REVIEW 35 cited by
MANTIS: Interleaved Multi-Image Instruction Tuning
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
MANTIS: Interleaved Multi-Image Instruction Tuning
read the original abstract
Large multimodal models (LMMs) have shown great results in single-image vision language tasks. However, their abilities to solve multi-image visual language tasks is yet to be improved. The existing LMMs like OpenFlamingo, Emu2, and Idefics gain their multi-image ability through pre-training on hundreds of millions of noisy interleaved image-text data from the web, which is neither efficient nor effective. In this paper, we aim to build strong multi-image LMMs via instruction tuning with academic-level resources. Therefore, we meticulously construct Mantis-Instruct containing 721K multi-image instruction data to train a family of Mantis models. The instruction tuning empowers Mantis with different multi-image skills like co-reference, comparison, reasoning, and temporal understanding. We evaluate Mantis on 8 multi-image benchmarks and 6 single-image benchmarks. Mantis-Idefics2 can achieve SoTA results on all the multi-image benchmarks and beat the strongest multi-image baseline, Idefics2-8B by an average of 13 absolute points. Notably, Idefics2-8B was pre-trained on 140M interleaved multi-image data, which is 200x larger than Mantis-Instruct. We observe that Mantis performs equivalently well on the held-in and held-out benchmarks, which shows its generalization ability. We further evaluate Mantis on single-image benchmarks and demonstrate that Mantis also maintains a strong single-image performance on par with CogVLM and Emu2. Our results show that multi-image abilities are not necessarily gained through massive pre-training, instead, they can be gained by low-cost instruction tuning. The training and evaluation of Mantis has paved the road for future work to improve LMMs' multi-image abilities.
Forward citations
Cited by 35 Pith papers
-
DataComp-VLM: Improved Open Datasets for Vision-Language Models
DataComp-VLM benchmark shows instruction-heavy data mixing outperforms filtering for VLM training, with DCVLM-Baseline achieving 63.6% on 33 tasks for 8B models (+5.4pp over FineVision).
-
Molmo and PixMo: Open Weights and Open Data for State-of-the-Art Vision-Language Models
Molmo VLMs trained on newly collected PixMo open datasets achieve state-of-the-art performance among open-weight models and surpass multiple proprietary VLMs including Claude 3.5 Sonnet and Gemini 1.5 Pro.
-
MMMU-Pro: A More Robust Multi-discipline Multimodal Understanding Benchmark
MMMU-Pro is a stricter multimodal benchmark that removes text-only solvable questions, augments options, and requires reading text from images, yielding substantially lower model scores of 16.8-26.9%.
-
Same Evidence, Different Answer: Auditing Order Sensitivity in Multimodal Large Language Models
All 18 audited MLLMs exhibit order sensitivity with per-facet flip rates of 24-50%, exceeding same-order decoder noise.
-
CGC: Compositional Grounded Contrast for Fine-Grained Multi-Image Understanding
CGC improves fine-grained multi-image understanding in MLLMs by constructing contrastive training instances from existing single-image annotations and adding a rule-based spatial reward, achieving SOTA on MIG-Bench an...
-
LLaVA-NeXT-Interleave: Tackling Multi-image, Video, and 3D in Large Multimodal Models
LLaVA-NeXT-Interleave unifies multi-image, video, and 3D capabilities in large multimodal models via a new 1.18M-sample interleaved dataset and benchmark, achieving leading results across those tasks while preserving ...
-
MuirBench: A Comprehensive Benchmark for Robust Multi-image Understanding
MuirBench is a new benchmark showing that top multimodal LLMs struggle with robust multi-image understanding, with GPT-4o at 68% and open-source models below 33% accuracy.
-
MVEI & EmObserver: Empowering MLLM-Oriented Visual Emotional Intelligence via Emotion Statement Judgement
ESJ statement verification, the INSETS-462k corpus, the MVEI benchmark, and the EmObserver model together provide a scalable way to evaluate and improve MLLM visual emotional intelligence.
-
DataComp-VLM: Improved Open Datasets for Vision-Language Models
DataComp-VLM benchmark shows instruction-heavy data mixtures outperform caption-heavy ones for VLM training, with DCVLM-Baseline reaching 63.6% on 33 tasks using 200B tokens, +5.4pp over FineVision.
-
Pareto LoRA: Mitigating Modality Imbalance in Unified Multimodal Models via Pareto-Optimal Gradient Integration
Pareto LoRA applies Pareto-optimal gradient integration to balance text and image objectives in LoRA-based fine-tuning of unified multimodal models, reporting up to 44.9% gains in image quality on the CoMM benchmark w...
-
HYDRA-X: Native Unified Multimodal Models with Holistic Visual Tokenizers
HYDRA-X presents the first unified multimodal model using a single ViT for holistic image-video tokenization, with ablations on attention and compression plus a latent-level editing improvement.
-
LoMo: Local Modality Substitution for Deeper Vision-Language Fusion
LoMo is a lightweight data curation technique that locally substitutes text with images in prompts to enforce cross-modal invariance, yielding 2.67-2.82 point gains over standard SFT on two VLMs across 13 benchmarks.
-
ROVER: Routing Object-Centric Visual Evidence for Grounded Multi-Image Reasoning
ROVER introduces a learnable routing plugin for object-centric visual evidence in MLLMs via token triplets and differential attention, reporting gains on MM-GCoT and VideoEspresso when integrated into Qwen2.5-VL-7B.
