Lucas Willems lcswillems Paris, France

RL starter files in order to immediatly train, visualize and evaluate an agent without writing any line of code

A simple and funny Python Brick Breaker with an easy level-maker

A Python handwritten digit classifier trained on MNIST with a UI for using it

Tests méchants pour le projet de compil'

Implementation of the SVM Convex and Combinatorial Optimization Courses of the ENS.

The report and slides of my presentation on the minimization of automatas

An interface to test a game of "coins on a chessboard" invented by Maxim Kontsevich

A simple set of Python classes to create statistics for CPGE classes.

startedtinacms/tinacms

started time in 5 days

startedapollographql/apollo-client

started time in 13 days

issue commentlcswillems/rl-starter-files

Thank you for the details!

So if it is not working without --memory, the issue seems related to CUDA. It will be hard for me to find where the issue comes from because I can't access a CUDA anymore.

Could you please give me the sequence of commands without --memory that leads to the error? (training et evaluation) Are they both executing on CUDA?

Thanks!

lpsantao

comment created time in 18 days

issue commentlcswillems/rl-starter-files

Hi Lili,

How did you train the network? Which command? On CUDA? Where did you evaluate it? On CUDA?

Lucas

lpsantao

comment created time in 18 days

delete branch : patch-3

delete time in a month

They should not be required because a bearer_token can be sufficient.

+2 -2

0 comment

1 changed file

pr created time in a month

commit sha 4074c44ba8127c8be1cc19886c004cb41642cfbc

Make consumer_key and consumer_secret optional They should not be required because a bearer_token can be sufficient.

push time in a month

commit sha 08b4410e65e238d93a6dc614dff3aef18fae8a13

Make consumer_key and consumer_secret optional They should not be required because a bearer_token can be sufficient.

push time in a month

issue commentmicrosoft/TypeScript-Website

I don't see the link with the issue. I gave you an example where b is of type Record<Page, PageInfo> but is not a set of properties Page of type PageInfo.

I think the issue should be reopened.

lcswillems

comment created time in a month

issue openedmicrosoft/TypeScript-Website

Page URL: https://www.typescriptlang.org/docs/handbook/utility-types.html#recordkt

Issue: It is written: "Constructs a type with a set of properties K of type T." which is wrong.

Because this code is correct TypeScript:

interface PageInfo {
title: string;
}

type Page = "home" | "about" | "contact";

const a = {
contact: { title: "contact" },
home: { title: "home" },
test: { title: "test" }
}

const b: Record<Page, PageInfo> = a;


However, b is not a set of properties K of type T.

Recommended Fix:

I don't know.

created time in a month

pull request commentlcswillems/python-MNIST-classifier

In this case, the packages installed by Tensorflow should not be added to the requirements.txt.

fotisk07

comment created time in a month

issue closedstyled-components/styled-components

## Reproduction

Click on "Miple" on the top left. The page is completely broken:

If the page is reloaded, it displays correctly:

## Environment

I use styled-components v5 with Next.js 9.3. The bug was not present in v4

closed time in 2 months

lcswillems

I close this issue as I don't have this issue anymore (I did nothing special except updating my code...). I will open a new issue if it happens again and have a real reproduction.

lcswillems

comment created time in 2 months

issue commentianstormtaylor/slate

And I am not sure that generics are the good way to do it too because, from my understanding, we will still have to provide generics at every call. Declaration merging might be easier to implement and easier to use.

exports

comment created time in 2 months

issue commentianstormtaylor/slate

How can you observe push back?

exports

comment created time in 2 months

issue commentianstormtaylor/slate

I agree with @thesunny . Currently, having to cast all the output of functions is very painful, and may prevent new users to adopt Slate. It would be better if it was [key: string]: any or [key: string]: unknown but with declaration merging.

