irplag_unixcoder_ep30_bs16_lr2e-05_l512_s42_ppn_loss

This model is a fine-tuned version of microsoft/unixcoder-base-nine on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0792
• Accuracy: 0.9565
• Recall: 0.9636
• Precision: 0.9815
• F1: 0.9725
• F Beta Score: 0.9691

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: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• 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	Accuracy	Recall	Precision	F1	F Beta Score
0.7212	1.0	21	0.5773	0.8261	0.8909	0.8909	0.8909	0.8909
0.4669	2.0	42	0.3188	0.8406	0.8	1.0	0.8889	0.8525
0.1992	3.0	63	0.2318	0.9565	0.9818	0.9643	0.9730	0.9764
0.0944	4.0	84	0.2172	0.9565	0.9818	0.9643	0.9730	0.9764
0.0882	5.0	105	0.1541	0.9855	1.0	0.9821	0.9910	0.9944
0.0177	6.0	126	0.0792	0.9565	0.9636	0.9815	0.9725	0.9691
0.0114	7.0	147	0.2376	0.9855	1.0	0.9821	0.9910	0.9944
0.0014	8.0	168	0.2135	0.9710	0.9818	0.9818	0.9818	0.9818
0.0023	9.0	189	0.1374	0.9710	0.9818	0.9818	0.9818	0.9818

Framework versions

• Transformers 4.51.3
• Pytorch 2.8.0+cu128
• Datasets 3.1.0
• Tokenizers 0.21.4