roberta-finetuned

This model is a fine-tuned version of IRIIS-RESEARCH/RoBERTa_Nepali_125M on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.8258
• Model Preparation Time: 0.0028

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: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Use OptimizerNames.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
• training_steps: 25000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Model Preparation Time
2.7798	0.02	500	2.5232	0.0028
2.4726	1.0066	1000	2.2927	0.0028
2.3473	1.0266	1500	2.1815	0.0028
2.2124	2.0133	2000	2.1117	0.0028
2.15	2.0333	2500	2.0413	0.0028
2.0669	3.0199	3000	2.0138	0.0028
2.0006	4.0066	3500	1.9784	0.0028
1.9425	4.0266	4000	1.9815	0.0028
1.9012	5.0132	4500	1.9378	0.0028
1.9179	5.0332	5000	1.9062	0.0028
1.8373	6.0198	5500	1.8962	0.0028
1.792	7.0065	6000	1.8855	0.0028
1.7343	7.0265	6500	1.8661	0.0028
1.8189	8.0131	7000	1.8380	0.0028
1.7639	8.0331	7500	1.8170	0.0028
1.7031	9.0198	8000	1.8233	0.0028
1.6691	10.0064	8500	1.8298	0.0028
1.6316	10.0264	9000	1.8258	0.0028

Framework versions

• Transformers 4.53.3
• Pytorch 2.6.0+cu124
• Datasets 4.4.1
• Tokenizers 0.21.2