resnet101-base_tobacco-cnn_tobacco3482_kd_MSE

This model is a fine-tuned version of bdpc/resnet101-base_tobacco on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1315
• Accuracy: 0.365
• Brier Loss: 0.7313
• Nll: 5.5846
• F1 Micro: 0.3650
• F1 Macro: 0.2369
• Ece: 0.2526
• Aurc: 0.4412

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: 256
• eval_batch_size: 256
• 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	4	1.1507	0.1	0.8998	9.9508	0.1000	0.0462	0.1728	0.9208
No log	2.0	8	0.9900	0.155	0.8924	9.6289	0.155	0.0268	0.2400	0.9571
No log	3.0	12	0.8441	0.155	0.9273	8.9944	0.155	0.0268	0.3276	0.9345
No log	4.0	16	1.4048	0.155	1.3149	8.9869	0.155	0.0268	0.6569	0.6091
No log	5.0	20	1.0761	0.155	1.1553	8.9185	0.155	0.0272	0.5441	0.6112
No log	6.0	24	1.1745	0.155	1.1386	9.2644	0.155	0.0304	0.4982	0.6120
No log	7.0	28	0.4686	0.225	0.8829	7.3879	0.225	0.0724	0.3173	0.5804
No log	8.0	32	0.3535	0.24	0.8393	7.0880	0.24	0.0797	0.2963	0.5518
No log	9.0	36	0.2519	0.295	0.8157	6.6738	0.295	0.1375	0.2944	0.4810
No log	10.0	40	0.2957	0.265	0.8432	6.8903	0.265	0.1030	0.3171	0.5807
No log	11.0	44	0.5224	0.21	0.8832	8.6128	0.2100	0.0987	0.2948	0.6814
No log	12.0	48	0.4088	0.18	0.8807	7.0533	0.18	0.0309	0.2966	0.7466
No log	13.0	52	0.5082	0.225	0.8732	8.3126	0.225	0.0606	0.2761	0.7285
No log	14.0	56	0.5253	0.18	0.8905	8.3229	0.18	0.0305	0.2973	0.7838
No log	15.0	60	0.5612	0.225	0.8579	7.9410	0.225	0.0642	0.2690	0.7108
No log	16.0	64	0.2805	0.28	0.8094	6.0275	0.28	0.1475	0.2633	0.5701
No log	17.0	68	0.3076	0.32	0.8151	6.1462	0.32	0.1641	0.2852	0.6162
No log	18.0	72	0.3824	0.29	0.8072	6.0214	0.29	0.1681	0.2900	0.6048
No log	19.0	76	0.5089	0.19	0.8701	8.9391	0.19	0.0418	0.2582	0.7152
No log	20.0	80	0.1490	0.335	0.7347	5.7349	0.335	0.1786	0.2500	0.4430
No log	21.0	84	0.3448	0.255	0.8455	6.6598	0.255	0.0998	0.3124	0.7183
No log	22.0	88	0.6254	0.22	0.8413	6.9926	0.22	0.0966	0.2654	0.7197
No log	23.0	92	0.5464	0.215	0.8909	8.4952	0.2150	0.0570	0.2931	0.7084
No log	24.0	96	0.4465	0.24	0.8445	7.2319	0.24	0.1396	0.2575	0.6667
No log	25.0	100	0.3967	0.215	0.8547	6.7234	0.2150	0.0962	0.2913	0.7053
No log	26.0	104	0.2459	0.295	0.8041	5.1627	0.295	0.1901	0.2525	0.6590
No log	27.0	108	0.4125	0.19	0.8595	7.1181	0.19	0.0551	0.2707	0.7087
No log	28.0	112	0.1686	0.36	0.7309	5.1322	0.36	0.2178	0.2296	0.4432
No log	29.0	116	0.3573	0.205	0.8664	6.6815	0.205	0.0523	0.2753	0.7131
No log	30.0	120	0.1634	0.32	0.7416	5.6798	0.32	0.1862	0.2473	0.4616
No log	31.0	124	0.1404	0.35	0.7295	5.6538	0.35	0.2152	0.2688	0.4389
No log	32.0	128	0.1435	0.325	0.7415	5.5376	0.325	0.1439	0.2567	0.4489
No log	33.0	132	0.1428	0.33	0.7292	5.5151	0.33	0.1791	0.2502	0.4403
No log	34.0	136	0.1602	0.33	0.7371	5.8829	0.33	0.1941	0.2542	0.4481
No log	35.0	140	0.1663	0.325	0.7398	5.6501	0.325	0.1880	0.2443	0.4564
No log	36.0	144	0.1637	0.35	0.7422	5.9440	0.35	0.2053	0.2748	0.4361
No log	37.0	148	0.1520	0.325	0.7317	5.3284	0.325	0.1787	0.2677	0.4531
No log	38.0	152	0.1585	0.335	0.7385	5.9712	0.335	0.1939	0.2648	0.4483
No log	39.0	156	0.1491	0.335	0.7334	5.6729	0.335	0.1912	0.2533	0.4404
No log	40.0	160	0.1367	0.32	0.7297	5.7350	0.32	0.1818	0.2512	0.4498
No log	41.0	164	0.2089	0.335	0.7583	5.2150	0.335	0.2073	0.2822	0.4712
No log	42.0	168	0.1612	0.335	0.7323	4.9145	0.335	0.2058	0.2696	0.4482
No log	43.0	172	0.1616	0.335	0.7349	5.4305	0.335	0.1916	0.2650	0.4493
No log	44.0	176	0.1477	0.335	0.7335	5.3482	0.335	0.1761	0.2478	0.4410
No log	45.0	180	0.1426	0.34	0.7321	5.4265	0.34	0.2018	0.2307	0.4483
No log	46.0	184	0.1531	0.345	0.7351	5.2269	0.345	0.2108	0.2812	0.4572
No log	47.0	188	0.1426	0.34	0.7299	5.1412	0.34	0.2040	0.2418	0.4443
No log	48.0	192	0.1321	0.335	0.7353	5.2955	0.335	0.2017	0.2515	0.4547
No log	49.0	196	0.1330	0.34	0.7332	5.5391	0.34	0.2065	0.2485	0.4524
No log	50.0	200	0.1315	0.365	0.7313	5.5846	0.3650	0.2369	0.2526	0.4412

Framework versions

• Transformers 4.36.0.dev0
• Pytorch 2.2.0.dev20231112+cu118
• Datasets 2.14.5
• Tokenizers 0.14.1