vit-hybrid-base-bit-384_rice-leaf-disease-augmented-v4_v5_pft

This model is a fine-tuned version of google/vit-hybrid-base-bit-384 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2706
• Accuracy: 0.9195

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: 0.0003
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• 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: cosine_with_restarts
• lr_scheduler_warmup_steps: 256
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.2177	0.5	64	2.0234	0.2248
1.7719	1.0	128	1.4579	0.5872
1.258	1.5	192	1.0302	0.7081
0.9323	2.0	256	0.7450	0.8154
0.7252	2.5	320	0.6527	0.7953
0.6134	3.0	384	0.5488	0.8154
0.5375	3.5	448	0.5004	0.8221
0.5028	4.0	512	0.4624	0.8557
0.471	4.5	576	0.4532	0.8557
0.44	5.0	640	0.4378	0.8557
0.4302	5.5	704	0.4446	0.8423
0.4267	6.0	768	0.4300	0.8591
0.4341	6.5	832	0.4302	0.8591
0.406	7.0	896	0.4174	0.8658
0.4077	7.5	960	0.3973	0.8523
0.3639	8.0	1024	0.3747	0.8792
0.3463	8.5	1088	0.3701	0.8859
0.343	9.0	1152	0.3682	0.8859
0.322	9.5	1216	0.3567	0.8792
0.3224	10.0	1280	0.3555	0.8859
0.3103	10.5	1344	0.3529	0.8859
0.314	11.0	1408	0.3531	0.8859
0.3153	11.5	1472	0.3546	0.8859
0.3033	12.0	1536	0.3434	0.8792
0.2905	12.5	1600	0.3326	0.8859
0.2857	13.0	1664	0.3323	0.8893
0.2693	13.5	1728	0.3238	0.8893
0.2683	14.0	1792	0.3273	0.9027
0.2582	14.5	1856	0.3243	0.9060
0.2544	15.0	1920	0.3181	0.8993
0.2478	15.5	1984	0.3167	0.8993
0.255	16.0	2048	0.3166	0.8993
0.2586	16.5	2112	0.3087	0.8993
0.24	17.0	2176	0.3126	0.9060
0.2351	17.5	2240	0.3032	0.9027
0.2302	18.0	2304	0.3005	0.9094
0.2229	18.5	2368	0.2993	0.9128
0.2185	19.0	2432	0.2982	0.9027
0.2138	19.5	2496	0.2968	0.9027
0.2128	20.0	2560	0.2952	0.9027
0.2134	20.5	2624	0.2946	0.9027
0.2107	21.0	2688	0.3014	0.8993
0.2077	21.5	2752	0.2885	0.9060
0.2073	22.0	2816	0.2911	0.9094
0.1943	22.5	2880	0.2853	0.9128
0.1979	23.0	2944	0.2806	0.9094
0.1907	23.5	3008	0.2793	0.9161
0.1848	24.0	3072	0.2794	0.9094
0.1884	24.5	3136	0.2780	0.9094
0.179	25.0	3200	0.2778	0.9094
0.1872	25.5	3264	0.2828	0.9128
0.181	26.0	3328	0.2749	0.9128
0.1779	26.5	3392	0.2752	0.9094
0.1783	27.0	3456	0.2730	0.9128
0.1777	27.5	3520	0.2720	0.9195
0.162	28.0	3584	0.2717	0.9128
0.1682	28.5	3648	0.2679	0.9128
0.1599	29.0	3712	0.2709	0.9128
0.1587	29.5	3776	0.2711	0.9161
0.164	30.0	3840	0.2706	0.9195

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1+cu124
• Datasets 3.3.2
• Tokenizers 0.21.1