segformer-finetuned-blueberries-3classes_final_w_aug

This model is a fine-tuned version of nvidia/mit-b0 on the sankarip/blue_berry_data_3classes_final_w_aug dataset. It achieves the following results on the evaluation set:

• Loss: 0.1885
• Mean Iou: 0.3832
• Mean Accuracy: 0.5196
• Overall Accuracy: 0.8996
• Accuracy Background: nan
• Accuracy Berry: 0.9462
• Accuracy Debris: 0.2946
• Accuracy Greenberry: 0.3179
• Iou Background: 0.0
• Iou Berry: 0.9393
• Iou Debris: 0.2765
• Iou Greenberry: 0.3171

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: 8
• eval_batch_size: 8
• seed: 1337
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: polynomial
• training_steps: 4000

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Background	Accuracy Berry	Accuracy Debris	Accuracy Greenberry	Iou Background	Iou Berry	Iou Debris	Iou Greenberry
No log	1.0	30	0.8151	0.3153	0.4411	0.8407	nan	0.8817	0.0032	0.4383	0.0	0.8616	0.0032	0.3967
No log	2.0	60	0.4668	0.3668	0.5138	0.8848	nan	0.9188	0.0077	0.6149	0.0	0.9107	0.0077	0.5490
No log	3.0	90	0.3764	0.3651	0.5148	0.8676	nan	0.8988	0.0050	0.6408	0.0	0.8934	0.0050	0.5621
0.6234	4.0	120	0.3360	0.3528	0.4941	0.8517	nan	0.8844	0.0039	0.5941	0.0	0.8796	0.0039	0.5276
0.6234	5.0	150	0.3083	0.3813	0.5264	0.9045	nan	0.9402	0.0368	0.6021	0.0	0.9306	0.0366	0.5581
0.6234	6.0	180	0.2831	0.3611	0.4920	0.8379	nan	0.8713	0.0627	0.5420	0.0	0.8675	0.0627	0.5143
0.3285	7.0	210	0.2698	0.3364	0.4530	0.8340	nan	0.8745	0.0727	0.4117	0.0	0.8712	0.0726	0.4017
0.3285	8.0	240	0.2419	0.3606	0.4869	0.8626	nan	0.9022	0.1045	0.4540	0.0	0.8966	0.1032	0.4427
0.3285	9.0	270	0.2431	0.3583	0.4833	0.8338	nan	0.8701	0.1090	0.4709	0.0	0.8672	0.1078	0.4580
0.2612	10.0	300	0.2555	0.3600	0.4849	0.8635	nan	0.9049	0.1337	0.4161	0.0	0.8995	0.1318	0.4088
0.2612	11.0	330	0.2347	0.3834	0.5180	0.8544	nan	0.8888	0.1458	0.5194	0.0	0.8855	0.1425	0.5057
0.2612	12.0	360	0.2266	0.3675	0.4948	0.8608	nan	0.9006	0.1514	0.4325	0.0	0.8963	0.1478	0.4258
0.2612	13.0	390	0.2207	0.3706	0.4988	0.8754	nan	0.9168	0.1581	0.4215	0.0	0.9121	0.1535	0.4169
0.2294	14.0	420	0.2194	0.3873	0.5234	0.8573	nan	0.8923	0.1762	0.5016	0.0	0.8890	0.1689	0.4913
0.2294	15.0	450	0.2048	0.3960	0.5349	0.8778	nan	0.9137	0.1781	0.5128	0.0	0.9091	0.1719	0.5031
