COHeN_2.0

This model is a fine-tuned version of gngpostalsrvc/BERiT_2.0 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.5952
• Accuracy: 0.7400

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: 0.0027431492469971175
• 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: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	300	0.6925	0.7776
0.7839	2.0	600	0.7892	0.7659
0.7839	3.0	900	0.8811	0.7960
0.4686	4.0	1200	0.9765	0.7517
0.3895	5.0	1500	0.9252	0.7542
0.3895	6.0	1800	0.8912	0.7801
0.3393	7.0	2100	1.0998	0.7375
0.3393	8.0	2400	0.8183	0.7952
0.3216	9.0	2700	0.8209	0.8052
0.3037	10.0	3000	1.0785	0.7993
0.3037	11.0	3300	0.8882	0.7977
0.268	12.0	3600	1.0609	0.7759
0.268	13.0	3900	1.0853	0.7584
0.2606	14.0	4200	1.0418	0.7742
0.2415	15.0	4500	1.1806	0.7667
0.2415	16.0	4800	1.0558	0.7692
0.2379	17.0	5100	1.1977	0.7600
0.2379	18.0	5400	1.3400	0.7550
0.2305	19.0	5700	1.3001	0.7550
0.216	20.0	6000	1.5855	0.7291
0.216	21.0	6300	1.4484	0.7525
0.2069	22.0	6600	1.3613	0.7483
0.2069	23.0	6900	1.3460	0.7517
0.2063	24.0	7200	1.4035	0.7617
0.1885	25.0	7500	1.5058	0.7533
0.1885	26.0	7800	1.2427	0.7868
0.1804	27.0	8100	1.4027	0.7383
0.1804	28.0	8400	1.2624	0.7492
0.1754	29.0	8700	1.2262	0.7692
0.1754	30.0	9000	1.4582	0.7433
0.1754	31.0	9300	1.4874	0.7258
0.1579	32.0	9600	1.3454	0.7492
0.1579	33.0	9900	1.3912	0.7475
0.1579	34.0	10200	1.4488	0.7441
0.1491	35.0	10500	1.4408	0.7333
0.1491	36.0	10800	1.5203	0.7299
0.1384	37.0	11100	1.3623	0.7617
0.1384	38.0	11400	1.4582	0.7408
0.1333	39.0	11700	1.4770	0.7333
0.1323	40.0	12000	1.5151	0.7358
0.1323	41.0	12300	1.3267	0.7450
0.119	42.0	12600	1.3765	0.7458
0.119	43.0	12900	1.5366	0.7349
0.123	44.0	13200	1.4966	0.7425
0.1217	45.0	13500	1.5868	0.7349
0.1217	46.0	13800	1.5798	0.7383
0.1142	47.0	14100	1.5946	0.7383
0.1142	48.0	14400	1.5240	0.7408
0.1172	49.0	14700	1.5858	0.7391
0.1077	50.0	15000	1.5952	0.7400

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu118
• Datasets 3.2.0
• Tokenizers 0.21.0