Retrieval Pipeline Experiment -- 2
Budget: $750 – $1,100 USD
The experimental protocol aims to measure efficiency metrics in the indexing, training, and
inference phases of neural information retrieval models based on transformers. All models are
for the passage ranking task using the MS MARCO Passage Ranking dataset. All models have in
common efficiency characteristics in their proposals and implementation in the Pytorch
framework. Below is the identification of each model and its role in the retrieve pipeline.
inference phases of neural information retrieval models based on transformers. All models are
for the passage ranking task using the MS MARCO Passage Ranking dataset. All models have in
common efficiency characteristics in their proposals and implementation in the Pytorch
framework. Below is the identification of each model and its role in the retrieve pipeline.