MatBERT_800abstracts_NER

This model is a fine-tuned version of Flamenco43/MatBERT on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1728
• Precision: 0.8484
• Recall: 0.8869
• F1: 0.8672
• Accuracy: 0.9497
• Per Tag Metrics: {'I-SMT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9967652080150957}, 'I-DSC': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9977086890106928}, 'B-SPL': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.999056519004403}, 'I-MAT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9949231736903585}, 'B-MAT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9944289693593314}, 'I-APL': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9973941953454938}, 'I-CMT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9952825950220145}, 'I-SPL': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.999865217000629}, 'O': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9656303351603918}, 'B-PRO': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9866564830622697}, 'B-CMT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9974391230119508}, 'B-SMT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9956869440201276}, 'I-PRO': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9864767723964417}, 'B-APL': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9975739060113218}, 'B-DSC': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9945637523587025}}

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 4

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy	Per Tag Metrics
No log	1.0	221	0.1873	0.8001	0.8524	0.8255	0.9376	{'I-SMT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9960463653517836}, 'I-DSC': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9962710036840686}, 'B-SPL': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9986970976727468}, 'I-MAT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9924521520352233}, 'B-MAT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9928115733668793}, 'I-APL': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9967652080150957}, 'I-CMT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9948782460239015}, 'I-SPL': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9995057956689729}, 'O': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9581274148620721}, 'B-PRO': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9846796657381616}, 'B-CMT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9971695570132088}, 'B-SMT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9954173780213855}, 'I-PRO': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9838260400754785}, 'B-APL': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9968999910144667}, 'B-DSC': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9916434540389972}}
No log	2.0	442	0.1639	0.8259	0.8718	0.8482	0.9464	{'I-SMT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9964956420163537}, 'I-DSC': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9970797016802947}, 'B-SPL': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9988318806721179}, 'I-MAT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9940695480276754}, 'B-MAT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9933956330308203}, 'I-APL': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9972144846796658}, 'I-CMT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9951028843561865}, 'I-SPL': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.999865217000629}, 'O': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9644622158325097}, 'B-PRO': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9865217000628987}, 'B-CMT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9971246293467517}, 'B-SMT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9954173780213855}, 'I-PRO': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9857579297331297}, 'B-APL': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9976188336777788}, 'B-DSC': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9938449096953904}}
0.2438	3.0	663	0.1699	0.8348	0.8784	0.8561	0.9487	{'I-SMT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9967202803486387}, 'I-DSC': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9972594123461227}, 'B-SPL': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.999056519004403}, 'I-MAT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9947434630245305}, 'B-MAT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9943391140264175}, 'I-APL': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9973043400125797}, 'I-CMT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9953275226884716}, 'I-SPL': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.999865217000629}, 'O': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9655404798274778}, 'B-PRO': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9870159043939257}, 'B-CMT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9973043400125797}, 'B-SMT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9956420163536706}, 'I-PRO': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9860724233983287}, 'B-APL': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9976637613442357}, 'B-DSC': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9936202713631054}}
0.2438	4.0	884	0.1728	0.8484	0.8869	0.8672	0.9497	{'I-SMT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9967652080150957}, 'I-DSC': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9977086890106928}, 'B-SPL': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.999056519004403}, 'I-MAT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9949231736903585}, 'B-MAT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9944289693593314}, 'I-APL': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9973941953454938}, 'I-CMT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9952825950220145}, 'I-SPL': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.999865217000629}, 'O': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9656303351603918}, 'B-PRO': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9866564830622697}, 'B-CMT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9974391230119508}, 'B-SMT': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9956869440201276}, 'I-PRO': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9864767723964417}, 'B-APL': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9975739060113218}, 'B-DSC': {'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'accuracy': 0.9945637523587025}}

Framework versions

• Transformers 4.41.2
• Pytorch 2.2.1+cu118
• Datasets 2.19.1
• Tokenizers 0.19.1