deeplabv3-mobilevit-small-corm

This model is a fine-tuned version of apple/deeplabv3-mobilevit-small on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3683
• Mean Iou: 0.5100
• Mean Accuracy: 0.8339
• Overall Accuracy: 0.8317
• Accuracy Background: nan
• Accuracy Corm: 0.8057
• Accuracy Damage: 0.8621
• Iou Background: 0.0
• Iou Corm: 0.7541
• Iou Damage: 0.7759

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: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Background	Accuracy Corm	Accuracy Damage	Iou Background	Iou Corm	Iou Damage
1.0605	2.2222	20	1.0716	0.2673	0.5409	0.5579	nan	0.7620	0.3198	0.0	0.5135	0.2884
1.0126	4.4444	40	1.0116	0.2753	0.5032	0.5099	nan	0.5897	0.4167	0.0	0.4581	0.3679
0.9674	6.6667	60	0.9641	0.2948	0.5118	0.5104	nan	0.4937	0.5298	0.0	0.4295	0.4548
0.9281	8.8889	80	0.9208	0.3102	0.5386	0.5337	nan	0.4750	0.6022	0.0	0.4233	0.5073
0.8893	11.1111	100	0.8746	0.3071	0.5235	0.5191	nan	0.4660	0.5811	0.0	0.4201	0.5013
0.8539	13.3333	120	0.8368	0.3205	0.5450	0.5394	nan	0.4718	0.6182	0.0	0.4335	0.5279
0.8164	15.5556	140	0.7937	0.3227	0.5465	0.5399	nan	0.4613	0.6316	0.0	0.4274	0.5407
0.7956	17.7778	160	0.7577	0.3383	0.5686	0.5634	nan	0.5010	0.6361	0.0	0.4620	0.5528
0.7716	20.0	180	0.7257	0.3475	0.5847	0.5779	nan	0.4960	0.6735	0.0	0.4636	0.5789
0.7278	22.2222	200	0.6851	0.3567	0.5920	0.5902	nan	0.5683	0.6157	0.0	0.5155	0.5546
0.6895	24.4444	220	0.6568	0.3787	0.6295	0.6261	nan	0.5853	0.6738	0.0	0.5408	0.5952
0.6668	26.6667	240	0.6265	0.3920	0.6505	0.6476	nan	0.6125	0.6885	0.0	0.5661	0.6100
0.6513	28.8889	260	0.6051	0.4073	0.6788	0.6769	nan	0.6542	0.7033	0.0	0.5976	0.6242
0.6197	31.1111	280	0.5823	0.4129	0.6843	0.6828	nan	0.6654	0.7032	0.0	0.6085	0.6301
0.6191	33.3333	300	0.5574	0.4217	0.7003	0.6976	nan	0.6657	0.7349	0.0	0.6136	0.6514
0.5747	35.5556	320	0.5463	0.4354	0.7230	0.7203	nan	0.6872	0.7588	0.0	0.6342	0.6719
0.567	37.7778	340	0.5256	0.4361	0.7293	0.7248	nan	0.6707	0.7878	0.0	0.6269	0.6813
0.5657	40.0	360	0.5091	0.4410	0.7344	0.7304	nan	0.6828	0.7860	0.0	0.6367	0.6862
0.5355	42.2222	380	0.4998	0.4493	0.7531	0.7483	nan	0.6915	0.8146	0.0	0.6478	0.7001
0.529	44.4444	400	0.4865	0.4547	0.7564	0.7528	nan	0.7096	0.8032	0.0	0.6606	0.7033
0.5019	46.6667	420	0.4756	0.4586	0.7666	0.7622	nan	0.7092	0.8241	0.0	0.6644	0.7114
0.4869	48.8889	440	0.4659	0.4656	0.7748	0.7714	nan	0.7305	0.8191	0.0	0.6799	0.7168
0.481	51.1111	460	0.4542	0.4646	0.7736	0.7696	nan	0.7221	0.8251	0.0	0.6762	0.7177
0.4795	53.3333	480	0.4441	0.4690	0.7790	0.7753	nan	0.7304	0.8276	0.0	0.6838	0.7232
0.4626	55.5556	500	0.4379	0.4774	0.7910	0.7880	nan	0.7512	0.8309	0.0	0.7003	0.7319
0.4679	57.7778	520	0.4296	0.4791	0.7941	0.7906	nan	0.7480	0.8403	0.0	0.7007	0.7366
0.478	60.0	540	0.4198	0.4821	0.7961	0.7931	nan	0.7563	0.8359	0.0	0.7066	0.7396
0.4562	62.2222	560	0.4142	0.4803	0.7976	0.7931	nan	0.7394	0.8558	0.0	0.6986	0.7424
0.4309	64.4444	580	0.4111	0.4857	0.8063	0.8021	nan	0.7513	0.8614	0.0	0.7090	0.7481
0.4239	66.6667	600	0.4040	0.4918	0.8105	0.8075	nan	0.7708	0.8503	0.0	0.7224	0.7529
0.4436	68.8889	620	0.3981	0.4922	0.8121	0.8086	nan	0.7669	0.8573	0.0	0.7220	0.7547
0.4253	71.1111	640	0.3924	0.4915	0.8101	0.8065	nan	0.7628	0.8574	0.0	0.7192	0.7553
0.4054	73.3333	660	0.3926	0.4991	0.8197	0.8172	nan	0.7869	0.8526	0.0	0.7361	0.7611
0.4368	75.5556	680	0.3861	0.5028	0.8241	0.8220	nan	0.7970	0.8511	0.0	0.7437	0.7648
0.4	77.7778	700	0.3836	0.5009	0.8245	0.8210	nan	0.7787	0.8704	0.0	0.7352	0.7675
0.3801	80.0	720	0.3815	0.4985	0.8231	0.8188	nan	0.7677	0.8784	0.0	0.7290	0.7664
0.3929	82.2222	740	0.3788	0.5048	0.8276	0.8249	nan	0.7935	0.8616	0.0	0.7447	0.7698
0.3923	84.4444	760	0.3752	0.5038	0.8283	0.8249	nan	0.7851	0.8715	0.0	0.7406	0.7708
0.412	86.6667	780	0.3735	0.5029	0.8286	0.8247	nan	0.7775	0.8798	0.0	0.7370	0.7718
0.3757	88.8889	800	0.3731	0.5082	0.8335	0.8306	nan	0.7962	0.8708	0.0	0.7492	0.7753
0.3863	91.1111	820	0.3729	0.5063	0.8336	0.8299	nan	0.7854	0.8819	0.0	0.7437	0.7751
0.3833	93.3333	840	0.3709	0.5044	0.8310	0.8269	nan	0.7777	0.8843	0.0	0.7391	0.7741
0.378	95.5556	860	0.3700	0.5073	0.8330	0.8297	nan	0.7896	0.8765	0.0	0.7462	0.7758
0.3942	97.7778	880	0.3679	0.5070	0.8316	0.8284	nan	0.7900	0.8732	0.0	0.7457	0.7752
0.3877	100.0	900	0.3683	0.5100	0.8339	0.8317	nan	0.8057	0.8621	0.0	0.7541	0.7759

Framework versions

• Transformers 4.49.0
• Pytorch 2.6.0+cu124
• Datasets 3.3.2
• Tokenizers 0.21.0