segformer-b0-finetuned-segments-sidewalk-2

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

• Loss: 0.0552
• Mean Iou: 0.4126
• Mean Accuracy: 0.8251
• Overall Accuracy: 0.8251
• Accuracy Background: nan
• Accuracy Amyloid: 0.8251
• Iou Background: 0.0
• Iou Amyloid: 0.8251

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: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Background	Accuracy Amyloid	Iou Background	Iou Amyloid
0.5486	0.2326	20	0.6709	0.1174	0.2349	0.2349	nan	0.2349	0.0	0.2349
0.3496	0.4651	40	0.6179	0.2210	0.4420	0.4420	nan	0.4420	0.0	0.4420
0.2825	0.6977	60	0.4770	0.4871	0.9743	0.9743	nan	0.9743	0.0	0.9743
0.2814	0.9302	80	0.2872	0.1395	0.2790	0.2790	nan	0.2790	0.0	0.2790
0.2083	1.1628	100	0.2249	0.3871	0.7742	0.7742	nan	0.7742	0.0	0.7742
0.1355	1.3953	120	0.2623	0.2753	0.5507	0.5507	nan	0.5507	0.0	0.5507
0.1671	1.6279	140	0.1225	0.2283	0.4565	0.4565	nan	0.4565	0.0	0.4565
0.1035	1.8605	160	0.0822	0.1782	0.3564	0.3564	nan	0.3564	0.0	0.3564
0.097	2.0930	180	0.0783	0.1628	0.3255	0.3255	nan	0.3255	0.0	0.3255
0.0641	2.3256	200	0.1464	0.4488	0.8976	0.8976	nan	0.8976	0.0	0.8976
0.06	2.5581	220	0.0819	0.3430	0.6860	0.6860	nan	0.6860	0.0	0.6860
0.0511	2.7907	240	0.0835	0.2525	0.5049	0.5049	nan	0.5049	0.0	0.5049
0.0377	3.0233	260	0.0681	0.2144	0.4288	0.4288	nan	0.4288	0.0	0.4288
0.0381	3.2558	280	0.0975	0.4299	0.8598	0.8598	nan	0.8598	0.0	0.8598
0.0377	3.4884	300	0.1277	0.4671	0.9342	0.9342	nan	0.9342	0.0	0.9342
0.1049	3.7209	320	0.0754	0.4208	0.8417	0.8417	nan	0.8417	0.0	0.8417
0.0252	3.9535	340	0.0757	0.2233	0.4467	0.4467	nan	0.4467	0.0	0.4467
0.022	4.1860	360	0.0758	0.3314	0.6628	0.6628	nan	0.6628	0.0	0.6628
0.0223	4.4186	380	0.0840	0.3403	0.6806	0.6806	nan	0.6806	0.0	0.6806
0.0323	4.6512	400	0.0490	0.4347	0.8694	0.8694	nan	0.8694	0.0	0.8694
0.0298	4.8837	420	0.0618	0.3695	0.7391	0.7391	nan	0.7391	0.0	0.7391
0.0172	5.1163	440	0.0918	0.3770	0.7541	0.7541	nan	0.7541	0.0	0.7541
0.0171	5.3488	460	0.0955	0.1962	0.3924	0.3924	nan	0.3924	0.0	0.3924
0.0134	5.5814	480	0.1008	0.2416	0.4832	0.4832	nan	0.4832	0.0	0.4832
0.0153	5.8140	500	0.0997	0.2675	0.5351	0.5351	nan	0.5351	0.0	0.5351
0.0114	6.0465	520	0.0788	0.3411	0.6823	0.6823	nan	0.6823	0.0	0.6823
0.0126	6.2791	540	0.0947	0.0264	0.0527	0.0527	nan	0.0527	0.0	0.0527
0.0134	6.5116	560	0.0728	0.4199	0.8399	0.8399	nan	0.8399	0.0	0.8399
0.0098	6.7442	580	0.0522	0.2287	0.4574	0.4574	nan	0.4574	0.0	0.4574
0.0094	6.9767	600	0.0869	0.2789	0.5578	0.5578	nan	0.5578	0.0	0.5578
0.0123	7.2093	620	0.1064	0.3969	0.7937	0.7937	nan	0.7937	0.0	0.7937
0.017	7.4419	640	0.0700	0.2119	0.4237	0.4237	nan	0.4237	0.0	0.4237
0.0092	7.6744	660	0.0471	0.1847	0.3695	0.3695	nan	0.3695	0.0	0.3695
0.0083	7.9070	680	0.0701	0.3242	0.6485	0.6485	nan	0.6485	0.0	0.6485
0.0102	8.1395	700	0.0793	0.3773	0.7547	0.7547	nan	0.7547	0.0	0.7547
0.0114	8.3721	720	0.0596	0.3937	0.7873	0.7873	nan	0.7873	0.0	0.7873
0.0069	8.6047	740	0.0646	0.3988	0.7976	0.7976	nan	0.7976	0.0	0.7976
0.0084	8.8372	760	0.0686	0.3288	0.6577	0.6577	nan	0.6577	0.0	0.6577
0.008	9.0698	780	0.0658	0.3455	0.6910	0.6910	nan	0.6910	0.0	0.6910
0.0065	9.3023	800	0.0185	0.1691	0.3382	0.3382	nan	0.3382	0.0	0.3382
0.0063	9.5349	820	0.0732	0.2238	0.4477	0.4477	nan	0.4477	0.0	0.4477
