segformer-b0-finetuned-segments-greenhouse-jun-24

This model is a fine-tuned version of nvidia/mit-b0 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.6502
• Mean Iou: 0.3640
• Mean Accuracy: 0.4319
• Overall Accuracy: 0.8283
• Accuracy Unlabeled: nan
• Accuracy Object: 0.0
• Accuracy Road: 0.9324
• Accuracy Plant: 0.8871
• Accuracy Iron: 0.0017
• Accuracy Wood: nan
• Accuracy Wall: 0.7226
• Accuracy Raw Road: 0.9465
• Accuracy Bottom Wall: 0.0
• Accuracy Roof: 0.0
• Accuracy Grass: nan
• Accuracy Mulch: 0.8289
• Accuracy Person: nan
• Accuracy Tomato: 0.0
• Iou Unlabeled: nan
• Iou Object: 0.0
• Iou Road: 0.7525
• Iou Plant: 0.7027
• Iou Iron: 0.0017
• Iou Wood: nan
• Iou Wall: 0.5584
• Iou Raw Road: 0.8998
• Iou Bottom Wall: 0.0
• Iou Roof: 0.0
• Iou Grass: nan
• Iou Mulch: 0.7252
• Iou Person: nan
• Iou Tomato: 0.0

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: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Unlabeled	Accuracy Object	Accuracy Road	Accuracy Plant	Accuracy Iron	Accuracy Wood	Accuracy Wall	Accuracy Raw Road	Accuracy Bottom Wall	Accuracy Roof	Accuracy Grass	Accuracy Mulch	Accuracy Person	Accuracy Tomato	Iou Unlabeled	Iou Object	Iou Road	Iou Plant	Iou Iron	Iou Wood	Iou Wall	Iou Raw Road	Iou Bottom Wall	Iou Roof	Iou Grass	Iou Mulch	Iou Person	Iou Tomato
1.9416	1.05	20	2.3650	0.1880	0.3464	0.6650	nan	0.0	0.7192	0.7931	0.2656	nan	0.0681	0.8201	0.0	0.0	nan	0.7950	nan	0.0029	nan	0.0	0.4874	0.5054	0.1242	0.0	0.0676	0.8065	0.0	0.0	0.0	0.4498	0.0	0.0027
1.4047	2.11	40	1.6208	0.2889	0.3699	0.7203	nan	0.0	0.7452	0.8135	0.0384	nan	0.4353	0.8655	0.0	0.0	nan	0.8014	nan	0.0	nan	0.0	0.4970	0.5407	0.0371	nan	0.4041	0.8614	0.0	0.0	nan	0.5489	nan	0.0
1.4998	3.16	60	1.2645	0.3150	0.3936	0.7522	nan	0.0	0.7532	0.8121	0.0174	nan	0.6304	0.9056	0.0	0.0	nan	0.8171	nan	0.0	nan	0.0	0.5316	0.5644	0.0174	nan	0.5346	0.8961	0.0	0.0	nan	0.6057	nan	0.0
1.0844	4.21	80	1.1551	0.3234	0.4083	0.7685	nan	0.0	0.8290	0.7952	0.0230	nan	0.6585	0.9033	0.0	0.0	nan	0.8740	nan	0.0	nan	0.0	0.5971	0.5910	0.0229	nan	0.5307	0.8905	0.0	0.0	nan	0.6020	nan	0.0
1.2949	5.26	100	1.0333	0.3363	0.4129	0.7841	nan	0.0	0.8274	0.8389	0.0140	nan	0.7114	0.9133	0.0	0.0	nan	0.8243	nan	0.0	nan	0.0	0.6211	0.6125	0.0140	nan	0.5854	0.8890	0.0	0.0	nan	0.6410	nan	0.0
