embc25_finetuned_30000_es_it

This model is a fine-tuned version of Kyungjin-Kim/mmc_roberta_500000_es_it on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5049
• Accuracy: 0.7852
• Precision: 0.8153
• Recall: 0.7373
• F1: 0.7744

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
0.5653	0.5926	500	0.5470	0.7152	0.7160	0.7133	0.7146
0.4475	1.1849	1000	0.5145	0.7495	0.7144	0.8313	0.7684
0.4051	1.7775	1500	0.4836	0.7797	0.7946	0.7543	0.7739
0.3338	2.3698	2000	0.5049	0.7852	0.8153	0.7373	0.7744
0.3319	2.9624	2500	0.5172	0.7815	0.8083	0.738	0.7716
0.2674	3.5547	3000	0.5799	0.7748	0.7886	0.751	0.7693
0.1836	4.1470	3500	0.6434	0.7767	0.7554	0.8183	0.7856
0.1793	4.7396	4000	0.6924	0.78	0.7871	0.7677	0.7773
0.1382	5.3319	4500	0.7945	0.7783	0.7736	0.787	0.7802
0.1378	5.9244	5000	0.7990	0.7792	0.7791	0.7793	0.7792
0.1044	6.5167	5500	0.9019	0.7805	0.8030	0.7433	0.7720
0.0802	7.1090	6000	0.9631	0.7737	0.7680	0.7843	0.7761
0.0729	7.7016	6500	1.0627	0.778	0.7878	0.761	0.7742
0.067	8.2939	7000	1.1077	0.7778	0.7723	0.788	0.7801
0.0539	8.8865	7500	1.1524	0.776	0.7855	0.7593	0.7722
0.0531	9.4788	8000	1.1908	0.779	0.7822	0.7733	0.7777

Framework versions

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