curelink-biomed-nli-v5

This model is a fine-tuned version of CureLink/curelink-biomed-nli-v2 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.6909
• Accuracy: 0.6933
• F1 Macro: 0.6887

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: 5e-06
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 8
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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: 50
• num_epochs: 4

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Macro
5.7900	0.2427	200	0.6736	0.6495	0.4193
4.9944	0.4854	400	0.6474	0.6610	0.4330
4.9430	0.7282	600	0.7358	0.6457	0.5942
4.7595	0.9709	800	0.6601	0.6762	0.6694
4.6664	1.2136	1000	0.6659	0.6724	0.6682
4.3585	1.4563	1200	0.6974	0.68	0.6692
4.4865	1.6990	1400	0.6515	0.6857	0.6744
4.4832	1.9417	1600	0.6747	0.6876	0.6751
4.5484	2.1845	1800	0.6741	0.68	0.6687
4.4410	2.4272	2000	0.6652	0.6876	0.6824
4.3377	2.6699	2200	0.6780	0.6857	0.6734
4.2296	2.9126	2400	0.6830	0.6857	0.6793
3.9448	3.1553	2600	0.6907	0.6952	0.6874
3.9506	3.3981	2800	0.7021	0.6876	0.6751
3.8357	3.6408	3000	0.6994	0.6952	0.6882
4.2035	3.8835	3200	0.6913	0.6933	0.6890
4.2035	4.0	3296	0.6909	0.6933	0.6887

Framework versions

• Transformers 5.4.0
• Pytorch 2.11.0
• Datasets 4.8.4
• Tokenizers 0.22.2