Modelpublic

Spatiallysaying/segformer_finetuned_rwy_obb_100epochs

sourceHugging Faceotherupdated 9mo agoView on Hugging Face

0likes7downloads

Model Card

segformerfinetunedrwyobb100epochs

This model is a fine-tuned version of nvidia/mit-b0 on the Spatiallysaying/rwy_obb-300-65-65 dataset. It achieves the following results on the evaluation set:

—Loss: 0.0390
—Mean Iou: 0.3599
—Mean Accuracy: 0.7198
—Overall Accuracy: 0.7198
—Accuracy Background : nan
—Accuracy Rwy Obb: 0.7198
—Iou Background : 0.0
—Iou Rwy Obb: 0.7198

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
—trainbatchsize: 2
—evalbatchsize: 2
—seed: 1337
—optimizer: Use OptimizerNames.ADAMWTORCHFUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
—lrschedulertype: cosine
—lrschedulerwarmup_steps: 0.1
—num_epochs: 100.0
—mixedprecisiontraining: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Background	Accuracy Rwy Obb	Iou Rwy Obb
0.5957	1.0	152	0.5023	0.2461	0.4923	0.4923	nan	0.4923	0.4923
0.4697	2.0	304	0.3524	0.2104	0.4208	0.4208	nan	0.4208	0.4208
0.3752	3.0	456	0.2361	0.2458	0.4915	0.4915	nan	0.4915	0.4915
0.1900	4.0	608	0.1412	0.1648	0.3296	0.3296	nan	0.3296	0.3296
0.1406	5.0	760	0.0964	0.1899	0.3797	0.3797	nan	0.3797	0.3797
0.0870	6.0	912	0.0790	0.2317	0.4634	0.4634	nan	0.4634	0.4634
0.0731	7.0	1064	0.0636	0.2109	0.4218	0.4218	nan	0.4218	0.4218
0.0588	8.0	1216	0.0586	0.2207	0.4413	0.4413	nan	0.4413	0.4413
0.0580	9.0	1368	0.0544	0.2900	0.5800	0.5800	nan	0.5800	0.5800
0.0514	10.0	1520	0.0497	0.2634	0.5267	0.5267	nan	0.5267	0.5267
0.0469	11.0	1672	0.0483	0.2826	0.5652	0.5652	nan	0.5652	0.5652
0.0444	12.0	1824	0.0467	0.3008	0.6015	0.6015	nan	0.6015	0.6015
0.0397	13.0	1976	0.0427	0.2877	0.5753	0.5753	nan	0.5753	0.5753
0.0388	14.0	2128	0.0454	0.2836	0.5673	0.5673	nan	0.5673	0.5673
0.0425	15.0	2280	0.0469	0.2433	0.4865	0.4865	nan	0.4865	0.4865
0.0345	16.0	2432	0.0406	0.2930	0.5860	0.5860	nan	0.5860	0.5860
0.0335	17.0	2584	0.0387	0.3416	0.6833	0.6833	nan	0.6833	0.6833
0.0352	18.0	2736	0.0406	0.2728	0.5456	0.5456	nan	0.5456	0.5456
0.0319	19.0	2888	0.0407	0.3050	0.6099	0.6099	nan	0.6099	0.6099
0.0316	20.0	3040	0.0441	0.2999	0.5998	0.5998	nan	0.5998	0.5998
0.0329	21.0	3192	0.0392	0.3560	0.7120	0.7120	nan	0.7120	0.7120
0.0291	22.0	3344	0.0382	0.3530	0.7059	0.7059	nan	0.7059	0.7059
0.0287	23.0	3496	0.0402	0.3648	0.7296	0.7296	nan	0.7296	0.7296
0.0257	24.0	3648	0.0415	0.3316	0.6632	0.6632	nan	0.6632	0.6632
