README.md

metadata

tags:
  - pyterrier
  - pyterrier-artifact
  - pyterrier-artifact.indexer_cache
  - pyterrier-artifact.indexer_cache.lz4pickle
task_categories:
  - text-retrieval
viewer: false

SPLADE vector cache for Vaswani

Description

A SPLADE vector cache for the Vaswani dataset.

Usage

>>> # Load the artifact
>>> from pyterrier_caching import IndexerCache
>>> cache = IndexerCache.from_hf('macavaney/vaswani.splade.cache')
>>> len(cache)
11429

Benchmarks

TODO: Provide benchmarks for the artifact.

Reproduction

import pyterrier as pt
pt.init()
from pyterrier_caching import IndexerCache
from pyt_splade import Splade
splade = Splade()
cache = IndexerCache('vaswani.splade.cache')
dataset = pt.get_dataset('irds:vaswani')
pipeline = splade.doc_encoder() >> cache.indexer()
pipeline.index(dataset.get_corpus_iter())

Metadata

{
  "type": "indexer_cache",
  "format": "lz4pickle",
  "record_count": 11429
}