| | --- |
| | license: cc-by-sa-3.0 |
| | datasets: |
| | - kilt_tasks |
| | metrics: |
| | - exact_match |
| | pipeline_tag: text-generation |
| | model-index: |
| | - name: results |
| | results: |
| | - task: |
| | name: Question Answering |
| | type: text-generation |
| | dataset: |
| | name: NQ KILT |
| | type: kilt_tasks |
| | args: nq |
| | metrics: |
| | - name: Exact Macth |
| | type: exact_match |
| | value: 51.55 |
| | language: |
| | - en |
| | --- |
| | |
| | # Fusion-In-Decoder Base on Natural Questions |
| |
|
| | This trained model is based on the [Fusion-In-Decoder](https://arxiv.org/abs/2007.01282) model, and trained on the [Natural Questions](https://huggingface.co/datasets/natural_questions) dataset. |
| |
|
| | # Model Details |
| |
|
| | Model is based on Fusion-In-Decoder, which in turn is based on the `google/flan-t5-base` checkpoint as the base model. For training, we utilized text retrieval for each query, which provides a collection of relevant passages for it. |
| |
|
| | We note that the passages were retrieved using a corpus based on [Wikipedia](https://huggingface.co/datasets/wiki_dpr). |
| |
|
| | # Evaluation |
| |
|
| | See model performance on Evaluation Results tab on the right side. |
