Image Captioning on Flickr30K (CIDEr)
87.12CIDEr ScoreLION-12B
Evaluation results
| Method | Links | |
|---|---|---|
| LION-12Bparameters=12B2023.11 | 87.12 | |
| BLIVA2023.11 | 87.1 | |
| MonkeyParameters=9.8B, Resolution=896x8962023.11 | 86.1 | |
| Qwen-VL + SC-TuneModel Category=Generalists, SC-Tune=true2024.03 | 86 | |
| Qwen-VLBackbone=Qwen-7B2023.11 | 85.8 | |
| InstructBLIP (T5XXL)backbone=T5XXL2023.11 | 85.79 | |
| LION-4Bparameters=4B2023.11 | 85.57 | |
| LyricsBackbone=Vicuna-13B, Evaluation protocol=0-shot2023.12 | 85.4 | |
| mPLUG-Owl2#Params=8.2B, Model Type=Generalist, Evaluation Protocol=Zero-Shot2023.11 | 85.1 | |
| mPLUG-Owl22023.11 | 85.1 | |
| InstructBLIP+ (T5XL)backbone=T5XL, uses_in-house_data=true2023.11 | 84.5 | |
| Specialists SOTAsModel Category=Specialists2024.03 | 84.5 | |
| CaMML-13Bshots=32024.01 | 84.5 | |
| InstructBLIP (T5XL)backbone=T5XL2023.11 | 83.71 | |
| InstructBLIP+ (T5XXL)backbone=T5XXL, uses_in-house_data=true2023.11 | 83.5 | |
| InstructBLIPModel Category=Generalists2024.03 | 82.8 | |
| InstructBLIPBackbone=Vicuna-13B, Evaluation protocol=0-shot2023.12 | 82.8 | |
| InstructBLIPBackbone LLM=Vicuna-13B2024.06 | 82.8 | |
| InstructBLIPBackbone=Vicuna-13B2023.11 | 82.8 | |
| InstructBLIPLLM=Vicuna-13B2023.12 | 82.8 | |
| CaMML-7Bshots=32024.01 | 82.7 | |
| ChatBridge-13BBackbone LLM=Vicuna-13B2024.06 | 82.5 | |
| ChatBridge-13BLLM=Vicuna-13B2023.12 | 82.5 | |
| InstructBLIP#Params=8.2B, Model Type=Generalist, Evaluation Protocol=Zero-Shot2023.11 | 82.4 | |
| InstructBLIPBackbone LLM=Vicuna-7B2024.06 | 82.4 | |
| InstructBLIPLLM=Vicuna-7B2023.12 | 82.4 | |
| Qwen-VL-Chat#Params=9.6B, Model Type=Generalist, Evaluation Protocol=Zero-Shot2023.11 | 81 | |
| Qwen-VLModel Category=Generalists, SC-Tune=false2024.03 | 81 | |
| Qwen-VLBackbone=Qwen-7B, Evaluation protocol=0-shot2023.12 | 81 | |
| Qwen-VLBackbone LLM=Qwen-7B2024.06 | 81 | |
| Qwen-VL-Chat2023.11 | 81 | |
| Qwen-VLLLM=Qwen-7B2023.12 | 81 | |
| MiniGPTV2evaluated_by_authors=true2023.11 | 80.75 | |
| Kosmos-22023.11 | 80.5 | |
| OneLLM-7BBackbone LLM=LLaMA2-7B2024.06 | 78.6 | |
| OneLLM-7BLLM=LLAMA2-7B2023.12 | 78.6 | |
| MiniGPT-v2 + SC-TuneModel Category=Generalists, SC-Tune=true2024.03 | 78.4 | |
| Emu-Izero-shot=true, instruction_tuned=true2023.07 | 77.4 | |
| CASP_QuIP#Bit=22025.03 | 77.2 | |
| MiniGPT-v2Model Category=Generalists, SC-Tune=false2024.03 | 77.1 | |
| MiCo-Chat-7BBackbone LLM=Vicuna-7B2024.06 | 76.3 | |
| Flamingo-80Bshots=322024.01 | 75.4 | |
| KOSMOS-1shots=42024.01 | 75.3 | |
| QuIP#Bit=22025.03 | 75.3 | |
| Flamingo-80Bshots=42024.01 | 75.1 | |
| Vanilla (LLaVA-1.5-7B)Retained Tokens=576, Reduction Rate=0.0%2026.01 | 75 | |
| BLIP-2#Params=8.2B, Model Type=Generalist, Evaluation Protocol=Zero-Shot2023.11 | 74.9 | |
| BLIP-2Backbone LLM=Vicuna-7B2024.06 | 74.9 | |
