Video Captioning on VATEX Chinese
33.2BLEUInternVL2.5-1B-Instruct
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| InternVL2.5-1B-InstructSize=1B, Mode=Instruct2026.03 | 33.2 | 22.5 | 4.4 | 18.8 | 19.1 | 31.5 | 24.2 | |
| v-Sonar w/ OmniSONAR DecoderDecoder=OmniSONAR2026.03 | 30.6 | 32.1 | 8.53 | 26.9 | 23.2 | 48.5 | 33.7 | |
| InternVL2-1B-InstructSize=1B, Mode=Instruct2026.03 | 22.3 | 14.1 | 2.9 | 11.8 | 8.7 | 18.2 | 12.6 |