robeczech_lr3e-05_bs16_train30

This model is a fine-tuned version of ufal/robeczech-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.7524
• Precision: 0.6957
• Recall: 0.7042
• F1: 0.6999
• Accuracy: 0.8581

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: 3e-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: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	2	2.1827	0.0789	0.0029	0.0056	0.5672
No log	2.0	4	1.8547	0.0	0.0	0.0	0.5666
No log	3.0	6	1.7184	0.0	0.0	0.0	0.5666
No log	4.0	8	1.6366	0.0	0.0	0.0	0.5666
No log	5.0	10	1.5608	0.0	0.0	0.0	0.5666
No log	6.0	12	1.4897	0.0	0.0	0.0	0.5666
No log	7.0	14	1.4198	0.5	0.0048	0.0096	0.5686
No log	8.0	16	1.3505	0.4098	0.0121	0.0235	0.5718
No log	9.0	18	1.2871	0.3481	0.0227	0.0426	0.5764
No log	10.0	20	1.2321	0.3907	0.0526	0.0928	0.5892
No log	11.0	22	1.1821	0.4436	0.1347	0.2067	0.6235
No log	12.0	24	1.1378	0.4340	0.2685	0.3317	0.6789
No log	13.0	26	1.0951	0.4228	0.3317	0.3718	0.7041
No log	14.0	28	1.0571	0.4412	0.3877	0.4127	0.7265
No log	15.0	30	1.0247	0.4677	0.4433	0.4551	0.7461
No log	16.0	32	0.9942	0.5116	0.4915	0.5014	0.7662
No log	17.0	34	0.9657	0.5523	0.5278	0.5398	0.7830
No log	18.0	36	0.9422	0.5946	0.5645	0.5791	0.7997
No log	19.0	38	0.9165	0.6177	0.6007	0.6091	0.8144
No log	20.0	40	0.8933	0.6319	0.6340	0.6329	0.8248
No log	21.0	42	0.8715	0.6493	0.6552	0.6522	0.8336
No log	22.0	44	0.8489	0.6654	0.6644	0.6649	0.8391
No log	23.0	46	0.8303	0.6819	0.6707	0.6762	0.8451
No log	24.0	48	0.8136	0.6873	0.6760	0.6816	0.8472
No log	25.0	50	0.8005	0.6881	0.6871	0.6876	0.8506
No log	26.0	52	0.7916	0.6884	0.6944	0.6913	0.8512
No log	27.0	54	0.7836	0.6913	0.6997	0.6955	0.8529
No log	28.0	56	0.7764	0.6934	0.7001	0.6968	0.8539
No log	29.0	58	0.7714	0.6961	0.7011	0.6986	0.8550
No log	30.0	60	0.7693	0.6984	0.7021	0.7002	0.8560

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 3.6.0
• Tokenizers 0.21.1