0.2294	16.0	480	0.2131	0.3771	0.5087	0.8739	nan	0.9145	0.1878	0.4238	0.0	0.9102	0.1788	0.4194
0.21	17.0	510	0.2055	0.4095	0.5540	0.9062	nan	0.9432	0.1909	0.5279	0.0	0.9377	0.1833	0.5169
0.21	18.0	540	0.2089	0.3820	0.5154	0.8596	nan	0.8975	0.2025	0.4462	0.0	0.8950	0.1921	0.4410
0.21	19.0	570	0.2051	0.3809	0.5126	0.8855	nan	0.9268	0.1798	0.4312	0.0	0.9214	0.1754	0.4269
0.1978	20.0	600	0.2054	0.3785	0.5095	0.8697	nan	0.9098	0.1930	0.4256	0.0	0.9066	0.1855	0.4217
0.1978	21.0	630	0.2101	0.3633	0.4876	0.8318	nan	0.8696	0.1746	0.4187	0.0	0.8673	0.1716	0.4142
0.1978	22.0	660	0.2048	0.3622	0.4865	0.8662	nan	0.9098	0.1873	0.3623	0.0	0.9061	0.1822	0.3606
0.1978	23.0	690	0.1930	0.3969	0.5352	0.8894	nan	0.9280	0.2049	0.4727	0.0	0.9237	0.1981	0.4657
0.1833	24.0	720	0.1921	0.4111	0.5571	0.8989	nan	0.9362	0.2409	0.4941	0.0	0.9310	0.2264	0.4872
0.1833	25.0	750	0.1964	0.3812	0.5129	0.8908	nan	0.9339	0.2079	0.3969	0.0	0.9287	0.2017	0.3945
0.1833	26.0	780	0.1931	0.3742	0.5018	0.8758	nan	0.9177	0.1820	0.4056	0.0	0.9139	0.1805	0.4024
0.1867	27.0	810	0.1958	0.3644	0.4890	0.8638	nan	0.9066	0.1890	0.3714	0.0	0.9033	0.1847	0.3695
0.1867	28.0	840	0.1892	0.3954	0.5318	0.8893	nan	0.9273	0.1743	0.4937	0.0	0.9229	0.1733	0.4855
0.1867	29.0	870	0.2003	0.3939	0.5308	0.9083	nan	0.9510	0.2187	0.4225	0.0	0.9441	0.2113	0.4200
0.1821	30.0	900	0.1854	0.3975	0.5336	0.8888	nan	0.9277	0.2062	0.4668	0.0	0.9239	0.2046	0.4612
0.1821	31.0	930	0.1889	0.3836	0.5154	0.8794	nan	0.9213	0.2309	0.3939	0.0	0.9181	0.2248	0.3914
0.1821	32.0	960	0.1874	0.4134	0.5588	0.8895	nan	0.9257	0.2542	0.4964	0.0	0.9210	0.2431	0.4897
0.1821	33.0	990	0.1993	0.3867	0.5208	0.8640	nan	0.9030	0.2403	0.4192	0.0	0.9003	0.2315	0.4150
0.1785	34.0	1020	0.1970	0.4020	0.5429	0.9142	nan	0.9564	0.2407	0.4316	0.0	0.9496	0.2306	0.4280
0.1785	35.0	1050	0.1925	0.3935	0.5290	0.9010	nan	0.9423	0.1994	0.4454	0.0	0.9365	0.1964	0.4411
0.1785	36.0	1080	0.1938	0.3867	0.5202	0.9006	nan	0.9449	0.2351	0.3805	0.0	0.9390	0.2293	0.3787
0.1693	37.0	1110	0.1852	0.4027	0.5436	0.8962	nan	0.9369	0.2717	0.4223	0.0	0.9321	0.2598	0.4191
0.1693	38.0	1140	0.1943	0.3970	0.5370	0.9019	nan	0.9441	0.2560	0.4110	0.0	0.9383	0.2424	0.4075
0.1693	39.0	1170	0.1874	0.4214	0.5702	0.9130	nan	0.9505	0.2554	0.5046	0.0	0.9439	0.2442	0.4975
0.1668	40.0	1200	0.1870	0.3924	0.5298	0.8884	nan	0.9302	0.2650	0.3941	0.0	0.9261	0.2515	0.3919
