💬┊general

[left rock]

Hi there,
Is there any developer working on toptal platform here?

[marsh glen]

Retired from the real world - just playing in cyber world.

[young depot]

Hi am Belvic

[steel glen]

#1130808226280132608 for all

#1130808226280132608

[stiff magnet]

Hi I'm Duane

[twin harness]

HI I'm Muhammad Asfar
A loading Data Scientist

I am here to learn new things and to explore data science deeply

[silk heart]

hi all

is anyone online? i am feeling very demotivated for data science

i was doing a certification, but now I am 3 weeks late 😦

i am doing this:
https://www.kaggle.com/competitions/titanic

[silk heart]

i scored 0.78229 in titanic predictions. how are people scoring 1.000???

[spice seal]

I got 0.74162 using a decision tree regressor. It's just a matter of figuring out which modeling technique works best and adjusting parameters. Just have fun experimenting I guess

Also when I got my predictions using my model, I got predictions that were between 0 and 1. I ended up just rounding to the nearest digit but is there a better way to deal with predictions that are between 0 and 1 (especially if they are 0.5)?

[unreal idol]

hey guys

i have a question

i need to look for a dataset with certain specifications, i have been on kaggle for around 4 hours trying to find one but im unable to find one

i cant seem to find a dataset that is: 1. uncleaned, 2. has numbers only, 3. has at least 20 columns, 4. needs to be under 100 MB

i need one like this one: https://www.kaggle.com/datasets/bobbyscience/league-of-legends-diamond-ranked-games-10-min

this one is literally perfect but i cant use this one because someone else has already used it

any help?

please it's urgent and i have a project due on this and i need any help

does anyone know where i could find a dataset like that or has a dataset like that?

[flint prawn]

hey there how you learn math for machine learning

[twin harness]

Check DM. I will tell you there.

[steel glen]

Print("some names")

Ravan

Anil

Narayana

Ok

[abstract leaf]

Hello. I am a university student in the information science department. I recently started studying machine learning. Nice to meet you.

[solemn plover]

Hello. I'm Data Scince student and I'm looking forward to meet people like me.

[deep sapphire]

hey everyone Ansh this side , I am Artificial intelligence and Machine learning undregrad in India , Looking forward to hands dirty in contests and for DS and AI job opportunities.
here is my linkedin: https://www.linkedin.com/in/ansh-ml/

[dawn oracle]

Hello. I am a university student in the information technology. I recently started studying machine learning. Nice to meet you.

[glass lagoon]

Hello, My name is Sakshi . I am Data Science Student and looking forward to explore and learn new things

Here is my Kaggle - https://www.kaggle.com/sakshisatre

[drowsy field]

Hi, My name is Pranav. Really excited to be part of this journey..

[vernal cloud]

hi, I am Vaibhav. I am a Data Science Student. Looking forward to learn new things in this field

[ripe abyss]

hihi

[last sphinx]

hey

[hushed night]

Rimviews guys

[weak gulch]

Hi

[coarse gulch]

Hello, I'm a Data Science Student and I'm excited to be part of the Kaggle team!

[topaz wind]

Hi, I am Sumit, what should be the ideal way for me to start learning data science and ai, I do competitive programming in python and have computer science background and I am good at math

[weary marlin]

Hi, my names Tariyel. I am software developer. I come here to be new member of the new generation.

[hushed night]

Hello, My name is Neel . I am a Data Science Student looking forward to explore and learn new things

Here is my Kaggle -
https://www.kaggle.com/noobidoobidoo/

[shell compass]

Hi I'm Carl I'm an Artificial intelligence enthusiast.....and I'm here to be a member of the new generation

[rich galleon]

My name is Robert and I`m interested in PyTorch and Tensorflow. Wanna learn how to program AI so that I can use that knowledge to go further in the field of neuromorphic. Here is my kaggle: https://www.kaggle.com/robertbz

[sonic flume]

Hi I'am a Business Informatics student and I'm here to begin my ML learning journey because i want to become a Data Scientist!

[real scarab]

Hello, I am William Wu. I am a software developer in China and I'm happy to meet you guys here!

