se_train_run_PARASITIC

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

• Loss: 3.8881
• Model Preparation Time: 0.0019
• F1: 0.9162
• Precision: 0.8951
• Recall: 0.9382
• Threshold: 0.6388
• Sim Ratio: 1.8568
• Pos Sim: 0.8121
• Neg Sim: 0.4374

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: 0.0001
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Use 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: 1

Training results

Training Loss	Epoch	Step	Validation Loss	Model Preparation Time	F1	Precision	Recall	Threshold	Sim Ratio	Pos Sim	Neg Sim
1.199	0.0763	5000	4.1294	0.0019	0.8763	0.8449	0.9102	0.5986	2.2839	0.8098	0.3546
1.1869	0.1526	10000	3.5058	0.0019	0.88	0.8739	0.8863	0.6131	2.0881	0.8009	0.3835
1.1298	0.2289	15000	3.6247	0.0019	0.8903	0.8642	0.918	0.6164	2.0653	0.8163	0.3952
1.1008	0.3052	20000	3.6331	0.0019	0.894	0.8832	0.9049	0.6290	2.1236	0.8072	0.3801
1.0824	0.3815	25000	3.5983	0.0019	0.8964	0.8782	0.9154	0.5923	2.1619	0.7958	0.3681
0.9867	0.4578	30000	3.7920	0.0019	0.9012	0.8778	0.9258	0.6112	2.1457	0.8089	0.377
0.9326	0.5341	35000	3.8601	0.0019	0.904	0.8896	0.9189	0.6040	2.2014	0.7974	0.3622
0.9779	0.6104	40000	3.8622	0.0019	0.906	0.896	0.9163	0.6329	2.0996	0.8071	0.3844
0.9159	0.6866	45000	3.8546	0.0019	0.9104	0.9016	0.9194	0.6348	2.0243	0.8059	0.3981
0.9443	0.7629	50000	3.9447	0.0019	0.9129	0.8941	0.9326	0.6321	1.9574	0.8088	0.4132
0.8501	0.8392	55000	3.8744	0.0019	0.9131	0.8948	0.9323	0.6371	1.9077	0.8094	0.4243
0.833	0.9155	60000	3.8991	0.0019	0.9166	0.8954	0.9389	0.6335	1.8906	0.8096	0.4282
0.8589	0.9918	65000	3.8864	0.0019	0.9159	0.8945	0.9384	0.6377	1.8594	0.8116	0.4365

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1
• Datasets 3.2.0
• Tokenizers 0.21.0