CeLLaTe3.1_Base_CRF_no_vague_adapted_pubmed

This model is a fine-tuned version of Mardiyyah/cellate2.0-tapt_base-LR_5e-05 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 3.1429
• Precision: 0.6820
• Recall: 0.8154
• F1: 0.7428
• Accuracy: 0.9829

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: 32
• eval_batch_size: 16
• seed: 3407
• 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_ratio: 0.01
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
40.5714	0.5848	100	8.0047	0.0776	0.0808	0.0792	0.9516
7.6643	1.1696	200	3.4953	0.6548	0.5275	0.5843	0.9804
4.1729	1.7544	300	2.9402	0.6559	0.7255	0.6889	0.9810
3.0949	2.3392	400	2.9402	0.6412	0.8171	0.7186	0.9810
2.2148	2.9240	500	2.8872	0.6915	0.7729	0.7299	0.9820
2.0593	3.5088	600	2.7382	0.6760	0.7893	0.7283	0.9824
1.7345	4.0936	700	2.9636	0.6863	0.7709	0.7262	0.9832
1.4038	4.6784	800	3.1429	0.6820	0.8154	0.7428	0.9829
1.1723	5.2632	900	4.0733	0.6382	0.7526	0.6907	0.9804
1.3177	5.8480	1000	4.8858	0.5487	0.7507	0.6340	0.9737
0.9404	6.4327	1100	4.0962	0.5945	0.7893	0.6782	0.9775

Framework versions

• Transformers 4.48.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.2
• Tokenizers 0.21.0