[runic sinew]

Hello, I am Aman Raushan.i am a student in india and i'm happey to meet you guys here!

[inland hawk]

Hi

[carmine plank]

Hellu! I am Sandhya. I'm an engineering student from India. Looking forward to learning from everyone here ^-^

[short pumice]

Hi

[grand wagon]

Hi guys, Please check https://www.kaggle.com/code/ayushtiwari2323/bilstm-rnn-delivers-99-accuracy

[smoky crater]

I am Govind Shaw . My Education is bachelor in electrical-Engineering stream. I have 3+ Years of experience as a Data Analytics using MS-excel,MS SQL Server,Tableau Desktop and Tableau Server.

[placid quartz]

is anyone familiar with distributed training? I am trying to use pytorch Data Parallel but still only 1 out of 4 GPUs is being used.

[rain ibex]

Hi

[visual pasture]

Hi everyone, I have a Data science assignment regrading A levels students performance in Universities.
So if you have done A levels and are in university, kindly fill this form.

https://forms.gle/ERkcSDvSnDWMkbAD6

[bronze flower]

Hey everyone! Just trying to get better and more comfortable with data science and Python in general 🙂

[open cedar]

Hi everyone, I'm a Data Scientist student.
Happy to be in this forum

[warped fog]

Hi everyone
I am a student aspiring to be a data scientist
Nice to meet y'all

[spice loom]

Hello Everyone, I am a CS grad exploring Data Science. Glad to be here

[open linden]

Hi everyone, I'm a CS grad exploring data science, machine learning and python

[dreamy mirage]

Hi everyone. I am CS grad trying to learn and explore ML/DS world

[cyan root]

hi everyone, i am a grad in Administration management exploring machine learning and data science😊

[proper bramble]

Hi there! I'm a experienced Software Engineer that is trying to learn more about the Data Science world

[lethal elm]

Hi I am a data analyst who has some experience in data science and machine learning. Cheers

[swift pollen]

hi Im an Information System student from Multimedia Nusantara University Indonesia and im excited to learn more about kaggle 😁

[humble jackal]

helo

merhaba bana yardımcı olabilecek biri varmı konu derin öğrenme dm ulaşırsa sevinirim

[leaden quarry]

Guys

can someone help me pleaseeee

after the model trains, i cant find it on the working dir

Here is my syntax im using:

Save the best estimator using pickle

    with open('/kaggle/working/model_rf_overpowered.pkl', 'wb') as f:
        pickle.dump(best_estimator, f)
    print("Best estimator saved using pickle.")

    # Save the best estimator using joblib
    joblib.dump(best_estimator, '/kaggle/working/model_rf_overpowered.joblib')
    print("Best estimator saved using joblib.")

    return best_estimator

Please help kaggle is the only free platform i can use

thank you guys!!

[full grove]

I am Shripad SK. Comp Science Undergrad.

[oak gulch]

I am Pratyush. Comp Science Undergrad

[wicked berry]

Hello, everyone. I hope you are well. I am new this channel and i want to become a talented ai engineer with your help. Thank you.

[shrewd urchin]

ensure that the paths '/kaggle/working/model_rf_overpowered.pkl' and '/kaggle/working/model_rf_overpowered.joblib' are correct and point to the directory where u expect the files to be saved cuz if u r running this code in a Kaggle notebook, the working directory might be different from what u expect

[leaden quarry]

is that caused cause i specified the model nametogether with the kaggle/working? how the fidn the working dir of a notebook

[shrewd urchin]

u sure that dir kaggle/working exists

u might consider using a relative path or a dynamically generated path based on the current working directory

[leaden quarry]

Okay i will try, do you how to specify the current working directory in kaggle? thanks

[leaden quarry]

This modification uses os.path.dirname(os.path.abspath(file)) to get the directory of the current script, and then constructs the relative path to the desired save location using os.path.join. This ensures that the model is saved relative to the script's location, making it easier to manage and more portable across different environments 25.

like this?

