mask2former

This model is a fine-tuned version of facebook/mask2former-swin-tiny-ade-semantic on an unknown dataset. It achieves the following results on the evaluation set:

Model description

More information needed

More information needed

More information needed

The following hyperparameters were used during training:

learning_rate: 5e-05
train_batch_size: 8
eval_batch_size: 8
seed: 42
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: linear
lr_scheduler_warmup_steps: 4
num_epochs: 100

Training Loss	Epoch	Step	Validation Loss
50.7018	0.1408	50	44.2435
40.5877	0.2817	100	39.6465
37.4102	0.4225	150	37.2471
35.7502	0.5634	200	36.3455
34.7067	0.7042	250	34.8824
34.0798	0.8451	300	34.8520
33.3503	0.9859	350	33.7321
32.3436	1.1268	400	33.1560
32.3845	1.2676	450	33.0411
30.8809	1.4085	500	32.7852
31.689	1.5493	550	31.9914
31.036	1.6901	600	32.7297
30.9795	1.8310	650	31.8848
30.7918	1.9718	700	31.5285
30.1432	2.1127	750	32.0634
29.7082	2.2535	800	31.1849
28.7869	2.3944	850	30.9022
29.4227	2.5352	900	30.5902
29.1865	2.6761	950	30.3818
29.2715	2.8169	1000	30.9196
29.1941	2.9577	1050	30.8163
28.5256	3.0986	1100	30.4730
28.0419	3.2394	1150	30.6531
28.0538	3.3803	1200	30.0779
27.9463	3.5211	1250	30.6114
27.4152	3.6620	1300	30.5519
27.7461	3.8028	1350	29.5641
27.5604	3.9437	1400	30.1296
27.381	4.0845	1450	30.5017
26.3816	4.2254	1500	29.6898
26.5218	4.3662	1550	29.9475
26.9798	4.5070	1600	29.3323
26.8186	4.6479	1650	29.5755
27.5111	4.7887	1700	30.7945
27.0839	4.9296	1750	29.4147
26.6393	5.0704	1800	28.7983
26.3564	5.2113	1850	29.2245
25.6174	5.3521	1900	28.9337
25.8777	5.4930	1950	29.4778
25.6848	5.6338	2000	28.4992
26.4625	5.7746	2050	29.6182
26.8448	5.9155	2100	29.5377
26.0681	6.0563	2150	29.2390
25.628	6.1972	2200	29.1112