plant-seedlings-model-swin

This model is a fine-tuned version of microsoft/swin-base-patch4-window7-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.2169
• Accuracy: 0.9474

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: 0.0002
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.8259	0.2	100	0.7181	0.7520
1.0121	0.39	200	0.7504	0.7092
0.5952	0.59	300	0.6254	0.7986
0.6031	0.79	400	0.4595	0.8438
0.637	0.98	500	0.5830	0.8080
0.5896	1.18	600	0.5042	0.8384
0.6758	1.38	700	0.4827	0.8325
0.543	1.57	800	0.4713	0.8433
0.3312	1.77	900	0.4752	0.8546
0.5559	1.96	1000	0.4578	0.8369
0.4303	2.16	1100	0.5034	0.8389
0.5705	2.36	1200	0.4322	0.8502
0.5369	2.55	1300	0.4646	0.8404
0.3628	2.75	1400	0.3984	0.8659
0.4071	2.95	1500	0.3872	0.8689
0.4988	3.14	1600	0.3543	0.8792
0.4607	3.34	1700	0.3933	0.8674
0.3342	3.54	1800	0.3883	0.8639
0.4141	3.73	1900	0.3886	0.8644
0.5513	3.93	2000	0.3335	0.8900
0.4659	4.13	2100	0.4286	0.8590
0.2263	4.32	2200	0.3587	0.8772
0.4518	4.52	2300	0.3332	0.8870
0.3422	4.72	2400	0.2723	0.9062
0.6113	4.91	2500	0.2811	0.9057
0.3636	5.11	2600	0.3157	0.8939
0.2794	5.3	2700	0.2773	0.9082
0.3486	5.5	2800	0.3099	0.8978
0.2563	5.7	2900	0.3077	0.9052
0.3709	5.89	3000	0.3650	0.8836
0.3732	6.09	3100	0.3132	0.8988
0.2218	6.29	3200	0.2947	0.9052
0.2488	6.48	3300	0.2737	0.9131
0.2689	6.68	3400	0.3471	0.8924
0.3212	6.88	3500	0.3447	0.8905
0.3604	7.07	3600	0.2974	0.9086
0.2492	7.27	3700	0.3057	0.8993
0.1674	7.47	3800	0.3241	0.9032
0.3248	7.66	3900	0.2952	0.9077
0.204	7.86	4000	0.2883	0.9111
0.2783	8.06	4100	0.3017	0.9047
0.3721	8.25	4200	0.2782	0.9136
0.2554	8.45	4300	0.2625	0.9170
0.1104	8.64	4400	0.2590	0.9190
0.247	8.84	4500	0.3021	0.9096
0.3316	9.04	4600	0.3190	0.8988
0.3214	9.23	4700	0.2883	0.9140
0.192	9.43	4800	0.2770	0.9155
0.3568	9.63	4900	0.2475	0.9229
0.3365	9.82	5000	0.2568	0.9229
0.1226	10.02	5100	0.2534	0.9204
0.2359	10.22	5200	0.2679	0.9131
0.1623	10.41	5300	0.3127	0.9204
0.2369	10.61	5400	0.2779	0.9170
0.1234	10.81	5500	0.2486	0.9273
0.1823	11.0	5600	0.2608	0.9239
0.2875	11.2	5700	0.2612	0.9190
0.1408	11.39	5800	0.2208	0.9298
0.1094	11.59	5900	0.2399	0.9332
0.213	11.79	6000	0.2636	0.9209
0.1599	11.98	6100	0.2458	0.9249
0.2565	12.18	6200	0.2698	0.9204
0.0773	12.38	6300	0.2348	0.9322
0.1515	12.57	6400	0.2370	0.9263
0.2308	12.77	6500	0.2185	0.9307
0.2009	12.97	6600	0.2211	0.9342
0.2126	13.16	6700	0.2552	0.9342
0.1348	13.36	6800	0.2206	0.9371
0.1473	13.56	6900	0.2199	0.9357
0.1861	13.75	7000	0.2512	0.9224
0.1136	13.95	7100	0.2803	0.9214
0.1726	14.15	7200	0.2201	0.9361
0.202	14.34	7300	0.2105	0.9371
0.2043	14.54	7400	0.2472	0.9263
0.1427	14.73	7500	0.2250	0.9381
0.1599	14.93	7600	0.2270	0.9391
0.1216	15.13	7700	0.2409	0.9307
0.2869	15.32	7800	0.2208	0.9386
0.1254	15.52	7900	0.2298	0.9332
0.1314	15.72	8000	0.1959	0.9416
0.1106	15.91	8100	0.2183	0.9342
0.2211	16.11	8200	0.2581	0.9337
0.1589	16.31	8300	0.2091	0.9381
0.0791	16.5	8400	0.1792	0.9455
0.0849	16.7	8500	0.2481	0.9298
0.089	16.9	8600	0.2143	0.9386
0.0609	17.09	8700	0.2020	0.9524
0.1509	17.29	8800	0.2039	0.9396
0.0934	17.49	8900	0.2242	0.9322
0.0398	17.68	9000	0.1891	0.9460
0.1106	17.88	9100	0.1939	0.9470
0.1742	18.07	9200	0.1965	0.9479
0.1015	18.27	9300	0.1886	0.9440
0.089	18.47	9400	0.1851	0.9479
0.1393	18.66	9500	0.1844	0.9484
0.0849	18.86	9600	0.2205	0.9396
0.0708	19.06	9700	0.1888	0.9435
0.1037	19.25	9800	0.2070	0.9450
0.1109	19.45	9900	0.2079	0.9460
0.0533	19.65	10000	0.2036	0.9489
0.0757	19.84	10100	0.2169	0.9474

Framework versions

• Transformers 4.28.1
• Pytorch 2.0.0+cu118
• Datasets 2.11.0
• Tokenizers 0.13.3