Image Parsing on 100 images from ICT domain
72CIDErDICModel
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| DICModelSize=7B2026.01 | 72 | 19 | 58 | |
| Intern2.5-VLSize=32B2026.01 | 67 | 20 | 48 | |
| Gemini-2.0Size=API2026.01 | 66 | 19 | 50 | |
| Claude-3.5Size=API2026.01 | 65 | 19 | 52 | |
| Qwen2.5-VLSize=32B2026.01 | 65 | 20 | 46 | |
| Qwen2.5-VLSize=7B2026.01 | 53 | 16 | 37 | |
| InternLM-XC-2.0Size=7B2026.01 | 50 | 17 | 35 | |
| LLAVA-1.5Size=7B2026.01 | 26 | 12 | 26 | |
| Phi-3.5-visionSize=4B2026.01 | 23 | 17 | 4 |