semeval_taskC_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: 0.8179
• Accuracy: 0.7361
• F1: 0.7389
• Precision: 0.7573
• Recall: 0.7361

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: 5
• mixed_precision_training: Native AMP
• label_smoothing_factor: 0.1

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
1.3856	0.0374	200	1.3809	0.2922	0.2097	0.3850	0.2922
1.3627	0.0748	400	1.3282	0.4083	0.3750	0.4449	0.4083
1.1939	0.1123	600	1.1515	0.5344	0.5196	0.5389	0.5344
1.1285	0.1497	800	1.0875	0.5656	0.5512	0.6013	0.5656
1.0717	0.1871	1000	1.0269	0.6065	0.5986	0.6265	0.6065
0.9975	0.2245	1200	0.9728	0.6462	0.6477	0.6601	0.6462
0.9667	0.2619	1400	0.9268	0.6718	0.6720	0.6725	0.6718
0.9366	0.2993	1600	0.9263	0.6752	0.6772	0.6862	0.6752
0.9133	0.3368	1800	0.9023	0.6850	0.6858	0.7001	0.6850
0.9058	0.3742	2000	0.8714	0.6997	0.6970	0.7056	0.6997
0.8807	0.4116	2200	0.8766	0.7013	0.7054	0.7171	0.7013
0.8804	0.4490	2400	0.8519	0.7148	0.7159	0.7227	0.7148
0.8287	0.4864	2600	0.8470	0.7170	0.7195	0.7290	0.7170
0.8347	0.5239	2800	0.8433	0.7207	0.7241	0.7319	0.7207
0.8395	0.5613	3000	0.8450	0.7168	0.7209	0.7339	0.7168
0.8348	0.5987	3200	0.8248	0.7296	0.7326	0.7446	0.7296
0.8082	0.6361	3400	0.8410	0.7201	0.7216	0.7452	0.7201
0.8001	0.6735	3600	0.8045	0.7409	0.7423	0.7449	0.7409
0.8011	0.7109	3800	0.8057	0.7394	0.7401	0.7441	0.7394
0.7946	0.7484	4000	0.8179	0.7361	0.7389	0.7573	0.7361

Framework versions

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