TorchUMM: A Unified Multimodal Model Codebase for Evaluation, Analysis, and Post-training

TL;DR

TorchUMM is a comprehensive, unified codebase for evaluating, analyzing, and post-training diverse multimodal models across various tasks, datasets, and architectures.

Contribution

It introduces the first unified framework supporting diverse UMM backbones, tasks, and datasets with standardized evaluation protocols.

Findings

Supports a broad spectrum of models and tasks.

Enables fair and reproducible comparisons.

Facilitates deeper insights into model strengths and limitations.

Abstract

Recent advances in unified multimodal models (UMMs) have led to a proliferation of architectures capable of understanding, generating, and editing across visual and textual modalities. However, developing a unified framework for UMMs remains challenging due to the diversity of model architectures and the heterogeneity of training paradigms and implementation details. In this paper, we present TorchUMM, the first unified codebase for comprehensive evaluation, analysis, and post-training across diverse UMM backbones, tasks, and datasets. TorchUMM supports a broad spectrum of models covering a wide range of scales and design paradigms. Our benchmark encompasses three core task dimensions: multimodal understanding, generation, and editing, and integrates both established and novel datasets to evaluate perception, reasoning, compositionality, and instruction-following abilities. By providing a unified interface and standardized evaluation protocols, TorchUMM enables fair and reproducible comparisons across heterogeneous models and fosters deeper insights into their strengths and limitations, facilitating the development of more capable unified multimodal systems. Code is available at: https://github.com/AIFrontierLab/TorchUMM.