1.3389	6.32	120	0.9260	0.3417	0.4155	0.7932	nan	0.0	0.8668	0.8408	0.0	nan	0.7105	0.9202	0.0	0.0	nan	0.8164	nan	0.0	nan	0.0	0.6489	0.6214	0.0	nan	0.6039	0.8936	0.0	0.0	nan	0.6495	nan	0.0
0.7833	7.37	140	0.9264	0.3357	0.4075	0.7871	nan	0.0	0.8811	0.8468	0.0	nan	0.6389	0.9125	0.0	0.0	nan	0.7963	nan	0.0	nan	0.0	0.6176	0.6285	0.0	nan	0.5777	0.8915	0.0	0.0	nan	0.6419	nan	0.0
1.0194	8.42	160	0.8761	0.3499	0.4231	0.8038	nan	0.0	0.8549	0.8586	0.0	nan	0.7365	0.9299	0.0	0.0	nan	0.8508	nan	0.0	nan	0.0	0.6797	0.6342	0.0	nan	0.6119	0.8995	0.0	0.0	nan	0.6738	nan	0.0
0.5558	9.47	180	0.8468	0.3458	0.4174	0.7981	nan	0.0	0.8533	0.8817	0.0	nan	0.6946	0.9063	0.0	0.0	nan	0.8381	nan	0.0	nan	0.0	0.6659	0.6338	0.0	nan	0.6155	0.8865	0.0	0.0	nan	0.6564	nan	0.0
1.2579	10.53	200	0.7776	0.3502	0.4184	0.8047	nan	0.0	0.8678	0.8680	0.0	nan	0.6966	0.9388	0.0	0.0	nan	0.8131	nan	0.0	nan	0.0	0.6432	0.6556	0.0	nan	0.6191	0.8990	0.0	0.0	nan	0.6852	nan	0.0
0.7671	11.58	220	0.7935	0.3579	0.4276	0.8152	nan	0.0	0.8816	0.8768	0.0	nan	0.7413	0.9356	0.0	0.0	nan	0.8410	nan	0.0	nan	0.0	0.6987	0.6610	0.0	nan	0.6315	0.9022	0.0	0.0	nan	0.6857	nan	0.0
0.5097	12.63	240	0.7718	0.3549	0.4262	0.8129	nan	0.0	0.9047	0.8658	0.0	nan	0.7146	0.9298	0.0	0.0	nan	0.8467	nan	0.0	nan	0.0	0.6773	0.6707	0.0	nan	0.6172	0.9016	0.0	0.0	nan	0.6818	nan	0.0
0.624	13.68	260	0.7270	0.3609	0.4282	0.8228	nan	0.0	0.8772	0.9219	0.0004	nan	0.7225	0.9308	0.0	0.0	nan	0.8291	nan	0.0	nan	0.0	0.7310	0.6897	0.0004	nan	0.5916	0.8975	0.0	0.0	nan	0.6988	nan	0.0
0.535	14.74	280	0.7681	0.3526	0.4243	0.8085	nan	0.0	0.9574	0.8230	0.0009	nan	0.7059	0.9289	0.0	0.0	nan	0.8268	nan	0.0	nan	0.0	0.6786	0.6512	0.0009	nan	0.6011	0.9014	0.0	0.0	nan	0.6930	nan	0.0
0.6093	15.79	300	0.6960	0.3636	0.4349	0.8257	nan	0.0	0.9296	0.8704	0.0102	nan	0.7227	0.9435	0.0	0.0	nan	0.8722	nan	0.0	nan	0.0	0.7270	0.6943	0.0102	nan	0.5991	0.9034	0.0	0.0	nan	0.7024	nan	0.0
0.5584	16.84	320	0.6886	0.3671	0.4368	0.8281	nan	0.0	0.9186	0.8889	0.0157	nan	0.7333	0.9371	0.0	0.0	nan	0.8739	nan	0.0	nan	0.0	0.7428	0.6928	0.0157	nan	0.6008	0.9040	0.0	0.0	nan	0.7148	nan	0.0
0.4421	17.89	340	0.6946	0.3644	0.4336	0.8238	nan	0.0	0.9061	0.8956	0.0308	nan	0.7280	0.9336	0.0	0.0	nan	0.8422	nan	0.0	nan	0.0	0.7217	0.6974	0.0308	nan	0.5717	0.9021	0.0	0.0	nan	0.7199	nan	0.0
