# Video Question Answering on AGQA 2.0 Source: [Video Question Answering on AGQA 2.0]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:54:09.554Z Evaluation of model performance on the AGQA 2.0 dataset for Video Question Answering, reporting accuracy across compositional question types such as Object-relation, Sequencing, and Duration comparison. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Video Question Answering | | Dataset (including any stated split) | AGQA 2.0 | | Metric | Object-relation Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 15 of 15 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Object-relation Accuracy](): Higher is better; unit: PERCENT. - [Relation-action Accuracy](): Higher is better; unit: PERCENT. - [Object-action Accuracy](): Higher is better; unit: PERCENT. - [Superlative Accuracy](): Higher is better; unit: PERCENT. - [Sequencing Accuracy](): Higher is better; unit: PERCENT. - [Exists Accuracy](): Higher is better; unit: PERCENT. - [Duration Comparison Accuracy](): Higher is better; unit: PERCENT. - [Action Recognition Accuracy](): Higher is better; unit: PERCENT. - [Overall Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | Object-relation Accuracy | Result date | Report paper | | --- | --- | --- | --- | | TGB-BLIP2 | 62.27 | Not recorded | 68fa75191e209131cea65c5b | | BLIP2 | 53.72 | Not recorded | 68fa75191e209131cea65c5b | | SINGULARITY + TGB | 52.33 | Not recorded | 68fa75191e209131cea65c5b | | VIOLET + TGB | 51.59 | Not recorded | 68fa75191e209131cea65c5b | | MIST-AIO | 51.43 | Not recorded | 68fa75191e209131cea65c5b | | SeViLA | 51.15 | Not recorded | 68fa75191e209131cea65c5b | | ALBEF + TGB | 51.05 | Not recorded | 68fa75191e209131cea65c5b | | VIOLET | 50.89 | Not recorded | 68fa75191e209131cea65c5b | | SINGULARITY | 50.87 | Not recorded | 68fa75191e209131cea65c5b | | ALBEF | 50.53 | Not recorded | 68fa75191e209131cea65c5b | | ATP | 50.15 | Not recorded | 68fa75191e209131cea65c5b | | AIO | 48.34 | Not recorded | 68fa75191e209131cea65c5b | | HCRN | 40.33 | Not recorded | 68fa75191e209131cea65c5b | | PSAC | 37.84 | Not recorded | 68fa75191e209131cea65c5b | | HME | 37.42 | Not recorded | 68fa75191e209131cea65c5b | ### TGB-BLIP2 Object-relation Accuracy: 62.27. Result ID: 69938dec3a6c31484916d13e. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Category":"Open-ended Video-Language Models","Augmented":"Temporal Grounding Bridge"}. This record does not assert independent reproduction. ### BLIP2 Object-relation Accuracy: 53.72. Result ID: 69938dec3a6c31484916d13d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Category":"Open-ended Video-Language Models"}. This record does not assert independent reproduction. ### SINGULARITY + TGB Object-relation Accuracy: 52.33. Result ID: 69938dec3a6c31484916d139. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Category":"Retrieval-based Video-Language Models","Augmented":"Temporal Grounding Bridge"}. This record does not assert independent reproduction. ### VIOLET + TGB Object-relation Accuracy: 51.59. Result ID: 69938dec3a6c31484916d13b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Category":"Retrieval-based Video-Language Models","Augmented":"Temporal Grounding Bridge"}. This record does not assert independent reproduction. ### MIST-AIO Object-relation Accuracy: 51.43. Result ID: 69938dec3a6c31484916d135. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Category":"Retrieval-based Video-Language Models"}. This record does not assert independent reproduction. ### SeViLA Object-relation Accuracy: 51.15. Result ID: 69938dec3a6c31484916d13c. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Category":"Open-ended Video-Language Models","Status":"Re-implementation result"}. This record does not assert independent reproduction. ### ALBEF + TGB Object-relation Accuracy: 51.05. Result ID: 69938dec3a6c31484916d137. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Category":"Retrieval-based Video-Language Models","Augmented":"Temporal Grounding Bridge"}. This record does not assert independent reproduction. ### VIOLET Object-relation Accuracy: 50.89. Result ID: 69938dec3a6c31484916d13a. [Reporting paper]() Recorded evaluation configuration: {"Model Category":"Retrieval-based Video-Language Models"}. This record does not assert independent reproduction. ### SINGULARITY Object-relation Accuracy: 50.87. Result ID: 69938dec3a6c31484916d138. [Reporting paper]() Recorded evaluation configuration: {"Model Category":"Retrieval-based Video-Language Models"}. This record does not assert independent reproduction. ### ALBEF Object-relation Accuracy: 50.53. Result ID: 69938dec3a6c31484916d136. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Category":"Retrieval-based Video-Language Models"}. This record does not assert independent reproduction. ### ATP Object-relation Accuracy: 50.15. Result ID: 69938dec3a6c31484916d134. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Category":"Retrieval-based Video-Language Models"}. This record does not assert independent reproduction. ### AIO Object-relation Accuracy: 48.34. Result ID: 69938dec3a6c31484916d133. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Category":"Retrieval-based Video-Language Models"}. This record does not assert independent reproduction. ### HCRN Object-relation Accuracy: 40.33. Result ID: 69938dec3a6c31484916d132. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Category":"Retrieval-based Video-Language Models"}. This record does not assert independent reproduction. ### PSAC Object-relation Accuracy: 37.84. Result ID: 69938dec3a6c31484916d131. [Reporting paper]() Recorded evaluation configuration: {"Model Category":"Retrieval-based Video-Language Models"}. This record does not assert independent reproduction. ### HME Object-relation Accuracy: 37.42. Result ID: 69938dec3a6c31484916d130. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Category":"Retrieval-based Video-Language Models"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.