0.1668	41.0	1230	0.1874	0.3730	0.5034	0.8887	nan	0.9351	0.2528	0.3224	0.0	0.9300	0.2407	0.3214
0.1668	42.0	1260	0.1895	0.4208	0.5693	0.9170	nan	0.9561	0.2749	0.4770	0.0	0.9484	0.2620	0.4727
0.1668	43.0	1290	0.1905	0.3731	0.5025	0.8973	nan	0.9445	0.2386	0.3244	0.0	0.9383	0.2308	0.3233
0.1693	44.0	1320	0.2006	0.4134	0.5608	0.9205	nan	0.9617	0.2750	0.4457	0.0	0.9519	0.2592	0.4423
0.1693	45.0	1350	0.1893	0.3796	0.5106	0.9038	nan	0.9498	0.2227	0.3594	0.0	0.9428	0.2179	0.3576
0.1693	46.0	1380	0.1935	0.3804	0.5127	0.9079	nan	0.9539	0.2160	0.3680	0.0	0.9459	0.2094	0.3662
0.1659	47.0	1410	0.1888	0.3650	0.4910	0.8868	nan	0.9337	0.2146	0.3246	0.0	0.9288	0.2079	0.3233
0.1659	48.0	1440	0.1884	0.3773	0.5084	0.9027	nan	0.9495	0.2355	0.3401	0.0	0.9428	0.2280	0.3384
0.1659	49.0	1470	0.1913	0.3944	0.5331	0.9063	nan	0.9507	0.2757	0.3730	0.0	0.9439	0.2623	0.3712
0.1638	50.0	1500	0.1820	0.4074	0.5510	0.8936	nan	0.9333	0.2896	0.4301	0.0	0.9283	0.2743	0.4270
0.1638	51.0	1530	0.1954	0.3931	0.5320	0.9166	nan	0.9611	0.2373	0.3977	0.0	0.9514	0.2253	0.3956
0.1638	52.0	1560	0.1939	0.4023	0.5504	0.9102	nan	0.9546	0.3453	0.3512	0.0	0.9472	0.3122	0.3498
0.1638	53.0	1590	0.1844	0.4046	0.5514	0.9008	nan	0.9430	0.3280	0.3832	0.0	0.9379	0.2989	0.3817
0.1614	54.0	1620	0.1852	0.3934	0.5316	0.8952	nan	0.9383	0.2780	0.3786	0.0	0.9326	0.2645	0.3767
0.1614	55.0	1650	0.1813	0.4034	0.5444	0.8977	nan	0.9391	0.2872	0.4069	0.0	0.9342	0.2750	0.4044
0.1614	56.0	1680	0.1927	0.3839	0.5199	0.9035	nan	0.9499	0.2752	0.3345	0.0	0.9424	0.2598	0.3335
0.1595	57.0	1710	0.1929	0.3878	0.5234	0.9130	nan	0.9584	0.2312	0.3806	0.0	0.9489	0.2237	0.3786
0.1595	58.0	1740	0.1840	0.3987	0.5417	0.8968	nan	0.9394	0.3068	0.3789	0.0	0.9341	0.2838	0.3769
0.1595	59.0	1770	0.1812	0.4011	0.5440	0.8935	nan	0.9355	0.3139	0.3828	0.0	0.9307	0.2928	0.3807
0.1583	60.0	1800	0.1852	0.3848	0.5214	0.8898	nan	0.9350	0.3050	0.3240	0.0	0.9300	0.2864	0.3230
0.1583	61.0	1830	0.1871	0.4036	0.5478	0.8934	nan	0.9350	0.3236	0.3849	0.0	0.9297	0.3016	0.3829
0.1583	62.0	1860	0.1884	0.3968	0.5377	0.9036	nan	0.9480	0.3098	0.3554	0.0	0.9415	0.2915	0.3542
0.1583	63.0	1890	0.1826	0.3930	0.5318	0.8935	nan	0.9372	0.2990	0.3594	0.0	0.9320	0.2819	0.3580
0.1579	64.0	1920	0.1809	0.4386	0.5969	0.9103	nan	0.9462	0.3475	0.4970	0.0	0.9402	0.3216	0.4924
