segformer-b5-finetuned-apple-dms-run3

This model is a fine-tuned version of nvidia/segformer-b5-finetuned-ade-640-640 on the AllanK24/apple-dms-materials dataset. It achieves the following results on the evaluation set:

Loss: 11.7839
Mean Iou: 0.4203
Mean Accuracy: 0.4907
Overall Accuracy: 0.7320
Iou No Label: 0.0
Accuracy Animal Skin: 0.5754
Iou Animal Skin: 0.3556
Accuracy Bone Teeth Horn: 0.0028
Iou Bone Teeth Horn: 0.0028
Accuracy Brickwork: 0.6283
Iou Brickwork: 0.5529
Accuracy Cardboard: 0.5400
Iou Cardboard: 0.4636
Accuracy Carpet Rug: 0.7694
Iou Carpet Rug: 0.7043
Accuracy Ceiling Tile: 0.7873
Iou Ceiling Tile: 0.7258
Accuracy Ceramic: 0.7227
Iou Ceramic: 0.6279
Accuracy Chalkboard Blackboard: 0.7058
Iou Chalkboard Blackboard: 0.6366
Accuracy Clutter: 0.0031
Iou Clutter: 0.0028
Accuracy Concrete: 0.4385
Iou Concrete: 0.3509
Accuracy Cork Corkboard: 0.0203
Iou Cork Corkboard: 0.0191
Accuracy Engineered Stone: 0.0238
Iou Engineered Stone: 0.0213
Accuracy Fabric Cloth: 0.8462
Iou Fabric Cloth: 0.7791
Accuracy Fiberglass Wool: 0.0
Iou Fiberglass Wool: 0.0
Accuracy Fire: 0.3918
Iou Fire: 0.3545
Accuracy Foliage: 0.9080
Iou Foliage: 0.8384
Accuracy Food: 0.8727
Iou Food: 0.7758
Accuracy Fur: 0.8913
Iou Fur: 0.8047
Accuracy Gemstone Quartz: 0.2157
Iou Gemstone Quartz: 0.1761
Accuracy Glass: 0.6599
Iou Glass: 0.5828
Accuracy Hair: 0.8424
Iou Hair: 0.7460
Accuracy Ice: 0.3504
Iou Ice: 0.2609
Accuracy Leather: 0.3762
Iou Leather: 0.3338
Accuracy Liquid Non-water: 0.2289
Iou Liquid Non-water: 0.2157
Accuracy Metal: 0.3665
Iou Metal: 0.3088
Accuracy Mirror: 0.4631
Iou Mirror: 0.4021
Accuracy Paint Plaster Enamel: 0.7616
Iou Paint Plaster Enamel: 0.6999
Accuracy Paper: 0.6700
Iou Paper: 0.5762
Accuracy Pearl: 0.0
Iou Pearl: 0.0
Accuracy Photograph Painting: 0.3697
Iou Photograph Painting: 0.3016
Accuracy Plastic Clear: 0.3195
Iou Plastic Clear: 0.2717
Accuracy Plastic Non-clear: 0.3865
Iou Plastic Non-clear: 0.3220
Accuracy Rubber Latex: 0.3122
Iou Rubber Latex: 0.2716
Accuracy Sand: 0.5707
Iou Sand: 0.4479
Accuracy Skin Lips: 0.8519
Iou Skin Lips: 0.7560
Accuracy Sky: 0.9708
Iou Sky: 0.9418
Accuracy Snow: 0.7146
Iou Snow: 0.6191
Accuracy Soap: 0.0
Iou Soap: 0.0
Accuracy Soil Mud: 0.5586
Iou Soil Mud: 0.4173
Accuracy Sponge: 0.0
Iou Sponge: 0.0
Accuracy Stone Natural: 0.6606
Iou Stone Natural: 0.5339
Accuracy Stone Polished: 0.2223
Iou Stone Polished: 0.1965
Accuracy Styrofoam: 0.0
Iou Styrofoam: 0.0
Accuracy Tile: 0.6934
Iou Tile: 0.6312
Accuracy Wallpaper: 0.5224
Iou Wallpaper: 0.4148
Accuracy Water: 0.9011
Iou Water: 0.8268
Accuracy Wax: 0.3265
Iou Wax: 0.3251
Accuracy Whiteboard: 0.7606
Iou Whiteboard: 0.6864
Accuracy Wicker: 0.4968
Iou Wicker: 0.4467
Accuracy Wood: 0.7456
Iou Wood: 0.6849
Accuracy Wood Tree: 0.4807
Iou Wood Tree: 0.3870
Accuracy Asphalt: 0.5909
Iou Asphalt: 0.4742

