atifmallo/rome

How atifmallo/rome works

RoME is a two-stage multimodal sentiment-analysis framework for CMU-MOSI and CMU-MOSEI. The principal workflow loads utterance features and labels, trains modality experts in Stage 1, then trains and evaluates a mask-aware multimodal model in Stage 2 across single-, pair-, and trimodal conditions. The model combines expert routing with CIT, MRR, RMC, and CM-DAE; training also computes masked regression losses and saves results. The README documents the model capabilities and evaluation conditions; detailed implementation of some Stage-2 internals was only partially sampled.

Main entry points: model.py RoME model · train_Rome.py Training and evaluation · utils.py Data and model setup · dataloader_cmumosi.py Utterance loader

Components 14

Data input

CMU features dataset
Path configuration config.py
Utterance loader Rome/dataloader_cmumosi.py

Model learning

RoME model Rome/model.py
Modality experts Rome/model.py
Expert attention Rome/modules/Attention_softmoe.py
Cross-modal interaction Rome/model.py
Reliability reweighting Rome/model.py
Latent reconstruction Rome/model.py
Residual correction Rome/model.py

Training and evaluation

Data and model setup Rome/utils.py
Masked losses Rome/loss.py
Training and evaluation Rome/train_Rome.py

Other components

Researcher

Connections 17

Researcher → Training and evaluation: starts runs README.md
Path configuration → Utterance loader: sets data paths Rome/utils.py
Path configuration → Training and evaluation: sets output paths config.py
CMU features → Utterance loader: provides features Rome/dataloader_cmumosi.py
Training and evaluation → Data and model setup: requests loaders Rome/train_Rome.py
Data and model setup → Utterance loader: builds dataset Rome/utils.py
Data and model setup → RoME model: constructs model Rome/utils.py
Training and evaluation → Masked losses: computes regression loss Rome/train_Rome.py
Training and evaluation → RoME model: runs forward pass Rome/train_Rome.py
RoME model → Modality experts: routes modalities Rome/model.py
Modality experts → Expert attention: uses attention blocks Rome/model.py
RoME model → Cross-modal interaction: interacts across modalities Rome/model.py
RoME model → Reliability reweighting: scales by reliability README.md
RoME model → Latent reconstruction: reconstructs latent features Rome/model.py
RoME model → Residual correction: refines fused output Rome/model.py
Training and evaluation → RoME model: saves checkpoints Rome/train_Rome.py
Training and evaluation → Training and evaluation: evaluates conditions README.md

Drawn by GitDiagram on November 14, 2026 from the repository's file tree, README and source files. Read as Markdown ·

How atifmallo/rome works

Main entry points: model.py RoME model · train_Rome.py Training and evaluation · utils.py Data and model setup · dataloader_cmumosi.py Utterance loader

Connections 17

Researcher → Training and evaluation: starts runs README.md
Path configuration → Utterance loader: sets data paths Rome/utils.py
Path configuration → Training and evaluation: sets output paths config.py
CMU features → Utterance loader: provides features Rome/dataloader_cmumosi.py
Training and evaluation → Data and model setup: requests loaders Rome/train_Rome.py
Data and model setup → Utterance loader: builds dataset Rome/utils.py
Data and model setup → RoME model: constructs model Rome/utils.py
Training and evaluation → Masked losses: computes regression loss Rome/train_Rome.py
Training and evaluation → RoME model: runs forward pass Rome/train_Rome.py
RoME model → Modality experts: routes modalities Rome/model.py
Modality experts → Expert attention: uses attention blocks Rome/model.py
RoME model → Cross-modal interaction: interacts across modalities Rome/model.py
RoME model → Reliability reweighting: scales by reliability README.md
RoME model → Latent reconstruction: reconstructs latent features Rome/model.py
RoME model → Residual correction: refines fused output Rome/model.py
Training and evaluation → RoME model: saves checkpoints Rome/train_Rome.py
Training and evaluation → Training and evaluation: evaluates conditions README.md

Drawn by GitDiagram on November 14, 2026 from the repository's file tree, README and source files. Read as Markdown ·