0.1579	65.0	1950	0.1861	0.4011	0.5459	0.9027	nan	0.9462	0.3275	0.3641	0.0	0.9403	0.3013	0.3628
0.1579	66.0	1980	0.2017	0.3952	0.5359	0.9201	nan	0.9660	0.2787	0.3628	0.0	0.9550	0.2640	0.3616
0.159	67.0	2010	0.1943	0.4073	0.5499	0.9142	nan	0.9570	0.2859	0.4067	0.0	0.9489	0.2752	0.4051
0.159	68.0	2040	0.1900	0.3703	0.4997	0.8970	nan	0.9451	0.2492	0.3049	0.0	0.9386	0.2384	0.3042
0.159	69.0	2070	0.1925	0.3961	0.5384	0.9103	nan	0.9548	0.2907	0.3696	0.0	0.9453	0.2706	0.3684
0.1592	70.0	2100	0.1822	0.4166	0.5648	0.9049	nan	0.9447	0.3153	0.4344	0.0	0.9387	0.2956	0.4320
0.1592	71.0	2130	0.1881	0.3927	0.5316	0.9038	nan	0.9485	0.2881	0.3582	0.0	0.9413	0.2728	0.3569
0.1592	72.0	2160	0.1855	0.3996	0.5398	0.9063	nan	0.9490	0.2670	0.4035	0.0	0.9423	0.2554	0.4007
0.1592	73.0	2190	0.1894	0.3988	0.5415	0.9107	nan	0.9546	0.2857	0.3841	0.0	0.9468	0.2665	0.3819
0.1576	74.0	2220	0.1885	0.4126	0.5585	0.9184	nan	0.9603	0.2908	0.4244	0.0	0.9525	0.2758	0.4221
0.1576	75.0	2250	0.1896	0.3922	0.5304	0.9074	nan	0.9520	0.2641	0.3753	0.0	0.9444	0.2511	0.3734
0.1576	76.0	2280	0.1906	0.4095	0.5567	0.9167	nan	0.9584	0.2846	0.4272	0.0	0.9502	0.2631	0.4246
0.1524	77.0	2310	0.1940	0.3895	0.5293	0.9078	nan	0.9535	0.2861	0.3483	0.0	0.9450	0.2662	0.3469
0.1524	78.0	2340	0.1923	0.3836	0.5209	0.9016	nan	0.9483	0.2942	0.3201	0.0	0.9412	0.2742	0.3192
0.1524	79.0	2370	0.1990	0.3812	0.5188	0.9007	nan	0.9479	0.3027	0.3058	0.0	0.9407	0.2790	0.3050
0.1495	80.0	2400	0.1871	0.3912	0.5302	0.9051	nan	0.9500	0.2813	0.3594	0.0	0.9431	0.2642	0.3576
0.1495	81.0	2430	0.1917	0.3673	0.4980	0.8866	nan	0.9348	0.2825	0.2767	0.0	0.9290	0.2640	0.2761
0.1495	82.0	2460	0.1853	0.4002	0.5448	0.8950	nan	0.9376	0.3282	0.3685	0.0	0.9327	0.3014	0.3665
0.1495	83.0	2490	0.1906	0.3866	0.5242	0.8916	nan	0.9362	0.2950	0.3413	0.0	0.9304	0.2756	0.3403
0.1546	84.0	2520	0.1865	0.3935	0.5336	0.8963	nan	0.9404	0.3079	0.3525	0.0	0.9344	0.2886	0.3511
0.1546	85.0	2550	0.1876	0.4078	0.5561	0.9047	nan	0.9473	0.3455	0.3754	0.0	0.9413	0.3161	0.3740
0.1546	86.0	2580	0.1862	0.4023	0.5470	0.9000	nan	0.9428	0.3258	0.3724	0.0	0.9373	0.3007	0.3712
0.1491	87.0	2610	0.1851	0.3995	0.5442	0.8899	nan	0.9321	0.3348	0.3656	0.0	0.9268	0.3070	0.3642
0.1491	88.0	2640	0.1858	0.3991	0.5408	0.8989	nan	0.9420	0.3071	0.3732	0.0	0.9360	0.2886	0.3716