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: 32
eval_batch_size: 16
seed: 42
distributed_type: multi-GPU
num_devices: 8
total_train_batch_size: 256
total_eval_batch_size: 128
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 0.2
num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Iou	Mean Accuracy	Overall Accuracy	Accuracy Animal Skin	Iou Animal Skin	Accuracy Bone Teeth Horn	Iou Bone Teeth Horn	Accuracy Brickwork	Iou Brickwork	Accuracy Cardboard	Iou Cardboard	Accuracy Carpet Rug	Iou Carpet Rug	Accuracy Ceiling Tile	Iou Ceiling Tile	Accuracy Ceramic	Iou Ceramic	Accuracy Chalkboard Blackboard	Iou Chalkboard Blackboard	Accuracy Clutter	Iou Clutter	Accuracy Concrete	Iou Concrete	Accuracy Cork Corkboard	Iou Cork Corkboard	Accuracy Engineered Stone	Iou Engineered Stone	Accuracy Fabric Cloth	Iou Fabric Cloth	Accuracy Fire	Iou Fire	Accuracy Foliage	Iou Foliage	Accuracy Food	Iou Food	Accuracy Fur	Iou Fur	Accuracy Gemstone Quartz	Iou Gemstone Quartz	Accuracy Glass	Iou Glass	Accuracy Hair	Iou Hair	Accuracy Ice	Iou Ice	Accuracy Leather	Iou Leather	Accuracy Liquid Non-water	Iou Liquid Non-water	Accuracy Metal	Iou Metal	Accuracy Mirror	Iou Mirror	Accuracy Paint Plaster Enamel	Iou Paint Plaster Enamel	Accuracy Paper	Iou Paper	Accuracy Photograph Painting	Iou Photograph Painting	Accuracy Plastic Clear	Iou Plastic Clear	Accuracy Plastic Non-clear	Iou Plastic Non-clear	Accuracy Rubber Latex	Iou Rubber Latex	Accuracy Sand	Iou Sand	Accuracy Skin Lips	Iou Skin Lips	Accuracy Sky	Iou Sky	Accuracy Snow	Iou Snow	Accuracy Soil Mud	Iou Soil Mud	Accuracy Stone Natural	Iou Stone Natural	Accuracy Stone Polished	Iou Stone Polished	Accuracy Tile	Iou Tile	Accuracy Wallpaper	Iou Wallpaper	Accuracy Water	Iou Water	Accuracy Wax	Iou Wax	Accuracy Whiteboard	Iou Whiteboard	Accuracy Wicker	Iou Wicker	Accuracy Wood	Iou Wood	Accuracy Wood Tree	Iou Wood Tree	Accuracy Asphalt	Iou Asphalt
23.2591	4.5455	400	13.8395	0.1470	0.1862	0.6536	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.7884	0.6442	0.6791	0.6385	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.8633	0.6991	0.0	0.0	0.9553	0.5771	0.2769	0.2726	0.0	0.0	0.0	0.0	0.5524	0.4723	0.1713	0.1690	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.8369	0.6781	0.5544	0.4156	0.0	0.0	0.0	0.0	0.0134	0.0131	0.0	0.0	0.0	0.0	0.6679	0.4211	0.9623	0.9005	0.0	0.0	0.0	0.0	0.0208	0.0199	0.0	0.0	0.6591	0.5641	0.0	0.0	0.8817	0.6566	0.0	0.0	0.0	0.0	0.0	0.0	0.7994	0.6479	0.0	0.0	0.0	0.0
