Hey data nerds, greetings from my hometown in Vietnam! 👋I forgot to mention in the video that if you use the extension for the first time, you'll need to enter a product key after launching Jupyterlab. You should receive the product key via email after you register your email here link.makinarocks.ai/start.html 😊
@@Thuvu5 We successfully reproduced your "mrx-link" demo using a relatively simple "Data Analysis" notebook. Everything worked as you demonstrated ! Next, we will create some varied pipelines with branches and test running different paths. We like the comment feature ! Thanks Again !
@@Thuvu5 Additional Update to our 'mrx-link' testing. Created a short "Advanced Numpy" pipeline with branches -- then closed original notebook -- opened new empty notebook -- AND Successfully uploaded and ran short pipeline with testing of all branches ! P.S. Still having some problems uploading importing saved caches .. will test more later ! ...
Another amazing video!!!! I can't wait to play around with that extension!!! Thank you so much for producing such helpful content!!! You're the best!!! 💯👏👏👏
We successfully reproduced your "mrx-link" demo using a relatively simple "Data Analysis" notebook. Everything worked as you demonstrated ! Next, we will create some varied pipelines with branches and test running different paths. We like the comment feature ! Thanks Again !
Hey Thu! I really loved your Witcher character network videos, they inspired me to learn more about NLP from Codebasics' channel so I can eventually test it out on some other book series and maybe scripts. I'm dying to see your follow-up, but I hope you enjoy your time in Vietnam too Thanks for the edutainment and motivation to keep learning!
Perhaps, the ability to include a hierarchy of branched components (cells) whereby the developer can compare running multiple alternative code branches within the same notebook is an added benefit. Plus the graphical pipeline presentation is helpful.
I regularly come to your channel and watch video's, but today I came here by promoting this of this video. I just wanted to say that your videos are great and informational, you don't need to promote them, you can get organic growth❤️ Just a personal opinion.
Haha yeah.. it's hard to imagine but under time pressure with lots of moving parts, it's easy to mess up the notebooks and lose track of what's being done.. all unintentionally 😅
you are solving a problem which does not exist! obviously it is your job to write good code (following rules like not using a defined variable in a previous cell... or forgetting to delete the experimentation cells) if you want to collaborate... ofcourse if it is only you, you can do whatever you want
I love your channel thanks for the videos. You are similar to me in making life easier with plugins and Ad ons and use your iPad efficiently. I learned data science on an old iPad mini and since my job I have been implementing technology things to make my life easier in learning. Have you ever thought about dabbling into VR for productivity, learning and working. It has a much lower cost of entry with the meta quest 2 and it allows you to have multiple screens in a virtual reality setting where everything else around you can me drowned out.
Hey Paul, thank you for sharing this! I agree tech things do make our lives much easier! The VR idea sounds really cool. I believe it will be the future of learning. I haven’t had the chance to look further into this. Where would do you suggest I should start with using VR for learning?
Great video - thanks! Edit: are you also in Amsterdam? I assumed VU was referencing Vue, but based on the “city center” comment, it sounds like VU might be referring to VU (the university).
How does the pipeline ensure that you get consistent results? Or are you referring to the saved cached result as proof that the model produced such a result?
The pipeline explicitly specifies the execution order of the cells, so it addresses the problem of inconsistency due to this. So if the code doesn't change, it should produce the same output. The cache surely helps too! But in my opinion the extension can't prevent inconsistencies due to differences in hardware/ package version etc.
Additional Update to our 'mrx-link' testing. Created a short "Advanced Numpy" pipeline with branches -- then closed original notebook -- opened new empty notebook -- AND Successfully uploaded and ran short pipeline with testing of all branches ! P.S. Still having some problems uploading importing saved caches .. will test more later ! ...
As U may have already determined, their website emails to U a product key when U register with the website. Do NOT know, however, how long key will be active and the "Pricing" page has difficulty loading from website. Good Luck !
Can you continue the video. Once shared, how to bring up the package in other environments. Also, it only runs on 3.10 and below. I had 3.12 and it wasn't running.
I guess I could get used to the idea of having each cell be a separate pipeline component, but it would be nice if Link supported a sequence of multiple code cells to be assigned to a single component.
