resnet101_rvl-cdip-cnn_rvl_cdip-NK1000_hint

This model is a fine-tuned version of microsoft/resnet-50 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 20.4893
• Accuracy: 0.7622
• Brier Loss: 0.3995
• Nll: 2.6673
• F1 Micro: 0.7622
• F1 Macro: 0.7619
• Ece: 0.1742
• Aurc: 0.0853

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: 64
• eval_batch_size: 64
• 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
No log	1.0	250	27.0152	0.144	0.9329	8.3774	0.144	0.1293	0.0760	0.8496
26.9201	2.0	500	25.8022	0.4547	0.8625	4.1098	0.4547	0.4194	0.3292	0.3673
26.9201	3.0	750	24.5485	0.5617	0.6135	3.0722	0.5617	0.5439	0.1557	0.2257
24.565	4.0	1000	23.9825	0.6388	0.5062	2.7343	0.6388	0.6354	0.1084	0.1537
24.565	5.0	1250	23.8483	0.6747	0.4518	2.5930	0.6747	0.6686	0.0597	0.1289
23.3904	6.0	1500	23.2280	0.7137	0.3953	2.4736	0.7138	0.7117	0.0486	0.0997
23.3904	7.0	1750	23.0275	0.725	0.3781	2.3823	0.7250	0.7238	0.0414	0.0911
22.6462	8.0	2000	22.8213	0.7358	0.3699	2.3745	0.7358	0.7351	0.0539	0.0881
22.6462	9.0	2250	22.6219	0.7468	0.3629	2.3056	0.7468	0.7465	0.0617	0.0852
22.0944	10.0	2500	22.4746	0.751	0.3593	2.3500	0.751	0.7523	0.0637	0.0846
22.0944	11.0	2750	22.3503	0.752	0.3624	2.4245	0.752	0.7533	0.0810	0.0834
21.6411	12.0	3000	22.2263	0.7545	0.3693	2.4277	0.7545	0.7547	0.0972	0.0885
21.6411	13.0	3250	22.1353	0.7522	0.3740	2.4647	0.7522	0.7532	0.1141	0.0862
21.2742	14.0	3500	22.1122	0.7475	0.3868	2.5369	0.7475	0.7495	0.1250	0.0922
21.2742	15.0	3750	22.0040	0.7508	0.3842	2.5364	0.7508	0.7501	0.1304	0.0911
20.9515	16.0	4000	21.8795	0.758	0.3772	2.5474	0.7580	0.7578	0.1324	0.0846
20.9515	17.0	4250	21.7554	0.754	0.3892	2.5498	0.754	0.7543	0.1420	0.0923
20.6695	18.0	4500	21.6863	0.749	0.3981	2.6337	0.749	0.7507	0.1510	0.0922
20.6695	19.0	4750	21.6123	0.7498	0.4007	2.5993	0.7498	0.7499	0.1551	0.0921
20.4239	20.0	5000	21.5128	0.7595	0.3845	2.5510	0.7595	0.7590	0.1498	0.0870
20.4239	21.0	5250	21.4770	0.7542	0.4005	2.6396	0.7542	0.7547	0.1623	0.0932
20.2131	22.0	5500	21.3497	0.7612	0.3892	2.5117	0.7612	0.7609	0.1539	0.0891
20.2131	23.0	5750	21.3489	0.7572	0.3956	2.5227	0.7572	0.7570	0.1608	0.0883
20.0332	24.0	6000	21.2609	0.7585	0.3939	2.5487	0.7585	0.7595	0.1629	0.0860
20.0332	25.0	6250	21.2046	0.7552	0.3982	2.6283	0.7552	0.7559	0.1663	0.0878
19.8699	26.0	6500	21.1515	0.7528	0.4038	2.6730	0.7528	0.7536	0.1721	0.0858
19.8699	27.0	6750	21.0789	0.7562	0.4003	2.6027	0.7562	0.7575	0.1683	0.0876
19.7228	28.0	7000	21.0357	0.7565	0.3996	2.6490	0.7565	0.7561	0.1707	0.0844
19.7228	29.0	7250	20.9975	0.758	0.3971	2.6300	0.7580	0.7574	0.1704	0.0835
19.589	30.0	7500	20.9221	0.7568	0.4007	2.5841	0.7568	0.7567	0.1714	0.0860
19.589	31.0	7750	20.8725	0.7562	0.3996	2.5775	0.7562	0.7562	0.1752	0.0847
19.4738	32.0	8000	20.8438	0.7572	0.3999	2.6441	0.7572	0.7570	0.1693	0.0877
19.4738	33.0	8250	20.8337	0.755	0.4052	2.6660	0.755	0.7555	0.1743	0.0868
19.3704	34.0	8500	20.7635	0.7575	0.4022	2.6885	0.7575	0.7583	0.1764	0.0868
19.3704	35.0	8750	20.7705	0.7608	0.4001	2.6415	0.7608	0.7601	0.1735	0.0856
19.2791	36.0	9000	20.7221	0.7632	0.3984	2.7139	0.7632	0.7640	0.1706	0.0857
19.2791	37.0	9250	20.6873	0.7622	0.3986	2.6743	0.7622	0.7625	0.1715	0.0838
19.2036	38.0	9500	20.6757	0.7618	0.3990	2.6225	0.7618	0.7620	0.1735	0.0852
19.2036	39.0	9750	20.6421	0.7588	0.4018	2.6342	0.7588	0.7579	0.1761	0.0870
19.1398	40.0	10000	20.6432	0.761	0.4057	2.6595	0.761	0.7610	0.1760	0.0868
19.1398	41.0	10250	20.5778	0.7672	0.3981	2.6180	0.7672	0.7674	0.1680	0.0850
19.0835	42.0	10500	20.5628	0.764	0.3981	2.6309	0.764	0.7625	0.1726	0.0851
19.0835	43.0	10750	20.5530	0.7632	0.3995	2.6470	0.7632	0.7628	0.1733	0.0868
19.0398	44.0	11000	20.5625	0.761	0.4029	2.6650	0.761	0.7608	0.1764	0.0864
19.0398	45.0	11250	20.5637	0.7628	0.4010	2.6709	0.7628	0.7623	0.1760	0.0850
19.0073	46.0	11500	20.5378	0.7628	0.3998	2.6522	0.7628	0.7631	0.1749	0.0859
19.0073	47.0	11750	20.5199	0.7615	0.4010	2.6406	0.7615	0.7619	0.1748	0.0867
18.9818	48.0	12000	20.5378	0.761	0.4031	2.6434	0.761	0.7616	0.1767	0.0856
18.9818	49.0	12250	20.4962	0.7652	0.3962	2.6250	0.7652	0.7653	0.1720	0.0853
18.9734	50.0	12500	20.4893	0.7622	0.3995	2.6673	0.7622	0.7619	0.1742	0.0853

Framework versions

• Transformers 4.33.3
• Pytorch 2.2.0.dev20231002
• Datasets 2.7.1
• Tokenizers 0.13.3