9.1725	9.0909	800	8.9969	0.2906	0.3404	0.7036	0.0	0.0	0.0	0.0	0.5746	0.4968	0.0877	0.0874	0.8142	0.7217	0.8229	0.7461	0.5568	0.5112	0.0	0.0	0.0	0.0	0.3627	0.3073	0.0	0.0	0.0	0.0	0.8609	0.7657	0.0	0.0	0.9110	0.8009	0.8586	0.7565	0.8423	0.7348	0.0	0.0	0.6727	0.5680	0.7639	0.6660	0.0	0.0	0.4000	0.3719	0.0	0.0	0.2393	0.2182	0.4064	0.3614	0.7567	0.6891	0.5823	0.5151	0.1694	0.1603	0.0	0.0	0.3700	0.3025	0.0	0.0	0.0	0.0	0.7966	0.6699	0.9678	0.9209	0.0	0.0	0.5291	0.3117	0.6160	0.4794	0.0	0.0	0.6652	0.6036	0.4365	0.3794	0.8898	0.7867	0.0	0.0	0.6625	0.4786	0.0	0.0	0.7186	0.6605	0.0	0.0	0.3648	0.3294
5.6198	13.6364	1200	9.1325	0.3624	0.4234	0.7157	0.4067	0.2758	0.0	0.0	0.6107	0.5291	0.4344	0.3906	0.6586	0.6181	0.8505	0.7557	0.6837	0.5871	0.7123	0.6449	0.0	0.0	0.4172	0.3205	0.0	0.0	0.0	0.0	0.8222	0.7590	0.0	0.0	0.9124	0.8205	0.8812	0.7713	0.8405	0.7616	0.0	0.0	0.6625	0.5726	0.8104	0.7068	0.0	0.0	0.2324	0.2197	0.0	0.0	0.3667	0.3067	0.4322	0.3839	0.7636	0.6965	0.6618	0.5549	0.3177	0.2574	0.2445	0.2072	0.3408	0.2888	0.0128	0.0128	0.3770	0.3158	0.8249	0.7035	0.9648	0.9258	0.6217	0.5548	0.4869	0.3403	0.6749	0.5090	0.2337	0.2010	0.6966	0.6221	0.5075	0.4223	0.8979	0.8136	0.0	0.0	0.6738	0.6149	0.4526	0.4111	0.7180	0.6636	0.3129	0.2699	0.4980	0.3955
4.2141	18.1818	1600	9.6546	0.3807	0.4547	0.7338	0.5327	0.3111	0.0	0.0	0.6684	0.5487	0.5404	0.4648	0.7323	0.6697	0.7930	0.7259	0.6798	0.5899	0.7361	0.6591	0.0	0.0	0.4286	0.3427	0.0	0.0	0.0	0.0	0.8493	0.7728	0.0	0.0	0.9188	0.8270	0.8819	0.7639	0.8914	0.7938	0.0	0.0	0.6376	0.5625	0.8344	0.7277	0.0012	0.0012	0.3937	0.3468	0.0	0.0	0.3476	0.2984	0.5061	0.3964	0.7746	0.7034	0.6598	0.5531	0.3919	0.3055	0.2772	0.2286	0.3813	0.3160	0.2416	0.2118	0.4754	0.3893	0.8287	0.7256	0.9666	0.9260	0.7738	0.5683	0.5030	0.3712	0.6483	0.5063	0.2448	0.2039	0.7022	0.6262	0.5501	0.4202	0.8853	0.8122	0.0	0.0	0.7265	0.6311	0.4756	0.4223	0.7654	0.6859	0.4262	0.3134	0.5736	0.4572
3.5156	22.7273	2000	10.1557	0.3948	0.4683	0.7236	0.6345	0.3247	0.0	0.0	0.6522	0.5595	0.5211	0.4645	0.7087	0.6577	0.7127	0.6737	0.7282	0.6211	0.6796	0.6168	0.0	0.0	0.4782	0.3628	0.0	0.0	0.0	0.0	0.8433	0.7735	0.3067	0.2605	0.9130	0.8356	0.8713	0.7633	0.8657	0.7822	0.0	0.0	0.6961	0.5917	0.8403	0.7350	0.3328	0.2865	0.5103	0.4189	0.0	0.0	0.3771	0.3129	0.4079	0.3679	0.7586	0.6939	0.6052	0.5408	0.3144	0.2552	0.2957	0.2502	0.3759	0.3118	0.2715	0.2310	0.6055	0.4525	0.8323	0.7353	0.9642	0.9363	0.7553	0.6282	0.5383	0.3959	0.6861	0.5398	0.2991	0.2084	0.6631	0.6003	0.5337	0.4122	0.8987	0.8216	0.0	0.0	0.6954	0.6182	0.4685	0.4142	0.6995	0.6522	0.4855	0.3605	0.5259	0.4586
