segformer-b0-finetuned-lipid-droplets-v2

This model is a fine-tuned version of nvidia/mit-b0 on the jhaberbe/lipid-droplets-v3 dataset. It achieves the following results on the evaluation set:

• Loss: 0.1451
• Mean Iou: 0.3767
• Mean Accuracy: 0.7533
• Overall Accuracy: 0.7533
• Accuracy Unlabeled: nan
• Accuracy Lipid: 0.7533
• Iou Unlabeled: 0.0
• Iou Lipid: 0.7533

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: 6e-05
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Unlabeled	Accuracy Lipid	Iou Unlabeled	Iou Lipid
0.5878	1.82	20	0.6215	0.2190	0.4381	0.4381	nan	0.4381	0.0	0.4381
0.4506	3.64	40	0.4553	0.2542	0.5084	0.5084	nan	0.5084	0.0	0.5084
0.3435	5.45	60	0.3476	0.2703	0.5405	0.5405	nan	0.5405	0.0	0.5405
0.2931	7.27	80	0.2191	0.2148	0.4296	0.4296	nan	0.4296	0.0	0.4296
0.2361	9.09	100	0.2199	0.2829	0.5659	0.5659	nan	0.5659	0.0	0.5659
0.2739	10.91	120	0.2493	0.2668	0.5336	0.5336	nan	0.5336	0.0	0.5336
0.1993	12.73	140	0.2123	0.3617	0.7234	0.7234	nan	0.7234	0.0	0.7234
0.1226	14.55	160	0.0776	0.1070	0.2140	0.2140	nan	0.2140	0.0	0.2140
0.1348	16.36	180	0.1453	0.3343	0.6686	0.6686	nan	0.6686	0.0	0.6686
0.0939	18.18	200	0.1973	0.3632	0.7265	0.7265	nan	0.7265	0.0	0.7265
0.1426	20.0	220	0.1698	0.2628	0.5256	0.5256	nan	0.5256	0.0	0.5256
0.0722	21.82	240	0.1061	0.3520	0.7040	0.7040	nan	0.7040	0.0	0.7040
0.0764	23.64	260	0.1653	0.2990	0.5979	0.5979	nan	0.5979	0.0	0.5979
0.0551	25.45	280	0.1072	0.3103	0.6206	0.6206	nan	0.6206	0.0	0.6206
0.0437	27.27	300	0.2012	0.3494	0.6988	0.6988	nan	0.6988	0.0	0.6988
0.0514	29.09	320	0.1825	0.3777	0.7553	0.7553	nan	0.7553	0.0	0.7553
0.1095	30.91	340	0.0897	0.3392	0.6785	0.6785	nan	0.6785	0.0	0.6785
0.0682	32.73	360	0.1785	0.3504	0.7008	0.7008	nan	0.7008	0.0	0.7008
0.0422	34.55	380	0.1167	0.3444	0.6887	0.6887	nan	0.6887	0.0	0.6887
0.0538	36.36	400	0.2332	0.4529	0.9057	0.9057	nan	0.9057	0.0	0.9057
0.0347	38.18	420	0.1698	0.3115	0.6231	0.6231	nan	0.6231	0.0	0.6231
0.0496	40.0	440	0.1201	0.3278	0.6555	0.6555	nan	0.6555	0.0	0.6555
0.0681	41.82	460	0.1830	0.3916	0.7831	0.7831	nan	0.7831	0.0	0.7831
0.0498	43.64	480	0.1848	0.4086	0.8172	0.8172	nan	0.8172	0.0	0.8172
0.0365	45.45	500	0.1234	0.3741	0.7481	0.7481	nan	0.7481	0.0	0.7481
0.0258	47.27	520	0.2000	0.4300	0.8599	0.8599	nan	0.8599	0.0	0.8599
0.0355	49.09	540	0.1273	0.3907	0.7814	0.7814	nan	0.7814	0.0	0.7814
0.0476	50.91	560	0.1827	0.4322	0.8644	0.8644	nan	0.8644	0.0	0.8644
0.0472	52.73	580	0.1014	0.3420	0.6839	0.6839	nan	0.6839	0.0	0.6839
0.0467	54.55	600	0.1330	0.3758	0.7516	0.7516	nan	0.7516	0.0	0.7516
0.0802	56.36	620	0.0698	0.3073	0.6147	0.6147	nan	0.6147	0.0	0.6147
0.0558	58.18	640	0.1291	0.4088	0.8176	0.8176	nan	0.8176	0.0	0.8176
0.0443	60.0	660	0.1097	0.4060	0.8119	0.8119	nan	0.8119	0.0	0.8119
0.0442	61.82	680	0.1112	0.3944	0.7887	0.7887	nan	0.7887	0.0	0.7887
0.0422	63.64	700	0.1837	0.4288	0.8576	0.8576	nan	0.8576	0.0	0.8576
0.0216	65.45	720	0.1140	0.3735	0.7470	0.7470	nan	0.7470	0.0	0.7470
0.0414	67.27	740	0.1017	0.3821	0.7643	0.7643	nan	0.7643	0.0	0.7643
0.0336	69.09	760	0.1458	0.3685	0.7370	0.7370	nan	0.7370	0.0	0.7370
0.0575	70.91	780	0.1392	0.3425	0.6851	0.6851	nan	0.6851	0.0	0.6851
0.0324	72.73	800	0.1162	0.3689	0.7377	0.7377	nan	0.7377	0.0	0.7377
0.0336	74.55	820	0.1366	0.4143	0.8287	0.8287	nan	0.8287	0.0	0.8287
0.0889	76.36	840	0.1604	0.3726	0.7452	0.7452	nan	0.7452	0.0	0.7452
0.0438	78.18	860	0.1528	0.3948	0.7895	0.7895	nan	0.7895	0.0	0.7895
0.0194	80.0	880	0.1360	0.3335	0.6671	0.6671	nan	0.6671	0.0	0.6671
0.0311	81.82	900	0.1832	0.4278	0.8555	0.8555	nan	0.8555	0.0	0.8555
0.0478	83.64	920	0.0718	0.2656	0.5312	0.5312	nan	0.5312	0.0	0.5312
0.0363	85.45	940	0.1281	0.3630	0.7259	0.7259	nan	0.7259	0.0	0.7259
0.0354	87.27	960	0.0986	0.3891	0.7782	0.7782	nan	0.7782	0.0	0.7782
0.0195	89.09	980	0.1582	0.4123	0.8247	0.8247	nan	0.8247	0.0	0.8247
0.0432	90.91	1000	0.1086	0.3217	0.6433	0.6433	nan	0.6433	0.0	0.6433
0.0288	92.73	1020	0.1669	0.4089	0.8179	0.8179	nan	0.8179	0.0	0.8179
0.0326	94.55	1040	0.1335	0.3908	0.7816	0.7816	nan	0.7816	0.0	0.7816
0.0384	96.36	1060	0.1415	0.3947	0.7895	0.7895	nan	0.7895	0.0	0.7895
0.0399	98.18	1080	0.1784	0.4204	0.8408	0.8408	nan	0.8408	0.0	0.8408
0.0214	100.0	1100	0.1451	0.3767	0.7533	0.7533	nan	0.7533	0.0	0.7533

Framework versions

• Transformers 4.37.2
• Pytorch 2.1.0+cu121
• Datasets 2.17.1
• Tokenizers 0.15.2