0.0303	25.0	3800	0.0366	0.3315	0.6630	0.6630	nan	0.6630	0.6630
0.0240	26.0	3952	0.0374	0.3441	0.6882	0.6882	nan	0.6882	0.6882
0.0253	27.0	4104	0.0383	0.3391	0.6783	0.6783	nan	0.6783	0.6783
0.0253	28.0	4256	0.0358	0.3507	0.7014	0.7014	nan	0.7014	0.7014
0.0256	29.0	4408	0.0372	0.3402	0.6804	0.6804	nan	0.6804	0.6804
0.0245	30.0	4560	0.0379	0.3541	0.7082	0.7082	nan	0.7082	0.7082
0.0228	31.0	4712	0.0392	0.3344	0.6689	0.6689	nan	0.6689	0.6689
0.0253	32.0	4864	0.0392	0.3057	0.6115	0.6115	nan	0.6115	0.6115
0.0267	33.0	5016	0.0363	0.3466	0.6932	0.6932	nan	0.6932	0.6932
0.0240	34.0	5168	0.0393	0.3259	0.6518	0.6518	nan	0.6518	0.6518
0.0223	35.0	5320	0.0419	0.3382	0.6763	0.6763	nan	0.6763	0.6763
0.0243	36.0	5472	0.0390	0.3217	0.6434	0.6434	nan	0.6434	0.6434
0.0210	37.0	5624	0.0364	0.3555	0.7109	0.7109	nan	0.7109	0.7109
0.0245	38.0	5776	0.0380	0.3554	0.7109	0.7109	nan	0.7109	0.7109
0.0228	39.0	5928	0.0386	0.3381	0.6762	0.6762	nan	0.6762	0.6762
0.0209	40.0	6080	0.0352	0.3594	0.7187	0.7187	nan	0.7187	0.7187
0.0187	41.0	6232	0.0372	0.3651	0.7301	0.7301	nan	0.7301	0.7301
0.0211	42.0	6384	0.0408	0.3293	0.6585	0.6585	nan	0.6585	0.6585
0.0213	43.0	6536	0.0359	0.3653	0.7306	0.7306	nan	0.7306	0.7306
0.0208	44.0	6688	0.0362	0.3747	0.7495	0.7495	nan	0.7495	0.7495
0.0197	45.0	6840	0.0375	0.3580	0.7161	0.7161	nan	0.7161	0.7161
0.0188	46.0	6992	0.0378	0.3651	0.7302	0.7302	nan	0.7302	0.7302
0.0204	47.0	7144	0.0365	0.3732	0.7465	0.7465	nan	0.7465	0.7465
0.0191	48.0	7296	0.0373	0.3509	0.7017	0.7017	nan	0.7017	0.7017
0.0181	49.0	7448	0.0363	0.3697	0.7395	0.7395	nan	0.7395	0.7395
0.0197	50.0	7600	0.0366	0.3601	0.7203	0.7203	nan	0.7203	0.7203
0.0194	51.0	7752	0.0406	0.3355	0.6710	0.6710	nan	0.6710	0.6710
0.0193	52.0	7904	0.0365	0.3655	0.7309	0.7309	nan	0.7309	0.7309
0.0186	53.0	8056	0.0385	0.3545	0.7090	0.7090	nan	0.7090	0.7090
0.0186	54.0	8208	0.0387	0.3808	0.7616	0.7616	nan	0.7616	0.7616
0.0195	55.0	8360	0.0412	0.3384	0.6768	0.6768	nan	0.6768	0.6768
0.0175	56.0	8512	0.0370	0.3625	0.7249	0.7249	nan	0.7249	0.7249
0.0172	57.0	8664	0.0370	0.3685	0.7369	0.7369	nan	0.7369	0.7369
0.0173	58.0	8816	0.0374	0.3581	0.7162	0.7162	nan	0.7162	0.7162
0.0177	59.0	8968	0.0374	0.3728	0.7456	0.7456	nan	0.7456	0.7456
0.0165	60.0	9120	0.0373	0.3588	0.7176	0.7176	nan	0.7176	0.7176
0.0177	61.0	9272	0.0375	0.3762	0.7523	0.7523	nan	0.7523	0.7523
0.0163	62.0	9424	0.0395	0.3651	0.7303	0.7303	nan	0.7303	0.7303
0.0159	63.0	9576	0.0357	0.3612	0.7224	0.7224	nan	0.7224	0.7224