Hi creator after installing and using this library certain shortcut keys like shift + tab is not working for me . Do you have any solution to make both work
I like Jupyter notebooks, especially if people are writing documentation, or some other text to accompany the code., but I have been caught out by notebooks being structured via some alien logic, where the numbering system starts at 23, followed by 11, and quickly degrading from there. Somehow it seems worse than "ordinary" code being poorly commented with a naming scheme from one of the hotter parts of hell.
Uh-oh... the Link extension hangs indefinitely at completion of a cell after loading a large dataset, with or without cacheing enabled. Without Link the cell completes in 7 minutes, with Link it takes "Infinity minutes"... Seems like they have a memory management overhead issue or something similar. Too bad, I really wanted to start using this tool :(
Hi, Jack! Thanks for your kind interest in Link. Our technical team has reviewed your concern and will contact you to provide support. In case the problem persists, please do not hesitate to contact us. Our technical team will assist you as soon as possible. It is greatly appreciated that you tried it out and shared your issues with us. Your report will help us make an even better Link. Thank you!
Error when performing an action. Reason given: Error: ERROR: Could not find a version that satisfies the requirement mrx-link (from versions: none) ERROR: No matching distribution found for mrx-link
I tried link and it seems like it's slowed down my python Kernel significantly while using Jupyter...my version looks different from yours so perhaps it's something with updates. In any case, will be uninstalling it and never looking back.
I think most of the problems mentioned in the intro is because of bad coding practices. When sending code to another person, the sender should clean up the code to make it readable, and use enough comments.
If only it were that easy! In a previous job, colleagues would maintain some sort of standards by enforcing penalties for stupid and or dangerous mistakes. 1 carton of beer (24 cans) per mistake. Payable at the end of the shift, or before going home after 4 weeks of work. There was always "free" beer to be had every night because of this, but people did make an effort to do things properly.
I know of a few reasons 1. the order of execution isn't enforced, maybe the previous person didn't run the cells in the same order 2. for machine learning in particular there may be a different random seed used if not explicitly specified. if doing cross validation for example if the exact same rows are not chosen for each fold then you get different results. 3. different hardware and operating systems sometimes generate different results. this can be differences in the underlying packages or binaries used.
The library did not work for me and dependencies as well as previously defined functions are not defined in later cells. Too bad, since it is an interesting idea.
Well... Jupyter notebooks are great for experimentation. But after expererimenting, people should clean up the notebook before giving it to another person. What you experience, is just lazy, messy people's unfinished work, of course it is frustrating.
this seems like a bandaid for how awful jupyter is when your experiments start using large enough amounts of code. do people use jupyter notebooks for anything other than experimentation and communication? "tensor_to_image" should just go into your src module that you install into your active environment. and then you just import it into a jupyter notebook that's run within that environment. seems silly. you're also not having all your data scraped by this unknown ai startup lol.
hi thu,i am using jupyterlab, i have one problem in latex,i install one modulus about the jupyter labextension install @jupyterlab/latex,when i using in show Latex Preivew (Output Result-Error)....This is XeTeX, Version 3.141592653-2.6-0.999994 (TeX Live 2022) (preloaded format=xelatex) restricted \write18 enabled.
I love your channel thanks for the videos. You are similar to me in making life easier with plugins and Ad ons and use your iPad efficiently. I learned data science on an old iPad mini and since my job I have been implementing technology things to make my life easier in learning. Have you ever thought about dabbling into VR for productivity, learning and working. It has a much lower cost of entry with the meta quest 2 and it allows you to have multiple screens in a virtual reality setting where everything else around you can me drowned out.
Hey Paul, thank you for sharing this! I agree tech things do make our lives much easier! The VR idea sounds really cool. I believe it will be the future of learning. I haven’t had the chance to look further into this. Where would do you suggest I should start with using VR for learning?
Hey data nerds, greetings from my hometown in Vietnam! 👋I forgot to mention in the video that if you use the extension for the first time, you'll need to enter a product key after launching Jupyterlab. You should receive the product key via email after you register your email here link.makinarocks.ai/start.html 😊
I was waiting for this extension my whole life!
