segformer-b1-from-scratch-run1

This model is a fine-tuned version of on the samitizerxu/kelp_data_rgbaa_swin_nir dataset. It achieves the following results on the evaluation set:

• Iou Kelp: 0.0067
• Loss: 0.9872

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: 0.1
• train_batch_size: 22
• eval_batch_size: 22
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• num_epochs: 40

Training results

Training Loss	Epoch	Step	Iou Kelp	Validation Loss
0.9999	0.15	30	0.0067	0.9872
1.0	0.29	60	0.0067	0.9872
0.9933	0.44	90	0.0067	0.9872
0.998	0.59	120	0.0067	0.9872
1.0	0.73	150	0.0067	0.9872
0.9998	0.88	180	0.0067	0.9872
0.9998	1.02	210	0.0067	0.9872
1.0	1.17	240	0.0082	0.9861
0.9976	1.32	270	0.0069	0.9869
0.9995	1.46	300	0.0070	0.9868
0.9967	1.61	330	0.0067	0.9872
0.9945	1.76	360	0.0067	0.9872
1.0	1.9	390	0.0067	0.9872
0.9992	2.05	420	0.0067	0.9872
0.9991	2.2	450	0.0067	0.9872
0.997	2.34	480	0.0067	0.9872
0.999	2.49	510	0.0067	0.9872
0.9999	2.63	540	0.0067	0.9872
0.9991	2.78	570	0.0067	0.9872
0.9987	2.93	600	0.0067	0.9872
0.9999	3.07	630	0.0067	0.9872
0.9983	3.22	660	0.0067	0.9872
0.9973	3.37	690	0.0067	0.9872
0.9987	3.51	720	0.0067	0.9872
0.9915	3.66	750	0.0067	0.9872
0.9984	3.8	780	0.0067	0.9872
0.9992	3.95	810	0.0067	0.9872
0.9993	4.1	840	0.0067	0.9872
1.0	4.24	870	0.0067	0.9872
0.9998	4.39	900	0.0067	0.9872
0.9999	4.54	930	0.0067	0.9872
0.9995	4.68	960	0.0067	0.9872
0.998	4.83	990	0.0067	0.9872
0.9989	4.98	1020	0.0067	0.9872
0.9975	5.12	1050	0.0067	0.9872
0.9993	5.27	1080	0.0067	0.9872
0.9971	5.41	1110	0.0067	0.9872
0.9944	5.56	1140	0.0067	0.9872
0.9967	5.71	1170	0.0067	0.9872
0.9986	5.85	1200	0.0067	0.9872
0.9994	6.0	1230	0.0067	0.9872
0.9997	6.15	1260	0.0067	0.9872
0.9998	6.29	1290	0.0067	0.9872
0.999	6.44	1320	0.0067	0.9872
0.9996	6.59	1350	0.0067	0.9872
1.0	6.73	1380	0.0067	0.9872
0.9999	6.88	1410	0.0067	0.9872
0.9933	7.02	1440	0.0067	0.9872
0.998	7.17	1470	0.0067	0.9872
0.9968	7.32	1500	0.0067	0.9872
0.997	7.46	1530	0.0067	0.9872
0.9981	7.61	1560	0.0067	0.9872
0.9992	7.76	1590	0.0067	0.9872
0.9999	7.9	1620	0.0067	0.9872
0.9964	8.05	1650	0.0067	0.9872
0.9999	8.2	1680	0.0067	0.9872
0.9941	8.34	1710	0.0067	0.9872
0.9963	8.49	1740	0.0067	0.9872
0.998	8.63	1770	0.0067	0.9872
0.9989	8.78	1800	0.0067	0.9872
