CNEC_2_0_Supertypes_Czert-B-base-cased

This model is a fine-tuned version of UWB-AIR/Czert-B-base-cased on the cnec dataset. It achieves the following results on the evaluation set:

• Loss: 0.2429
• Precision: 0.8320
• Recall: 0.8860
• F1: 0.8582
• Accuracy: 0.9590

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: 2e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	1.0	113	0.2231	0.7053	0.7472	0.7256	0.9363
No log	2.0	226	0.1791	0.7584	0.8170	0.7866	0.9490
No log	3.0	339	0.1746	0.7742	0.8385	0.8051	0.9508
No log	4.0	452	0.1783	0.7836	0.8509	0.8158	0.9512
0.2584	5.0	565	0.1742	0.7902	0.8558	0.8217	0.9541
0.2584	6.0	678	0.1653	0.8044	0.8645	0.8334	0.9565
0.2584	7.0	791	0.1694	0.8103	0.8715	0.8398	0.9579
0.2584	8.0	904	0.1838	0.8001	0.8678	0.8326	0.9556
0.0804	9.0	1017	0.1804	0.8204	0.8753	0.8469	0.9571
0.0804	10.0	1130	0.1918	0.8196	0.8761	0.8469	0.9576
0.0804	11.0	1243	0.2018	0.8169	0.8790	0.8468	0.9578
0.0804	12.0	1356	0.2067	0.8220	0.8815	0.8507	0.9579
0.0804	13.0	1469	0.2060	0.8285	0.8876	0.8570	0.9585
0.049	14.0	1582	0.2084	0.8271	0.8815	0.8534	0.9589
0.049	15.0	1695	0.2171	0.8257	0.8806	0.8523	0.9585
0.049	16.0	1808	0.2246	0.8307	0.8839	0.8565	0.9586
0.049	17.0	1921	0.2225	0.8288	0.8881	0.8574	0.9590
0.0338	18.0	2034	0.2272	0.8351	0.8889	0.8611	0.9598
0.0338	19.0	2147	0.2307	0.8337	0.8864	0.8593	0.9593
0.0338	20.0	2260	0.2387	0.8302	0.8864	0.8574	0.9588
0.0338	21.0	2373	0.2387	0.8338	0.8868	0.8595	0.9585
0.0338	22.0	2486	0.2400	0.8343	0.8881	0.8603	0.9592
0.0261	23.0	2599	0.2422	0.8319	0.8872	0.8587	0.9590
0.0261	24.0	2712	0.2431	0.8317	0.8860	0.858	0.9589
0.0261	25.0	2825	0.2429	0.8320	0.8860	0.8582	0.9590

Framework versions

• Transformers 4.36.2
• Pytorch 2.1.2+cu121
• Datasets 2.16.1
• Tokenizers 0.15.0