-
Notifications
You must be signed in to change notification settings - Fork 2.2k
French translation
If "Translation status" indicates WIP, the translator is still working on his translation. Even if a first version is available online. The translator may still have some things to adjust, such as the presentation of formulas or images, or does a proofreading of his work (it is common to make mistakes when writing in Markdown and you can't see what it does without have put a first version online). It is therefore useless to do PR to correct the work since it is still in progress. It is only once the status thinking of "OK" that a reviewer can start the revision.
Similarly, if the "Revision Status" says "WIP", the reviewer is still working. So there is no point in doing PR. A translation is considered finished once the "Revision Status" indicates "OK".
File name | Translator | Translation statut | Reviewer | Review statut |
---|---|---|---|---|
README.md |
lbourdois | OK | lbourdois | OK |
index.md |
lbourdois | OK | lbourdois | OK |
01.md |
lbourdois | OK | lbourdois | OK |
01-1.md |
lbourdois | OK | lbourdois | OK |
01-2.md |
lbourdois | OK | lbourdois | OK |
01-3.md |
lbourdois | OK | lbourdois | OK |
02.md |
lbourdois | OK | lbourdois | OK |
02-1.md |
lbourdois | OK | lbourdois | OK |
02-2.md |
lbourdois | OK | lbourdois | OK |
02-3.md |
lbourdois | OK | lbourdois | OK |
03.md |
lbourdois | OK | lbourdois | OK |
03-1.md |
lbourdois | OK | lbourdois | OK |
03-2.md |
lbourdois | OK | lbourdois | OK |
03-3.md |
lbourdois | OK | lbourdois | OK |
04.md |
lbourdois | OK | lbourdois | OK |
04-1.md |
lbourdois | OK | lbourdois | OK |
05.md |
lbourdois | OK | lbourdois | OK |
05-1.md |
lbourdois | OK | lbourdois | OK |
05-2.md |
lbourdois | OK | lbourdois | OK |
05-3.md |
lbourdois | OK | lbourdois | OK |
06.md |
lbourdois | OK | lbourdois | OK |
06-1.md |
lbourdois | OK | lbourdois | OK |
06-2.md |
lbourdois | OK | lbourdois | OK |
06-3.md |
lbourdois | OK | lbourdois | OK |
07.md |
lbourdois | OK | lbourdois | OK |
07-1.md |
lbourdois | OK | lbourdois | OK |
07-2.md |
lbourdois | OK | lbourdois | OK |
07-3.md |
lbourdois | OK | lbourdois | OK |
08.md |
lbourdois | OK | lbourdois | OK |
08-1.md |
lbourdois | OK | lbourdois | OK |
08-2.md |
lbourdois | OK | lbourdois | OK |
08-3.md |
lbourdois | OK | lbourdois | OK |
09.md |
lbourdois | OK | lbourdois | OK |
09-1.md |
lbourdois | OK | lbourdois | OK |
09-2.md |
lbourdois | OK | lbourdois | OK |
09-3.md |
lbourdois | OK | lbourdois | OK |
10.md |
lbourdois | OK | lbourdois | OK |
10-1.md |
lbourdois | OK | lbourdois | OK |
10-2.md |
lbourdois | OK | lbourdois | OK |
10-3.md |
lbourdois | OK | lbourdois | OK |
11.md |
lbourdois | OK | lbourdois | OK |
11-1.md |
lbourdois | OK | lbourdois | OK |
11-2.md |
lbourdois | OK | lbourdois | OK |
11-3.md |
lbourdois | OK | lbourdois | OK |
12.md |
lbourdois | OK | lbourdois | OK |
12-1.md |
lbourdois | OK | lbourdois | OK |
12-2.md |
lbourdois | OK | lbourdois | OK |
12-3.md |
lbourdois | OK | lbourdois | OK |
13.md |
lbourdois | OK | lbourdois | OK |
13-1.md |
lbourdois | OK | lbourdois | OK |
13-2.md |
lbourdois | OK | lbourdois | OK |
13-3.md |
lbourdois | OK | lbourdois | OK |
14.md |
lbourdois | OK | lbourdois | OK |
14-1.md |
lbourdois | OK | lbourdois | OK |
14-2.md |
lbourdois | OK | lbourdois | OK |
14-3.md |
lbourdois | OK | lbourdois | OK |
15.md |
lbourdois | OK | lbourdois | OK |
15-1.md |
lbourdois | OK | lbourdois | OK |
15-2.md |
lbourdois | OK | lbourdois | OK |
lecture01.sbv |
lbourdois | OK | ||
practicum01.sbv |
lbourdois | OK | ||
lecture02.sbv |
lbourdois | OK | ||
practicum02.sbv |
lbourdois | OK | ||
lecture03.sbv |
lbourdois | OK | ||
practicum03.sbv |
lbourdois | OK | ||
practicum04.sbv |
lbourdois | OK | ||
lecture05.sbv |
lbourdois | OK | ||
practicum05.sbv |
lbourdois | OK | ||
lecture06.sbv |
lbourdois | OK | ||
practicum06.sbv |
lbourdois | OK | ||
lecture07.sbv |
lbourdois | OK | ||
practicum07.sbv |
lbourdois | OK | ||
lecture08.sbv |
lbourdois | OK | ||
practicum08.sbv |
lbourdois | OK | ||
lecture09.sbv |
lbourdois | OK | ||
practicum09.sbv |
lbourdois | OK | ||
lecture10.sbv |
lbourdois | OK | ||
practicum10.sbv |
lbourdois | OK | ||
lecture11.sbv |
lbourdois | OK | ||
practicum11.sbv |
lbourdois | OK | ||
lecture12.sbv |
lbourdois | OK | ||
practicum12.sbv |
lbourdois | OK | ||
lecture13.sbv |
lbourdois | OK | ||
practicum13.sbv |
lbourdois | OK | ||
lecture14.sbv |
lbourdois | OK | ||
practicum14.sbv |
lbourdois | OK | ||
practicum15A.sbv |
lbourdois | OK | ||
practicum15B.sbv |
lbourdois | OK |
In the table above, you will find the list of files for which the French translation is currently missing. Or which need a review.
