SegForCoral-2025_05_05_72001-bs16_refine

This model is a fine-tuned version of nvidia/mit-b0 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4708
• Learning Rate: 0.0000

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: 1e-05
• train_batch_size: 16
• eval_batch_size: 16
• 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
• num_epochs: 100
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rate
0.8419	1.0	115	0.7590	1e-05
0.7569	2.0	230	0.6943	1e-05
0.7101	3.0	345	0.6526	1e-05
0.6722	4.0	460	0.6274	1e-05
0.6506	5.0	575	0.6015	1e-05
0.6282	6.0	690	0.5865	1e-05
0.6166	7.0	805	0.5672	1e-05
0.6008	8.0	920	0.5533	1e-05
0.593	9.0	1035	0.5361	1e-05
0.5804	10.0	1150	0.5343	1e-05
0.5749	11.0	1265	0.5341	1e-05
0.5662	12.0	1380	0.5229	1e-05
0.559	13.0	1495	0.5178	1e-05
0.5559	14.0	1610	0.5028	1e-05
0.5455	15.0	1725	0.5070	1e-05
0.545	16.0	1840	0.4901	1e-05
0.5422	17.0	1955	0.4999	1e-05
0.5378	18.0	2070	0.5047	1e-05
0.5345	19.0	2185	0.4924	1e-05
0.531	20.0	2300	0.4915	1e-05
0.5305	21.0	2415	0.4855	1e-05
0.5255	22.0	2530	0.4804	1e-05
0.5247	23.0	2645	0.4915	1e-05
0.5229	24.0	2760	0.4845	1e-05
0.5192	25.0	2875	0.4880	1e-05
0.5219	26.0	2990	0.4804	1e-05
0.521	27.0	3105	0.4767	1e-05
0.5165	28.0	3220	0.4729	1e-05
0.5176	29.0	3335	0.4870	1e-05
0.5157	30.0	3450	0.4736	1e-05
0.517	31.0	3565	0.4801	1e-05
0.5126	32.0	3680	0.4774	1e-05
0.5171	33.0	3795	0.4762	1e-05
0.513	34.0	3910	0.4755	1e-05
0.5121	35.0	4025	0.4698	0.0000
0.5115	36.0	4140	0.4696	0.0000
0.5097	37.0	4255	0.4767	0.0000
0.5128	38.0	4370	0.4686	0.0000
0.5094	39.0	4485	0.4694	0.0000
0.511	40.0	4600	0.4698	0.0000
0.5088	41.0	4715	0.4709	0.0000
0.509	42.0	4830	0.4720	0.0000
0.5091	43.0	4945	0.4702	0.0000
0.5078	44.0	5060	0.4692	0.0000
0.5076	45.0	5175	0.4720	0.0000
0.5076	46.0	5290	0.4726	0.0000
0.5049	47.0	5405	0.4732	0.0000
0.5104	48.0	5520	0.4708	0.0000

Framework versions

• Transformers 4.49.0
• Pytorch 2.3.1+cu121
• Datasets 3.5.0
• Tokenizers 0.21.1