embc25_finetuned_30000_en

This model is a fine-tuned version of Kyungjin-Kim/mmc_roberta_500000_en on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.6683
• Accuracy: 0.8843
• Precision: 0.8893
• Recall: 0.878
• F1: 0.8836

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: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• 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
• lr_scheduler_warmup_steps: 500
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
0.4117	1.1848	500	0.4034	0.819	0.8600	0.762	0.8081
0.2613	2.3697	1000	0.3137	0.866	0.8938	0.8307	0.8611
0.1669	3.5545	1500	0.3362	0.876	0.8944	0.8527	0.8730
0.1286	4.7393	2000	0.3396	0.8823	0.8736	0.894	0.8837
0.0798	5.9242	2500	0.4340	0.8843	0.8688	0.9053	0.8867
0.0429	7.1090	3000	0.5296	0.8843	0.8722	0.9007	0.8862
0.0286	8.2938	3500	0.6355	0.881	0.9033	0.8533	0.8776
0.022	9.4787	4000	0.6657	0.8823	0.8883	0.8747	0.8814

Framework versions

• Transformers 4.48.1
• Pytorch 2.3.1
• Datasets 3.2.0
• Tokenizers 0.21.0