|
---
|
|
pipeline_tag: sentence-similarity
|
|
tags:
|
|
- feature-extraction
|
|
- sentence-similarity
|
|
language:
|
|
- de
|
|
- en
|
|
- es
|
|
- fr
|
|
- it
|
|
- nl
|
|
- ja
|
|
- pt
|
|
- zh
|
|
---
|
|
|
|
# Model Card for `vectorizer.raspberry`
|
|
|
|
This model is a vectorizer developed by Sinequa. It produces an embedding vector given a passage or a query. The passage vectors are stored in our vector index and the query vector is used at query time to look up relevant passages in the index.
|
|
|
|
Model name: `vectorizer.raspberry`
|
|
|
|
## Supported Languages
|
|
|
|
The model was trained and tested in the following languages:
|
|
|
|
- English
|
|
- French
|
|
- German
|
|
- Spanish
|
|
- Italian
|
|
- Dutch
|
|
- Japanese
|
|
- Portuguese
|
|
- Chinese (simplified)
|
|
|
|
Besides these languages, basic support can be expected for additional 91 languages that were used during the pretraining of the base model (see Appendix A of XLM-R paper).
|
|
|
|
## Scores
|
|
|
|
| Metric | Value |
|
|
|:-----------------------|------:|
|
|
| Relevance (Recall@100) | 0.613 |
|
|
|
|
Note that the relevance score is computed as an average over 14 retrieval datasets (see
|
|
[details below](#evaluation-metrics)).
|
|
|
|
## Inference Times
|
|
|
|
| GPU | Quantization type | Batch size 1 | Batch size 32 |
|
|
|:------------------------------------------|:------------------|---------------:|---------------:|
|
|
| NVIDIA A10 | FP16 | 1 ms | 5 ms |
|
|
| NVIDIA A10 | FP32 | 2 ms | 18 ms |
|
|
| NVIDIA T4 | FP16 | 1 ms | 12 ms |
|
|
| NVIDIA T4 | FP32 | 3 ms | 52 ms |
|
|
| NVIDIA L4 | FP16 | 2 ms | 5 ms |
|
|
| NVIDIA L4 | FP32 | 4 ms | 24 ms |
|
|
|
|
## 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](https://developer.nvidia.com/cuda-gpus): above 5.0 (above 6.0 for FP16 use)
|
|
|
|
## Model Details
|
|
|
|
### Overview
|
|
|
|
- Number of parameters: 107 million
|
|
- Base language
|
|
model: [mMiniLMv2-L6-H384-distilled-from-XLMR-Large](https://huggingface.co/nreimers/mMiniLMv2-L6-H384-distilled-from-XLMR-Large) ([Paper](https://arxiv.org/abs/2012.15828), [GitHub](https://github.com/microsoft/unilm/tree/master/minilm))
|
|
- Insensitive to casing and accents
|
|
- Output dimensions: 256 (reduced with an additional dense layer)
|
|
- Training procedure: Query-passage-negative triplets for datasets that have mined hard negative data, Query-passage
|
|
pairs for the rest. Number of negatives is augmented with in-batch negative strategy
|
|
|
|
### Training Data
|
|
|
|
The model have been trained using all datasets that are cited in the [all-MiniLM-L6-v2](https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2) model. In addition to that, this model has been trained on the datasets cited in [this paper](https://arxiv.org/pdf/2108.13897.pdf) on the 9 aforementioned languages.
|
|
|
|
### Evaluation Metrics
|
|
|
|
To determine the relevance score, we averaged the results that we obtained when evaluating on the datasets of the
|
|
[BEIR benchmark](https://github.com/beir-cellar/beir). Note that all these datasets are in English.
|
|
|
|
| Dataset | Recall@100 |
|
|
|:------------------|-----------:|
|
|
| Average | 0.613 |
|
|
| | |
|
|
| Arguana | 0.957 |
|
|
| CLIMATE-FEVER | 0.468 |
|
|
| DBPedia Entity | 0.377 |
|
|
| FEVER | 0.820 |
|
|
| FiQA-2018 | 0.639 |
|
|
| HotpotQA | 0.560 |
|
|
| MS MARCO | 0.845 |
|
|
| NFCorpus | 0.287 |
|
|
| NQ | 0.756 |
|
|
| Quora | 0.992 |
|
|
| SCIDOCS | 0.456 |
|
|
| SciFact | 0.906 |
|
|
| TREC-COVID | 0.100 |
|
|
| Webis-Touche-2020 | 0.413 |
|
|
|
|
We evaluated the model on the datasets of the [MIRACL benchmark](https://github.com/project-miracl/miracl) to test its multilingual capacities. Note that not all training languages are part of the benchmark, so we only report the metrics for the existing languages.
|
|
|
|
| Language | Recall@100 |
|
|
|:----------------------|-----------:|
|
|
| French | 0.650 |
|
|
| German | 0.528 |
|
|
| Spanish | 0.602 |
|
|
| Japanese | 0.614 |
|
|
| Chinese (simplified) | 0.680 |
|
|
|
