JorSenti-LoRA
Handle Jordanian and Levantine sentiment analysis where slang, sarcasm, and neutral sentiment routinely break generic Arabic classifiers.
Local project explainer
Ask about this system
A local, predefined assistant for the documented details in JorSenti-LoRA.
Problem
What this project had to solve
Generic Arabic sentiment models struggle with Jordanian slang, sarcasm, and class imbalance, especially when the neutral class gets ignored. This project focuses on making sentiment classification behave better on local Talabat-style language.
Role / contribution
What I owned in the system
Project author responsible for the adaptation strategy, training setup, experiment tracking, and deployment packaging described in the repository.
Technical implementation
How the system is built
The model stack adapts MARBERTv2 with PEFT/LoRA so the training path stays practical on local hardware while remaining specialized for the task.
A weighted cross-entropy loss is used to push the model to respect the neutral class instead of collapsing toward majority sentiment behavior.
The repository documents experiment tracking with MLflow and packages the live interface for deployment on Hugging Face Spaces.
Architecture flow
System flow from input to output
Review pipeline
What happens
Talabat-style review data is prepared to expose dialect-heavy customer language.
Why it matters
The input pipeline determines whether the model actually sees the slang-heavy language the project is trying to handle.
Related tools
No explicit tool is called out for this step in the shared project data.
Engineering decisions
Tradeoffs that shaped the build
Use LoRA instead of full fine-tuning
Parameter-efficient adaptation keeps the training path practical while still specializing the model.
Bias optimization toward neutral recall
Weighted loss is a targeted fix for the exact failure mode the README calls out.
Treat deployment as part of the project
The repo goes beyond training by packaging a live interface and MLOps workflow.
Stack
Core tools and system layers
What I would improve next
Next technical upgrades
- Add a cleaner production API boundary beyond the current interactive app flow.
- Expand evaluation reporting around slang, sarcasm, and neutral edge cases.
- Introduce drift checks for new review language over time.