resnet101-base_tobacco-cnn_tobacco3482_kd_CEKD_t5.0_a0.9

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: 0.8809
• Accuracy: 0.7
• Brier Loss: 0.4126
• Nll: 2.4279
• F1 Micro: 0.7
• F1 Macro: 0.6279
• Ece: 0.2569
• Aurc: 0.1111

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	13	2.1185	0.165	0.8967	8.5399	0.165	0.1130	0.2151	0.8331
No log	2.0	26	2.1127	0.13	0.8958	8.1152	0.13	0.0842	0.1816	0.8392
No log	3.0	39	2.0781	0.165	0.8888	6.8828	0.165	0.0878	0.2150	0.8082
No log	4.0	52	2.0197	0.22	0.8762	5.7578	0.22	0.1155	0.2521	0.7521
No log	5.0	65	1.9499	0.205	0.8601	6.0641	0.205	0.0951	0.2567	0.7355
No log	6.0	78	1.9019	0.25	0.8483	5.8930	0.25	0.1178	0.2728	0.6862
No log	7.0	91	1.8252	0.28	0.8301	5.8062	0.28	0.1660	0.2890	0.6982
No log	8.0	104	1.8194	0.28	0.8275	5.2642	0.28	0.1625	0.2874	0.6935
No log	9.0	117	1.7671	0.355	0.8109	5.1326	0.3550	0.2211	0.3018	0.5678
No log	10.0	130	1.6582	0.355	0.7774	5.2226	0.3550	0.2200	0.2991	0.5305
No log	11.0	143	1.5849	0.395	0.7422	5.0239	0.395	0.2436	0.2979	0.3974
No log	12.0	156	1.4908	0.46	0.7001	4.2790	0.46	0.3169	0.3091	0.3003
No log	13.0	169	1.6016	0.395	0.7496	4.2149	0.395	0.2793	0.2929	0.4640
No log	14.0	182	1.4714	0.475	0.6971	4.0742	0.4750	0.3299	0.3177	0.3613
No log	15.0	195	1.5007	0.46	0.7119	3.8252	0.46	0.3145	0.3111	0.3954
No log	16.0	208	1.4352	0.515	0.6776	3.4028	0.515	0.3948	0.3376	0.2993
No log	17.0	221	1.2890	0.575	0.6104	3.4453	0.575	0.4478	0.2940	0.2119
No log	18.0	234	1.2190	0.595	0.5719	3.2413	0.595	0.4662	0.2608	0.1981
No log	19.0	247	1.2287	0.59	0.5764	3.2303	0.59	0.4857	0.2811	0.2020
No log	20.0	260	1.1726	0.64	0.5494	2.9544	0.64	0.5307	0.2993	0.1708
No log	21.0	273	1.1305	0.61	0.5384	2.9557	0.61	0.5170	0.2771	0.1949
No log	22.0	286	1.1256	0.645	0.5295	2.7934	0.645	0.5381	0.3181	0.1629
No log	23.0	299	1.1209	0.645	0.5217	2.8697	0.645	0.5432	0.3055	0.1687
No log	24.0	312	1.2513	0.685	0.5917	2.7262	0.685	0.5639	0.3779	0.1833
No log	25.0	325	1.0321	0.695	0.4819	2.7202	0.695	0.5896	0.2810	0.1280
No log	26.0	338	1.0405	0.645	0.4957	2.6116	0.645	0.5661	0.2515	0.1700
No log	27.0	351	1.0580	0.695	0.4933	2.7436	0.695	0.5996	0.2967	0.1339
No log	28.0	364	0.9740	0.65	0.4575	2.5682	0.65	0.5731	0.2513	0.1384
No log	29.0	377	0.9934	0.695	0.4651	2.5753	0.695	0.6108	0.2775	0.1171
No log	30.0	390	0.9900	0.645	0.4695	2.6280	0.645	0.5668	0.2459	0.1558
No log	31.0	403	0.9671	0.695	0.4504	2.8174	0.695	0.6094	0.2505	0.1188
No log	32.0	416	0.9327	0.715	0.4324	2.5285	0.715	0.6415	0.2565	0.1086
No log	33.0	429	0.9628	0.71	0.4464	2.5876	0.7100	0.6435	0.2709	0.1152
No log	34.0	442	0.9316	0.715	0.4353	2.7111	0.715	0.6334	0.2361	0.1078
No log	35.0	455	0.9275	0.7	0.4364	2.5226	0.7	0.6251	0.2586	0.1207
No log	36.0	468	0.9301	0.7	0.4346	2.6464	0.7	0.6232	0.2482	0.1142
No log	37.0	481	0.9013	0.695	0.4194	2.5575	0.695	0.6197	0.2554	0.1098
No log	38.0	494	0.9008	0.695	0.4196	2.6270	0.695	0.6156	0.2246	0.1063
1.0903	39.0	507	0.9185	0.71	0.4311	2.6290	0.7100	0.6362	0.2626	0.1165
1.0903	40.0	520	0.9053	0.685	0.4254	2.5057	0.685	0.6239	0.2210	0.1171
1.0903	41.0	533	0.8955	0.7	0.4189	2.4823	0.7	0.6291	0.1995	0.1103
1.0903	42.0	546	0.9012	0.69	0.4223	2.5377	0.69	0.6195	0.2486	0.1119
1.0903	43.0	559	0.8894	0.71	0.4138	2.6167	0.7100	0.6382	0.2459	0.1022
1.0903	44.0	572	0.8846	0.695	0.4132	2.5130	0.695	0.6265	0.2198	0.1093
1.0903	45.0	585	0.8946	0.69	0.4190	2.6357	0.69	0.6230	0.2375	0.1145
1.0903	46.0	598	0.8931	0.705	0.4168	2.6306	0.705	0.6342	0.2555	0.1102
1.0903	47.0	611	0.8842	0.71	0.4160	2.3021	0.7100	0.6347	0.2096	0.1120
1.0903	48.0	624	0.8805	0.695	0.4140	2.3447	0.695	0.6237	0.2181	0.1128
1.0903	49.0	637	0.8816	0.7	0.4142	2.4358	0.7	0.6295	0.2550	0.1112
1.0903	50.0	650	0.8809	0.7	0.4126	2.4279	0.7	0.6279	0.2569	0.1111

Framework versions

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