Modelpublic

quatatak/kazroberta-finetuned-pos-halved

sourceHugging Faceapache-2.0updated 1y agoView on Hugging Face

0likes4downloads

Model Card

kazroberta-finetuned-pos-halved

This model is a fine-tuned version of kz-transformers/kaz-roberta-conversational on the None dataset. It achieves the following results on the evaluation set:

—Loss: 0.0254
—Accuracy: 0.9961
—Precision: 0.9927
—Recall: 0.9921
—F1: 0.9924

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
—trainbatchsize: 32
—evalbatchsize: 32
—seed: 42
—optimizer: Use OptimizerNames.ADAMWTORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
—lrschedulertype: linear
—lrschedulerwarmup_steps: 1000
—num_epochs: 7

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
0.0855	1.0	1238	0.0586	0.9809	0.9600	0.9647	0.9622
0.0435	2.0	2476	0.0345	0.9886	0.9745	0.9759	0.9752
0.0167	3.0	3714	0.0234	0.9932	0.9863	0.9878	0.9870
0.0075	4.0	4952	0.0229	0.9949	0.9902	0.9900	0.9901
0.0034	5.0	6190	0.0241	0.9954	0.9910	0.9910	0.9910
0.0019	6.0	7428	0.0247	0.9959	0.9928	0.9910	0.9919
0.0009	7.0	8666	0.0254	0.9961	0.9927	0.9921	0.9924

Framework versions

—Transformers 4.51.3
—Pytorch 2.7.0+cu126
—Datasets 3.5.1
—Tokenizers 0.21.1