xkaska02/robeczech_lr5e-05_bs8_train287

trainer_output

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: 0.1311
• Precision: 0.9525
• Recall: 0.9664
• F1: 0.9594
• Accuracy: 0.9755

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• 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	36	0.6632	0.7342	0.7161	0.7250	0.8684
No log	2.0	72	0.3053	0.8813	0.8962	0.8887	0.9435
No log	3.0	108	0.2345	0.9286	0.9174	0.9230	0.9594
No log	4.0	144	0.1869	0.9350	0.9377	0.9364	0.9663
No log	5.0	180	0.1581	0.9459	0.9459	0.9459	0.9701
No log	6.0	216	0.1654	0.9353	0.9430	0.9392	0.9678
No log	7.0	252	0.1515	0.9439	0.9507	0.9473	0.9707
No log	8.0	288	0.1396	0.9471	0.9590	0.9530	0.9730
No log	9.0	324	0.1487	0.9426	0.9440	0.9433	0.9701
No log	10.0	360	0.1610	0.9395	0.9450	0.9422	0.9686
No log	11.0	396	0.1345	0.9556	0.9561	0.9558	0.9753

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 3.5.0
• Tokenizers 0.21.1