Architecture
ViT-B/16, BanglaBERT and LoRA pipeline draft.
ARCHITECTURE.md
Research documentation
Working documentation for an undergraduate thesis framework studying flood-damage classification and Bengali caption generation.
Status and scope
These are transparent working drafts. Final metrics, exact implementation details, dataset license information and publication status will be added only after verification from the code, experiment logs and official sources.
ViT-B/16, BanglaBERT and LoRA pipeline draft.
ARCHITECTURE.md
Intended use, limitations and responsible-use guidance.
MODEL_CARD.md
CrisisMMD context, Bengali annotation and preparation checklist.
DATASET.md
Environment, training and evaluation checklist.
REPRODUCIBILITY.md
Translation bias, privacy, safety and accountability notes.
ETHICS.md
Draft software citation metadata for the project.
CITATION.cff
The documented pipeline combines a ViT-B/16 vision component and BanglaBERT language component, with LoRA-based parameter-efficient adaptation for two target tasks:
The exact fusion, trainable layers and decoder behavior remain implementation-verification items rather than unsupported claims.
The current project materials document 3,755 CrisisMMD flood images and machine-translated Bengali annotations.
This is a research prototype, not an emergency-response authority or verified field-reporting service.
Reproducibility instructions are prepared as a checklist and will be completed from the final experiment configuration.