moe_train_run

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

• Loss: 3.9874
• Model Preparation Time: 0.0047
• F1: 0.8876
• Precision: 0.8509
• Recall: 0.9275
• Threshold: 0.7668
• Sim Ratio: 1.4762
• Pos Sim: 0.8878
• Neg Sim: 0.6014

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
0.7183	0.0821	10000	3.5469	0.0047	0.8386	0.7972	0.8846	0.8755	1.2415	0.9408	0.7578
0.7053	0.1643	20000	3.6924	0.0047	0.8496	0.7963	0.9104	0.8043	1.383	0.9156	0.6621
0.6003	0.2464	30000	3.9111	0.0047	0.862	0.8148	0.9151	0.7832	1.437	0.9048	0.6296
0.5856	0.3286	40000	3.9771	0.0047	0.8628	0.822	0.9079	0.7718	1.4877	0.894	0.6009
0.5801	0.4107	50000	3.9434	0.0047	0.8704	0.8277	0.9178	0.7749	1.4477	0.8995	0.6214
0.562	0.4929	60000	3.6962	0.0047	0.8685	0.8232	0.9192	0.7930	1.4037	0.9064	0.6457
0.5307	0.5750	70000	3.8964	0.0047	0.875	0.839	0.9142	0.7807	1.4542	0.8973	0.617
0.4793	0.6572	80000	4.0046	0.0047	0.8779	0.8429	0.916	0.7706	1.4946	0.8912	0.5963
0.4978	0.7393	90000	4.0062	0.0047	0.8796	0.8395	0.9239	0.7598	1.4979	0.8879	0.5927
0.4934	0.8215	100000	3.9771	0.0047	0.885	0.8522	0.9204	0.7734	1.478	0.89	0.6022
0.4757	0.9036	110000	4.0861	0.0047	0.884	0.8489	0.9221	0.7636	1.5028	0.8859	0.5895
0.4773	0.9858	120000	3.9877	0.0047	0.8874	0.8558	0.9215	0.7711	1.4765	0.8877	0.6012

Framework versions

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