semeval_taskB_model

This model is a fine-tuned version of microsoft/codebert-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.5445
• Accuracy: 0.5481
• F1: 0.4930
• Precision: 0.5257
• Recall: 0.4959

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: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 10
• mixed_precision_training: Native AMP
• label_smoothing_factor: 0.1

Training results

Training Loss	Epoch	Step	Accuracy	F1	Validation Loss	Precision	Recall
2.3078	0.2220	200	0.1833	0.0469	2.2934	0.0442	0.1049
2.1351	0.4440	400	0.3066	0.1719	2.0790	0.1697	0.2195
1.9606	0.6659	600	0.3693	0.2361	1.9354	0.2789	0.2798
1.8952	0.8879	800	0.3702	0.2573	1.9162	0.3782	0.2944
1.8292	1.1099	1000	0.4239	0.2996	1.7953	0.3586	0.3279
1.7478	1.3319	1200	0.4338	0.3133	1.7717	0.3553	0.3469
1.7098	1.5538	1400	0.4656	0.3473	1.7012	0.4297	0.3709
1.6912	1.7758	1600	0.4800	0.3853	1.6791	0.4966	0.4019
1.6572	1.9978	1800	0.4827	0.3918	1.6679	0.4640	0.4162
1.6072	2.2198	2000	0.5071	0.4194	1.6191	0.4700	0.4456
1.5818	2.4417	2200	0.5120	0.4279	1.6171	0.4822	0.4489
1.567	2.6637	2400	0.5171	0.4395	1.5871	0.4954	0.4480
1.5508	2.8857	2600	0.5263	0.4491	1.5767	0.4882	0.4658
1.5085	3.1077	2800	0.5144	0.4494	1.6068	0.4878	0.4583
1.5294	3.3296	3000	0.5328	0.4738	1.5578	0.4982	0.4835
1.4993	3.5516	3200	0.5347	0.4720	1.5506	0.5013	0.4821
1.4949	3.7736	3400	0.5213	0.4550	1.5899	0.4916	0.4708
1.4849	3.9956	3600	0.5317	0.4748	1.5539	0.5055	0.4864
1.4235	4.2175	3800	1.5294	0.5452	0.4888	0.5124	0.4973
1.4319	4.4395	4000	1.5267	0.5512	0.4941	0.5158	0.4987
1.4273	4.6615	4200	1.5444	0.5436	0.4851	0.5139	0.4923
1.4374	4.8835	4400	1.5213	0.5499	0.4948	0.5215	0.5027
1.4083	5.1054	4600	1.5253	0.5534	0.5039	0.5338	0.5045
1.3523	5.3274	4800	1.5098	0.5564	0.5005	0.5219	0.5092
1.3833	5.5494	5000	1.5407	0.5481	0.4925	0.5222	0.5007
1.3915	5.7714	5200	1.5445	0.5481	0.4930	0.5257	0.4959

Framework versions

• Transformers 4.45.0
• Pytorch 2.10.0+cu128
• Datasets 4.8.3
• Tokenizers 0.20.3