0.0163	64.0	9728	0.0371	0.3586	0.7173	0.7173	nan	0.7173	0.7173
0.0172	65.0	9880	0.0383	0.3500	0.6999	0.6999	nan	0.6999	0.6999
0.0145	66.0	10032	0.0383	0.3650	0.7299	0.7299	nan	0.7299	0.7299
0.0142	67.0	10184	0.0366	0.3698	0.7396	0.7396	nan	0.7396	0.7396
0.0153	68.0	10336	0.0381	0.3648	0.7295	0.7295	nan	0.7295	0.7295
0.0162	69.0	10488	0.0356	0.3726	0.7453	0.7453	nan	0.7453	0.7453
0.0148	70.0	10640	0.0386	0.3572	0.7144	0.7144	nan	0.7144	0.7144
0.0153	71.0	10792	0.0370	0.3671	0.7342	0.7342	nan	0.7342	0.7342
0.0144	72.0	10944	0.0370	0.3613	0.7225	0.7225	nan	0.7225	0.7225
0.0152	73.0	11096	0.0392	0.3503	0.7005	0.7005	nan	0.7005	0.7005
0.0144	74.0	11248	0.0379	0.3623	0.7246	0.7246	nan	0.7246	0.7246
0.0153	75.0	11400	0.0385	0.3681	0.7362	0.7362	nan	0.7362	0.7362
0.0139	76.0	11552	0.0381	0.3602	0.7205	0.7205	nan	0.7205	0.7205
0.0145	77.0	11704	0.0378	0.3626	0.7252	0.7252	nan	0.7252	0.7252
0.0166	78.0	11856	0.0387	0.3596	0.7193	0.7193	nan	0.7193	0.7193
0.0151	79.0	12008	0.0395	0.3634	0.7269	0.7269	nan	0.7269	0.7269
0.0165	80.0	12160	0.0393	0.3582	0.7163	0.7163	nan	0.7163	0.7163
0.0144	81.0	12312	0.0393	0.3535	0.7071	0.7071	nan	0.7071	0.7071
0.0156	82.0	12464	0.0391	0.3587	0.7173	0.7173	nan	0.7173	0.7173
0.0144	83.0	12616	0.0390	0.3707	0.7415	0.7415	nan	0.7415	0.7415
0.0137	84.0	12768	0.0385	0.3641	0.7282	0.7282	nan	0.7282	0.7282
0.0147	85.0	12920	0.0376	0.3622	0.7244	0.7244	nan	0.7244	0.7244
0.0159	86.0	13072	0.0382	0.3581	0.7163	0.7163	nan	0.7163	0.7163
0.0147	87.0	13224	0.0374	0.3645	0.7289	0.7289	nan	0.7289	0.7289
0.0142	88.0	13376	0.0388	0.3629	0.7257	0.7257	nan	0.7257	0.7257
0.0141	89.0	13528	0.0372	0.3652	0.7305	0.7305	nan	0.7305	0.7305
0.0142	90.0	13680	0.0378	0.3597	0.7194	0.7194	nan	0.7194	0.7194
0.0137	91.0	13832	0.0386	0.3587	0.7174	0.7174	nan	0.7174	0.7174
0.0140	92.0	13984	0.0387	0.3624	0.7249	0.7249	nan	0.7249	0.7249
0.0143	93.0	14136	0.0388	0.3608	0.7215	0.7215	nan	0.7215	0.7215
0.0144	94.0	14288	0.0384	0.3634	0.7269	0.7269	nan	0.7269	0.7269
0.0137	95.0	14440	0.0382	0.3595	0.7190	0.7190	nan	0.7190	0.7190
0.0142	96.0	14592	0.0394	0.3565	0.7131	0.7131	nan	0.7131	0.7131
0.0150	97.0	14744	0.0388	0.3577	0.7154	0.7154	nan	0.7154	0.7154
0.0147	98.0	14896	0.0383	0.3598	0.7197	0.7197	nan	0.7197	0.7197
0.0140	99.0	15048	0.0391	0.3620	0.7240	0.7240	nan	0.7240	0.7240
0.0132	100.0	15200	0.0390	0.3599	0.7198	0.7198	nan	0.7198	0.7198

Framework versions

—Transformers 5.0.0.dev0
—Pytorch 2.9.0+cu126
—Datasets 4.0.0
—Tokenizers 0.22.1