ania3000/kuosbert-morph

trainer_output

This model is a fine-tuned version of ania3000/kuosbert-from_multilingual on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.1742
• Accuracy: 89.6340
• Sentence accuracy: 32.1484

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_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Sentence accuracy
1.5179	1.0	596	1.2402	81.9319	14.5286
0.5915	2.0	1192	1.0153	85.5108	21.6383
0.3952	3.0	1788	0.9749	87.0280	27.2025
0.2929	4.0	2384	0.9609	87.7114	27.9753
0.2345	5.0	2980	0.9658	88.1631	29.5209
0.1419	6.0	3576	0.9037	88.5569	29.3663
0.1123	7.0	4172	0.9257	88.4410	29.2117
0.0913	8.0	4768	0.9203	88.8001	31.2210
0.0742	9.0	5364	0.9467	88.8117	29.2117
0.0616	10.0	5960	0.9320	88.5685	30.4482
0.0377	11.0	6556	0.9668	89.1128	32.4575
0.029	12.0	7152	1.0205	89.1476	31.3756
0.025	13.0	7748	1.0496	89.1012	31.0665
0.0196	14.0	8344	1.0772	89.2750	32.7666
0.016	15.0	8940	1.0822	89.2865	31.8393
0.0099	16.0	9536	1.0900	89.4371	32.4575
0.0079	17.0	10132	1.0938	89.4024	31.9938
0.0068	18.0	10728	1.1089	89.4255	33.0757
0.0059	19.0	11324	1.1300	89.1939	31.9938
0.005	20.0	11920	1.1247	89.5413	33.3849
0.0033	21.0	12516	1.1441	89.4950	31.6847
0.0032	22.0	13112	1.1512	89.5298	32.1484
0.0025	23.0	13708	1.1675	89.7151	32.7666
0.0019	24.0	14304	1.1738	89.6108	32.1484
0.0021	25.0	14900	1.1742	89.6340	32.1484

Framework versions

• Transformers 4.57.3
• Pytorch 2.11.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.2