• Using multiple workers of the default type 'sync' in gunicorn is not working on Kubernetes
• Workers constantly terminated due to signal 9
• Try gevent to see if it works out

• Added additional tutorial on how to use UFD to train and evaluate on custom dataset
• Bug fix for UFD parse_args_and_load_config util function
• Added feature to create folder if folder doesn't exist
• Added some train args param in eval args param to improve usability
• Made caching optional
• Added validation to make debugging easier
• Added links to config file examples for reccon models

Latest SenticGCN implementation for the Dev branch. In the dataclass.py, post_init method in SenticGCNTrainArgs, there are the following assertions,

assert self.repeats > 1, "Repeats value must be at least 1."
assert self.patience > 1, "Patience value must be at least 1." 

The comparison operator should be >= instead.

• Create a centralized logger for 'sgnlp' base logger
• 'sgnlp' logger is created from a config json and is init a the 'sgnlp' module init.py
• Replace all logging method call with their own script specific logger

• [x] Create a module named sgnlp.base
• [x] Add abstractmethods for preprocess, save, load
• [x] Add batch iteration to parent __call__
• [x] Parent __call__ should return a dictionary

• Add multi-word aspect support
• Update documentation to reflect multi-word support
• Update unit tests
• Update usage example to include multi-word support

The current implementation of preprocessor matches a single aspect index for the purpose of matching postprocessor output. The aspect index field for process_input payload should be expended to handle aspects with multiple indexes.

Close #43

This pull request is to add Sentic-GCN demo_api models to sgnlp. Includes the follow components:

• model_card
• api.py
• dockerfiles
• requirements.txt
• usage.py

close #41

This pull request is to add Sentic-GCN models to sgnlp. Includes the follow components:

• Models
• Configs
• Tokenizers
• Embedding models
• Trainer/Evaluator
• Unit test
• documentation

Does not include demo_api as it is covered in another issue tickets.

To allow the containers to start up quicker, models and files were downloaded and cached during build time.

Recent changes in the huggingface transformers package has broken this functionality:

• Released in v4.22.0
• Issue

Possible choices moving forward:

• Write a simple caching utility function
• Stick to versions of transformers before 4.22.0

Paper: https://aclanthology.org/2020.emnlp-main.108.pdf

Prof: Jiang Jing from SMU

Repo: GitHub - jefferyYu/DualHierarchicalTransformer: Predicting Stance and Rumor Veracity via Dual Hierarchical Transformer