3.0893	27.2727	2400	10.5690	0.3958	0.4659	0.7267	0.6126	0.3430	0.0	0.0	0.6235	0.5588	0.5181	0.4529	0.7599	0.6980	0.6838	0.6419	0.7167	0.6208	0.7022	0.6319	0.0	0.0	0.4359	0.3399	0.0	0.0	0.0045	0.0043	0.8301	0.7675	0.3497	0.2765	0.9015	0.8328	0.8690	0.7721	0.8676	0.7842	0.0168	0.0165	0.6729	0.5880	0.8457	0.7426	0.3620	0.2570	0.4232	0.3742	0.0	0.0	0.3515	0.2973	0.4294	0.3835	0.7767	0.7065	0.6667	0.5710	0.3710	0.2935	0.3306	0.2682	0.3258	0.2817	0.2801	0.2522	0.5744	0.4381	0.8406	0.7441	0.9755	0.9352	0.7163	0.5920	0.5348	0.4076	0.6487	0.5306	0.2212	0.1966	0.6759	0.6141	0.5072	0.4124	0.8816	0.8108	0.0	0.0	0.7437	0.6559	0.4440	0.4076	0.7220	0.6678	0.4594	0.3545	0.5561	0.4552
2.7957	31.8182	2800	11.0041	0.3994	0.4689	0.7228	0.6209	0.3772	0.0	0.0	0.6268	0.5629	0.4829	0.4400	0.7616	0.6984	0.8295	0.7525	0.7106	0.6199	0.6899	0.6288	0.0	0.0	0.4257	0.3365	0.0	0.0	0.0123	0.0110	0.8358	0.7708	0.3749	0.3116	0.9164	0.8389	0.8562	0.7649	0.8689	0.7808	0.0682	0.0512	0.6356	0.5654	0.8377	0.7455	0.3943	0.2643	0.3618	0.3256	0.0000	0.0000	0.3654	0.3078	0.4544	0.3935	0.7504	0.6909	0.6411	0.5556	0.3419	0.2757	0.3435	0.2769	0.3968	0.3235	0.3186	0.2756	0.5221	0.4145	0.8541	0.7497	0.9672	0.9390	0.7496	0.6174	0.5484	0.4113	0.6662	0.5300	0.2195	0.1908	0.6858	0.6244	0.4828	0.3995	0.8847	0.8189	0.0	0.0	0.7189	0.6406	0.4404	0.3973	0.7410	0.6780	0.4740	0.3658	0.5034	0.4458
2.5825	36.3636	3200	11.2353	0.4077	0.4785	0.7315	0.5945	0.3841	0.0	0.0	0.6424	0.5588	0.5305	0.4550	0.7683	0.6996	0.8011	0.7344	0.7088	0.6199	0.7103	0.6457	0.0	0.0	0.4373	0.3468	0.0	0.0	0.0434	0.0340	0.8433	0.7757	0.3865	0.3507	0.9046	0.8369	0.8771	0.7722	0.8795	0.7884	0.0503	0.0299	0.6320	0.5653	0.8392	0.7483	0.3621	0.2783	0.4448	0.3863	0.0970	0.0957	0.3473	0.2979	0.4391	0.3873	0.7739	0.7057	0.6514	0.5623	0.3540	0.2948	0.2827	0.2454	0.3545	0.3015	0.3201	0.2655	0.5504	0.4626	0.8503	0.7523	0.9710	0.9405	0.7998	0.6609	0.5126	0.3917	0.6944	0.5310	0.2704	0.2230	0.6933	0.6269	0.5628	0.4228	0.9002	0.8290	0.0	0.0	0.7569	0.6757	0.4504	0.4096	0.7410	0.6817	0.4752	0.3809	0.5766	0.4536
