andaba/TEMPURA-VER
TEMPURA-VER: Video Event Reasoning data for TEMPURA VER (Video Event Reasoning) is the training corpus of TEMPURA: Temporal Event Masked Prediction and Understanding for Reasoning in Action (COLM 2026). It pairs YouTube videos with temporally aligned, non-overlapping event descriptions and with step-by-step causal reasoning about masked events. TEMPURA uses it in a two-stage recipe: Masked Event Prediction (MEP) – the model sees a video with one segment masked out and must… See the full description on the dataset page: https://huggingface.co/datasets/andaba/TEMPURA-VER.
| Author / organization | andaba |
|---|---|
| License (as declared) | cc-by-4.0 |
| Downloads | 166 (as reported by Hugging Face) |
| Likes | 1 |
| Created | 1 May 2025 |
| Last updated at source | 24 September 2026 |
Popularity, last 3 days
166 now (+11 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
Connected in the knowledge graph
- mradermacher/TEMPURA-Qwen2.5-VL-3B-s2-GGUFModel · used to train
- mradermacher/TEMPURA-Qwen2.5-VL-7B-GGUFModel · used to train
- mradermacher/TEMPURA-Qwen2.5-VL-3B-GGUFModel · used to train
- mradermacher/TEMPURA-Qwen2.5-VL-3B-s1-GGUFModel · used to train
More from andaba
- andaba/RT-PoseDataset
- andaba/CRUW3DDataset
- andaba/so100_draw_curves_0Dataset
- andaba/so100_draw_curves_1Dataset
- andaba/so100_test2Dataset
- andaba/so100_test3Dataset
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.