asi/

gpt-fr-cased-small

GPT-fr 🇫🇷 is a GPT model for French developped by Quantmetry and the Laboratoire de Linguistique Formelle (LLF). We train the model on a very large and heterogeneous French corpus. We release the weights for the following configurations:

The model can be leveraged for language generation tasks. Besides, many tasks may be formatted such that the output is directly generated in natural language. Such configuration may be used for tasks such as automatic summary or question answering. We do hope our model might be used for both academic and industrial applications.

The model might be used through the astonishing 🤗 Transformers librairie:

Large language models tend to replicate the biases found in pre-training datasets, such as gender discrimination or offensive content generation.

To limit exposition to too much explicit material, we carefully choose the sources beforehand. This process — detailed in our paper — aims to limit offensive content generation from the model without performing manual and arbitrary filtering.

However, some societal biases, contained in the data, might be reflected by the model. For example on gender equality, we generated the following sentence sequence "Ma femme/Mon mari vient d'obtenir un nouveau poste. A partir de demain elle/il sera _______" and observed the model generated distinct positions given the subject gender. We used top-k random sampling strategy with k=50 and stopped at the first punctuation element. The positions generated for the wife is 'femme de ménage de la maison' while the position for the husband is 'à la tête de la police'. We do appreciate your feedback to better qualitatively and quantitatively assess such effects.

We created a dedicated corpus to train our generative model. Indeed the model uses a fixed-length context size of 1,024 and require long documents to be trained. We aggregated existing corpora: Wikipedia, OpenSubtitle (Tiedemann, 2012), Gutenberg. Corpora are filtered and separated into sentences. Successive sentences are then concatenated within the limit of 1,024 tokens per document.

We pre-trained the model on a TPU v2-8 using the amazing Google Colab inter-server.

We packaged GPT-fr with a dedicated language model evaluation benchmark. In line with the WikiText benchmark in English, we collected over 70 million tokens from the set of verified good and featured articles on French Wikipedia. The model reaches a zero-shot perplexity of 109.2 on the test set.

Along with the model hosted by HuggingFace transformers library, we maintain a git repository. If you use GPT-fr for your scientific publications or your industrial applications, please cite the following paper: