murat_chem_model

This model is a fine-tuned version of alvaroalon2/biobert_chemical_ner on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3827
• Chemical: {'precision': 0.8843176605504587, 'recall': 0.8439124487004104, 'f1': 0.863642727145457, 'number': 7310}
• Overall Precision: 0.8843
• Overall Recall: 0.8439
• Overall F1: 0.8636
• Overall Accuracy: 0.9447

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-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Chemical	Overall Precision	Overall Recall	Overall F1	Overall Accuracy
1.6817	1.2048	100	0.3090	{'precision': 0.8702002710435175, 'recall': 0.7905608755129959, 'f1': 0.828471077342126, 'number': 7310}	0.8702	0.7906	0.8285	0.9204
1.6817	2.4096	200	0.2900	{'precision': 0.8720861611094718, 'recall': 0.8086183310533516, 'f1': 0.8391538898353209, 'number': 7310}	0.8721	0.8086	0.8392	0.9312
1.6817	3.6145	300	0.3071	{'precision': 0.8830255057167986, 'recall': 0.824076607387141, 'f1': 0.8525332578545147, 'number': 7310}	0.8830	0.8241	0.8525	0.9398
1.6817	4.8193	400	0.3099	{'precision': 0.882646325363063, 'recall': 0.8231190150478797, 'f1': 0.8518439866921499, 'number': 7310}	0.8826	0.8231	0.8518	0.9411
0.1018	6.0241	500	0.3891	{'precision': 0.8816456613066782, 'recall': 0.8325581395348837, 'f1': 0.8563990712727785, 'number': 7310}	0.8816	0.8326	0.8564	0.9402
0.1018	7.2289	600	0.3672	{'precision': 0.8851640513552068, 'recall': 0.8488372093023255, 'f1': 0.8666201117318435, 'number': 7310}	0.8852	0.8488	0.8666	0.9451
0.1018	8.4337	700	0.3459	{'precision': 0.8812949640287769, 'recall': 0.8378932968536251, 'f1': 0.8590462833099579, 'number': 7310}	0.8813	0.8379	0.8590	0.9449
0.1018	9.6386	800	0.3601	{'precision': 0.880656108597285, 'recall': 0.8519835841313269, 'f1': 0.8660826032540675, 'number': 7310}	0.8807	0.8520	0.8661	0.9462
0.1018	10.8434	900	0.3711	{'precision': 0.881471972614463, 'recall': 0.8454172366621067, 'f1': 0.8630682214929124, 'number': 7310}	0.8815	0.8454	0.8631	0.9443
0.0038	12.0482	1000	0.3779	{'precision': 0.8816542644533486, 'recall': 0.8428180574555404, 'f1': 0.8617988529864317, 'number': 7310}	0.8817	0.8428	0.8618	0.9437
0.0038	13.2530	1100	0.3829	{'precision': 0.8837275985663082, 'recall': 0.8432284541723666, 'f1': 0.863003150157508, 'number': 7310}	0.8837	0.8432	0.8630	0.9447
0.0038	14.4578	1200	0.3827	{'precision': 0.8843176605504587, 'recall': 0.8439124487004104, 'f1': 0.863642727145457, 'number': 7310}	0.8843	0.8439	0.8636	0.9447

Framework versions

• Transformers 4.44.2
• Pytorch 2.3.0+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1