Model Card for passage-ranker.chocolate

This model is a passage ranker developed by Sinequa. It produces a relevance score given a query-passage pair and is used to order search results.

Model name: passage-ranker.chocolate

Supported Languages

The model was trained and tested in the following languages:

• English

Scores

Metric	Value
Relevance (NDCG@10)	0.484

Note that the relevance score is computed as an average over 14 retrieval datasets (see details below).

Inference Times

GPU	Quantization type	Batch size 1	Batch size 32
NVIDIA A10	FP16	1 ms	5 ms
NVIDIA A10	FP32	2 ms	22 ms
NVIDIA T4	FP16	1 ms	13 ms
NVIDIA T4	FP32	3 ms	66 ms
NVIDIA L4	FP16	2 ms	6 ms
NVIDIA L4	FP32	3 ms	30 ms

Gpu Memory usage

Quantization type	Memory
FP16	300 MiB
FP32	550 MiB

Note that GPU memory usage only includes how much GPU memory the actual model consumes on an NVIDIA T4 GPU with a batch size of 32. It does not include the fix amount of memory that is consumed by the ONNX Runtime upon initialization which can be around 0.5 to 1 GiB depending on the used GPU.

Requirements

• Minimal Sinequa version: 11.10.0
• Minimal Sinequa version for using FP16 models and GPUs with CUDA compute capability of 8.9+ (like NVIDIA L4): 11.11.0
• Cuda compute capability: above 5.0 (above 6.0 for FP16 use)

Model Details

Overview

• Number of parameters: 23 million
• Base language model: MiniLM-L6-H384-uncased (Paper, GitHub)
• Insensitive to casing and accents
• Training procedure: MonoBERT

Training Data

• MS MARCO Passage Ranking (Paper, Official Page, dataset on HF hub)

Evaluation Metrics

To determine the relevance score, we averaged the results that we obtained when evaluating on the datasets of the BEIR benchmark. Note that all these datasets are in English.

Dataset	NDCG@10
Average	0.486
Arguana	0.554
CLIMATE-FEVER	0.209
DBPedia Entity	0.367
FEVER	0.744
FiQA-2018	0.339
HotpotQA	0.685
MS MARCO	0.412
NFCorpus	0.352
NQ	0.454
Quora	0.818
SCIDOCS	0.158
SciFact	0.658
TREC-COVID	0.674
Webis-Touche-2020	0.345