AI Paper Atlas

Research map / Benchmarks / NoCaps

NoCaps leaderboard

31 papers at ICML, NeurIPS, ICLR, CVPR, AAAI, ICCV and ECCV, and the journals TPAMI, IJCV and JMLR report results on NoCaps. Best CIDEr: 124.4 (PaLI-17B, ICLR 2023).

Numbers are as each paper reports them, read by AI from its results tables and not checked by hand. Papers can differ in setting (training data, model size, extra data); the topic column shows what each paper is about. Open the paper before you compare or quote a number.

NoCaps · CIDEr

20 papers, each with its best reported number for its own method; higher is better.

#CIDEr ↑MethodPaperTopic
1124.4PaLI-17B
mT5-XXL + ViT-e
PaLI: A Jointly-Scaled Multilingual Language-Image Model
ICLR 2023 · Table 1
Human and embodied modeling
2124.2DDT-LLaMA
LLaMA3-8B
Generative Multimodal Pretraining with Discrete Diffusion Timestep Tokens
CVPR 2025 · Table 3
Multimodal LLMs
3124Morph-TokenAuto-Encoding Morph-Tokens for Multimodal LLM
ICML 2024 · Table 1
Multimodal LLMs
4123.7NExT-GPT
Vicuna-7B
NExT-GPT: Any-to-Any Multimodal LLM
ICML 2024 · Image captioning table
Multimodal LLMs
5123.1InstructBLIP (Vicuna-7B)
Vicuna-7B
InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning
NeurIPS 2023 · Table 1
CLIP and zero-shot
6121.4Qwen-VL + SC-Tune
Qwen-VL
SC-Tune: Unleashing Self-Consistent Referential Comprehension in Large Vision Language Models
CVPR 2024 · Table: Image-level evaluation
Multimodal LLMs
7117.2ASM-FT
Husky-7B
The All-Seeing Project: Towards Panoptic Visual Recognition and Understanding of the Open World
ICLR 2024 · Table: Zero-shot performance o
Visual question answering
8115.9OneLLM-7B
LLaMA2-7B
OneLLM: One Framework to Align All Modalities with Language
CVPR 2024 · Table 1
Multimodal LLMs
9114.2LaVIT (LLaMA-7B)
LLaMA-7B
Unified Language-Vision Pretraining in LLM with Dynamic Discrete Visual Tokenization
ICLR 2024 · zero-shot understanding table
Human and embodied modeling
10111.4AuroraCap-7B
Vicuna-1.5-7B
AuroraCap: Efficient, Performant Video Detailed Captioning and a New Benchmark
ICLR 2025 · Image-captioning comparison ta
Video-language understanding
10 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

NoCaps · CIDEr (val)

11 papers, each with its best reported number for its own method; higher is better.

#CIDEr ↑MethodPaperTopic
1151.1Florence-2-LFlorence-2: Advancing a Unified Representation for a Variety of Vision Tasks
CVPR 2024 · Table 4
CLIP and zero-shot
2126.3PaLI-X (55B)On Scaling Up a Multilingual Vision and Language Model
CVPR 2024 · Table 1
Visual question answering
3123.8Libra
LLaMA2-7B-Chat
Libra: Building Decoupled Vision System on Large Language Models
ICML 2024 · Table 1
Cross-modal learning
4123REVEAL
T5-Large, ViT-g/14
REVEAL: Retrieval-Augmented Visual-Language Pre-Training With Multi-Source Multimodal Knowledge Memory
CVPR 2023 · Table: Image Captioning
Visual question answering
5119.3EVCAP (attentive fusion)
Vicuna-13B
EVCap: Retrieval-Augmented Image Captioning with External Visual-Name Memory for Open-World Comprehension
CVPR 2024 · Table 3
CLIP and zero-shot
6113.1ToVE
CLIP ViT-B/16
ToVE: Efficient Vision-Language Learning via Knowledge Transfer from Vision Experts
ICLR 2025 · Table: Fine-tuned caption perf
Multimodal LLMs
7111.9BLIP-2 OPT2.7B w/ ICCC
BLIP-2 OPT2.7B
Learning by Correction: Efficient Tuning Task for Zero-Shot Generative Vision-Language Reasoning
CVPR 2024 · Table 1
CLIP and zero-shot
8110.3SimVLM hugeSimVLM: Simple Visual Language Model Pretraining with Weak Supervision
ICLR 2022 · Table 1
CLIP and zero-shot
9110.2ECoFLaP zeroth-order
BLIP-2
ECoFLaP: Efficient Coarse-to-Fine Layer-Wise Pruning for Vision-Language Models
ICLR 2024 · Table 1
Multimodal LLMs
10103.4ViLTA-BASEViLTA: Enhancing Vision-Language Pre-training through Textual Augmentation
ICCV 2023 · Table 3
CLIP and zero-shot
1 more papers on this leaderboard, the baselines reported on it and a CSV download are free with a scime account.
Sign in free to see all

Not affiliated with any conference, journal or the benchmark's authors. Corrections: contact@latenticlabs.com.