bert-base-uncased-finetuned-srl_arg

This model is a baseline fine-tuned version of bert-base-uncased on the English Universal Propbank dataset for the Semantics Role Labeling (SRL) task. It achieves the following results on the evaluation set:

• Loss: 0.1094
• Precision: 0.8207
• Recall: 0.8310
• F1: 0.8259
• Accuracy: 0.9722

Model description

The appraoch used for the baseline model is basically converting the sentence into the following form:

[CLS] This is the sentence content [SEP] is [SEP].

And this is realized by simply using the logic of the auto tokenizer: tokenizer(list1,list2) will return [CLS] list1 content [SEP] list2 content [SEP].

Usages

The model labels semantics roles given input sentences. See usage examples at https://github.com/dannashao/bertsrl/blob/main/Evaluation.ipynb

Training and evaluation data

The English Universal Proposition Bank v1.0 data. See details at https://github.com/UniversalPropositions/UP-1.0

Training procedure

See details at https://github.com/chuqiaog/Advanced_NLP_group_1/blob/main/A3/A3_main.ipynb

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: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
0.1082	1.0	2655	0.1236	0.7783	0.8158	0.7966	0.9671
0.0772	2.0	5310	0.1089	0.8055	0.8277	0.8165	0.9708
0.0609	3.0	7965	0.1094	0.8207	0.8310	0.8259	0.9722

Framework versions

• Transformers 4.37.0
• Pytorch 2.0.1+cu117
• Datasets 2.16.1
• Tokenizers 0.15.1