Interleaved text-mask generation on GCG (test)
17.4METEORQwen25VL-SAMTok (ft)
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Qwen25VL-SAMTok (ft)Size=7B, further_finetuned_on_GCG=true2026.01 | 17.4 | 53.7 | 37.5 | 71.2 | 48.7 | |
| Qwen25VL-SAMTok (ft)Size=3B, further_finetuned_on_GCG=true2026.01 | 17 | 53.9 | 36.3 | 71.4 | 48 | |
| Qwen25VL-SAMTokSize=7B2026.01 | 17 | 54.5 | 37 | 71.7 | 47.9 | |
| Qwen25VL-SAMTokSize=3B2026.01 | 16.7 | 52.5 | 36.1 | 70.9 | 47 | |
| Qwen25VL-SAMTok (rl)Size=3B, GRPO_trained=true2026.01 | 16.5 | 51 | 40.4 | 73.2 | 53.5 | |
| Sa2VASize=8B2026.01 | 16.2 | 49 | 32.2 | 66.8 | 44.5 | |
| Qwen3VL-SAMTokSize=4B2026.01 | 16.1 | 52 | 37.4 | 70.6 | 48.2 | |
| GLaMMSize=7B2026.01 | 15.8 | 43.5 | 29.2 | 65.6 | 40.8 | |
| OMG-LLaVASize=7B2026.01 | 14.5 | 38.5 | 28.6 | 64.7 | — | |
| LISASize=7B2026.01 | 12.9 | 32.2 | 24.8 | 61.7 | 35.5 |