CNEC_2_0_robeczech-base

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

• Loss: 0.3306
• Precision: 0.8531
• Recall: 0.8848
• F1: 0.8687
• Accuracy: 0.9545

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.4499	2.22	2000	0.3871	0.7163	0.7099	0.7131	0.9222
0.2342	4.44	4000	0.2576	0.8149	0.8251	0.8200	0.9451
0.1449	6.67	6000	0.2407	0.8231	0.8523	0.8375	0.9492
0.1027	8.89	8000	0.2267	0.8362	0.8748	0.8551	0.9527
0.0751	11.11	10000	0.2429	0.8394	0.8712	0.8550	0.9522
0.0473	13.33	12000	0.2633	0.8439	0.8720	0.8577	0.9535
0.0369	15.56	14000	0.2821	0.8468	0.8755	0.8609	0.9541
0.0286	17.78	16000	0.2797	0.8534	0.8827	0.8678	0.9558
0.0234	20.0	18000	0.2860	0.8550	0.8834	0.8690	0.9558
0.0168	22.22	20000	0.3146	0.8471	0.8795	0.8630	0.9531
0.0142	24.44	22000	0.3165	0.8488	0.8816	0.8649	0.9530
0.011	26.67	24000	0.3291	0.8518	0.8816	0.8664	0.9537
0.0092	28.89	26000	0.3306	0.8531	0.8848	0.8687	0.9545

Framework versions

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