SPICE

Selective Pretraining for Informed Context Extraction (SPICE)- a small but powerful retriever.

The future is a thing that is not given. It is a thing that is made." - Leto II (Dune)

Using the model

from sentence_transformers import SentenceTransformer

model = SentenceTransformer("iamgroot42/spice")
sentences = [
  "Beware of manuals. Manuals create habits.",
  "How easy it is to mistake clear reasoning from correct reasoning!",
  "Show me a completely smooth operation and I'll show you someone who's covering mistakes. Real boats rock."
]

embeddings = model.encode(sentences)
similarities = model.similarity(embeddings, embeddings)

Model training details and data will be uploaded soon!

Downloads last month: 659

Safetensors

Model size

33.4M params

Tensor type

F32

Inference Providers NEW

Feature Extraction

This model is not currently available via any of the supported third-party Inference Providers, and the model is not deployed on the HF Inference API.

Evaluation results

ndcg_at_1 on MTEB ArguAna (default)
test set self-reported

49.075
ndcg_at_3 on MTEB ArguAna (default)
test set self-reported

66.383
ndcg_at_5 on MTEB ArguAna (default)
test set self-reported

70.342
ndcg_at_10 on MTEB ArguAna (default)
test set self-reported

72.962
ndcg_at_20 on MTEB ArguAna (default)
test set self-reported

73.723
ndcg_at_100 on MTEB ArguAna (default)
test set self-reported

73.826
ndcg_at_1000 on MTEB ArguAna (default)
test set self-reported

73.826
map_at_1 on MTEB ArguAna (default)
test set self-reported

49.075
map_at_3 on MTEB ArguAna (default)
test set self-reported

62.138
map_at_5 on MTEB ArguAna (default)
test set self-reported

64.354

View on Papers With Code