1.1264	0.8627
0.8591	107.0	642	1.1238	0.8627
0.8591	108.0	648	1.1227	0.8627
0.8586	109.0	654	1.1208	0.8627
0.8415	110.0	660	1.1194	0.8627
0.8415	111.0	666	1.1185	0.8627
0.8465	112.0	672	1.1178	0.8529
0.8465	113.0	678	1.1184	0.8529
0.8503	114.0	684	1.1183	0.8431
0.8332	115.0	690	1.1174	0.8431
0.8332	116.0	696	1.1165	0.8431
0.8476	117.0	702	1.1153	0.8529
0.8476	118.0	708	1.1142	0.8529
0.8382	119.0	714	1.1137	0.8627
0.8527	120.0	720	1.1135	0.8627

Framework versions

Transformers 4.37.0.dev0
Pytorch 2.1.0+cu121
Datasets 2.15.0
Tokenizers 0.15.0

nicolasdupuisroy/vit-letter-identification-v2

vit-letter-identification-v2

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

nicolasdupuisroy/
vit-letter-identification-v2