| BLIP-2LLM=Vicuna-7B2023.12 | 74.9 | |
| OriginalBit=162025.03 | 74.4 | |
| FastVRetained Tokens=1922026.01 | 74.1 | |
| Shikra#Params=7.2B, Model Type=Generalist, Evaluation Protocol=Zero-Shot2023.11 | 73.9 | |
| Shikra2023.11 | 73.9 | |
| ShikraModel Category=Generalists2024.03 | 73.9 | |
| ShikraBackbone=Vicuna-13B, Evaluation protocol=0-shot2023.12 | 73.9 | |
| ShikraBackbone=Vicuna-13B2023.11 | 73.9 | |
| VisionZipRetained Tokens=1922026.01 | 73.7 | |
| ViTCoPRetained Tokens=1922026.01 | 73.5 | |
| PyramidDropRetained Tokens=1922026.01 | 73.4 | |
| Flamingo-9Bshots=322024.01 | 72.8 | |
| Flamingo-9Bshots=42024.01 | 72.6 | |
| ViTCoPRetained Tokens=1282026.01 | 72.4 | |
| Emuzero-shot=true, prompt_strategy=Chain-of-Thought2023.07 | 72 | |
| SparseVLMRetained Tokens=1922026.01 | 72 | |
| FastVRetained Tokens=1282026.01 | 71.9 | |
| BLIP-2Model Category=Generalists2024.03 | 71.6 | |
| BLIP-2Backbone=Vicuna-13B, Evaluation protocol=0-shot2023.12 | 71.6 | |
| BLIP-2Backbone LLM=Vicuna-13B2024.06 | 71.6 | |
| BLIP-2Backbone=Vicuna-13B2023.11 | 71.6 | |
| BLIP-2LLM=Vicuna-13B2023.12 | 71.6 | |
| VisionZipRetained Tokens=1282026.01 | 71.3 | |
| ViTCoPRetained Tokens=642026.01 | 69.6 | |
| PyramidDropRetained Tokens=1282026.01 | 69.2 | |
| VisInContextText=Rendered Image, ICL Tokens=2048, Shots=322024.06 | 68.4 | |
| CASP_AQLMBit=22025.03 | 68.3 | |
| Flamingo-80BBackbone=80B, Evaluation protocol=0-shot2023.12 | 67.2 | |
| FlamingoParameters=80B2023.11 | 67.2 | |
| Flamingo-80BLLM=Chinchilla-70B2023.12 | 67.2 | |
| Kosmos-12023.11 | 67.1 | |
| Kosmos-1zero-shot=true2023.07 | 67.1 | |
| Open-Flamingo MOEText=Raw Text, ICL Tokens=256, Shots=322024.06 | 66.2 | |
| VisInContextText=Rendered Image, ICL Tokens=2048, Shots=42024.06 | 65.3 | |
| VisionZipRetained Tokens=642026.01 | 65.1 | |
| CASP_GPTQBit=2.22025.03 | 64.1 | |
| Open-Flamingo MOEText=Raw Text, ICL Tokens=256, Shots=42024.06 | 63.5 | |
| Flamingo-9Bparameters=9B2023.11 | 61.5 | |
| Flamingo-9BBackbone=9B, Evaluation protocol=0-shot2023.12 | 61.5 | |
| Flamingo-9Bzero-shot=true, prompt_strategy=2-example text-only prompt2023.07 | 61.5 | |
| Flamingo-9BLLM=Chinchilla-7B2023.12 | 61.5 | |
| Flamingo-3Bparameters=3B2023.11 | 60.6 | |
| VisInContextText=Rendered Image, ICL Tokens=2048, Shots=02024.06 | 60.5 | |
| Open-Flamingo MOEText=Raw Text, ICL Tokens=256, Shots=02024.06 | 59.4 | |
| SparseVLMRetained Tokens=1282026.01 | 58.3 | |
| AQLMBit=22025.03 | 57.3 | |
| IDEFICS-80BBackbone=LLaMA-65B, Evaluation protocol=0-shot2023.12 | 53.7 | |
| IDEFICS-80BLLM=LLaMA-65B2023.12 | 53.7 | |
| ReViLMshots=02024.01 | 52.1 | |
| FastVRetained Tokens=642026.01 | 51.1 | |
| GIT2#Params=5.1B, Model Type=Specialist, Evaluation Protocol=Fine-tuned2023.11 | 50.7 | |
| GIT#Params=0.7B, Model Type=Specialist, Evaluation Protocol=Fine-tuned2023.11 | 49.6 |