metadata

license: other
base_model: nvidia/mit-b0
tags:
  - generated_from_trainer
model-index:
  - name: mit-b0-finetuned-sidewalk-semantic
    results: []

mit-b0-finetuned-sidewalk-semantic

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

Loss: 0.7584
Mean Iou: 0.1932
Mean Accuracy: 0.2342
Overall Accuracy: 0.8034
Accuracy Unlabeled: nan
Accuracy Flat-road: 0.7939
Accuracy Flat-sidewalk: 0.9534
Accuracy Flat-crosswalk: 0.0000
Accuracy Flat-cyclinglane: 0.7630
Accuracy Flat-parkingdriveway: 0.3561
Accuracy Flat-railtrack: 0.0
Accuracy Flat-curb: 0.2446
Accuracy Human-person: 0.0018
Accuracy Human-rider: 0.0
Accuracy Vehicle-car: 0.9007
Accuracy Vehicle-truck: 0.0
Accuracy Vehicle-bus: 0.0
Accuracy Vehicle-tramtrain: 0.0
Accuracy Vehicle-motorcycle: 0.0
Accuracy Vehicle-bicycle: 0.0
Accuracy Vehicle-caravan: 0.0
Accuracy Vehicle-cartrailer: 0.0
Accuracy Construction-building: 0.9194
Accuracy Construction-door: 0.0
Accuracy Construction-wall: 0.0706
Accuracy Construction-fenceguardrail: 0.0
Accuracy Construction-bridge: 0.0
Accuracy Construction-tunnel: nan
Accuracy Construction-stairs: 0.0
Accuracy Object-pole: 0.0018
Accuracy Object-trafficsign: 0.0
Accuracy Object-trafficlight: 0.0
Accuracy Nature-vegetation: 0.9304
Accuracy Nature-terrain: 0.8467
Accuracy Sky: 0.9472
Accuracy Void-ground: 0.0
Accuracy Void-dynamic: 0.0
Accuracy Void-static: 0.0004
Accuracy Void-unclear: 0.0
Iou Unlabeled: nan
Iou Flat-road: 0.6083
Iou Flat-sidewalk: 0.8277
Iou Flat-crosswalk: 0.0000
Iou Flat-cyclinglane: 0.6702
Iou Flat-parkingdriveway: 0.2581
Iou Flat-railtrack: 0.0
Iou Flat-curb: 0.2113
Iou Human-person: 0.0018
Iou Human-rider: 0.0
Iou Vehicle-car: 0.7545
Iou Vehicle-truck: 0.0
Iou Vehicle-bus: 0.0
Iou Vehicle-tramtrain: 0.0
Iou Vehicle-motorcycle: 0.0
Iou Vehicle-bicycle: 0.0
Iou Vehicle-caravan: 0.0
Iou Vehicle-cartrailer: 0.0
Iou Construction-building: 0.6531
Iou Construction-door: 0.0
Iou Construction-wall: 0.0652
Iou Construction-fenceguardrail: 0.0
Iou Construction-bridge: 0.0
Iou Construction-tunnel: nan
Iou Construction-stairs: 0.0
Iou Object-pole: 0.0018
Iou Object-trafficsign: 0.0
Iou Object-trafficlight: 0.0
Iou Nature-vegetation: 0.8060
Iou Nature-terrain: 0.6540
Iou Sky: 0.8649
Iou Void-ground: 0.0
Iou Void-dynamic: 0.0
Iou Void-static: 0.0003
Iou Void-unclear: 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: 5e-05
train_batch_size: 2
eval_batch_size: 2
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 8
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Unlabeled	Accuracy Flat-road	Accuracy Flat-sidewalk	Accuracy Flat-crosswalk	Accuracy Flat-cyclinglane	Accuracy Flat-parkingdriveway	Accuracy Flat-curb	Accuracy Human-person	Accuracy Vehicle-car	Accuracy Construction-building	Accuracy Construction-wall	Accuracy Construction-tunnel	Accuracy Object-pole	Accuracy Nature-vegetation	Accuracy Nature-terrain	Accuracy Sky	Accuracy Void-static	Iou Unlabeled	Iou Flat-road	Iou Flat-sidewalk	Iou Flat-crosswalk	Iou Flat-cyclinglane	Iou Flat-parkingdriveway	Iou Flat-curb	Iou Human-person	Iou Vehicle-car	Iou Construction-building	Iou Construction-wall	Iou Construction-tunnel	Iou Object-pole	Iou Nature-vegetation	Iou Nature-terrain	Iou Sky	Iou Void-static
