dit-base-small_rvl_cdip-NK1000_hint

This model is a fine-tuned version of WinKawaks/vit-small-patch16-224 on the None dataset. It achieves the following results on the evaluation set:

Loss: 2.2813
Accuracy: 0.841
Brier Loss: 0.2948
Nll: 1.9789
F1 Micro: 0.841
F1 Macro: 0.8409
Ece: 0.1435
Aurc: 0.0496

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: 0.0001
train_batch_size: 16
eval_batch_size: 16
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.1
num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Brier Loss	Nll	F1 Micro	F1 Macro	Ece	Aurc
2.5147	1.0	1000	2.3332	0.659	0.4557	2.1792	0.659	0.6434	0.0525	0.1283
2.0194	2.0	2000	1.9888	0.7225	0.3793	2.1426	0.7225	0.7275	0.0522	0.0910
1.7429	3.0	3000	1.8097	0.759	0.3332	1.9995	0.7590	0.7599	0.0480	0.0713
1.5665	4.0	4000	1.8189	0.7652	0.3417	2.0345	0.7652	0.7676	0.0920	0.0706
1.4286	5.0	5000	1.7489	0.7588	0.3435	2.0233	0.7588	0.7639	0.0654	0.0745
1.3258	6.0	6000	1.7223	0.7837	0.3212	1.9911	0.7837	0.7818	0.0939	0.0595
1.1929	7.0	7000	1.6777	0.8015	0.3090	1.9764	0.8015	0.8045	0.0991	0.0554
1.0707	8.0	8000	1.8240	0.79	0.3365	1.9708	0.79	0.7907	0.1334	0.0588
1.0054	9.0	9000	1.9355	0.7853	0.3482	2.0445	0.7853	0.7861	0.1404	0.0637
0.924	10.0	10000	2.0125	0.799	0.3399	2.0334	0.799	0.7993	0.1471	0.0584
0.9043	11.0	11000	1.9868	0.8075	0.3252	2.0171	0.8075	0.8078	0.1429	0.0526
0.8442	12.0	12000	2.1626	0.8087	0.3339	2.0839	0.8087	0.8075	0.1523	0.0534
0.8158	13.0	13000	2.1123	0.7903	0.3567	2.0193	0.7903	0.7922	0.1592	0.0580
0.8087	14.0	14000	2.1296	0.814	0.3269	2.0417	0.8140	0.8152	0.1482	0.0533
0.7924	15.0	15000	2.0768	0.816	0.3181	2.1244	0.816	0.8152	0.1435	0.0500
0.7574	16.0	16000	2.3326	0.797	0.3596	2.1292	0.797	0.7960	0.1668	0.0590
0.7501	17.0	17000	2.3190	0.8067	0.3475	2.0635	0.8067	0.8106	0.1617	0.0588
0.7097	18.0	18000	2.3543	0.8067	0.3468	2.0987	0.8067	0.8063	0.1636	0.0561
0.6939	19.0	19000	2.3151	0.8067	0.3405	2.0927	0.8067	0.8047	0.1617	0.0546
0.6972	20.0	20000	2.3289	0.8087	0.3409	2.0291	0.8087	0.8113	0.1628	0.0550
0.6869	21.0	21000	2.3878	0.797	0.3567	2.0736	0.797	0.7995	0.1727	0.0599
0.6621	22.0	22000	2.3796	0.8137	0.3323	2.1080	0.8137	0.8138	0.1570	0.0553
0.6496	23.0	23000	2.4264	0.8143	0.3342	2.0650	0.8143	0.8182	0.1598	0.0562
0.6478	24.0	24000	2.3300	0.8167	0.3313	2.0509	0.8167	0.8170	0.1564	0.0552
0.6184	25.0	25000	2.3143	0.8283	0.3178	1.9842	0.8283	0.8284	0.1498	0.0509
0.6203	26.0	26000	2.3665	0.825	0.3181	2.0647	0.825	0.8246	0.1501	0.0530
0.6151	27.0	27000	2.2673	0.8335	0.3042	1.9961	0.8335	0.8351	0.1431	0.0501
0.6054	28.0	28000	2.2927	0.8283	0.3100	2.0229	0.8283	0.8281	0.1490	0.0508
0.5965	29.0	29000	2.4001	0.8223	0.3235	2.0393	0.8223	0.8266	0.1553	0.0536
0.5869	30.0	30000	2.4167	0.821	0.3269	1.9890	0.821	0.8213	0.1576	0.0553
0.5852	31.0	31000	2.4030	0.83	0.3120	2.0384	0.83	0.8304	0.1500	0.0539
0.5741	32.0	32000	2.4800	0.8233	0.3259	2.1183	0.8233	0.8212	0.1561	0.0550
0.5743	33.0	33000	2.4718	0.821	0.3269	2.0593	0.821	0.8191	0.1582	0.0578
0.5694	34.0	34000	2.3638	0.8297	0.3135	2.0746	0.8297	0.8293	0.1484	0.0534
0.5567	35.0	35000	2.3320	0.8313	0.3054	2.0498	0.8313	0.8311	0.1470	0.0524
0.5547	36.0	36000	2.3902	0.8293	0.3138	2.0001	0.8293	0.8294	0.1506	0.0513
0.55	37.0	37000	2.3812	0.822	0.3239	2.0567	0.822	0.8226	0.1560	0.0541
0.5482	38.0	38000	2.3680	0.8333	0.3063	2.0447	0.8333	0.8332	0.1460	0.0496
0.546	39.0	39000	2.3250	0.8323	0.3051	1.9994	0.8323	0.8315	0.1478	0.0504
0.541	40.0	40000	2.3498	0.837	0.3003	2.0228	0.8370	0.8367	0.1456	0.0493
0.5448	41.0	41000	2.3504	0.833	0.3053	1.9875	0.833	0.8323	0.1492	0.0532
0.5365	42.0	42000	2.3421	0.8323	0.3077	2.0985	0.8323	0.8314	0.1477	0.0501
0.5324	43.0	43000	2.2976	0.84	0.2929	1.9862	0.8400	0.8403	0.1418	0.0507
0.5326	44.0	44000	2.3270	0.838	0.2993	2.0043	0.838	0.8384	0.1438	0.0521
0.5303	45.0	45000	2.2919	0.839	0.2957	1.9625	0.839	0.8395	0.1430	0.0494
0.5276	46.0	46000	2.2861	0.8397	0.2973	1.9382	0.8397	0.8402	0.1442	0.0512
0.5279	47.0	47000	2.2930	0.8387	0.2962	1.9738	0.8387	0.8384	0.1456	0.0499
0.5251	48.0	48000	2.2841	0.84	0.2950	1.9888	0.8400	0.8399	0.1441	0.0490
0.5257	49.0	49000	2.2802	0.84	0.2950	1.9827	0.8400	0.8400	0.1443	0.0491
0.5232	50.0	50000	2.2813	0.841	0.2948	1.9789	0.841	0.8409	0.1435	0.0496

Framework versions

Transformers 4.26.1
Pytorch 1.13.1.post200
Datasets 2.9.0
Tokenizers 0.13.2

jordyvl/dit-base-small_rvl_cdip-NK1000_hint

dit-base-small_rvl_cdip-NK1000_hint

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions