dit-base_tobacco_crl_allv2

This model is a fine-tuned version of jordyvl/dit-base_tobacco on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3885
• Accuracy: 0.945
• Brier Loss: 0.1018
• Nll: 0.7205
• F1 Micro: 0.945
• F1 Macro: 0.9429
• Ece: 0.0554
• Aurc: 0.0107

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 16
• 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	Brier Loss	Nll	F1 Micro	F1 Macro	Ece	Aurc
No log	0.96	3	0.3286	0.925	0.1161	1.0901	0.925	0.9187	0.0807	0.0113
No log	1.96	6	0.3344	0.93	0.1123	1.0906	0.93	0.9219	0.0760	0.0120
No log	2.96	9	0.3363	0.935	0.1092	1.0909	0.935	0.9319	0.0711	0.0135
No log	3.96	12	0.3530	0.935	0.1135	1.0883	0.935	0.9320	0.0753	0.0158
No log	4.96	15	0.3673	0.93	0.1170	1.0778	0.93	0.9247	0.0811	0.0162
No log	5.96	18	0.3583	0.93	0.1167	1.0706	0.93	0.9247	0.0708	0.0159
No log	6.96	21	0.3469	0.93	0.1121	1.0675	0.93	0.9247	0.0783	0.0148
No log	7.96	24	0.3357	0.935	0.1071	1.0654	0.935	0.9279	0.0724	0.0136
No log	8.96	27	0.3329	0.935	0.1048	1.0488	0.935	0.9279	0.0615	0.0127
No log	9.96	30	0.3354	0.94	0.1027	1.0178	0.94	0.9391	0.0636	0.0130
No log	10.96	33	0.3350	0.94	0.1018	1.0054	0.94	0.9418	0.0616	0.0136
No log	11.96	36	0.3342	0.94	0.1012	1.0160	0.94	0.9418	0.0632	0.0133
No log	12.96	39	0.3341	0.935	0.1002	1.0132	0.935	0.9318	0.0692	0.0135
No log	13.96	42	0.3427	0.93	0.1039	1.0032	0.93	0.9275	0.0644	0.0137
No log	14.96	45	0.3393	0.945	0.0985	0.9986	0.945	0.9406	0.0581	0.0122
No log	15.96	48	0.3304	0.94	0.0995	0.9934	0.94	0.9390	0.0575	0.0124
No log	16.96	51	0.3372	0.94	0.1010	0.9796	0.94	0.9390	0.0597	0.0127
No log	17.96	54	0.3399	0.94	0.1023	0.9591	0.94	0.9459	0.0603	0.0123
No log	18.96	57	0.3443	0.94	0.1044	0.9473	0.94	0.9459	0.0588	0.0122
No log	19.96	60	0.3491	0.94	0.1064	0.9401	0.94	0.9398	0.0617	0.0122
No log	20.96	63	0.3510	0.94	0.1081	0.9288	0.94	0.9398	0.0681	0.0131
No log	21.96	66	0.3485	0.94	0.1074	0.9111	0.94	0.9398	0.0628	0.0132
No log	22.96	69	0.3481	0.935	0.1056	0.8993	0.935	0.9382	0.0616	0.0132
No log	23.96	72	0.3605	0.935	0.1131	0.9013	0.935	0.9378	0.0684	0.0120
No log	24.96	75	0.3738	0.935	0.1159	0.9113	0.935	0.9377	0.0683	0.0117
No log	25.96	78	0.3657	0.935	0.1108	0.8932	0.935	0.9394	0.0690	0.0124
No log	26.96	81	0.3511	0.94	0.1060	0.8761	0.94	0.9446	0.0563	0.0120
No log	27.96	84	0.3375	0.94	0.1025	0.8662	0.94	0.9446	0.0602	0.0108
No log	28.96	87	0.3369	0.94	0.1019	0.8654	0.94	0.9446	0.0558	0.0090
No log	29.96	90	0.3423	0.94	0.1055	0.8602	0.94	0.9446	0.0610	0.0076
No log	30.96	93	0.3458	0.945	0.1065	0.8525	0.945	0.9474	0.0605	0.0078
No log	31.96	96	0.3436	0.945	0.1035	0.8390	0.945	0.9490	0.0591	0.0082
No log	32.96	99	0.3436	0.94	0.1025	0.8294	0.94	0.9397	0.0574	0.0086
No log	33.96	102	0.3481	0.94	0.0990	0.8225	0.94	0.9398	0.0579	0.0102
No log	34.96	105	0.3519	0.945	0.0965	0.8203	0.945	0.9491	0.0576	0.0109
No log	35.96	108	0.3551	0.945	0.0939	0.8213	0.945	0.9491	0.0547	0.0116
No log	36.96	111	0.3611	0.95	0.0945	0.8193	0.9500	0.9519	0.0556	0.0117
No log	37.96	114	0.3678	0.94	0.1037	0.8166	0.94	0.9446	0.0591	0.0116
No log	38.96	117	0.3740	0.94	0.1086	0.8226	0.94	0.9446	0.0588	0.0112
No log	39.96	120	0.3754	0.94	0.1106	0.8328	0.94	0.9446	0.0631	0.0114
No log	40.96	123	0.3699	0.94	0.1097	0.8241	0.94	0.9446	0.0584	0.0116
No log	41.96	126	0.3606	0.94	0.1051	0.8010	0.94	0.9446	0.0550	0.0122
No log	42.96	129	0.3548	0.94	0.0970	0.7939	0.94	0.9447	0.0603	0.0130
No log	43.96	132	0.3533	0.95	0.0948	0.7902	0.9500	0.9522	0.0589	0.0131
No log	44.96	135	0.3588	0.945	0.0973	0.7818	0.945	0.9478	0.0540	0.0128
No log	45.96	138	0.3634	0.945	0.1011	0.7802	0.945	0.9478	0.0566	0.0124
No log	46.96	141	0.3642	0.945	0.1018	0.7813	0.945	0.9478	0.0564	0.0108
No log	47.96	144	0.3624	0.945	0.1018	0.7858	0.945	0.9478	0.0568	0.0104
No log	48.96	147	0.3653	0.945	0.1011	0.7949	0.945	0.9478	0.0570	0.0110
No log	49.96	150	0.3697	0.945	0.1022	0.8296	0.945	0.9478	0.0565	0.0110
No log	50.96	153	0.3705	0.945	0.1025	0.8677	0.945	0.9478	0.0558	0.0111
No log	51.96	156	0.3753	0.945	0.1042	0.7933	0.945	0.9478	0.0553	0.0106
No log	52.96	159	0.3763	0.945	0.1038	0.7869	0.945	0.9478	0.0580	0.0112
No log	53.96	162	0.3735	0.95	0.1007	0.7751	0.9500	0.9522	0.0551	0.0114
No log	54.96	165	0.3713	0.95	0.0995	0.7660	0.9500	0.9522	0.0551	0.0112
No log	55.96	168	0.3709	0.95	0.0985	0.7592	0.9500	0.9522	0.0540	0.0108
No log	56.96	171	0.3747	0.95	0.0985	0.7580	0.9500	0.9522	0.0530	0.0119
No log	57.96	174	0.3793	0.95	0.0992	0.7567	0.9500	0.9522	0.0532	0.0121
No log	58.96	177	0.3802	0.95	0.0992	0.7519	0.9500	0.9522	0.0540	0.0115
No log	59.96	180	0.3815	0.95	0.1007	0.7485	0.9500	0.9522	0.0545	0.0108
No log	60.96	183	0.3873	0.945	0.1069	0.7489	0.945	0.9478	0.0572	0.0098
No log	61.96	186	0.3883	0.945	0.1070	0.7477	0.945	0.9478	0.0564	0.0097
No log	62.96	189	0.3818	0.945	0.1053	0.7451	0.945	0.9478	0.0561	0.0096
No log	63.96	192	0.3745	0.945	0.1048	0.7446	0.945	0.9478	0.0586	0.0101
No log	64.96	195	0.3762	0.945	0.1030	0.8090	0.945	0.9478	0.0576	0.0103
No log	65.96	198	0.3822	0.95	0.1025	0.8092	0.9500	0.9522	0.0564	0.0104
No log	66.96	201	0.3896	0.95	0.1030	0.8112	0.9500	0.9522	0.0566	0.0103
No log	67.96	204	0.3914	0.945	0.1036	0.8095	0.945	0.9490	0.0586	0.0102
No log	68.96	207	0.3900	0.945	0.1043	0.8060	0.945	0.9490	0.0585	0.0097
No log	69.96	210	0.3903	0.945	0.1059	0.7370	0.945	0.9490	0.0586	0.0099
No log	70.96	213	0.3923	0.94	0.1069	0.7327	0.94	0.9446	0.0568	0.0096
No log	71.96	216	0.3894	0.94	0.1070	0.7316	0.94	0.9446	0.0611	0.0094
No log	72.96	219	0.3847	0.94	0.1053	0.7318	0.94	0.9446	0.0607	0.0100
No log	73.96	222	0.3833	0.94	0.1043	0.7315	0.94	0.9446	0.0603	0.0105
No log	74.96	225	0.3822	0.935	0.1041	0.7310	0.935	0.9353	0.0620	0.0101
No log	75.96	228	0.3771	0.945	0.1026	0.7314	0.945	0.9429	0.0552	0.0100
No log	76.96	231	0.3748	0.945	0.1014	0.7322	0.945	0.9429	0.0569	0.0100
No log	77.96	234	0.3759	0.945	0.1010	0.7332	0.945	0.9429	0.0556	0.0104
No log	78.96	237	0.3775	0.945	0.1009	0.7346	0.945	0.9429	0.0546	0.0107
No log	79.96	240	0.3784	0.94	0.1012	0.7343	0.94	0.9398	0.0558	0.0108
No log	80.96	243	0.3797	0.94	0.1013	0.7340	0.94	0.9398	0.0559	0.0109
No log	81.96	246	0.3821	0.94	0.1012	0.7359	0.94	0.9398	0.0578	0.0109
No log	82.96	249	0.3836	0.94	0.1011	0.7332	0.94	0.9398	0.0576	0.0108
No log	83.96	252	0.3844	0.94	0.1009	0.7318	0.94	0.9398	0.0574	0.0106
No log	84.96	255	0.3859	0.94	0.1009	0.7316	0.94	0.9398	0.0572	0.0106
No log	85.96	258	0.3885	0.94	0.1012	0.7312	0.94	0.9398	0.0546	0.0106
No log	86.96	261	0.3898	0.945	0.1015	0.7292	0.945	0.9429	0.0546	0.0106
No log	87.96	264	0.3905	0.945	0.1018	0.7265	0.945	0.9429	0.0560	0.0108
No log	88.96	267	0.3909	0.945	0.1020	0.7239	0.945	0.9429	0.0558	0.0106
No log	89.96	270	0.3903	0.945	0.1018	0.7219	0.945	0.9429	0.0559	0.0105
No log	90.96	273	0.3895	0.945	0.1017	0.7208	0.945	0.9429	0.0559	0.0105
No log	91.96	276	0.3891	0.945	0.1017	0.7202	0.945	0.9429	0.0562	0.0104
No log	92.96	279	0.3890	0.945	0.1017	0.7201	0.945	0.9429	0.0564	0.0106
No log	93.96	282	0.3889	0.945	0.1018	0.7202	0.945	0.9429	0.0554	0.0105
No log	94.96	285	0.3883	0.945	0.1016	0.7206	0.945	0.9429	0.0555	0.0105
No log	95.96	288	0.3880	0.945	0.1016	0.7210	0.945	0.9429	0.0556	0.0107
No log	96.96	291	0.3880	0.945	0.1016	0.7209	0.945	0.9429	0.0555	0.0107
No log	97.96	294	0.3882	0.945	0.1017	0.7207	0.945	0.9429	0.0555	0.0107
No log	98.96	297	0.3884	0.945	0.1017	0.7205	0.945	0.9429	0.0554	0.0107
No log	99.96	300	0.3885	0.945	0.1018	0.7205	0.945	0.9429	0.0554	0.0107

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.1.post200
• Datasets 2.9.0
• Tokenizers 0.13.2