Welcome back. Yes, we will try this Jupyter Extension, thank you.
Hey David, thank you so much! I'd be curious about your experience with this extension 🧐🙌🏽
@@Thuvu5 We successfully reproduced your "mrx-link" demo using a relatively simple "Data Analysis"
notebook. Everything worked as you demonstrated ! Next, we will create some varied
pipelines with branches and test running different paths. We like the comment feature !
Thanks Again !
@@Thuvu5 Additional Update to our 'mrx-link' testing. Created a short "Advanced Numpy" pipeline
with branches -- then closed original notebook -- opened new empty notebook -- AND
Successfully uploaded and ran short pipeline with testing of all branches !
P.S. Still having some problems uploading importing saved caches .. will test more later !
...
🎉🎉 incredible🎉🎉 !! This looks so useful, specially to people who use Jupyterlab in engineering
You definitley deserve a lot subscriber. To be honest.
Aww thank you 🙌🏽
Great video, thanks for sharing. I needed an extension to organize the large notebooks. Now it's done.
Thank you for sharing this Thu! 🙌🏼 Def a problem i've experienced in the past with sharing notebooks
Thanks for watching and sharing Luke! 🤗🙌
Yeah, indeed it can get really crazy with notebooks..
Another amazing video!!!! I can't wait to play around with that extension!!! Thank you so much for producing such helpful content!!! You're the best!!! 💯👏👏👏
We successfully reproduced your "mrx-link" demo using a relatively simple "Data Analysis"
notebook. Everything worked as you demonstrated ! Next, we will create some varied
pipelines with branches and test running different paths. We like the comment feature !
Thanks Again !
That's awesome, David! Thank you for testing it and commenting 🙌🏽🙌🏽
Great video!! I am using Link now in my notebooks. Showing the high level and low level whenever needed in a clear and concise way!!!
Hey Thu! I really loved your Witcher character network videos, they inspired me to learn more about NLP from Codebasics' channel so I can eventually test it out on some other book series and maybe scripts. I'm dying to see your follow-up, but I hope you enjoy your time in Vietnam too
Thanks for the edutainment and motivation to keep learning!
wow that looks handy, thanks!
Very good explanation and super tool!
Love the video Thu, Thanks
That‘s why you re-run the notebook from the beginning before checking it into git.
Haha absolutely!
Perhaps, the ability to include a hierarchy of branched components (cells) whereby
the developer can compare running multiple alternative code branches within the
same notebook is an added benefit. Plus the graphical pipeline presentation is helpful.
@@Thuvu5jzqbuh
@@Thuvu5 Ghbkj
just want to say a great video!! The info is very clear and easy to follow. thank you
Awesome, I was looking for this kinda tool
Thank you so much
I love your channel. Keep up the pace!
That’s so kind of you! Thank you so much 💛🙌
thank you so much for sharing, something that i was looking for
Very nice. Thank you!
I regularly come to your channel and watch video's, but today I came here by promoting this of this video. I just wanted to say that your videos are great and informational, you don't need to promote them, you can get organic growth❤️
Just a personal opinion.
Thanks for this great demo
Thank you so much for checking the video🙌🙌
@@Thuvu5 I am usually checking the latest videos to update my skills. Thanks
Great video as always. I never considered working on a team who have notebook cells that are out of order. Sounds terrifying.
Haha yeah.. it's hard to imagine but under time pressure with lots of moving parts, it's easy to mess up the notebooks and lose track of what's being done.. all unintentionally 😅
Nice video, and helpful tool, thanks for sharing!
We missed you alot
Aw thank you Domingos 💜☺️. I'm happy to get back slowly with making videos!
Amazing and convenient too!😀
Thank you for checking out the vid!
WOW! Thu you break down concepts very well. have you thought about making lessons on youtube?
you are solving a problem which does not exist! obviously it is your job to write good code (following rules like not using a defined variable in a previous cell... or forgetting to delete the experimentation cells) if you want to collaborate... ofcourse if it is only you, you can do whatever you want
Wow nice thu its a really useful extension gonna use it .