0.1491	89.0	2670	0.1905	0.3802	0.5140	0.9009	nan	0.9476	0.2657	0.3287	0.0	0.9400	0.2530	0.3278
0.1497	90.0	2700	0.1848	0.4055	0.5504	0.9061	nan	0.9487	0.3135	0.3891	0.0	0.9420	0.2928	0.3872
0.1497	91.0	2730	0.1848	0.3820	0.5162	0.8942	nan	0.9400	0.2801	0.3285	0.0	0.9344	0.2660	0.3275
0.1497	92.0	2760	0.1848	0.3930	0.5329	0.8932	nan	0.9371	0.3115	0.3502	0.0	0.9317	0.2912	0.3491
0.1497	93.0	2790	0.1870	0.3861	0.5208	0.9005	nan	0.9453	0.2520	0.3652	0.0	0.9383	0.2424	0.3637
0.1514	94.0	2820	0.1859	0.3767	0.5093	0.8899	nan	0.9369	0.2909	0.3001	0.0	0.9314	0.2760	0.2993
0.1514	95.0	2850	0.1890	0.3694	0.4989	0.8941	nan	0.9425	0.2629	0.2913	0.0	0.9364	0.2505	0.2906
0.1514	96.0	2880	0.1896	0.3852	0.5223	0.9015	nan	0.9475	0.2847	0.3349	0.0	0.9398	0.2672	0.3339
0.1479	97.0	2910	0.1882	0.3834	0.5203	0.9015	nan	0.9483	0.2948	0.3179	0.0	0.9419	0.2748	0.3170
0.1479	98.0	2940	0.1907	0.3706	0.5011	0.9003	nan	0.9490	0.2579	0.2965	0.0	0.9412	0.2453	0.2957
0.1479	99.0	2970	0.1915	0.3766	0.5098	0.9011	nan	0.9485	0.2628	0.3180	0.0	0.9413	0.2481	0.3172
0.1498	100.0	3000	0.1854	0.3950	0.5350	0.9018	nan	0.9457	0.2901	0.3690	0.0	0.9394	0.2732	0.3676
0.1498	101.0	3030	0.1885	0.3856	0.5230	0.8986	nan	0.9444	0.2926	0.3321	0.0	0.9377	0.2735	0.3311
0.1498	102.0	3060	0.1863	0.3970	0.5375	0.9032	nan	0.9466	0.2829	0.3830	0.0	0.9395	0.2668	0.3815
0.1498	103.0	3090	0.1933	0.3840	0.5218	0.9060	nan	0.9530	0.2910	0.3215	0.0	0.9448	0.2708	0.3206
0.1497	104.0	3120	0.1885	0.3784	0.5113	0.8983	nan	0.9451	0.2657	0.3231	0.0	0.9382	0.2528	0.3224
0.1497	105.0	3150	0.1882	0.3827	0.5196	0.8948	nan	0.9407	0.2936	0.3245	0.0	0.9346	0.2726	0.3239
0.1497	106.0	3180	0.1865	0.3922	0.5317	0.9056	nan	0.9510	0.2974	0.3466	0.0	0.9434	0.2797	0.3456
0.1511	107.0	3210	0.1890	0.3874	0.5245	0.9022	nan	0.9478	0.2814	0.3443	0.0	0.9406	0.2659	0.3432
0.1511	108.0	3240	0.1856	0.3831	0.5188	0.8972	nan	0.9435	0.2921	0.3209	0.0	0.9375	0.2747	0.3201
0.1511	109.0	3270	0.1878	0.3912	0.5309	0.9023	nan	0.9474	0.3005	0.3447	0.0	0.9409	0.2804	0.3436
0.1476	110.0	3300	0.1921	0.3856	0.5217	0.9081	nan	0.9547	0.2736	0.3367	0.0	0.9464	0.2603	0.3357
0.1476	111.0	3330	0.1888	0.3838	0.5204	0.9019	nan	0.9486	0.2902	0.3224	0.0	0.9412	0.2723	0.3216
