robeczech_lr1e-05_bs8_train287

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: 0.1992
• Precision: 0.9427
• Recall: 0.9523
• F1: 0.9475
• Accuracy: 0.9694

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: 1e-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	1.4794	0.0	0.0	0.0	0.5666
No log	2.0	72	1.0189	0.4090	0.4104	0.4097	0.7275
No log	3.0	108	0.7386	0.7125	0.7050	0.7087	0.8612
No log	4.0	144	0.5270	0.8403	0.8387	0.8395	0.9215
No log	5.0	180	0.4245	0.8699	0.8909	0.8802	0.9372
No log	6.0	216	0.3600	0.8917	0.8909	0.8913	0.9443
No log	7.0	252	0.3263	0.8931	0.8957	0.8944	0.9462
No log	8.0	288	0.3094	0.8892	0.8986	0.8939	0.9456
No log	9.0	324	0.2963	0.8948	0.8991	0.8969	0.9470
No log	10.0	360	0.2797	0.8947	0.9025	0.8986	0.9483
No log	11.0	396	0.2715	0.8991	0.9039	0.9015	0.9498
No log	12.0	432	0.2535	0.9033	0.9116	0.9075	0.9525
No log	13.0	468	0.2393	0.9261	0.9256	0.9259	0.9607
0.5667	14.0	504	0.2368	0.9305	0.9252	0.9278	0.9617
0.5667	15.0	540	0.2184	0.9350	0.9382	0.9366	0.9657
0.5667	16.0	576	0.2142	0.9341	0.9382	0.9362	0.9655
0.5667	17.0	612	0.2131	0.9332	0.9372	0.9352	0.9642
0.5667	18.0	648	0.2027	0.9456	0.9401	0.9429	0.9686
0.5667	19.0	684	0.1993	0.9408	0.9440	0.9424	0.9686
0.5667	20.0	720	0.1966	0.9427	0.9445	0.9436	0.9694
0.5667	21.0	756	0.1928	0.9468	0.9459	0.9464	0.9701

Framework versions

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