2.4282	40.9091	3600	11.5373	0.4132	0.4845	0.7272	0.5829	0.3698	0.0	0.0	0.6306	0.5552	0.5290	0.4568	0.7755	0.7076	0.7498	0.7086	0.7273	0.6309	0.7192	0.6447	0.0001	0.0001	0.4090	0.3263	0.0447	0.0423	0.0444	0.0377	0.8414	0.7785	0.4448	0.3992	0.9062	0.8385	0.8763	0.7731	0.8858	0.8058	0.1693	0.1139	0.6577	0.5815	0.8423	0.7411	0.3545	0.2498	0.4091	0.3627	0.2365	0.2140	0.4018	0.3260	0.4800	0.4102	0.7566	0.6961	0.6409	0.5622	0.3895	0.3156	0.3329	0.2735	0.3817	0.3187	0.3173	0.2736	0.5735	0.4458	0.8479	0.7531	0.9708	0.9382	0.7379	0.6305	0.5262	0.4022	0.6644	0.5274	0.2263	0.1998	0.6694	0.6145	0.4785	0.3923	0.9003	0.8277	0.0056	0.0056	0.7635	0.6862	0.4754	0.4292	0.7357	0.6771	0.4783	0.3823	0.6010	0.4719
2.3205	45.4545	4000	11.7263	0.4159	0.4866	0.7328	0.5663	0.3607	0.0018	0.0018	0.6217	0.5483	0.5611	0.4707	0.7732	0.7045	0.8032	0.7357	0.7234	0.6292	0.7014	0.6347	0.0039	0.0036	0.4299	0.3461	0.0157	0.0149	0.0177	0.0158	0.8511	0.7809	0.3897	0.3384	0.9069	0.8385	0.8682	0.7738	0.8900	0.8054	0.1158	0.0913	0.6518	0.5789	0.8454	0.7460	0.3505	0.2713	0.3802	0.3371	0.2234	0.2104	0.3517	0.3007	0.4591	0.3993	0.7630	0.7004	0.6742	0.5780	0.3805	0.3088	0.3165	0.2702	0.3783	0.3165	0.3215	0.2760	0.5748	0.4497	0.8513	0.7557	0.9698	0.9429	0.7618	0.6583	0.5656	0.4170	0.6628	0.5355	0.2233	0.1971	0.6894	0.6289	0.5044	0.4070	0.8995	0.8282	0.1616	0.1615	0.7561	0.6793	0.5033	0.4504	0.7543	0.6892	0.4970	0.3871	0.5889	0.4651
2.2572	50.0	4400	11.7839	0.4203	0.4907	0.7320	0.5754	0.3556	0.0028	0.0028	0.6283	0.5529	0.5400	0.4636	0.7694	0.7043	0.7873	0.7258	0.7227	0.6279	0.7058	0.6366	0.0031	0.0028	0.4385	0.3509	0.0203	0.0191	0.0238	0.0213	0.8462	0.7791	0.3918	0.3545	0.9080	0.8384	0.8727	0.7758	0.8913	0.8047	0.2157	0.1761	0.6599	0.5828	0.8424	0.7460	0.3504	0.2609	0.3762	0.3338	0.2289	0.2157	0.3665	0.3088	0.4631	0.4021	0.7616	0.6999	0.6700	0.5762	0.3697	0.3016	0.3195	0.2717	0.3865	0.3220	0.3122	0.2716	0.5707	0.4479	0.8519	0.7560	0.9708	0.9418	0.7146	0.6191	0.5586	0.4173	0.6606	0.5339	0.2223	0.1965	0.6934	0.6312	0.5224	0.4148	0.9011	0.8268	0.3265	0.3251	0.7606	0.6864	0.4968	0.4467	0.7456	0.6849	0.4807	0.3870	0.5909	0.4742

Framework versions

Transformers 5.0.0
Pytorch 2.9.1+cu128
Datasets 4.5.0
Tokenizers 0.22.2

Downloads last month: 7

Safetensors

Model size

84.6M params

Tensor type

F32

Model tree for AllanK24/segformer-b5-finetuned-apple-dms-run3

Base model

nvidia/segformer-b5-finetuned-ade-640-640

Finetuned

(22)

this model