xkaska02/robeczech_lr3e-05_bs8_ep10

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: 0.1513
• Precision: 0.9488
• Recall: 0.9620
• F1: 0.9554
• Accuracy: 0.9732

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: 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	0.9171	0.5049	0.4940	0.4994	0.7710
No log	2.0	72	0.4115	0.8764	0.8899	0.8831	0.9406
No log	3.0	108	0.2925	0.8940	0.8962	0.8951	0.9466
No log	4.0	144	0.2443	0.9208	0.9213	0.9211	0.9586
No log	5.0	180	0.2048	0.9364	0.9377	0.9370	0.9657
No log	6.0	216	0.1897	0.9384	0.9411	0.9397	0.9661
No log	7.0	252	0.1745	0.9418	0.9454	0.9436	0.9688
No log	8.0	288	0.1659	0.9413	0.9527	0.9470	0.9699
No log	9.0	324	0.1704	0.9441	0.9454	0.9448	0.9699
No log	10.0	360	0.1570	0.9474	0.9479	0.9476	0.9711

Framework versions

• Transformers 4.51.1
• Pytorch 2.6.0+cu124
• Datasets 3.5.0
• Tokenizers 0.21.1