resnet101-base_tobacco-cnn_tobacco3482_simkd

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: 13.1229
• Accuracy: 0.295
• Brier Loss: 0.7636
• Nll: 6.8757
• F1 Micro: 0.295
• F1 Macro: 0.1150
• Ece: 0.2446
• Aurc: 0.4919

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: 128
• eval_batch_size: 128
• 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	7	0.2512	0.18	0.9617	7.0686	0.18	0.0305	0.3439	0.7810
No log	2.0	14	0.3629	0.18	1.0943	7.0153	0.18	0.0305	0.4345	0.8186
No log	3.0	21	0.4745	0.18	1.1577	6.9805	0.18	0.0305	0.5034	0.8029
No log	4.0	28	0.6953	0.18	1.1290	6.9352	0.18	0.0305	0.4731	0.8367
No log	5.0	35	173.4450	0.18	1.1346	6.8314	0.18	0.0305	0.4615	0.8814
No log	6.0	42	412.7549	0.18	1.1098	6.8364	0.18	0.0305	0.4420	0.8716
No log	7.0	49	148.0839	0.18	1.0291	6.9271	0.18	0.0305	0.3960	0.7698
No log	8.0	56	61.2696	0.18	0.9674	6.9593	0.18	0.0305	0.3413	0.7924
No log	9.0	63	175.4512	0.18	0.9708	6.9854	0.18	0.0305	0.3549	0.8252
No log	10.0	70	139.2036	0.18	0.9400	6.9022	0.18	0.0305	0.3300	0.7760
No log	11.0	77	12.5605	0.295	0.8656	6.9766	0.295	0.1138	0.3093	0.5354
No log	12.0	84	2.3147	0.18	0.9363	6.9778	0.18	0.0305	0.3084	0.7507
No log	13.0	91	75.2050	0.18	0.9543	9.1566	0.18	0.0305	0.2990	0.7716
No log	14.0	98	37.4873	0.18	0.9410	9.1473	0.18	0.0305	0.3029	0.7517
No log	15.0	105	8.5750	0.18	0.9304	9.1440	0.18	0.0305	0.3033	0.7718
No log	16.0	112	21.5310	0.18	0.9232	9.1349	0.18	0.0305	0.3122	0.7717
No log	17.0	119	66.9546	0.18	0.9287	9.1376	0.18	0.0305	0.2920	0.7715
No log	18.0	126	2.6525	0.285	0.8357	7.0773	0.285	0.1143	0.3156	0.5306
No log	19.0	133	7.7253	0.24	0.8574	7.0190	0.24	0.0880	0.2948	0.7186
No log	20.0	140	30.0305	0.285	0.8086	6.9862	0.285	0.1133	0.3001	0.5273
No log	21.0	147	3.9243	0.18	0.8680	7.4799	0.18	0.0306	0.2739	0.7704
No log	22.0	154	4.4660	0.18	0.8831	8.9935	0.18	0.0308	0.2652	0.7313
No log	23.0	161	3.9728	0.18	0.8719	8.9609	0.18	0.0308	0.2600	0.7651
No log	24.0	168	2.6913	0.285	0.8089	6.9969	0.285	0.1146	0.2873	0.5122
No log	25.0	175	1.3141	0.29	0.8086	7.0227	0.29	0.1156	0.3154	0.5256
No log	26.0	182	13.5853	0.29	0.7782	6.8763	0.29	0.1168	0.2735	0.5045
No log	27.0	189	11.9763	0.3	0.7730	6.8499	0.3	0.1171	0.2740	0.4971
No log	28.0	196	1.6467	0.285	0.8067	7.1641	0.285	0.1144	0.2870	0.5193
No log	29.0	203	30.5306	0.285	0.8424	7.1576	0.285	0.1129	0.2686	0.6662
No log	30.0	210	13.5964	0.18	0.8584	7.0972	0.18	0.0305	0.2704	0.7307
No log	31.0	217	98.3061	0.29	0.8274	7.0330	0.29	0.1167	0.3163	0.5653
No log	32.0	224	53.0911	0.29	0.7984	6.9311	0.29	0.1167	0.2911	0.5181
No log	33.0	231	2.2010	0.265	0.8291	6.9883	0.265	0.1037	0.2945	0.6039
No log	34.0	238	3.6255	0.295	0.7836	6.8954	0.295	0.1176	0.2636	0.5025
No log	35.0	245	0.9640	0.3	0.7571	6.7913	0.3	0.1170	0.2388	0.4746
No log	36.0	252	1.1935	0.295	0.7711	6.7993	0.295	0.1175	0.2619	0.4779
No log	37.0	259	12.7465	0.305	0.7650	6.8142	0.305	0.1205	0.2512	0.4798
No log	38.0	266	56.6876	0.305	0.7840	6.8750	0.305	0.1205	0.2835	0.4985
No log	39.0	273	122.6602	0.295	0.7919	6.9220	0.295	0.1116	0.2493	0.5312
No log	40.0	280	14.4685	0.295	0.7757	6.8232	0.295	0.1162	0.2575	0.4988
No log	41.0	287	3.9605	0.295	0.7601	6.7809	0.295	0.1138	0.2437	0.4911
No log	42.0	294	7.9424	0.295	0.7567	6.7609	0.295	0.1138	0.2398	0.4883
No log	43.0	301	17.7810	0.295	0.7713	6.8075	0.295	0.1175	0.2503	0.5090
No log	44.0	308	30.8773	0.295	0.7747	6.8248	0.295	0.1127	0.2651	0.5149
No log	45.0	315	16.3877	0.29	0.7736	6.8888	0.29	0.1117	0.2641	0.5026
No log	46.0	322	7.4195	0.29	0.7674	6.8179	0.29	0.1117	0.2621	0.4991
No log	47.0	329	9.6560	0.295	0.7694	6.8960	0.295	0.1138	0.2604	0.4963
No log	48.0	336	6.6040	0.29	0.7622	6.7835	0.29	0.1117	0.2271	0.4958
No log	49.0	343	10.3365	0.29	0.7640	6.8293	0.29	0.1117	0.2583	0.4941
No log	50.0	350	13.1229	0.295	0.7636	6.8757	0.295	0.1150	0.2446	0.4919

Framework versions

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