2.1654	1.0	100	1.6435	0.1291	0.1770	0.6919	nan	0.6823	0.9206	0.0	0.0022	0.0053	0.0	0.0	0.8850	0.8425	0.0	nan	0.0	0.9138	0.7723	0.8186	0.0000	nan	0.4039	0.7194	0.0	0.0022	0.0053	0.0	0.0	0.5851	0.5722	0.0	nan	0.0	0.7101	0.4854	0.7772	0.0000
1.5984	2.0	200	1.2905	0.1517	0.1973	0.7396	nan	0.7415	0.9557	0.0	0.3711	0.0160	0.0	0.0	0.9108	0.8769	0.0	nan	0.0	0.8585	0.8699	0.9088	0.0	nan	0.4974	0.7465	0.0	0.3665	0.0159	0.0	0.0	0.6673	0.5889	0.0	nan	0.0	0.7531	0.5509	0.8181	0.0
1.2146	3.0	300	1.0847	0.1626	0.2071	0.7634	nan	0.7918	0.9561	0.0	0.5225	0.0828	0.0000	0.0	0.9191	0.8607	0.0	nan	0.0	0.8959	0.8726	0.9324	0.0	nan	0.5364	0.7830	0.0	0.5010	0.0768	0.0000	0.0	0.6566	0.6147	0.0	nan	0.0	0.7682	0.5979	0.8319	0.0
1.0528	4.0	400	0.9593	0.1726	0.2148	0.7818	nan	0.8491	0.9467	0.0	0.6248	0.1582	0.0028	0.0	0.9208	0.8971	0.0	nan	0.0	0.9124	0.8446	0.9320	0.0	nan	0.5559	0.8209	0.0	0.5799	0.1382	0.0028	0.0	0.6851	0.6297	0.0	nan	0.0	0.7887	0.6506	0.8429	0.0
1.151	5.0	500	0.8917	0.1781	0.2211	0.7856	nan	0.7689	0.9449	0.0	0.7350	0.2627	0.0394	0.0	0.8992	0.9103	0.0001	nan	0.0	0.9213	0.8734	0.9404	0.0	nan	0.5697	0.8051	0.0	0.6312	0.2044	0.0386	0.0	0.7368	0.6341	0.0001	nan	0.0	0.7860	0.6232	0.8482	0.0
0.8834	6.0	600	0.8488	0.1833	0.2253	0.7930	nan	0.8031	0.9523	0.0	0.6739	0.3297	0.1109	0.0	0.9189	0.9135	0.0097	nan	0.0	0.9249	0.8625	0.9365	0.0	nan	0.5945	0.8182	0.0	0.6215	0.2340	0.1044	0.0	0.7278	0.6417	0.0096	nan	0.0	0.7977	0.6393	0.8609	0.0
0.7373	7.0	700	0.7922	0.1870	0.2305	0.7991	nan	0.8274	0.9366	0.0	0.7533	0.3386	0.1780	0.0001	0.9212	0.9144	0.0184	nan	0.0009	0.9324	0.8420	0.9422	0.0000	nan	0.5946	0.8448	0.0	0.6344	0.2432	0.1632	0.0001	0.7258	0.6472	0.0181	nan	0.0009	0.8002	0.6380	0.8615	0.0000
0.8252	8.0	800	0.7820	0.1888	0.2303	0.7987	nan	0.8024	0.9520	0.0	0.7312	0.3178	0.1998	0.0003	0.9060	0.9255	0.0366	nan	0.0005	0.9177	0.8774	0.9322	0.0002	nan	0.5972	0.8253	0.0	0.6590	0.2429	0.1763	0.0003	0.7464	0.6439	0.0351	nan	0.0005	0.8022	0.6324	0.8689	0.0002
0.6442	9.0	900	0.7593	0.1923	0.2328	0.8030	nan	0.8301	0.9488	0.0	0.7388	0.3596	0.2196	0.0013	0.9120	0.9158	0.0559	nan	0.0009	0.9261	0.8226	0.9516	0.0001	nan	0.5992	0.8342	0.0	0.6607	0.2569	0.1954	0.0013	0.7445	0.6532	0.0523	nan	0.0009	0.8115	0.6722	0.8643	0.0001
0.8239	10.0	1000	0.7584	0.1932	0.2342	0.8034	nan	0.7939	0.9534	0.0000	0.7630	0.3561	0.2446	0.0018	0.9007	0.9194	0.0706	nan	0.0018	0.9304	0.8467	0.9472	0.0004	nan	0.6083	0.8277	0.0000	0.6702	0.2581	0.2113	0.0018	0.7545	0.6531	0.0652	nan	0.0018	0.8060	0.6540	0.8649	0.0003

Framework versions

Transformers 4.38.2
Pytorch 2.2.1+cu121
Datasets 2.18.0
Tokenizers 0.15.2