param-bharat/ModernBERT-base-nli-scorer

ModernBERT-base-nli-clf

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

• Loss: 0.0101
• F1: 0.8717
• Accuracy: 0.8717
• Precision: 0.8717
• Recall: 0.8717

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: 0.0003
• train_batch_size: 128
• eval_batch_size: 128
• seed: 2024
• distributed_type: multi-GPU
• num_devices: 8
• total_train_batch_size: 1024
• total_eval_batch_size: 1024
• 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: cosine_with_restarts
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Accuracy	Precision	Recall
No log	0	0	0.0185	0.5044	0.5297	0.5418	0.5297
0.0135	0.4999	6630	0.0150	0.7539	0.755	0.7582	0.755
0.0108	0.9998	13260	0.0108	0.8539	0.8539	0.8540	0.8539
0.0109	1.4998	19890	0.0113	0.8492	0.8493	0.8496	0.8493
0.0103	1.9997	26520	0.0103	0.8641	0.8641	0.8641	0.8641
0.0099	2.4996	33150	0.0109	0.8575	0.8579	0.8630	0.8579
0.0095	2.9995	39780	0.0103	0.8686	0.8686	0.8686	0.8686
0.0092	3.4995	46410	0.0101	0.8700	0.87	0.8700	0.87
0.0094	3.9994	53040	0.0097	0.8751	0.8751	0.8751	0.8751
0.0095	4.4993	59670	0.0105	0.8664	0.8664	0.8664	0.8664
0.0086	4.9992	66300	0.0101	0.8717	0.8717	0.8717	0.8717

Framework versions

• Transformers 4.48.0.dev0
• Pytorch 2.5.1
• Datasets 3.2.0
• Tokenizers 0.21.0