Text-to-Video Generation on EvalCrafter (Count, Color, Action, Others Splits)
51.27T2V Alignment (Count)VIDEOREPAIR
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| VIDEOREPAIRBase T2V Model=T2V-turbo, Refinement Framework=VIDEOREPAIR2024.11 | 51.27 | 46.66 | 45.81 | 45.45 | 47.3 | 63.2 | 57.9 | 61.8 | |
| VIDEOREPAIRBase T2V Model=Wan 2.1-1.3B, Refinement Framework=VIDEOREPAIR2024.11 | 50.03 | 51.97 | 46.01 | 48.3 | 49.01 | 65.1 | 61.6 | 62 | |
| VIDEOREPAIRBase T2V Model=VideoCrafter2, Refinement Framework=VIDEOREPAIR2024.11 | 49.84 | 51.57 | 45.78 | 48.12 | 48.83 | 62.1 | 62.4 | 62 | |
| VIDEOREPAIRBase T2V Model=CogVideoX-5B, Refinement Framework=VIDEOREPAIR2024.11 | 49.63 | 49.94 | 40.69 | 45.36 | 46.41 | 64.8 | 61.1 | 61.9 | |
| LLM paraphrasingBase T2V Model=T2V-turbo, Refinement Framework=LLM paraphrasing2024.11 | 49.49 | 43.16 | 41.32 | 44.75 | 44.68 | 62.9 | 52.9 | 61.9 | |
| OPT2IBase T2V Model=Wan 2.1-1.3B, Refinement Framework=OPT2I2024.11 | 49.1 | 51.86 | 45.88 | 47.92 | 48.69 | 64.3 | 61 | 62 | |
| OPT2IBase T2V Model=CogVideoX-5B, Refinement Framework=OPT2I2024.11 | 48.62 | 48.89 | 41.39 | 43.62 | 45.63 | 59.7 | 60.9 | 61.9 | |
| SLDBase T2V Model=Wan 2.1-1.3B, Refinement Framework=SLD2024.11 | 48.24 | 49.77 | 43.64 | 46.78 | 47.11 | 49.1 | 58.3 | 32.5 | |
| CogVideoX-5BBase T2V Model=CogVideoX-5B, Refinement Framework=None (Base)2024.11 | 47.88 | 49.63 | 37.76 | 44.78 | 45.01 | 65.8 | 61 | 61.8 | |
| SLDBase T2V Model=CogVideoX-5B, Refinement Framework=SLD2024.11 | 47.73 | 46.27 | 39.55 | 43.75 | 44.33 | 49.6 | 51.2 | 21 | |
| OPT2IBase T2V Model=VideoCrafter2, Refinement Framework=OPT2I2024.11 | 47.69 | 47.67 | 45.04 | 44.65 | 46.26 | 62.1 | 62.6 | 62.8 | |
| VideoCrafter2Base T2V Model=VideoCrafter2, Refinement Framework=None (Base)2024.11 | 47.52 | 46.28 | 44.07 | 46.02 | 45.97 | 61.8 | 62.6 | 62.9 | |
| OPT2IBase T2V Model=T2V-turbo, Refinement Framework=OPT2I2024.11 | 47.44 | 45 | 44.64 | 45.54 | 45.66 | 63.3 | 56.4 | 48.9 | |
| SLDBase T2V Model=T2V-turbo, Refinement Framework=SLD2024.11 | 47.39 | 43.99 | 42.13 | 43.28 | 44.2 | 56.6 | 58.2 | 49.2 | |
| T2V-turboBase T2V Model=T2V-turbo, Refinement Framework=None (Base)2024.11 | 46.14 | 43.06 | 41.42 | 43.16 | 43.94 | 63.3 | 57.8 | 61.6 | |
| LLM paraphrasingBase T2V Model=VideoCrafter2, Refinement Framework=LLM paraphrasing2024.11 | 45.87 | 47.81 | 44.41 | 45.16 | 45.81 | 62.4 | 62.7 | 62.7 | |
| LLM paraphrasingBase T2V Model=CogVideoX-5B, Refinement Framework=LLM paraphrasing2024.11 | 45.58 | 46.56 | 37.17 | 43.18 | 43.12 | 58.4 | 61.1 | 61.7 | |
| Wan 2.1-1.3BBase T2V Model=Wan 2.1-1.3B, Refinement Framework=None (Base)2024.11 | 45.06 | 48.18 | 41.06 | 45 | 44.83 | 63.2 | 61 | 62.1 | |
| SLDBase T2V Model=VideoCrafter2, Refinement Framework=SLD2024.11 | 44.47 | 46.45 | 39.89 | 44.06 | 43.72 | 52.5 | 62.2 | 44.4 | |
| LLM paraphrasingBase T2V Model=Wan 2.1-1.3B, Refinement Framework=LLM paraphrasing2024.11 | 44.38 | 47.19 | 42.35 | 44.03 | 44.49 | 63.5 | 61.2 | 62.2 |