Model Card for answer-finder-v1-L-multilingual

This model is a question answering model developed by Sinequa. It produces two lists of logit scores corresponding to the start token and end token of an answer.

Model name: answer-finder-v1-L-multilingual

Supported Languages

The model was trained and tested in the following languages:

• English
• French
• German
• Spanish

Scores

Metric	Value
F1 Score on SQuAD v2 EN with Hugging Face evaluation pipeline	75
F1 Score on SQuAD v2 EN with Haystack evaluation pipeline	75
F1 Score on SQuAD v2 FR with Haystack evaluation pipeline	73.4
F1 Score on SQuAD v2 DE with Haystack evaluation pipeline	90.8
F1 Score on SQuAD v2 ES with Haystack evaluation pipeline	67.1

Inference Time

GPU	Quantization type	Batch size 1	Batch size 32
NVIDIA A10	FP16	2 ms	30 ms
NVIDIA A10	FP32	4 ms	83 ms
NVIDIA T4	FP16	3 ms	65 ms
NVIDIA T4	FP32	14 ms	373 ms
NVIDIA L4	FP16	2 ms	38 ms
NVIDIA L4	FP32	5 ms	124 ms

Note that the Answer Finder models are only used at query time.

Gpu Memory usage

Quantization type	Memory
FP16	550 MiB
FP32	1050 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: 110 million
• Base language model: bert-base-multilingual-cased pre-trained by Sinequa in English, French, German and Spanish
• Insensitive to casing and accents

Training Data

• SQuAD v2
• French-SQuAD + French translation of SQuAD v2 "impossible" query-passage pairs
• GermanQuAD + German translation of SQuAD v2 "impossible" query-passage pairs
• SQuAD-es-v2