The .sbv
files correspond to the translation of the subtitles of the course videos.
As an indication, it takes at least 4 hours of work for a 1h30 video (this can be more for technical videos where mathematical formulas are developed for example). Keep this in mind when registering !
Attention! English subtitles have been artificially generated. They may therefore contain errors. A manual verification of the English texts is in progress (see the Caption Videos Wiki or the Slack for more information). Before translating, make sure that the .sbv
file in question has been manually checked!
The .md
files correspond to the translation of the lecture notes taken by the students. It takes about 1 hour of work for one file (except for those presenting the content of the week's courses, which are much smaller). Attention! To preserve the homogeneity of the translation, it is requested that for a given file, the translator commits to do all the others of the week (so 4 files). The working time is then counted between 3 and 4 hours.
Here is the list of choices that have been retained to carry out the translation. It is the result of a discussion with Yann Le Cun.
The texts found on the website are course notes from students who attended Yann Le Cun's classes (and from the four guest speakers). For each course, several students were asked to put their own notes to be added to the site. Thus the notes in English came from about 130 different people, who did not work together to unify their text. Some people write in the past tense, while others write in the present tense. The abbreviations used are not necessarily always the same, people are not always designated in the same way, etc. The variety of authors also means that the way of writing varies from one file to another.
Thus the choices that have been made for French are as follows:
- use of the present tense in all the files (if the course is about a past event, as for example when Yann tells episodes about the history of artificial intelligence dating back to the 40's, the past tense is obviously used)
- Some students tend to write very long sentences of 5 or 6 lines. In order to keep the text from ending in Proust and to keep it understandable, the sentences have been simplified as much as possible (three sentences of 2 lines or instead of one of 6), and repetitions in the text have been removed.
- Concerning the technical vocabulary, it is detailed in the following sub-section.
Videos are not translated word for word. Indeed, in order to fluidify the translation and the comprehension, it was decided not to transcribe the "parasitic" filling and transition words ("you know", "sort of", "right", "so", etc.).
The translations that have been selected for the following list of words are the result of a discussion with Yann Le Cun :
Term | Translation | Reason |
---|---|---|
CNN | ConvNet | Yann is particularly keen to respect this translation |
Convolutional sparse auto-encoder | Auto-encodeur épars convolutif | Yann's choice |
Contrastive | Constrative au féminin, Contrastif au masculin | Yann's choice |
Dropout | Dropout | Yann thought of the term "decimation" but it is already used in signal processing to mean "subsampling". |
Energy-Based Models | Modèles à base d’énergie | Yann on the subject: "Not very satisfying, but I don't have better! " |
Energy based factor graphs | Graphes de facteurs à base d'énergie | Yann's choice |
Embedding | Enchâssement | Used by Yann in his book |
Gated Graph ConvNets | Réseau à rupteur convolutif sur graphe | Yann's choice (it is an invention on his part) |
Graph Attention Networks | Réseau à graphe d'attention | Yann's choice |
Graph Convolution Networks | Réseau convolutif sur graphe (abréviation : GCN) | Yann's choice |
Graph Domain | Domaine du graphe (comme le domaine de Fourier) | Yann's choice |
Group Sparsity | Eparsité de groupe | Yann's choice |
One hot | One hot | The notion of "canonical base vectors" could be used but seems a bit technical and the expression is rather long to translate barely 2 words. The choice was therefore made to let the English expression |
Overfitting | Surentraînement | Yann's choice |
Pooling | Pooling | Translations into French as "aggrégation", "agglomération", were not satisfactory |
Residual Gated Graph Convolutional Network | Réseau à rupteur convolutif à résidu sur graphe | Yann's choice |
Simple Energy-Based Factor Graphs with “Shallow” Factors | Graphes de facteur simples à base d'énergie avec facteurs peu profonds | Yann's choice |
Sparse | Epars | Yann's choice |
Sparsity | Eparsité | Yann's choice |
Spectral graph convolutional neural networks | Réseau convolutif spectral sur graphe | Yann's choice |
SplineGCNs / LapGCNs / CayleyNets / ChebNets / MoNets / GraphSage | SplineGCNs / LapGCNs / CayleyNets / ChebNets / MoNets / GraphSage | Easier to keep the English abbreviations according to Yann |
Vanilla Spatial GCNs | Graphe ConvNet spatial ordinaire | Yann's choice |
Yann LeCun | Yann Le Cun ou Yann | For the writing of the family name, the Breton form is adopted. Yann explains in his book why in Anglo-Saxon countries he is obliged to write his surname all attached. In the notes of the students in English, it is possible to find "Mr Yann LeCun", "Mr LeCun", "Doctor Yann LeCun", "Professor LeCun", etc. Yann wishes that in these texts, he is simply called Yann. |