vit-base-patch16-384-finetuned-humid-classes-4

This model is a fine-tuned version of google/vit-base-patch16-384 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.1728
• Accuracy: 0.9589
• F1 Macro: 0.9401
• Precision Macro: 0.9476
• Recall Macro: 0.9415
• Precision Dry: 1.0
• Recall Dry: 0.9474
• F1 Dry: 0.9730
• Precision Firm: 1.0
• Recall Firm: 0.96
• F1 Firm: 0.9796
• Precision Humid: 1.0
• Recall Humid: 0.8
• F1 Humid: 0.8889
• Precision Lump: 0.9048
• Recall Lump: 1.0
• F1 Lump: 0.95
• Precision Rockies: 0.8333
• Recall Rockies: 1.0
• F1 Rockies: 0.9091

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• 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
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Macro	Precision Macro	Recall Macro	Precision Dry	Recall Dry	F1 Dry	Precision Firm	Recall Firm	F1 Firm	Precision Humid	Recall Humid	F1 Humid	Precision Lump	Recall Lump	F1 Lump	Precision Rockies	Recall Rockies	F1 Rockies
No log	1.0	7	1.4385	0.4247	0.2252	0.2357	0.2783	0.3333	0.0526	0.0909	0.4524	0.76	0.5672	0.0	0.0	0.0	0.3929	0.5789	0.4681	0.0	0.0	0.0
1.6349	2.0	14	0.9416	0.7945	0.51	0.4834	0.5524	0.6552	1.0	0.7917	1.0	0.92	0.9583	0.0	0.0	0.0	0.7619	0.8421	0.8	0.0	0.0	0.0
0.9402	3.0	21	0.6032	0.8082	0.5225	0.4944	0.5655	0.9	0.9474	0.9231	0.9167	0.88	0.8980	0.0	0.0	0.0	0.6552	1.0	0.7917	0.0	0.0	0.0
0.9402	4.0	28	0.3723	0.8493	0.6441	0.6607	0.6535	1.0	0.9474	0.9730	0.9583	0.92	0.9388	0.0	0.0	0.0	0.6786	1.0	0.8085	0.6667	0.4	0.5
0.4561	5.0	35	0.3619	0.8767	0.8169	0.7978	0.8808	1.0	0.9474	0.9730	1.0	0.92	0.9583	0.5	0.8	0.6154	0.9333	0.7368	0.8235	0.5556	1.0	0.7143
0.2823	6.0	42	0.3235	0.8904	0.7194	0.7062	0.7415	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.0	0.0	0.0	0.7308	1.0	0.8444	0.8	0.8	0.8
0.2823	7.0	49	0.2951	0.8493	0.7928	0.7949	0.8598	1.0	0.9474	0.9730	1.0	0.92	0.9583	0.3846	1.0	0.5556	0.9231	0.6316	0.75	0.6667	0.8	0.7273
0.1837	8.0	56	0.2225	0.9178	0.8482	0.9081	0.8535	1.0	0.9474	0.9730	1.0	0.92	0.9583	1.0	0.4	0.5714	0.8261	1.0	0.9048	0.7143	1.0	0.8333
0.1518	9.0	63	0.2789	0.9315	0.8975	0.8829	0.9229	1.0	0.9474	0.9730	1.0	0.92	0.9583	0.7143	1.0	0.8333	0.9	0.9474	0.9231	0.8	0.8	0.8
0.1076	10.0	70	0.1728	0.9589	0.9401	0.9476	0.9415	1.0	0.9474	0.9730	1.0	0.96	0.9796	1.0	0.8	0.8889	0.9048	1.0	0.95	0.8333	1.0	0.9091
0.1076	11.0	77	0.2158	0.9178	0.8599	0.8734	0.8509	1.0	0.9474	0.9730	0.96	0.96	0.96	0.75	0.6	0.6667	0.8571	0.9474	0.9	0.8	0.8	0.8
0.0371	12.0	84	0.1969	0.9589	0.9298	0.9323	0.9415	1.0	1.0	1.0	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	1.0	0.8	0.8889
0.0299	13.0	91	0.2114	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0299	14.0	98	0.1714	0.9589	0.9272	0.9167	0.9415	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.8333	1.0	0.9091	0.95	1.0	0.9744	0.8	0.8	0.8
0.0091	15.0	105	0.2330	0.9315	0.8970	0.8837	0.9204	1.0	0.9474	0.9730	0.96	0.96	0.96	0.7143	1.0	0.8333	0.9444	0.8947	0.9189	0.8	0.8	0.8
0.0044	16.0	112	0.3195	0.9315	0.8975	0.8829	0.9229	1.0	0.9474	0.9730	1.0	0.92	0.9583	0.7143	1.0	0.8333	0.9	0.9474	0.9231	0.8	0.8	0.8
0.0044	17.0	119	0.2089	0.9315	0.8806	0.8733	0.8909	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.6667	0.8	0.7273	0.9	0.9474	0.9231	0.8	0.8	0.8
0.0026	18.0	126	0.2572	0.9589	0.9436	0.9228	0.9709	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.8333	1.0	0.9091	0.9474	0.9474	0.9474	0.8333	1.0	0.9091
0.0029	19.0	133	0.2714	0.9315	0.9081	0.8981	0.9229	1.0	0.9474	0.9730	1.0	0.92	0.9583	0.8	0.8	0.8	0.8571	0.9474	0.9	0.8333	1.0	0.9091
0.0012	20.0	140	0.2290	0.9452	0.9169	0.9067	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.8	0.8	0.8	0.9	0.9474	0.9231	0.8333	1.0	0.9091
0.0012	21.0	147	0.2409	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0009	22.0	154	0.2617	0.9315	0.8970	0.8837	0.9204	1.0	0.9474	0.9730	0.96	0.96	0.96	0.7143	1.0	0.8333	0.9444	0.8947	0.9189	0.8	0.8	0.8
0.0009	23.0	161	0.2603	0.9315	0.8970	0.8837	0.9204	1.0	0.9474	0.9730	0.96	0.96	0.96	0.7143	1.0	0.8333	0.9444	0.8947	0.9189	0.8	0.8	0.8
0.0009	24.0	168	0.2516	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0007	25.0	175	0.2556	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0006	26.0	182	0.2653	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0006	27.0	189	0.2760	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0006	28.0	196	0.2804	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0005	29.0	203	0.2809	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0005	30.0	210	0.2781	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0005	31.0	217	0.2782	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0005	32.0	224	0.2784	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0004	33.0	231	0.2775	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0004	34.0	238	0.2773	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0004	35.0	245	0.2780	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0004	36.0	252	0.2817	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0004	37.0	259	0.2817	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0004	38.0	266	0.2796	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0004	39.0	273	0.2773	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0004	40.0	280	0.2752	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0004	41.0	287	0.2753	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0004	42.0	294	0.2785	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0004	43.0	301	0.2797	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0004	44.0	308	0.2793	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0004	45.0	315	0.2783	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0003	46.0	322	0.2774	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0003	47.0	329	0.2770	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0004	48.0	336	0.2768	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0004	49.0	343	0.2768	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8
0.0003	50.0	350	0.2768	0.9452	0.9067	0.8923	0.9309	1.0	0.9474	0.9730	1.0	0.96	0.9796	0.7143	1.0	0.8333	0.9474	0.9474	0.9474	0.8	0.8	0.8

Framework versions

• Transformers 4.56.1
• Pytorch 2.5.1+cu124
• Datasets 4.0.0
• Tokenizers 0.22.0