1.0	8.93	1830	0.0067	0.9872
1.0	9.07	1860	0.0067	0.9872
0.9974	9.22	1890	0.0067	0.9872
0.9989	9.37	1920	0.0067	0.9872
0.9989	9.51	1950	0.0067	0.9872
0.996	9.66	1980	0.0067	0.9872
0.9995	9.8	2010	0.0067	0.9872
0.9973	9.95	2040	0.0067	0.9872
0.9957	10.1	2070	0.0067	0.9872
0.9996	10.24	2100	0.0067	0.9872
1.0	10.39	2130	0.0067	0.9872
0.9967	10.54	2160	0.0067	0.9872
0.9989	10.68	2190	0.0067	0.9872
0.9989	10.83	2220	0.0067	0.9872
0.9994	10.98	2250	0.0067	0.9872
0.9992	11.12	2280	0.0067	0.9872
0.9973	11.27	2310	0.0067	0.9872
0.9993	11.41	2340	0.0067	0.9872
0.9973	11.56	2370	0.0067	0.9872
0.9996	11.71	2400	0.0067	0.9872
1.0	11.85	2430	0.0067	0.9872
0.9989	12.0	2460	0.0067	0.9872
1.0	12.15	2490	0.0067	0.9872
0.9987	12.29	2520	0.0067	0.9872
0.9914	12.44	2550	0.0067	0.9872
0.9974	12.59	2580	0.0067	0.9872
1.0	12.73	2610	0.0067	0.9872
0.999	12.88	2640	0.0067	0.9872
1.0	13.02	2670	0.0067	0.9872
0.9991	13.17	2700	0.0067	0.9872
0.9979	13.32	2730	0.0067	0.9872
1.0	13.46	2760	0.0067	0.9872
0.9973	13.61	2790	0.0067	0.9872
0.9995	13.76	2820	0.0067	0.9872
0.9973	13.9	2850	0.0067	0.9872
0.9961	14.05	2880	0.0067	0.9872
0.9907	14.2	2910	0.0067	0.9872
0.9984	14.34	2940	0.0067	0.9872
0.9986	14.49	2970	0.0067	0.9872
0.9935	14.63	3000	0.0067	0.9872
0.998	14.78	3030	0.0067	0.9872
0.9982	14.93	3060	0.0067	0.9872
0.9956	15.07	3090	0.0067	0.9872
0.9991	15.22	3120	0.0067	0.9872
0.9985	15.37	3150	0.0067	0.9872
0.9958	15.51	3180	0.0067	0.9872
0.9998	15.66	3210	0.0067	0.9872
0.9972	15.8	3240	0.0067	0.9872
0.9996	15.95	3270	0.0067	0.9872
0.9965	16.1	3300	0.0067	0.9872
0.9983	16.24	3330	0.0067	0.9872
0.9993	16.39	3360	0.0067	0.9872
0.9962	16.54	3390	0.0067	0.9872
0.9985	16.68	3420	0.0067	0.9872
0.9998	16.83	3450	0.0067	0.9872
0.9993	16.98	3480	0.0067	0.9872
0.9993	17.12	3510	0.0067	0.9872
0.9998	17.27	3540	0.0067	0.9872
1.0	17.41	3570	0.0067	0.9872
0.9999	17.56	3600	0.0067	0.9872
0.9993	17.71	3630	0.0067	0.9872
0.999	17.85	3660	0.0067	0.9872
0.9975	18.0	3690	0.0067	0.9872
0.9993	18.15	3720	0.0067	0.9872
1.0	18.29	3750	0.0067	0.9872
0.9983	18.44	3780	0.0067	0.9872
0.9994	18.59	3810	0.0067	0.9872
0.9993	18.73	3840	0.0067	0.9872
0.9982	18.88	3870	0.0067	0.9872
0.9997	19.02	3900	0.0067	0.9872
0.9955	19.17	3930	0.0067	0.9872
0.9992	19.32	3960	0.0067	0.9872
0.9592	19.46	3990	0.0067	0.9872
