Michaël Benesty – Medium

Michaël Benesty

Pinned

Published in
TDS Archive

Hugging Face Transformer Inference Under 1 Millisecond Latency

Go to production with Microsoft and Nvidia open source tooling

Nov 5, 2021

Hugging Face Transformer Inference Under 1 Millisecond Latency

Nov 5, 2021

Published in
TDS Archive

Divide Hugging Face Transformers training time by 2 or more

…with dynamic padding and uniform length batching

May 20, 2020

Divide Hugging Face Transformers training time by 2 or more

May 20, 2020

Published in
TDS Archive

NER algo benchmark: spaCy, Flair, m-BERT and camemBERT on anonymizing French commercial legal cases

Does (model) size matters?

Dec 10, 2019

NER algo benchmark: spaCy, Flair, m-BERT and camemBERT on anonymizing French commercial legal cases

Dec 10, 2019

Published in
TDS Archive

Why we switched from Spacy to Flair to anonymize French legal cases

… and why you should always review your options

Sep 26, 2019

Why we switched from Spacy to Flair to anonymize French legal cases

Sep 26, 2019

Michaël Benesty

Michaël Benesty

Head of R&D, applied ML engineer (Lefebvre Sarrut, Paris). Former: tax lawyer (Deloitte, Paris), CPA (Constantin, NYC) https://twitter.com/pommedeterre33

Following

Help
Status
About
Careers
Press
Blog
Privacy
Terms
Text to speech
Teams