Scene Captioning on COCO RGBP seen scenes (test)
127.58CIDErQwen3-VL-8B-Instruct RGBP
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Qwen3-VL-8B-Instruct RGBPBackbone=Qwen3-VL-8B, Dataset=RGBP2026.04 | 127.58 | 21.14 | |
| ProCapBackbone=Vicuna-1.5-7B2026.04 | 78.99 | 13.83 | |
| ProCapBackbone=OpenLLaMA-3B2026.04 | 76.98 | 13.58 | |
| ProCapBackbone=TinyLlama-1.1B2026.04 | 70.27 | 21.92 | |
| ProCapBackbone=OpenLLaMA-3B2026.04 | 69.43 | 22.01 | |
| ProCapBackbone=OPT-2.7B2026.04 | 57.72 | 10.21 | |
| ProCapBackbone=TinyLlama-1.1B2026.04 | 54.37 | 9.75 | |
| Qwen3-VL-8B-Instruct RGBPBackbone=Qwen3-VL-8B, Dataset=RGBP2026.04 | 37.81 | 23.22 | |
| ProCapBackbone=Vicuna-1.5-7B2026.04 | 36.17 | 22.75 | |
| ProCapBackbone=OPT-2.7B2026.04 | 28.01 | 24.34 | |
| Qwen3-VL-8B-InstructParameters=8B2026.04 | 11.56 | 11.63 | |
| Qwen3-VL-2B-InstructParameters=2B2026.04 | 10.73 | 12.47 | |
| Qwen3-VL-4B-InstructParameters=4B2026.04 | 10.59 | 11.72 | |
| FastVLM-7BParameters=7B2026.04 | 7.65 | 7.33 | |
| FastVLM-1.5BParameters=1.5B2026.04 | 7.21 | 7.31 | |
| FastVLM-0.5BParameters=0.5B2026.04 | 7.17 | 7.43 | |
| Qwen3-VL-4B-InstructParameters=4B2026.04 | 2.38 | 11.68 | |
| Qwen3-VL-8B-InstructParameters=8B2026.04 | 2.38 | 13.03 | |
| Qwen3-VL-2B-InstructParameters=2B2026.04 | 2.34 | 12.17 | |
| FastVLM-7BParameters=7B2026.04 | 2.31 | 10.31 | |
| FastVLM-1.5BParameters=1.5B2026.04 | 1.92 | 8.77 | |
| FastVLM-0.5BParameters=0.5B2026.04 | 1.72 | 9.2 |