LLMsForDisinformationAnalysis/supporting/RAGAS_Service at 624d45bc5392da83fd1ec703f3b8467d24d8585f - LLMsForDisinformationAnalysis - Gitea on Jillbox

jill/LLMsForDisinformationAnalysis

Files

T

History

William Jeynes 624d45bc53 Re-allow multithreading on service. Add results table

2026-03-24 18:29:40 +00:00

..

.gitignore

add linear regression model initial version

2026-03-24 12:25:15 +00:00

ensemble_service.py

Re-allow multithreading on service. Add results table

2026-03-24 18:29:40 +00:00

flan_service.py

Add training scripts for distilled, flan. Add run service for flan

2026-03-23 22:43:59 +00:00

generate_adversarial2.py

Working on making the classifier harsher on unseen data

2026-03-17 22:19:03 +00:00

generate_adversarial.py

Refine scoring to allow for better iteration on frontend. Update generate_adversarial.py

2026-03-22 16:04:38 +00:00

prepare_data.py

Add initial version of ROBERTA classifier, add ability for multi pi charts

2026-03-11 22:02:31 +00:00

ragas_service.py

cleanup requirements.txt for ragas service

2026-02-09 21:45:56 +00:00

README.md

Re-allow multithreading on service. Add results table

2026-03-24 18:29:40 +00:00

regression_service.py

Add downloading from hugging face

2026-03-24 13:23:08 +00:00

requirements.txt

Add initial version of ROBERTA classifier, add ability for multi pi charts

2026-03-11 22:02:31 +00:00

roberta_service.py

Add training scripts for distilled, flan. Add run service for flan

2026-03-23 22:43:59 +00:00

train_distilled.py

Add training scripts for distilled, flan. Add run service for flan

2026-03-23 22:43:59 +00:00

train_flan.py

Add training scripts for distilled, flan. Add run service for flan

2026-03-23 22:43:59 +00:00

train_regression.py

Increase dropout on regression model to cut down on overfitting

2026-03-24 13:16:18 +00:00

train_roberta.py

Make the model less overfitting. Make it harder for an event to be classed as "perfect"

2026-03-18 01:05:24 +00:00

README.md

Model	% Correct	% Valid taken forward	Used in ensemble	Link
Original	53.22	61.72
Original (RAGAS)	56.01	57.73
Roberta (base)	75	70
Roberta (Generated Data)	76	71
Roberta (Generated Data + Back Translation)	74	71
Roberta (Generated Data + Back Translation + Thresholding)	77	90	Y	Here
Distilled Roberta	72.73	69.57
Flan	79.17	85.71	Y	Here
Simple Regression Model	74.77	85.71	Y	Here
Ensemble Model (weighted confidence score sum)	84.21	83.33
Ensemble Model (majority voting)	80.2	95.12