roberta-base-nsp-1000000-1e-06-32-negcommonsensebalanced-1e-06-64

This model is a fine-tuned version of mhr2004/roberta-base-nsp-1000000-1e-06-32 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.4028

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: 1e-06
• train_batch_size: 256
• eval_batch_size: 1024
• 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: 30

Training results

Training Loss	Epoch	Step	Validation Loss
0.5812	1.0	795	0.5384
0.5273	2.0	1590	0.5009
0.5051	3.0	2385	0.4822
0.4865	4.0	3180	0.4731
0.4723	5.0	3975	0.4582
0.4616	6.0	4770	0.4602
0.4526	7.0	5565	0.4426
0.44	8.0	6360	0.4402
0.4329	9.0	7155	0.4316
0.4288	10.0	7950	0.4282
0.4174	11.0	8745	0.4234
0.4134	12.0	9540	0.4205
0.4101	13.0	10335	0.4203
0.4076	14.0	11130	0.4152
0.4028	15.0	11925	0.4094
0.3963	16.0	12720	0.4103
0.393	17.0	13515	0.4088
0.3899	18.0	14310	0.4120
0.3897	19.0	15105	0.4050
0.3842	20.0	15900	0.4051
0.3839	21.0	16695	0.4050
0.3791	22.0	17490	0.4015
0.3801	23.0	18285	0.4035
0.3807	24.0	19080	0.4016
0.3738	25.0	19875	0.4028

Framework versions

• Transformers 4.48.3
• Pytorch 2.6.0+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0