0.7997	18.95	360	0.7025	0.3580	0.4266	0.8172	nan	0.0	0.8983	0.8901	0.0075	nan	0.6955	0.9330	0.0	0.0	nan	0.8415	nan	0.0	nan	0.0	0.7140	0.6754	0.0075	nan	0.5592	0.9020	0.0	0.0	nan	0.7216	nan	0.0
0.8388	20.0	380	0.6959	0.3632	0.4366	0.8242	nan	0.0	0.9513	0.8467	0.0120	nan	0.7460	0.9393	0.0	0.0	nan	0.8710	nan	0.0	nan	0.0	0.7218	0.6943	0.0120	nan	0.5799	0.9040	0.0	0.0	nan	0.7199	nan	0.0
0.6424	21.05	400	0.6728	0.3651	0.4285	0.8280	nan	0.0	0.8680	0.9419	0.0007	nan	0.7148	0.9412	0.0	0.0	nan	0.8186	nan	0.0	nan	0.0	0.7527	0.6967	0.0007	nan	0.5737	0.9026	0.0	0.0	nan	0.7249	nan	0.0
0.3287	22.11	420	0.6786	0.3621	0.4314	0.8247	nan	0.0	0.9357	0.8771	0.0053	nan	0.7122	0.9410	0.0	0.0	nan	0.8427	nan	0.0	nan	0.0	0.7335	0.6949	0.0053	nan	0.5626	0.9025	0.0	0.0	nan	0.7222	nan	0.0
0.386	23.16	440	0.6603	0.3667	0.4354	0.8295	nan	0.0	0.9165	0.9030	0.0122	nan	0.7266	0.9361	0.0	0.0	nan	0.8593	nan	0.0	nan	0.0	0.7526	0.7050	0.0122	nan	0.5635	0.9033	0.0	0.0	nan	0.7301	nan	0.0
0.3378	24.21	460	0.6791	0.3644	0.4331	0.8265	nan	0.0	0.9426	0.8772	0.0103	nan	0.7197	0.9405	0.0	0.0	nan	0.8403	nan	0.0	nan	0.0	0.7441	0.6939	0.0103	nan	0.5636	0.9039	0.0	0.0	nan	0.7284	nan	0.0
0.3678	25.26	480	0.6915	0.3633	0.4342	0.8227	nan	0.0	0.9479	0.8577	0.0234	nan	0.7165	0.9384	0.0	0.0	nan	0.8579	nan	0.0	nan	0.0	0.7171	0.6910	0.0234	nan	0.5647	0.9051	0.0	0.0	nan	0.7320	nan	0.0
0.328	26.32	500	0.6879	0.3662	0.4360	0.8259	nan	0.0	0.9434	0.8741	0.0266	nan	0.7189	0.9346	0.0	0.0	nan	0.8627	nan	0.0	nan	0.0	0.7357	0.6927	0.0266	nan	0.5712	0.9042	0.0	0.0	nan	0.7316	nan	0.0
0.8502	27.37	520	0.6593	0.3644	0.4332	0.8270	nan	0.0	0.9414	0.8739	0.0066	nan	0.7263	0.9446	0.0	0.0	nan	0.8390	nan	0.0	nan	0.0	0.7449	0.6962	0.0066	nan	0.5647	0.9020	0.0	0.0	nan	0.7294	nan	0.0
0.3528	28.42	540	0.6777	0.3626	0.4305	0.8238	nan	0.0	0.9439	0.8717	0.0114	nan	0.7046	0.9429	0.0	0.0	nan	0.8307	nan	0.0	nan	0.0	0.7364	0.6872	0.0114	nan	0.5563	0.9029	0.0	0.0	nan	0.7320	nan	0.0
0.5908	29.47	560	0.6502	0.3640	0.4319	0.8283	nan	0.0	0.9324	0.8871	0.0017	nan	0.7226	0.9465	0.0	0.0	nan	0.8289	nan	0.0	nan	0.0	0.7525	0.7027	0.0017	nan	0.5584	0.8998	0.0	0.0	nan	0.7252	nan	0.0

Framework versions

• Transformers 4.33.2
• Pytorch 2.0.1
• Datasets 2.15.0
• Tokenizers 0.13.3