Cool! Thank you 🙌🏽
Quality as always
Wow interesting.... Thanks
I love your channel thanks for the videos. You are similar to me in making life easier with plugins and Ad ons and use your iPad efficiently. I learned data science on an old iPad mini and since my job I have been implementing technology things to make my life easier in learning.
Have you ever thought about dabbling into VR for productivity, learning and working. It has a much lower cost of entry with the meta quest 2 and it allows you to have multiple screens in a virtual reality setting where everything else around you can me drowned out.
Hey Paul, thank you for sharing this! I agree tech things do make our lives much easier! The VR idea sounds really cool. I believe it will be the future of learning. I haven’t had the chance to look further into this. Where would do you suggest I should start with using VR for learning?
Thanks for making this video. Do you still use this extension? It doesn't visualize the flow in vs code studio.
Great video - thanks!
Edit: are you also in Amsterdam? I assumed VU was referencing Vue, but based on the “city center” comment, it sounds like VU might be referring to VU (the university).
Thank you! Yes I am ☺️
@@Thuvu5 right on! Have fun with the allergies!
Good tool but reqs old Lab 3. Modern Lab 4 has quick render and tons optimization.
How does the pipeline ensure that you get consistent results? Or are you referring to the saved cached result as proof that the model produced such a result?
The pipeline explicitly specifies the execution order of the cells, so it addresses the problem of inconsistency due to this. So if the code doesn't change, it should produce the same output. The cache surely helps too! But in my opinion the extension can't prevent inconsistencies due to differences in hardware/ package version etc.
I have been missing your videos, I hope you are doing well
Thank you Ahmed! I'm doing great over here. Hope you're doing well too! 🤗
Additional Update to our 'mrx-link' testing. Created a short "Advanced Numpy" pipeline
with branches -- then closed original notebook -- opened new empty notebook -- AND
Successfully uploaded and ran short pipeline with testing of all branches !
P.S. Still having some problems uploading importing saved caches .. will test more later !
...
thank you very much
Great video Thu !!.. Don't we require product key for this extension to work?
As U may have already determined, their website emails to U a product key
when U register with the website. Do NOT know, however, how long key
will be active and the "Pricing" page has difficulty loading from website.
Good Luck !
Use Spyder, it's awesome IMO
Can you continue the video. Once shared, how to bring up the package in other environments. Also, it only runs on 3.10 and below. I had 3.12 and it wasn't running.
Excellent .
I wonder if this is the same thing as papermill?
How can an external separate LaTex file be included/embedded into a jupyter document?
How can documents be nested/included in JupyterLab, being updated when changed?
I guess I could get used to the idea of having each cell be a separate pipeline component, but it would be nice if Link supported a sequence of multiple code cells to be assigned to a single component.
Yeah I was thinking the same thing! But not sure if it would be a good idea.. it could potentially defeat the original purposes of the extension 🤔
Which one did you download/install: the python package or the desktop application?
Hi creator after installing and using this library certain shortcut keys like shift + tab is not working for me . Do you have any solution to make both work
Nice 😊
I like Jupyter notebooks, especially if people are writing documentation, or some other text to accompany the code., but I have been caught out by notebooks being structured via some alien logic, where the numbering system starts at 23, followed by 11, and quickly degrading from there. Somehow it seems worse than "ordinary" code being poorly commented with a naming scheme from one of the hotter parts of hell.
Uh-oh... the Link extension hangs indefinitely at completion of a cell after loading a large dataset, with or without cacheing enabled. Without Link the cell completes in 7 minutes, with Link it takes "Infinity minutes"... Seems like they have a memory management overhead issue or something similar. Too bad, I really wanted to start using this tool :(
Hi, Jack!
Thanks for your kind interest in Link.
Our technical team has reviewed your concern and will contact you to provide support.
In case the problem persists, please do not hesitate to contact us.
Our technical team will assist you as soon as possible.
It is greatly appreciated that you tried it out and shared your issues with us.
Your report will help us make an even better Link. Thank you!
happens to me too, i had to remove it , i'm sorry because it seems really nice idea
Where can I find the option to activate the Cell toolbar such that I can add tags? jupyterlab
This definitely seems like a useful feature. But it feels like it's really just putting a bandage on the underlying problem, bad coding patterns.
how do we import?