-
SpatialForge: Bootstrapping 3D-Aware Spatial Reasoning from Open-World 2D Images
SpatialForge synthesizes 10 million spatial QA pairs from in-the-wild 2D images to train VLMs for better depth ordering, layout, and viewpoint-dependent reasoning.
-
MiniCPM-o 4.5: Towards Real-Time Full-Duplex Omni-Modal Interaction
MiniCPM-o 4.5 uses the Omni-Flow streaming framework to deliver real-time full-duplex omni-modal interaction with proactive behavior in a 9B model that approaches Gemini 2.5 Flash performance.
-
Decoding the Pulse of Reasoning VLMs in Multi-Image Understanding Tasks
PulseFocus improves multi-image reasoning in VLMs by interleaving planning and attention-gated focus blocks during chain-of-thought, achieving gains on BLINK and MuirBench.
-
ReMoT: Reinforcement Learning with Motion Contrast Triplets
Training a 4B vision-language model on rule-generated motion-contrast triplets with GRPO lifts spatio-temporal QA accuracy by about 17 points on the authors' own benchmark and by smaller margins on standard benchmarks.
-
MiMo-Embodied: X-Embodied Foundation Model Technical Report
MiMo-Embodied is a single foundation model that achieves state-of-the-art results on 17 embodied AI benchmarks and 12 autonomous driving benchmarks through multi-stage learning, curated data, and CoT/RL fine-tuning th...
-
CARES: Context-Aware Resolution Selector for VLMs
A lightweight selector trained on target-VLM rollouts routes each image-query pair to the minimal sufficient resolution, cutting prefill compute by 63-80% across five benchmarks at roughly matched accuracy.
-
InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency
InternVL3.5 advances open-source multimodal models with Cascade RL for +16% reasoning gains and ViR for 4x inference speedup, with the 241B model reaching SOTA among open-source MLLMs on multimodal, reasoning, and age...
-
SEAM: Semantically Equivalent Across Modalities Benchmark for Vision-Language Models
SEAM measures VLM reasoning consistency across modalities using paired semantically equivalent textual and visual notations, and finds systematic vision-language imbalance.
-
InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Models
InternVL3-78B sets a new open-source SOTA of 72.2 on MMMU via native joint multimodal pre-training, V2PE, MPO, and test-time scaling while remaining competitive with proprietary models.
-
Improving Video Generation with Human Feedback
A human preference dataset and VideoReward model enable Flow-DPO and Flow-NRG to produce smoother, better-aligned videos from text prompts in flow-based generators.
-
Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scaling
InternVL 2.5 is the first open-source MLLM to surpass 70% on the MMMU benchmark via model, data, and test-time scaling, with a 3.7-point gain from chain-of-thought reasoning.
-
CLBench-V: Evaluating Multimodal Context Learning from Grounding to Knowledge Acquisition
Multimodal context-learning benchmark CLBench-V separates grounding, information application, and knowledge acquisition; the best evaluated model scores 0.2847.
-
CrossView Suite: Harnessing Cross-view Spatial Intelligence of MLLMs with Dataset, Model and Benchmark
CrossView Suite supplies a 1.6M-sample dataset, scene-disjoint benchmark, and explicit-alignment framework to advance MLLMs from single-view perception to cross-view spatial intelligence.
-
MiniCPM-V 4.5: Cooking Efficient MLLMs via Architecture, Data, and Training Recipe
An 8B MLLM reaches state-of-the-art efficiency and performance under 30B by combining a unified 3D resampler, joint document-text training, and hybrid RL for reasoning modes.
-
Reinforced Visual Perception with Tools
ReVPT uses GRPO reinforcement learning with a cold-start SFT phase to make Qwen2.5-VL models call visual tools, improving perception benchmarks over SFT and text-only RL baselines.
-
CTA-Flux: Integrating Chinese Cultural Semantics into High-Quality English Text-to-Image Communities
The paper claims a Chinese-prompt adapter for the Flux text-to-image model, but the manuscript body is an unrelated LLM self-recognition paper, leaving the claim unsupported.
-
mPLUG-Owl3: Towards Long Image-Sequence Understanding in Multi-Modal Large Language Models
mPLUG-Owl3 introduces hyper attention blocks to integrate vision and language for long image-sequence understanding and reports SOTA results on single-image, multi-image, and video benchmarks.
-
LLaVA-OneVision: Easy Visual Task Transfer
LLaVA-OneVision is the first single open LMM to simultaneously achieve strong performance in single-image, multi-image, and video scenarios with cross-scenario transfer capabilities.
-
Inference-Time Agentic Decision Rules Beat Longer Evolving Search for Multi-Image Medical Reasoning
On MedFrameQA, order-vote (57.89%) beats fixed prompting (52.73%) and order-rerank (55.79%), and a single 100-generation run drops final-test accuracy to 56.02%.
-
DFLOP: A Data-driven Framework for Multimodal LLM Training Pipeline Optimization
DFLOP is a data-driven framework that profiles data-induced computation variance and uses predictive scheduling to balance workloads in multimodal LLM training pipelines, claiming up to 3.6x faster training than exist...
-
Improving the Reasoning of Multi-Image Grounding in MLLMs via Reinforcement Learning
A pipeline of chain-of-thought data synthesis, LoRA-based supervised fine-tuning, rejection sampling, and rule-based reinforcement learning raises multi-image grounding accuracy by 9.04% on MIG-Bench and 4.41% on aver...
-
Pixel-Space Diffusion Transformers
A systematic review of pixel-space diffusion transformers, categorizing architectures and challenges for end-to-end image generation without latent compression.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.