[leaden quarry]

here, Here, i checked the current directory, and its the same i set to save the model

[leaden quarry]

best_estimator = self.create_model(X, y).best_estimator_

    # Save the best estimator using pickle
    with open('/kaggle/working/model_rf_overpowered.pkl', 'wb') as f:
        pickle.dump(best_estimator, f)
    print("Best estimator saved using pickle.")

    # Save the best estimator using joblib
    joblib.dump(best_estimator, '/kaggle/working/model_rf_overpowered.joblib')
    print("Best estimator saved using joblib.")

    return best_estimator

[primal mulch]

🌟 Hello everyone! I'm Roman Letodiani, a software developer specializing in full-stack web development. Currently, working as Frontend Web Developer and Assistant Lecturer.

🚀 As an educator in software development, I'm passionate about sharing my knowledge with others and helping them embark on their coding journeys. Whether it's through workshops, tutorials, or mentoring sessions, I love empowering people to unleash their creativity through code.

👩🏻‍💻 I'm a firm believer in collaboration and enjoy teaming up with fellow developers to work on exciting projects. Whether it's building a new app, contributing to open-source initiatives, or tackling challenging problems, I thrive in environments where creativity and teamwork flourish.

Feel free to reach out if you're interested in collaborating on projects or just want to chat about all things tech! You can connect with me on LinkedIn: https://www.linkedin.com/in/letodiani-roma

[arctic plume]

[hardy cedar]

Hey guys if anyone is interested in learning Deep Learning with me with the data set of Digit Recognition, just Dm me.

[ocean roost]

Hey all, anyone in the US (Chicago even better) that wants to team up? I’ve been a data scientist for a couple years but we don’t do much model building. Came to kaggle to get reacquainted with it all and see where I left off

[thorny berry]

Can someone tell me how to view your remaining time left for the week as far as using a gpu? I know its limited to like 30 hours but I dont know how to view the amount of hours i have left

[strong cedar]

Hello, my name is Bilguun Tegshbayar. I am a software engineering student at McGill University, and I am coding enthusiast.

[warm crow]

Which competition would I learn the most as practice if I am going to build a machine learning model for predicting how likely horses are to win a race in horse racing?

And would i learn more from an active or completed competition?

[hexed flax]

Hello everyone !
I'm Ghaith Mechi, data scientist from Tunisia. Here is my LinkedIn link: https://tn.linkedin.com/in/ghaith-mechi. Let's connect and share knowledge

[thorny berry]

Hey i think i broke it XD
my gpu's dissapeared

[coral harness]

Hello, this is Emmanuel.I'm Msc Data Science Student and looking to explore Kaggle to build on my Data Science Career. Nice to meet all.

[remote slate]

https://tenor.com/view/hello-dustin-henderson-gaten-matarazzo-robin-buckley-maya-hawke-gif-25962630

[toxic bronze]

I'm 37 yo, studying data science in college as my second graduatin course (fisrt one was computer networks) interested in learning data science and changing carriers inside my current job. Some oportunities are already showing up, so I have to speed up my learning.

[brittle token]

User
0 Image 3762 non-null object
1 Class 3762 non-null int64
2 Mean 3762 non-null float64
3 Variance 3762 non-null float64
4 Standard Deviation 3762 non-null float64
5 Entropy 3762 non-null float64
6 Skewness 3762 non-null float64
7 Kurtosis 3762 non-null float64
8 Contrast 3762 non-null float64
9 Energy 3762 non-null float64
10 ASM 3762 non-null float64
11 Homogeneity 3762 non-null float64
12 Dissimilarity 3762 non-null float64
13 Correlation 3762 non-null float64
14 Coarseness 3762 non-null float64

These are the columns in a dataset. Predict the class using ml algorithm, 0 means no brain tumor, 1 means there is brain tumor.

Can someone explain these columns in details.
Link- https://www.kaggle.com/datasets/jakeshbohaju/brain-tumor/data

[upbeat bramble]

hi there, are there anyone familiar with kaggle system?, I wanna install the required package to help pdf2image work but idk how to allow it to continue installing

[jovial musk]

Hello, has anyone ever did the Tensorflow Professional Developer Certificate exam ?