0.0058	9.7674	840	0.0457	0.2353	0.4705	0.4705	nan	0.4705	0.0	0.4705
0.0063	10.0	860	0.0777	0.4177	0.8354	0.8354	nan	0.8354	0.0	0.8354
0.0218	10.2326	880	0.0761	0.2859	0.5718	0.5718	nan	0.5718	0.0	0.5718
0.0064	10.4651	900	0.0651	0.3579	0.7159	0.7159	nan	0.7159	0.0	0.7159
0.0052	10.6977	920	0.0776	0.3691	0.7383	0.7383	nan	0.7383	0.0	0.7383
0.0064	10.9302	940	0.0615	0.2193	0.4386	0.4386	nan	0.4386	0.0	0.4386
0.01	11.1628	960	0.0513	0.2410	0.4820	0.4820	nan	0.4820	0.0	0.4820
0.0106	11.3953	980	0.0739	0.3586	0.7172	0.7172	nan	0.7172	0.0	0.7172
0.0173	11.6279	1000	0.0189	0.3301	0.6601	0.6601	nan	0.6601	0.0	0.6601
0.0059	11.8605	1020	0.0555	0.3845	0.7689	0.7689	nan	0.7689	0.0	0.7689
0.0087	12.0930	1040	0.0496	0.3700	0.7399	0.7399	nan	0.7399	0.0	0.7399
0.0046	12.3256	1060	0.0709	0.3949	0.7899	0.7899	nan	0.7899	0.0	0.7899
0.0047	12.5581	1080	0.0866	0.3646	0.7292	0.7292	nan	0.7292	0.0	0.7292
0.0044	12.7907	1100	0.0595	0.3145	0.6291	0.6291	nan	0.6291	0.0	0.6291
0.005	13.0233	1120	0.0922	0.3124	0.6248	0.6248	nan	0.6248	0.0	0.6248
0.0043	13.2558	1140	0.0658	0.3818	0.7636	0.7636	nan	0.7636	0.0	0.7636
0.004	13.4884	1160	0.0800	0.3649	0.7297	0.7297	nan	0.7297	0.0	0.7297
0.0067	13.7209	1180	0.0537	0.3581	0.7161	0.7161	nan	0.7161	0.0	0.7161
0.0043	13.9535	1200	0.0735	0.2332	0.4663	0.4663	nan	0.4663	0.0	0.4663
0.0047	14.1860	1220	0.0647	0.3186	0.6373	0.6373	nan	0.6373	0.0	0.6373
0.0138	14.4186	1240	0.0663	0.3183	0.6366	0.6366	nan	0.6366	0.0	0.6366
0.0052	14.6512	1260	0.0504	0.3514	0.7027	0.7027	nan	0.7027	0.0	0.7027
0.0037	14.8837	1280	0.0839	0.3544	0.7087	0.7087	nan	0.7087	0.0	0.7087
0.0075	15.1163	1300	0.0708	0.4158	0.8316	0.8316	nan	0.8316	0.0	0.8316
0.0122	15.3488	1320	0.0835	0.3907	0.7813	0.7813	nan	0.7813	0.0	0.7813
0.0034	15.5814	1340	0.0808	0.1924	0.3848	0.3848	nan	0.3848	0.0	0.3848
0.0037	15.8140	1360	0.0619	0.3453	0.6907	0.6907	nan	0.6907	0.0	0.6907
0.0091	16.0465	1380	0.0918	0.4024	0.8048	0.8048	nan	0.8048	0.0	0.8048
0.0034	16.2791	1400	0.0614	0.3826	0.7652	0.7652	nan	0.7652	0.0	0.7652
0.0034	16.5116	1420	0.0661	0.3468	0.6936	0.6936	nan	0.6936	0.0	0.6936
0.0037	16.7442	1440	0.0442	0.2780	0.5560	0.5560	nan	0.5560	0.0	0.5560
0.0058	16.9767	1460	0.0694	0.3634	0.7267	0.7267	nan	0.7267	0.0	0.7267
0.0046	17.2093	1480	0.0433	0.3618	0.7237	0.7237	nan	0.7237	0.0	0.7237
0.0068	17.4419	1500	0.0718	0.3741	0.7481	0.7481	nan	0.7481	0.0	0.7481
0.0053	17.6744	1520	0.0786	0.4243	0.8485	0.8485	nan	0.8485	0.0	0.8485
0.0065	17.9070	1540	0.0759	0.3885	0.7771	0.7771	nan	0.7771	0.0	0.7771
0.0142	18.1395	1560	0.0707	0.3476	0.6952	0.6952	nan	0.6952	0.0	0.6952
0.0036	18.3721	1580	0.0543	0.3193	0.6386	0.6386	nan	0.6386	0.0	0.6386
0.003	18.6047	1600	0.0595	0.3440	0.6879	0.6879	nan	0.6879	0.0	0.6879
0.006	18.8372	1620	0.0515	0.3808	0.7616	0.7616	nan	0.7616	0.0	0.7616
0.0031	19.0698	1640	0.0533	0.3440	0.6879	0.6879	nan	0.6879	0.0	0.6879
0.0033	19.3023	1660	0.0758	0.3302	0.6605	0.6605	nan	0.6605	0.0	0.6605
0.0176	19.5349	1680	0.0598	0.2267	0.4534	0.4534	nan	0.4534	0.0	0.4534
0.0055	19.7674	1700	0.0354	0.3215	0.6430	0.6430	nan	0.6430	0.0	0.6430
0.0032	20.0	1720	0.0552	0.4126	0.8251	0.8251	nan	0.8251	0.0	0.8251

Framework versions

• Transformers 4.44.2
• Pytorch 2.5.0+cu121
• Datasets 3.0.2
• Tokenizers 0.19.1