JaesonGu/flare-body-vit

charger-classif-model

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2678
• Accuracy: 0.9231

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: 5e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 4
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 8

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.4057	0.0769	1	0.5508	0.6923
0.5194	0.1538	2	0.5735	0.6923
0.4141	0.2308	3	0.5007	0.7692
0.5442	0.3077	4	0.5160	0.8462
0.43	0.3846	5	0.5931	0.7692
0.4126	0.4615	6	0.5228	0.7692
0.4151	0.5385	7	0.5552	0.7692
0.3753	0.6154	8	0.5825	0.6154
0.3468	0.6923	9	0.5637	0.6923
0.3467	0.7692	10	0.5148	0.6923
0.5188	0.8462	11	0.4735	0.7692
0.4342	0.9231	12	0.5058	0.7692
0.3888	1.0	13	0.5176	0.6923
0.3977	1.0769	14	0.4865	0.7692
0.1799	1.1538	15	0.5299	0.6923
0.4628	1.2308	16	0.5614	0.6923
0.8787	1.3077	17	0.5826	0.6923
0.3396	1.3846	18	0.5337	0.7692
0.2144	1.4615	19	0.5531	0.6923
0.242	1.5385	20	0.5317	0.6923
1.1866	1.6154	21	0.5042	0.6923
0.2689	1.6923	22	0.4067	0.8462
0.3953	1.7692	23	0.4513	0.8462
0.1978	1.8462	24	0.5103	0.6923
0.3293	1.9231	25	0.4829	0.6923
0.3324	2.0	26	0.4915	0.8462
0.2096	2.0769	27	0.5136	0.8462
0.4142	2.1538	28	0.4490	0.7692
0.4267	2.2308	29	0.4697	0.7692
0.1871	2.3077	30	0.4744	0.7692
0.3145	2.3846	31	0.5596	0.6923
0.3417	2.4615	32	0.4589	0.6923
0.1548	2.5385	33	0.5245	0.6923
0.3131	2.6154	34	0.4507	0.6923
0.1974	2.6923	35	0.4068	0.8462
0.3148	2.7692	36	0.5019	0.6923
0.5036	2.8462	37	0.4761	0.6923
0.2178	2.9231	38	0.4132	0.9231
0.4536	3.0	39	0.4745	0.7692
0.3118	3.0769	40	0.4869	0.7692
0.3465	3.1538	41	0.4473	0.7692
0.096	3.2308	42	0.4376	0.8462
0.1726	3.3077	43	0.5971	0.7692
0.1685	3.3846	44	0.4768	0.7692
0.2046	3.4615	45	0.3595	0.8462
0.1297	3.5385	46	0.4701	0.7692
0.4597	3.6154	47	0.4054	0.7692
0.3474	3.6923	48	0.3927	0.8462
0.4476	3.7692	49	0.5063	0.8462
0.1062	3.8462	50	0.4741	0.7692
0.5484	3.9231	51	0.4950	0.6923
0.0945	4.0	52	0.4647	0.7692
0.1053	4.0769	53	0.3743	0.8462
0.4122	4.1538	54	0.4350	0.8462
0.2825	4.2308	55	0.4246	0.8462
0.2912	4.3077	56	0.5250	0.6923
0.3193	4.3846	57	0.3639	0.8462
0.066	4.4615	58	0.3574	0.9231
0.0888	4.5385	59	0.4897	0.6923
0.1046	4.6154	60	0.3032	0.9231
0.2573	4.6923	61	0.5662	0.6154
0.368	4.7692	62	0.3699	0.8462
0.1484	4.8462	63	0.3517	0.8462
0.1444	4.9231	64	0.2988	0.9231
0.1492	5.0	65	0.3523	0.8462
0.112	5.0769	66	0.4245	0.8462
0.0711	5.1538	67	0.4451	0.6923
0.2455	5.2308	68	0.4774	0.7692
0.3981	5.3077	69	0.5084	0.7692
0.1682	5.3846	70	0.4053	0.8462
0.2809	5.4615	71	0.4574	0.6923
0.1929	5.5385	72	0.3242	0.7692
0.161	5.6154	73	0.3854	0.7692
0.1475	5.6923	74	0.3935	0.7692
0.1058	5.7692	75	0.5751	0.6923
0.1103	5.8462	76	0.3874	0.8462
0.1057	5.9231	77	0.3984	0.7692
0.1593	6.0	78	0.3299	0.8462
0.1154	6.0769	79	0.4778	0.7692
0.3131	6.1538	80	0.4863	0.7692
0.0791	6.2308	81	0.4897	0.7692
0.0635	6.3077	82	0.5831	0.7692
0.0704	6.3846	83	0.4384	0.8462
0.0597	6.4615	84	0.5519	0.7692
0.1117	6.5385	85	0.4525	0.7692
0.1542	6.6154	86	0.5354	0.8462
0.5737	6.6923	87	0.5034	0.7692
0.4216	6.7692	88	0.4514	0.7692
0.3276	6.8462	89	0.5688	0.7692
0.119	6.9231	90	0.3433	0.9231
0.1519	7.0	91	0.4454	0.7692
0.1155	7.0769	92	0.3323	0.7692
0.1264	7.1538	93	0.4030	0.6923
0.0585	7.2308	94	0.3404	0.8462
0.1404	7.3077	95	0.3507	0.8462
0.0417	7.3846	96	0.4860	0.7692
0.0873	7.4615	97	0.4896	0.8462
0.0801	7.5385	98	0.4383	0.7692
0.2163	7.6154	99	0.3764	0.8462
0.1823	7.6923	100	0.4258	0.8462
0.1832	7.7692	101	0.2890	0.8462
0.0879	7.8462	102	0.2909	0.8462
0.2345	7.9231	103	0.3617	0.8462
0.1096	8.0	104	0.2678	0.9231

Framework versions

• Transformers 4.48.3
• Pytorch 2.6.0+cpu
• Datasets 3.2.0
• Tokenizers 0.21.0