[elfin stag]

fuck

Lol what's a bad word

[inland jolt]

@upbeat bramble Hello !! you can apply

apt-get install poppler-utils -y

Apologizes . I wrote enter direct nstad of shift+enter , so it was posted . I am first time here !!

[upbeat bramble]

:ping:

umm, apply for what?

oh, so it will auto confirm for me, wow

it is ok, i am just cautious toward newleaf as there is a potential spam bot (in other server)

[inland jolt]

Hello everyone !! I am ayush , from kaggle .
kaggle.com/ayushkhaire

[upbeat bramble]

thank you for the guidance!!

[short wedge]

It's our spam bot prevention, it blocks people trying to tag everyone or spam external links.

[wicked berry]

hello, i want to become a talented ai engineer. can anyone help me?

[unborn compass]

I am beginner in this field for better understanding I started first with ML topics (supervised currently going on). Can someone guide me is my approach fine or will I get struck. I am good knowlegde on maths and python

[austere portal]

I started machine learning and data science about 5-6 months ago. In the next 6 months, I need to create a recommendation system for a startup platform, but I don't know how to do it. Can you give me suggestions and guidance?

[urban gyro]

I’m not so much of a beginner in this field. I was learning ML via videos on YouTube but apparently, my python skills was shaky so I decided to improve that first before coming back to ML/data science / AI cos I’ve got some personal projects I’ll like to embark on.

[woeful juniper]

just started my first kaggle project, cant wait to become a pro

[half salmon]

better use joblib or onnx

[queen grove]

heya, is it acceptable to write sql queries in all small case for a purpose of presentation of queries and such and to showcase a code? all courses use key words capitalized, but tbh it's really pain to use shift or even capslock all the time while writing a lot of code, so is it ok to write it all in small caps?

[vapid garnet]

Omw to india to record random birds for my database

[wicked elm]

https://tenor.com/view/ping-pingeveryone-gif-18070386

[broken sparrow]

Guys, have you ever had a problem with submitting for a competition where in the "Logs" it says Run Successful, the submission file is correctly formatted and under submissions, Kaggle says "Notebook Threw Exception," but the whole notebook executed successfully? Why is this happening?

[wicked berry]

Just started with kaggle 🙂 great to be here! Thank you!

[slate forge]

Just started with kaggle great to be here! Thank you!

[sonic sigil]

am getting an error

Resource 'corpora/wordnet' not found.  Please use the NLTK
  Downloader to obtain the resource:  >>> nltk.download()

i even used nltk.download('wordnet')
still the error persists

[cyan trail]

Just starting with kaggle. Happy to be here and look forward to controbuting and learning from the community

[glacial axle]

Hey everyone!

I'm thrilled to present my latest notebook: "Gemma, LLM, Kaggle Solution Writeup Agent"!

Ever wondered how top performers crack the code in Kaggle competitions? Join me as we'll unlock the secrets behind winning solutions, powered by Gemma, Google's Open Model.
We'll unravel the strategies, techniques, and secrets behind those winning submissions. With Gemma we'll dive deep into the heart of data science excellence.

Throughout the notebook, we'll:

🔍 Deep dive into Kaggle solution write-ups, dissecting the approaches of top performers.
🎓 Explore innovative methods for data generation, few-shot prompting, and fine-tuning LLMs to boost model performance.
💡 Break down complex concepts with real-life examples, making them accessible to everyone interested in data science.
Discover how to upload your fine-tuned models to the Kaggle Model Hub, opening up opportunities for collaboration and knowledge-sharing within the community.

Check out the notebook right here: https://www.kaggle.com/code/ianakoto/gemma-llm-kaggle-solution-writeup-agent

[old oxide]

Just started with Kaggle, great to be here! Thank you!

[wicked berry]

Hello, everyone. I hope you are well and i want to become a talented ai engineer. Can anyone help me?

[pliant nest]

Hi all, I've written a comprehensive review of variational inference in one concise book. It's been a lot of dedicated time and effort, so I hope you stay tuned https://www.thevariationalbook.com/ Happy to chat with anyone interested in the field. And on twitter, posting tidbits of knowledge https://twitter.com/TheVariational

The Variational Book (@TheVariational) on X

The topic of Variational Inference unites key concepts in machine learning. Follows us to learn more about all things AI.
https://t.co/F56BECcy1N

Twitter

[wicked berry]

I can help you about that

[wicked berry]

Thank you!