exports

comment created time in 2 months

issue commentianstormtaylor/slate

@majelbstoat I am pretty sure this issue is because of this PR https://github.com/ianstormtaylor/slate/pull/3620 (c.f. the comments I and people made to the PR)

evasteingrims

comment created time in 2 months

issue commentKaTeX/KaTeX

Indeed, I get this error. But it is okay for me for the moment

lcswillems

comment created time in 2 months

issue commentKaTeX/KaTeX

I found how to make it working. It should be:

macros: {
'\\color': '\\textcolor{#cc0000}{\\text{\\textbackslash color}}',
'\\textcolor': '\\textcolor{#cc0000}{\\text{\\textbackslash textcolor}}',
}


The macro's name should start with \\.

lcswillems

comment created time in 2 months

issue openedianstormtaylor/slate

#### What's the current behavior?

On the mentions example, if I click on a mention and I hit "backspace", the mention is not deleted. It would be nice to allow to delete them.

If somebody has a hint to implement this, I can do a PR.

Slate: 0.58.2 Browser: Chrome OS: Linux

created time in 2 months

issue commentKaTeX/KaTeX

@edemaine The solution you gave me with

macros: {
color: '\\textcolor{#cc0000}{\\text{\\textbackslash color}}',
texcolor: '\\textcolor{#cc0000}{\\text{\\textbackslash textcolor}}',
}


is not working.

If you try this code with the latest version of KaTeX:

console.log(
katex.renderToString('\\color{blue} x^2', {
output: 'html',
macros: {
color: '\\textcolor{#cc0000}{\\text{\\textbackslash color}}',
},
})
);


You will get this:

<span class="katex"><span class="katex-html" aria-hidden="true"><span class="base"><span class="strut" style="height:0.8141079999999999em;vertical-align:0em;"></span><span class="mord" style="color:blue;"><span class="mord mathdefault" style="color:blue;">x</span><span class="msupsub"><span class="vlist-t"><span class="vlist-r"><span class="vlist" style="height:0.8141079999999999em;"><span style="top:-3.063em;margin-right:0.05em;"><span class="pstrut" style="height:2.7em;"></span><span class="sizing reset-size6 size3 mtight" style="color:blue;"><span class="mord mtight" style="color:blue;">2</span></span></span></span></span></span></span></span></span></span></span>


Hence, the text is still blue. It seems like macros cannot be overloaded.

lcswillems

comment created time in 2 months

issue commentKaTeX/KaTeX

Would it be complicated to implement what I suggest? If it is okay, I could do a PR.

lcswillems

comment created time in 2 months

issue commentKaTeX/KaTeX

If there could be something like allow(context) { ... } as there is for trust trust(context) { ... }, it would be great!

lcswillems

comment created time in 2 months

issue commentKaTeX/KaTeX

Okay, thanks! It is quite nice, and it would be sufficient for me for the moment, but I would prefer showing the command in red like as it is the case for untrusted commands.

Anyway, thank you for all your work!!

lcswillems

comment created time in 2 months

issue openedKaTeX/KaTeX

Is your feature request related to a problem? Please describe. I am building a platform that allows persons to write maths, but I would like to disable some commands such as \color or \textcolor.

Describe the solution you'd like: I would like an option similar to trust.

created time in 2 months

pull request commentlcswillems/python-MNIST-classifier

Thank you for the time you took to make this PR!

If you can remove this data.py file and only put the required packages in requirements.txt, it would be great!

fotisk07

comment created time in 2 months

pull request commentlcswillems/python-MNIST-classifier

Thank you for your PR. I don't get why there is an extra data.py file added? Also, why is there this much packages in the requirements.txt? Only a few of them are currently required.

fotisk07

comment created time in 2 months

issue commentlcswillems/python-MNIST-classifier

Okay, thank you for reporting this! I will try to have a look. Again, if you mind doing a PR, it would be great!

fotisk07

comment created time in 2 months

issue commentlcswillems/python-MNIST-classifier

Thank you for your issue. Would you mind making a PR for this?

fotisk07

comment created time in 2 months

issue commentianstormtaylor/slate

Thank you for your answer! Now, I get it!

