This commit is contained in:
Per Stark
2025-12-08 20:39:12 +01:00
parent 1fccf3ab59
commit e77f2d51e8
13 changed files with 405 additions and 160 deletions
+32 -2
View File
@@ -20,7 +20,7 @@ datasets:
category: "Natural Questions"
entity_suffix: "Natural Questions"
source_prefix: "nq"
raw: "data/raw/nq/dev-all.jsonl"
raw: "data/raw/nq-dev/dev-all.jsonl"
converted: "data/converted/nq-dev-minne.json"
include_unanswerable: true
slices:
@@ -42,7 +42,7 @@ datasets:
slices:
- id: beir-mix-600
label: "BEIR mix (600)"
description: "Balanced slice across FEVER, FiQA, HotpotQA, NFCorpus, Quora, TREC-COVID"
description: "Balanced slice across FEVER, FiQA, HotpotQA, NFCorpus, Quora, TREC-COVID, SciFact, NQ-BEIR"
limit: 600
corpus_limit: 6000
seed: 0x5eed2025
@@ -136,3 +136,33 @@ datasets:
limit: 200
corpus_limit: 5000
seed: 0x5eed2025
- id: scifact
label: "SciFact (BEIR)"
category: "SciFact"
entity_suffix: "SciFact"
source_prefix: "scifact"
raw: "data/raw/scifact"
converted: "data/converted/scifact-minne.json"
include_unanswerable: false
slices:
- id: scifact-test-200
label: "SciFact test (200)"
description: "200-case slice from BEIR test qrels"
limit: 200
corpus_limit: 3000
seed: 0x5eed2025
- id: nq-beir
label: "Natural Questions (BEIR)"
category: "Natural Questions"
entity_suffix: "Natural Questions"
source_prefix: "nq-beir"
raw: "data/raw/nq"
converted: "data/converted/nq-beir-minne.json"
include_unanswerable: false
slices:
- id: nq-beir-test-200
label: "NQ (BEIR) test (200)"
description: "200-case slice from BEIR test qrels"
limit: 200
corpus_limit: 5000
seed: 0x5eed2025