resnet101-base_tobacco-cnn_tobacco3482_hint

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: 24.6607
• Accuracy: 0.57
• Brier Loss: 0.6012
• Nll: 2.9238
• F1 Micro: 0.57
• F1 Macro: 0.5344
• Ece: 0.2496
• Aurc: 0.2274

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	27.2504	0.07	0.9006	8.5876	0.07	0.0131	0.1634	0.9646
No log	2.0	14	27.1186	0.155	0.9229	12.2960	0.155	0.0268	0.2967	0.8769
No log	3.0	21	27.9163	0.155	1.3722	11.2040	0.155	0.0268	0.6887	0.5963
No log	4.0	28	28.2724	0.155	1.4334	9.3615	0.155	0.0273	0.7029	0.6185
No log	5.0	35	26.9699	0.175	1.0316	5.3928	0.175	0.0465	0.4168	0.5989
No log	6.0	42	26.1797	0.23	0.8746	3.9558	0.23	0.0993	0.3120	0.5627
No log	7.0	49	25.8507	0.25	0.8299	3.2357	0.25	0.1721	0.2686	0.6618
No log	8.0	56	25.7515	0.24	0.8336	2.7738	0.24	0.1579	0.2670	0.6619
No log	9.0	63	25.3041	0.39	0.7346	2.5881	0.39	0.2914	0.2649	0.4362
No log	10.0	70	25.1996	0.375	0.7406	2.7338	0.375	0.2616	0.2903	0.4923
No log	11.0	77	25.0418	0.44	0.6756	3.2534	0.44	0.3173	0.2520	0.3197
No log	12.0	84	25.3664	0.35	0.8231	3.6209	0.35	0.2628	0.2924	0.5484
No log	13.0	91	25.0353	0.44	0.6927	3.5523	0.44	0.3230	0.2842	0.3332
No log	14.0	98	25.2980	0.36	0.8265	3.3953	0.36	0.2859	0.3158	0.5347
No log	15.0	105	24.8521	0.425	0.6604	3.0888	0.425	0.3379	0.2641	0.3096
No log	16.0	112	24.8368	0.46	0.6622	2.7863	0.46	0.3626	0.2771	0.3429
No log	17.0	119	25.0490	0.355	0.7909	2.9342	0.3550	0.2764	0.3300	0.5313
No log	18.0	126	24.9950	0.4	0.7521	3.5010	0.4000	0.3467	0.2801	0.4721
No log	19.0	133	24.7232	0.505	0.6259	2.9709	0.505	0.4017	0.2799	0.2807
No log	20.0	140	24.7500	0.5	0.6408	3.1274	0.5	0.4278	0.2398	0.2752
No log	21.0	147	24.5976	0.54	0.5922	2.7847	0.54	0.4872	0.2422	0.2319
No log	22.0	154	24.9329	0.42	0.7518	2.9924	0.4200	0.3777	0.3094	0.4446
No log	23.0	161	24.6088	0.535	0.6089	2.8494	0.535	0.5067	0.2756	0.2770
No log	24.0	168	25.1851	0.39	0.8175	3.7625	0.39	0.3513	0.3211	0.5049
No log	25.0	175	24.5058	0.585	0.5754	2.6524	0.585	0.5707	0.2296	0.2227
No log	26.0	182	25.2073	0.435	0.7812	3.0365	0.435	0.3839	0.3190	0.5012
No log	27.0	189	24.7752	0.54	0.6558	2.9071	0.54	0.4667	0.2669	0.2898
No log	28.0	196	24.8546	0.515	0.6697	2.5989	0.515	0.4397	0.2943	0.3817
No log	29.0	203	24.5759	0.56	0.5969	2.6234	0.56	0.5342	0.2609	0.2493
No log	30.0	210	24.7052	0.53	0.6198	2.9462	0.53	0.4811	0.2779	0.2766
No log	31.0	217	24.5828	0.545	0.6038	2.7967	0.545	0.4979	0.2455	0.2369
No log	32.0	224	24.6622	0.545	0.6220	2.8878	0.545	0.4925	0.2854	0.2682
No log	33.0	231	24.6253	0.57	0.5991	3.1607	0.57	0.5327	0.2869	0.2518
No log	34.0	238	24.6230	0.535	0.6351	2.5626	0.535	0.5245	0.2766	0.3077
No log	35.0	245	24.5803	0.59	0.5900	2.8215	0.59	0.5564	0.2724	0.2563
No log	36.0	252	24.5679	0.57	0.5709	3.1573	0.57	0.5089	0.2523	0.2222
No log	37.0	259	24.5375	0.575	0.5631	2.9349	0.575	0.5381	0.2279	0.2007
No log	38.0	266	24.6423	0.565	0.6072	2.6772	0.565	0.5340	0.2587	0.2247
No log	39.0	273	24.6706	0.575	0.6139	2.9241	0.575	0.5291	0.2318	0.2416
No log	40.0	280	24.6007	0.575	0.5774	2.9918	0.575	0.5323	0.2575	0.2138
No log	41.0	287	24.7587	0.565	0.6231	2.9588	0.565	0.5023	0.2685	0.2665
No log	42.0	294	24.5681	0.56	0.5786	2.9999	0.56	0.5153	0.2558	0.2093
No log	43.0	301	24.5971	0.59	0.5687	3.0595	0.59	0.5365	0.2532	0.2004
No log	44.0	308	24.6424	0.58	0.5918	2.9073	0.58	0.5432	0.2470	0.2113
No log	45.0	315	24.5998	0.58	0.5705	3.0442	0.58	0.5488	0.2769	0.2011
No log	46.0	322	24.5625	0.62	0.5561	2.9855	0.62	0.5869	0.2492	0.2069
No log	47.0	329	24.6409	0.57	0.5817	2.8587	0.57	0.5400	0.2480	0.2239
No log	48.0	336	24.6218	0.57	0.5958	2.8299	0.57	0.5426	0.2725	0.2251
No log	49.0	343	24.5568	0.585	0.5762	2.9178	0.585	0.5590	0.2374	0.2102
No log	50.0	350	24.6607	0.57	0.6012	2.9238	0.57	0.5344	0.2496	0.2274

Framework versions

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