A simple sentence would be: a range is hanging if it starts at the end of a node or ends at the start of a node.

lcswillems

comment created time in 2 months

startedgraphql-nexus/nexus

started time in 2 months

pull request commentianstormtaylor/slate

No, I didn't have the time...

majelbstoat

comment created time in 3 months

+2 -0

0 comment

1 changed file

pr created time in 3 months

Missing type for bearer_token

push time in 3 months

A tiny, full-featured, flexible client / server library for the Twitter API

fork in 3 months

startedzbeyens/slate-plugins-next

started time in 3 months

issue commentianstormtaylor/slate

I don't get it either... Is somebody able to explain this concept with a sentence?

lcswillems

comment created time in 3 months

issue commentFyrd/caniuse

That's great!! Thank you for your work. It works perfectly!

lcswillems

comment created time in 3 months

issue commentianstormtaylor/slate

Okay great! Let me know if there is any way I can help :)

lcswillems

comment created time in 3 months

I am now thinking there is no typo... textState refers to this:

const textState = atom({
key: 'textState', // unique ID (with respect to other atoms/selectors)
default: '', // default value (aka initial value)
});


I got confused with 'textState' string for key. I thought this was the way to refer to the atom.

lcswillems

comment created time in 3 months

I think it should 'textState' instead of textState. Otherwise, I don't see where the variable textState is coming.

+1 -1

0 comment

1 changed file

pr created time in 3 months

push eventlcswillems/Recoil

commit sha c35bbeec790c6047961810a93770450651f8ae19

Update getting-started.mdx I think it should 'textState' instead of textState. Otherwise, I don't see where the variable textState is coming.

push time in 3 months

Recoil is an experimental state management library for React apps. It provides several capabilities that are difficult to achieve with React alone, while being compatible with the newest features of React.

fork in 3 months

issue commentianstormtaylor/slate

Thank you for your answer. It works with 0.57.2... Do you have any clue of why it is not working with 0.57.3?

lcswillems

comment created time in 3 months

started time in 3 months

issue commentianstormtaylor/slate

I have updated the issue. Why not allowing Github Discussions on this repo? The big advantage is that they will be indexed by Google.

lcswillems

comment created time in 3 months

pull request commentianstormtaylor/slate

I will try to write one tonight, but not sure I will succeed... Thanks for your quick answers :)

majelbstoat

comment created time in 3 months

pull request commentianstormtaylor/slate

Copy/pasting of void nodes works correctly for me on Slate 0.57.2 but not on 0.57.3, and only this commit might have altered copy/pasting

majelbstoat

comment created time in 3 months

pull request commentianstormtaylor/slate

Because of this commit, copy-pasting void nodes doesn't work anymore for me.

@majelbstoat Did you change the logic of the setFragmentData and getFragment? I can't identify it from the diff

majelbstoat

comment created time in 3 months

startedkeroxp/servest

started time in 3 months

issue commentianstormtaylor/slate

0.13 is a very very old version that I don't know. I can't help you and encourage you to upgrade since it will surely help you solve your issue.

596121109

comment created time in 3 months

issue commentianstormtaylor/slate

I think it is intended because Slate loses focus.

Have you tried to use ReactEditor.focus(editor) to focus editor again?

596121109

comment created time in 3 months

issue commentianstormtaylor/slate

I think it will be hard to understand what you mean with such elliptic title + description.

Adding a more accurate text + description is not making you waste time. On the contrary, it makes you win time because it is much more likely others will understand you and be able to help you.

What do you mean by "re-rendering"? I don't have the issue you describe on my side. Which version are you using? This is asked in the template...

596121109

comment created time in 3 months

issue openedianstormtaylor/slate

#### Question

From the source code and the documentation, I can't understand what "hanging" means. Could somebody explain?

created time in 3 months

issue openedianstormtaylor/slate

Bug

#### What's the current behavior?

On this example : https://www.slatejs.org/examples/richtext

When I add a list on Safari, it moves the cursor at the top of the document.

Slate: 0.52.3 Browser: Safari OS: Linux

created time in 3 months

more