0.9897	19.61	4020	0.0067	0.9872
0.9994	19.76	4050	0.0067	0.9872
0.9989	19.9	4080	0.0067	0.9872
0.9995	20.05	4110	0.0067	0.9872
0.9995	20.2	4140	0.0067	0.9872
0.9938	20.34	4170	0.0067	0.9872
0.9987	20.49	4200	0.0067	0.9872
0.9999	20.63	4230	0.0067	0.9872
0.9994	20.78	4260	0.0067	0.9872
0.9954	20.93	4290	0.0067	0.9872
0.9975	21.07	4320	0.0067	0.9872
0.9997	21.22	4350	0.0067	0.9872
0.9978	21.37	4380	0.0067	0.9872
0.9994	21.51	4410	0.0067	0.9872
0.9985	21.66	4440	0.0067	0.9872
0.9998	21.8	4470	0.0067	0.9872
0.998	21.95	4500	0.0067	0.9872
0.9983	22.1	4530	0.0067	0.9872
0.9989	22.24	4560	0.0067	0.9872
0.9973	22.39	4590	0.0067	0.9872
0.9961	22.54	4620	0.0067	0.9872
0.9984	22.68	4650	0.0067	0.9872
1.0	22.83	4680	0.0067	0.9872
0.9949	22.98	4710	0.0067	0.9872
0.9989	23.12	4740	0.0067	0.9872
0.9998	23.27	4770	0.0067	0.9872
0.9999	23.41	4800	0.0067	0.9872
0.9996	23.56	4830	0.0067	0.9872
0.9974	23.71	4860	0.0067	0.9872
0.9997	23.85	4890	0.0067	0.9872
0.9999	24.0	4920	0.0067	0.9872
0.9962	24.15	4950	0.0067	0.9872
0.9996	24.29	4980	0.0067	0.9872
0.9999	24.44	5010	0.0067	0.9872
0.9973	24.59	5040	0.0067	0.9872
0.9996	24.73	5070	0.0067	0.9872
0.9995	24.88	5100	0.0067	0.9872
0.9999	25.02	5130	0.0067	0.9872
0.9988	25.17	5160	0.0067	0.9872
1.0	25.32	5190	0.0067	0.9872
1.0	25.46	5220	0.0067	0.9872
0.9996	25.61	5250	0.0067	0.9872
0.9965	25.76	5280	0.0067	0.9872
0.9976	25.9	5310	0.0067	0.9872
1.0	26.05	5340	0.0067	0.9872
0.9989	26.2	5370	0.0067	0.9872
0.9996	26.34	5400	0.0067	0.9872
0.9998	26.49	5430	0.0067	0.9872
1.0	26.63	5460	0.0067	0.9872
0.9996	26.78	5490	0.0067	0.9872
0.9972	26.93	5520	0.0067	0.9872
0.9984	27.07	5550	0.0067	0.9872
0.9961	27.22	5580	0.0067	0.9872
1.0	27.37	5610	0.0067	0.9872
0.9977	27.51	5640	0.0067	0.9872
0.9969	27.66	5670	0.0067	0.9872
0.9971	27.8	5700	0.0067	0.9872
0.9986	27.95	5730	0.0067	0.9872
0.9995	28.1	5760	0.0067	0.9872
0.9992	28.24	5790	0.0067	0.9872
0.9976	28.39	5820	0.0067	0.9872
0.9994	28.54	5850	0.0067	0.9872
0.998	28.68	5880	0.0067	0.9872
0.9952	28.83	5910	0.0067	0.9872
0.9998	28.98	5940	0.0067	0.9872
0.9937	29.12	5970	0.0067	0.9872
0.9989	29.27	6000	0.0067	0.9872
0.9993	29.41	6030	0.0067	0.9872
0.9989	29.56	6060	0.0067	0.9872
0.999	29.71	6090	0.0067	0.9872
0.9939	29.85	6120	0.0067	0.9872
