segformer-b1-finetuned-segments-pv_v1_normalized_p100

This model is a fine-tuned version of nvidia/segformer-b1-finetuned-ade-512-512 on the mouadenna/satellite_PV_dataset_train_test_v1 dataset. It achieves the following results on the evaluation set:

• Loss: 0.0092
• Mean Iou: 0.8705
• Precision: 0.9201

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: 0.0001
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Precision
0.0078	1.0	3666	0.0068	0.8054	0.9187
0.0029	2.0	7332	0.0058	0.8298	0.8778
0.002	3.0	10998	0.0058	0.8274	0.8961
0.0013	4.0	14664	0.0060	0.8388	0.8774
0.0	5.0	18330	0.0055	0.8405	0.8901
0.0059	6.0	21996	0.0050	0.8547	0.9004
0.0048	7.0	25662	0.0060	0.8364	0.8668
0.0	8.0	29328	0.0063	0.8278	0.8776
0.0031	9.0	32994	0.0060	0.8547	0.9188
0.0017	10.0	36660	0.0061	0.8489	0.9105
0.0029	11.0	40326	0.0058	0.8572	0.9066
0.0	12.0	43992	0.0059	0.8525	0.9105
0.0031	13.0	47658	0.0057	0.8514	0.9035
0.0012	14.0	51324	0.0056	0.8567	0.9058
0.0042	15.0	54990	0.0058	0.8463	0.8898
0.0026	16.0	58656	0.0067	0.8607	0.9196
0.0012	17.0	62322	0.0050	0.8632	0.9224
0.0031	18.0	65988	0.0066	0.8404	0.9155
0.0018	19.0	69654	0.0059	0.8598	0.9115
0.002	20.0	73320	0.0065	0.8561	0.9210
0.0033	21.0	76986	0.0070	0.8580	0.9118
0.0014	22.0	80652	0.0066	0.8597	0.9306
0.0	23.0	84318	0.0066	0.8623	0.9014
0.0	24.0	87984	0.0062	0.8709	0.9217
0.0022	25.0	91650	0.0067	0.8644	0.9204
0.0013	26.0	95316	0.0063	0.8680	0.9214
0.0015	27.0	98982	0.0073	0.8520	0.8918
0.0	28.0	102648	0.0071	0.8674	0.9144
0.0015	29.0	106314	0.0069	0.8716	0.9261
0.0	30.0	109980	0.0068	0.8715	0.9246
0.0012	31.0	113646	0.0073	0.8682	0.9128
0.0009	32.0	117312	0.0071	0.8717	0.9260
0.0022	33.0	120978	0.0071	0.8715	0.9172
0.0019	34.0	124644	0.0075	0.8674	0.9127
0.0	35.0	128310	0.0078	0.8660	0.9140
0.0009	36.0	131976	0.0079	0.8720	0.9214
0.0007	37.0	135642	0.0087	0.8689	0.9206
0.0014	38.0	139308	0.0077	0.8697	0.9161
0.0	39.0	142974	0.0091	0.8682	0.9243
0.0025	40.0	146640	0.0091	0.8660	0.9161
0.0019	41.0	150306	0.0089	0.8722	0.9190
0.0009	42.0	153972	0.0087	0.8727	0.9233
0.0017	43.0	157638	0.0091	0.8721	0.9196
0.0	44.0	161304	0.0093	0.8737	0.9181
0.0012	45.0	164970	0.0093	0.8727	0.9237
0.0	46.0	168636	0.0094	0.8724	0.9230
0.0005	47.0	172302	0.0102	0.8675	0.9137
0.0	48.0	175968	0.0094	0.8631	0.9066
0.0009	49.0	179634	0.0103	0.8700	0.9165
0.0008	50.0	183300	0.0092	0.8705	0.9201

Framework versions

• Transformers 4.42.3
• Pytorch 2.1.2
• Datasets 2.20.0
• Tokenizers 0.19.1