Interesting tool but I can't install it at my anaconda3 command prompt. Receiving error.
I want to install it in my conda environment, but getting some error. How to properly install it in conda env
How about trying the code below?
python3 -m pip install --upgrade mrx-link
Nice
Error when performing an action.
Reason given:
Error: ERROR: Could not find a version that satisfies the requirement mrx-link (from versions: none)
ERROR: No matching distribution found for mrx-link
Have you tried Jupyter extension within VS Code?
Yes, thank them for that.
Once you get into vs code with jupyter, jupyter lab feels slow and clunky, leave alone jupyter notebook.
Did you mention if it was free?
Yes, the extension is free, at least for now! 😊
I tried link and it seems like it's slowed down my python Kernel significantly while using Jupyter...my version looks different from yours so perhaps it's something with updates. In any case, will be uninstalling it and never looking back.
I think most of the problems mentioned in the intro is because of bad coding practices. When sending code to another person, the sender should clean up the code to make it readable, and use enough comments.
If only it were that easy!
In a previous job, colleagues would maintain some sort of standards by enforcing penalties for stupid and or dangerous mistakes. 1 carton of beer (24 cans) per mistake. Payable at the end of the shift, or before going home after 4 weeks of work. There was always "free" beer to be had every night because of this, but people did make an effort to do things properly.
Do we know why it leads to inconsistency?
I know of a few reasons
1. the order of execution isn't enforced, maybe the previous person didn't run the cells in the same order
2. for machine learning in particular there may be a different random seed used if not explicitly specified. if doing cross validation for example if the exact same rows are not chosen for each fold then you get different results.
3. different hardware and operating systems sometimes generate different results. this can be differences in the underlying packages or binaries used.
The library did not work for me and dependencies as well as previously defined functions are not defined in later cells. Too bad, since it is an interesting idea.
Hi, you know any genuine website to search for remote data analyst job or paid data analyst project to work?
All's well. I liked it. But it doesn't support Python 3.10!!! And I'm using that version only. How to use it with Py 10?
It provides python package installation of the python 3.10 version.
PLEASE SOLVE THIS PROBLEMS,I REQUEST TO THU
this extension doesn't support python 3.10 and above
Well... Jupyter notebooks are great for experimentation. But after expererimenting, people should clean up the notebook before giving it to another person. What you experience, is just lazy, messy people's unfinished work, of course it is frustrating.
"Run all"? You are describing a non problem.
Use same seed 😂
So much for not writing Python scripts lol
OMG !
this seems like a bandaid for how awful jupyter is when your experiments start using large enough amounts of code. do people use jupyter notebooks for anything other than experimentation and communication? "tensor_to_image" should just go into your src module that you install into your active environment. and then you just import it into a jupyter notebook that's run within that environment. seems silly. you're also not having all your data scraped by this unknown ai startup lol.
hi thu,i am using jupyterlab, i have one problem in latex,i install one modulus about the jupyter labextension install @jupyterlab/latex,when i using in show Latex Preivew (Output Result-Error)....This is XeTeX, Version 3.141592653-2.6-0.999994 (TeX Live 2022) (preloaded format=xelatex) restricted \write18 enabled.
Hi Thu, great content as always. Is it possible to connect with you in LinkedIn?
Hi Adrian, thank you so much 🙌! Yes sure, my LinkedIn is on the About page of my channrl
@@Thuvu5 sent a connection request. You are a genius!
I love your channel thanks for the videos. You are similar to me in making life easier with plugins and Ad ons and use your iPad efficiently. I learned data science on an old iPad mini and since my job I have been implementing technology things to make my life easier in learning.
Have you ever thought about dabbling into VR for productivity, learning and working. It has a much lower cost of entry with the meta quest 2 and it allows you to have multiple screens in a virtual reality setting where everything else around you can me drowned out.
Hey Paul, thank you for sharing this! I agree tech things do make our lives much easier! The VR idea sounds really cool. I believe it will be the future of learning. I haven’t had the chance to look further into this. Where would do you suggest I should start with using VR for learning?