1.0	30.0	6150	0.0067	0.9872
0.9996	30.15	6180	0.0067	0.9872
0.9994	30.29	6210	0.0067	0.9872
0.999	30.44	6240	0.0067	0.9872
1.0	30.59	6270	0.0067	0.9872
0.9956	30.73	6300	0.0067	0.9872
0.9971	30.88	6330	0.0067	0.9872
0.9985	31.02	6360	0.0067	0.9872
1.0	31.17	6390	0.0067	0.9872
0.9987	31.32	6420	0.0067	0.9872
0.9992	31.46	6450	0.0067	0.9872
0.9996	31.61	6480	0.0067	0.9872
0.9998	31.76	6510	0.0067	0.9872
0.9989	31.9	6540	0.0067	0.9872
1.0	32.05	6570	0.0067	0.9872
0.9966	32.2	6600	0.0067	0.9872
0.9994	32.34	6630	0.0067	0.9872
0.9987	32.49	6660	0.0067	0.9872
0.9993	32.63	6690	0.0067	0.9872
0.9971	32.78	6720	0.0067	0.9872
0.9971	32.93	6750	0.0067	0.9872
0.9929	33.07	6780	0.0067	0.9872
0.9997	33.22	6810	0.0067	0.9872
0.9978	33.37	6840	0.0067	0.9872
1.0	33.51	6870	0.0067	0.9872
0.9991	33.66	6900	0.0067	0.9872
0.9971	33.8	6930	0.0067	0.9872
0.9999	33.95	6960	0.0067	0.9872
0.9999	34.1	6990	0.0067	0.9872
0.9997	34.24	7020	0.0067	0.9872
1.0	34.39	7050	0.0067	0.9872
0.9986	34.54	7080	0.0067	0.9872
0.9996	34.68	7110	0.0067	0.9872
0.9994	34.83	7140	0.0067	0.9872
0.9997	34.98	7170	0.0067	0.9872
0.9999	35.12	7200	0.0067	0.9872
0.9969	35.27	7230	0.0067	0.9872
1.0	35.41	7260	0.0067	0.9872
0.9984	35.56	7290	0.0067	0.9872
0.9961	35.71	7320	0.0067	0.9872
0.9988	35.85	7350	0.0067	0.9872
0.9985	36.0	7380	0.0067	0.9872
0.9997	36.15	7410	0.0067	0.9872
1.0	36.29	7440	0.0067	0.9872
0.9987	36.44	7470	0.0067	0.9872
0.9991	36.59	7500	0.0067	0.9872
0.9992	36.73	7530	0.0067	0.9872
0.9999	36.88	7560	0.0067	0.9872
0.9996	37.02	7590	0.0067	0.9872
0.9995	37.17	7620	0.0067	0.9872
0.9998	37.32	7650	0.0067	0.9872
0.9969	37.46	7680	0.0067	0.9872
0.9989	37.61	7710	0.0067	0.9872
0.9992	37.76	7740	0.0067	0.9872
0.9959	37.9	7770	0.0067	0.9872
0.9987	38.05	7800	0.0067	0.9872
0.998	38.2	7830	0.0067	0.9872
0.9992	38.34	7860	0.0067	0.9872
0.9992	38.49	7890	0.0067	0.9872
0.9993	38.63	7920	0.0067	0.9872
0.9997	38.78	7950	0.0067	0.9872
0.9976	38.93	7980	0.0067	0.9872
1.0	39.07	8010	0.0067	0.9872
0.9959	39.22	8040	0.0067	0.9872
0.9973	39.37	8070	0.0067	0.9872
0.9996	39.51	8100	0.0067	0.9872
1.0	39.66	8130	0.0067	0.9872
0.9986	39.8	8160	0.0067	0.9872
0.9999	39.95	8190	0.0067	0.9872

Framework versions

• Transformers 4.37.1
• Pytorch 2.1.2
• Datasets 2.16.1
• Tokenizers 0.15.1