segformer-b0-scene-parse-150-lr-4-e-30-new-9img_FFT

This model is a fine-tuned version of DiTo97/binarization-segformer-b3 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1542
• Mean Iou: 0.5040
• Mean Accuracy: 0.5233
• Overall Accuracy: 0.9668
• Per Category Iou: [0.04118617051522149, 0.9667182726602117]
• Per Category Accuracy: [0.05601009748261377, 0.9905862883445109]

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: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Per Category Iou	Per Category Accuracy
No log	1.0	131	0.1528	0.4873	0.5	0.9745	[0.0, 0.9745119730631511]	[0.0, 1.0]
No log	2.0	262	0.1889	0.4873	0.5	0.9745	[0.0, 0.9745119730631511]	[0.0, 1.0]
No log	3.0	393	0.1197	0.4873	0.5	0.9745	[0.0, 0.9745119730631511]	[0.0, 1.0]
0.4149	4.0	524	0.1201	0.4873	0.5000	0.9745	[0.00010223979970973872, 0.9745065024545415]	[0.00010227193358809854, 0.9999917796143308]
0.4149	5.0	655	0.1441	0.4873	0.5000	0.9745	[2.7437511068541397e-05, 0.9745114466111445]	[2.7438811450465463e-05, 0.999998760418034]
0.4149	6.0	786	0.1309	0.4873	0.5001	0.9745	[0.0001570390902700574, 0.9744979472713902]	[0.00015714955648902945, 0.9999816019939785]
0.4149	7.0	917	0.1319	0.4874	0.5002	0.9745	[0.0003957310820631776, 0.9744648577316526]	[0.0003966155473294553, 0.9999415439241303]
0.1817	8.0	1048	0.1295	0.4881	0.5009	0.9745	[0.0018197547805853048, 0.9744708577898599]	[0.0018259281801582471, 0.9999112720276977]
0.1817	9.0	1179	0.1309	0.4920	0.5047	0.9742	[0.00972514515778975, 0.9741753903361423]	[0.009947816369496024, 0.9994011514281158]
0.1817	10.0	1310	0.1343	0.4895	0.5022	0.9743	[0.004785912791126633, 0.9742689722328164]	[0.004854175189327799, 0.9996269510693939]
0.1817	11.0	1441	0.1212	0.4894	0.5022	0.9740	[0.0048241029726707255, 0.9740276066246742]	[0.004938986061083783, 0.9993771426826694]
0.1615	12.0	1572	0.1346	0.4931	0.5059	0.9737	[0.012490729722833617, 0.9736626335700124]	[0.013065863125230736, 0.9987957134994652]
0.1615	13.0	1703	0.1258	0.4912	0.5040	0.9740	[0.008481020628593294, 0.9739639804750839]	[0.008735519790866368, 0.9992151488920616]
0.1615	14.0	1834	0.1365	0.4989	0.5125	0.9727	[0.025065780362830633, 0.9726777607699496]	[0.02754108338405356, 0.9974171678709749]
0.1615	15.0	1965	0.1296	0.4969	0.5101	0.9732	[0.02060107372759099, 0.9731805906599663]	[0.022120670903884337, 0.9980707537728308]
0.144	16.0	2096	0.1436	0.5050	0.5219	0.9697	[0.040424047311196594, 0.9696216140306383]	[0.05014567514442793, 0.9937100348987993]
0.144	17.0	2227	0.1446	0.5023	0.5177	0.9707	[0.03386431595215697, 0.970717796701266]	[0.04022779202378696, 0.9950852532283281]
0.144	18.0	2358	0.1420	0.5028	0.5179	0.9712	[0.03438079444116839, 0.971131946623355]	[0.040285164084092474, 0.9955083421256665]
0.144	19.0	2489	0.1403	0.5018	0.5171	0.9706	[0.032975986996447486, 0.9706187216781221]	[0.03926992806042525, 0.9950080076995003]
0.1327	20.0	2620	0.1390	0.4983	0.5121	0.9720	[0.024489823088375932, 0.9720288138083142]	[0.02753110563443521, 0.9967519690433324]
0.1327	21.0	2751	0.1469	0.5044	0.5221	0.9686	[0.04024195335923191, 0.9685753252351375]	[0.051627370962753064, 0.9926002175923039]
0.1327	22.0	2882	0.1419	0.4993	0.5137	0.9714	[0.027247862213791583, 0.9713509036761706]	[0.031459844546660946, 0.9959570055561978]
0.1223	23.0	3013	0.1452	0.5014	0.5171	0.9700	[0.03283676275044511, 0.9699570020594283]	[0.03997834828332818, 0.9943116888403959]
0.1223	24.0	3144	0.1492	0.5028	0.5202	0.9684	[0.03726474521737445, 0.9683817614075229]	[0.04795805354060445, 0.9924947878840388]
0.1223	25.0	3275	0.1432	0.4988	0.5133	0.9711	[0.026574581969862024, 0.9710595860086739]	[0.030973429252766332, 0.9956706621220547]
0.1223	26.0	3406	0.1483	0.5022	0.5190	0.9688	[0.035495434136009404, 0.9688123464095088]	[0.04497969527952666, 0.9930115630815434]
0.1163	27.0	3537	0.1511	0.5033	0.5211	0.9681	[0.03854419663037566, 0.9680458059790112]	[0.05019556389251968, 0.9920938157386199]
0.1163	28.0	3668	0.1509	0.5034	0.5218	0.9675	[0.03933910306845004, 0.9674300294951071]	[0.05225846362611376, 0.9914105451107365]
0.1163	29.0	3799	0.1559	0.5050	0.5268	0.9649	[0.04520760305755583, 0.9648560029666593]	[0.06517715494447382, 0.9884467046300605]
0.1163	30.0	3930	0.1542	0.5040	0.5233	0.9668	[0.04118617051522149, 0.9667182726602117]	[0.05601009748261377, 0.9905862883445109]

Framework versions

• Transformers 4.39.3
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2