Text-to-Video (T2V) Temporally Grounded Generation on ChronusAV
1.4BLEU-4PadCaptioner
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| PadCaptionerModel Parameters=3B2026.07 | 1.4 | — | — | 3.9 | 2.7 | 5.6 | |
| ChronusOmniModel Parameters=7B2026.07 | 1.2 | — | — | 3.4 | 2.1 | 5.1 | |
| LongVALE-LLMModel Parameters=7B2026.07 | 0.4 | — | — | 2 | 1 | 1.4 | |
| Qwen2.5-OmniModel Parameters=7B2026.07 | 0.4 | — | — | 1.9 | 1 | 0.8 | |
| Qwen3-OmniModel Parameters=30B-A3B2026.07 | 0.4 | — | — | 2.1 | 1.6 | 0.9 | |
| OlaModel Parameters=7B2026.07 | 0.3 | — | — | 1.7 | 1 | 0.5 | |
| ARC-HY-VideoModel Parameters=7B2026.07 | 0.2 | — | — | 1.5 | 1.1 | 0.5 | |
| VideoLLaMAModel Parameters=7B2026.07 | 0.1 | — | — | 1.5 | 0.9 | 0 | |
| AVicunaModel Parameters=7B2026.07 | 0 | — | — | 1 | 0.3 | 0.1 |