robeczech_lr3e-05_bs16_train75

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: 0.3039
• Precision: 0.8942
• Recall: 0.9062
• F1: 0.9001
• Accuracy: 0.9484

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	5	1.7660	0.0	0.0	0.0	0.5666
No log	2.0	10	1.5595	0.0	0.0	0.0	0.5666
No log	3.0	15	1.3682	0.4364	0.0116	0.0226	0.5716
No log	4.0	20	1.2001	0.4255	0.0855	0.1423	0.6030
No log	5.0	25	1.0674	0.4077	0.3201	0.3587	0.7005
No log	6.0	30	0.9535	0.5608	0.5258	0.5427	0.7865
No log	7.0	35	0.8478	0.6840	0.6741	0.6790	0.8470
No log	8.0	40	0.7533	0.7179	0.7030	0.7104	0.8625
No log	9.0	45	0.6674	0.7420	0.7470	0.7445	0.8771
No log	10.0	50	0.6049	0.7620	0.7697	0.7658	0.8870
No log	11.0	55	0.5427	0.7914	0.7842	0.7878	0.8981
No log	12.0	60	0.5077	0.8333	0.8469	0.8400	0.9192
No log	13.0	65	0.4678	0.8465	0.8411	0.8438	0.9226
No log	14.0	70	0.4384	0.8567	0.8634	0.8600	0.9293
No log	15.0	75	0.4112	0.8652	0.8648	0.8650	0.9324
No log	16.0	80	0.3910	0.8679	0.8759	0.8719	0.9345
No log	17.0	85	0.3803	0.8730	0.8764	0.8747	0.9362
No log	18.0	90	0.3703	0.8755	0.8759	0.8757	0.9364
No log	19.0	95	0.3635	0.8739	0.8836	0.8788	0.9374
No log	20.0	100	0.3544	0.8804	0.8783	0.8794	0.9380
No log	21.0	105	0.3439	0.8780	0.8860	0.8820	0.9397
No log	22.0	110	0.3402	0.8770	0.8885	0.8827	0.9393
No log	23.0	115	0.3394	0.8829	0.8807	0.8818	0.9391
No log	24.0	120	0.3335	0.8822	0.8822	0.8822	0.9397
No log	25.0	125	0.3281	0.8819	0.8870	0.8844	0.9412
No log	26.0	130	0.3265	0.8799	0.8880	0.8839	0.9410
No log	27.0	135	0.3253	0.8827	0.8904	0.8865	0.9420
No log	28.0	140	0.3256	0.8862	0.8909	0.8885	0.9427
No log	29.0	145	0.3246	0.8865	0.8904	0.8885	0.9424
No log	30.0	150	0.3238	0.8848	0.8899	0.8873	0.9422

Framework versions

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