Skip to content

[AAAI 2025] Adaptive Prompting for Continual Relation Extraction: A Within-Task Variance Perspective

Notifications You must be signed in to change notification settings

Fsoft-AIC/WAVE-CRE

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Requirements

Install the necessary libraries using the following commands:

pip install transformers
pip install gpytorch

Note: Ensure you have Python and pip installed before running the commands.

Run the Code

Use the following command to execute the script with the specified hyperparameters:

python run.py \
    --max_length 256 \
    --dataname TACRED \
    --encoder_epochs 30 \
    --encoder_lr 2e-5 \
    --prompt_pool_epochs 25 \
    --prompt_pool_lr 1e-4 \
    --classifier_epochs 250 \
    --seed 2021 \
    --bert_path bert-base-uncased \
    --data_path datasets \
    --prompt_length 8 \
    --prompt_top_k 4 \
    --batch_size 16 \
    --prompt_pool_size 20 \
    --replay_s_e_e 100 \
    --replay_epochs 200 \
    --classifier_lr 5e-5 \
    --prompt_type only_prompt

Parameter Variations

  • Seeds: 2021, 2121, 2221, 2321, 2421
  • Dataname Options: TACRED, FewRel

Tip: To run experiments with different seeds or datasets, modify the --seed and --dataname arguments accordingly.

About

[AAAI 2025] Adaptive Prompting for Continual Relation Extraction: A Within-Task Variance Perspective

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published