AI models
CamemBERT
Training compute
8.3×10²⁰ FLOP
Parameters
335M
Published
Nov 10, 2019
CamemBERT is an AI model developed by Facebook, INRIA and Sorbonne University (United States and France), first published in November 2019. It works in the language domain, on tasks such as language modeling/generation, part-of-speech tagging and named entity recognition (ner).
Training it took an estimated 8.3×10²⁰ FLOP of compute (estimation method: hardware,operation counting). The model has 335,000,000 parameters. It was trained on roughly 28.6B datapoints. Training ran on NVIDIA V100 for about 24 hours. The compute alone is estimated at $2K in 2023 dollars.
Access: Open weights (unrestricted). Its weights are openly available. The reference paper has 1,083 citations. Epoch AI rates the confidence of this record as confident.
Full record
- Organization
- Facebook, INRIA, Sorbonne University
- Country of organization
- United States, France
- Domain
- Language
- Task
- Language modeling/generation, Part-of-speech tagging, Named entity recognition (NER)
- Training compute
- 8.3×10²⁰ FLOP
- Compute estimation method
- Hardware, Operation counting
- Parameters
- 335,000,000
- Dataset size
- 28.6B
- Training hardware
- NVIDIA V100
- Training time
- 24 h
- Training cost (2023 USD)
- $2K
- Model accessibility
- Open weights (unrestricted)
- Open weights
- Yes
- Citations
- 1,083
- Epoch confidence
- Confident
SourceEpoch AI, 'AI Models'. Published online at epoch.ai. Retrieved 2026-07-29 from https://epoch.ai/data/ai-models. Licensed under CC BY 4.0.
← All ai models