evacun-lemmatization-vanilla

This model is a fine-tuned version of bowphs/evacun2025-1 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4795
• Exact Match: 0.8701

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: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 16
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Exact Match
0.4563	1.0	4572	0.5323	0.6769
0.2377	2.0	9144	0.3433	0.8097
0.1578	3.0	13716	0.2905	0.7674
0.1118	4.0	18288	0.2715	0.8336
0.0733	5.0	22860	0.2761	0.8431
0.0534	6.0	27432	0.2985	0.8317
0.038	7.0	32004	0.3253	0.8465
0.0263	8.0	36576	0.3477	0.8485
0.0202	9.0	41148	0.3713	0.8468
0.0165	10.0	45720	0.3877	0.8123
0.0119	11.0	50292	0.4046	0.8492
0.0104	12.0	54864	0.4154	0.8396
0.0095	13.0	59436	0.4326	0.8415
0.0071	14.0	64008	0.4367	0.8458
0.006	15.0	68580	0.4485	0.8615
0.0055	16.0	73152	0.4544	0.8623
0.0044	17.0	77724	0.4618	0.8547
0.0037	18.0	82296	0.4613	0.8643
0.0034	19.0	86868	0.4720	0.8631
0.0033	20.0	91440	0.4731	0.8623
0.0026	21.0	96012	0.4849	0.861
0.0026	22.0	100584	0.4795	0.8701
0.002	23.0	105156	0.4890	0.8657
0.0021	24.0	109728	0.4876	0.8691
0.0016	25.0	114300	0.4916	0.8684
0.0017	26.0	118872	0.4931	0.8566
0.001	27.0	123444	0.4985	0.8606

Framework versions

• Transformers 4.45.1
• Pytorch 2.4.1+cu121
• Datasets 3.0.0
• Tokenizers 0.20.0