0.1476	112.0	3360	0.1873	0.3848	0.5215	0.8992	nan	0.9454	0.2946	0.3244	0.0	0.9391	0.2765	0.3235
0.1476	113.0	3390	0.1902	0.3814	0.5168	0.9033	nan	0.9505	0.2829	0.3170	0.0	0.9429	0.2667	0.3162
0.1489	114.0	3420	0.1876	0.3858	0.5224	0.8996	nan	0.9453	0.2859	0.3361	0.0	0.9386	0.2697	0.3351
0.1489	115.0	3450	0.1885	0.3864	0.5242	0.8990	nan	0.9448	0.2973	0.3305	0.0	0.9385	0.2775	0.3295
0.1489	116.0	3480	0.1855	0.3909	0.5298	0.8980	nan	0.9426	0.2992	0.3477	0.0	0.9366	0.2806	0.3465
0.1433	117.0	3510	0.1869	0.3916	0.5314	0.8980	nan	0.9423	0.2974	0.3547	0.0	0.9360	0.2773	0.3531
0.1433	118.0	3540	0.1897	0.3818	0.5174	0.8978	nan	0.9441	0.2827	0.3254	0.0	0.9373	0.2655	0.3245
0.1433	119.0	3570	0.1879	0.3930	0.5341	0.9040	nan	0.9490	0.3061	0.3474	0.0	0.9415	0.2844	0.3462
0.1457	120.0	3600	0.1897	0.3840	0.5219	0.8976	nan	0.9440	0.3084	0.3133	0.0	0.9372	0.2863	0.3124
0.1457	121.0	3630	0.1887	0.3880	0.5262	0.9007	nan	0.9464	0.2964	0.3359	0.0	0.9392	0.2779	0.3349
0.1457	122.0	3660	0.1856	0.3953	0.5360	0.9030	nan	0.9473	0.3000	0.3607	0.0	0.9405	0.2816	0.3592
0.1457	123.0	3690	0.1909	0.3888	0.5270	0.9054	nan	0.9515	0.2943	0.3353	0.0	0.9439	0.2770	0.3343
0.1454	124.0	3720	0.1866	0.3905	0.5285	0.9024	nan	0.9473	0.2822	0.3559	0.0	0.9404	0.2672	0.3545
0.1454	125.0	3750	0.1886	0.3853	0.5224	0.9018	nan	0.9482	0.2927	0.3263	0.0	0.9409	0.2751	0.3254
0.1454	126.0	3780	0.1876	0.3872	0.5254	0.8998	nan	0.9456	0.3014	0.3291	0.0	0.9387	0.2817	0.3282
0.1459	127.0	3810	0.1885	0.3910	0.5302	0.9050	nan	0.9506	0.2988	0.3413	0.0	0.9435	0.2804	0.3402
0.1459	128.0	3840	0.1879	0.3865	0.5244	0.9010	nan	0.9473	0.3040	0.3218	0.0	0.9406	0.2844	0.3209
0.1459	129.0	3870	0.1894	0.3790	0.5135	0.8984	nan	0.9456	0.2849	0.3099	0.0	0.9384	0.2685	0.3092
0.1476	130.0	3900	0.1880	0.3867	0.5244	0.8996	nan	0.9456	0.2996	0.3280	0.0	0.9387	0.2812	0.3271
0.1476	131.0	3930	0.1890	0.3859	0.5231	0.9011	nan	0.9473	0.2948	0.3271	0.0	0.9402	0.2773	0.3262
0.1476	132.0	3960	0.1903	0.3867	0.5245	0.9047	nan	0.9513	0.2980	0.3244	0.0	0.9438	0.2794	0.3235
0.1476	133.0	3990	0.1888	0.3863	0.5237	0.9027	nan	0.9491	0.2982	0.3239	0.0	0.9421	0.2799	0.3230
0.1454	133.3333	4000	0.1885	0.3832	0.5196	0.8996	nan	0.9462	0.2946	0.3179	0.0	0.9393	0.2765	0.3171

Framework versions

• Transformers 4.48.0
• Pytorch 2.4.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0