[wicked berry]

hello, everyone. i just entered in pii-detection-remove competition and i saw some codes. but all training data is from test data. can anyone explaint this?

[uncut mortar]

Hi...can I dm someone regarding an kaggle competition...I m new to this platform and want to take part in this competition where you need to convert 2d to 3d images...pls your guidance would help

[pulsar bane]

i dont like tensorflow

[wintry aurora]

Can I form a team in this community for the kaggle competition?
https://www.kaggle.com/competitions/llm-prompt-recovery

[rapid rain]

I have register for learning Agency Lab Competition , this is my first contest please someone guide me ,
how to begin ? how to approach ,where to code , what should be my output

[limpid surge]

1. how to approach - read highly voted notebooks from Code Section. If you're not able to understand something google it.

2. where to code - kaggle notebook / google colab

3. what should be the output - read the data section of the competition

[dull orbit]

hello,i am glad to join the kaggle's discord.this month i am plan to take a competition -birdclef2024 by myself.It's my first time to take program not train test.wish good luck.and i hope to recive some suggstions to get a former grade

[ornate saddle]

Hello, everyone.Is there a channel to discuss and find dataset?

[short wedge]

@ornate saddle You can ask here or in #💾┊data

[ornate saddle]

thanks,god bless u

[rigid fox]

Hello my name is Javier Chacon, I'm Systems Engineering student in Colombia and a Data Science enthusiast!

[odd walrus]

Hello, my name is Nadaa and I'm a fullstack software engineer and Computer Application Student

[robust sphinx]

Hi, my name is Vikram and I am a tech evangalist

[tawny lagoon]

📢 Exciting news! Ever wondered what vectors in a vector space look like? 💭 Check out my latest Kaggle Notebook where I delve into the visual representation of word vectors using the** Word2Vec model on the Game of Thrones dataset**! 🐉📚

🔗 Link to the Notebook: Visual Representation of Word2Vec on GoT Dataset

In this notebook, I've explored the fascinating world of word embeddings and how they can be represented visually in a vector space. Using the popular Word2Vec model, I've** analysed the rich language of George R.R. Martin's iconic series, Game of Thrones.**

👀 Dive in to:

🔍 Understand the concept of word vectors and their significance in natural language processing.
🎨 Visualise word embeddings in a high-dimensional vector space.
📊 Explore relationships between words and their vectors!
🔀 Investigate semantic similarities and analogies between words.
💡 Gain insights into how vector representations capture linguistic nuances and contexts.

Whether you're a seasoned data scientist or just curious about the magic behind language models, this notebook offers a captivating journey into the world of vectors and their applications.

Don't miss out! Click the link above to explore the notebook, and feel free to leave your comments, questions, and feedback. Let's dive into the realm of word vectors together! 🚀✨

#DataScience #WordEmbeddings #nlp #GameOfThrones #Kaggle #Word2Vec #VectorSpace #DataVisualization

[fallow slate]

Hi, my name is Timmy and I am a research mathematician

[harsh canopy]

Hi guys

Is there a bioinformatics guy here? I just want to know if it's possible to combine bioinformatics and computer vision 😭😭

[harsh canopy]

Well I will answer mysefl

I already knew about bioinspired algorithms so... the only way that comes to my mind is, using a bioinspired algorithm to find the better architecture of a NN

for computer vision

[proven lagoon]

Hey guys, I am looking for teammates to take part in Math LLM competition. Quite new at the field, so it's mostly going to be exploratory. Hopefully, it will be a learning experience. Do connect, Thankss!!🤗

[summer mauve]

Hi guys

I am Rishi , BTech Electronics Engineer IITBHU

[cunning thorn]

Hi guys
I am Paulo Marcos.

[indigo current]

I have one serious doubt

Regarding stable diffusion architecture

There are two things written denoising unet and and denoising step

Are these both identical or both different

Because in the SD paper there is no clear information related to it

[eternal bolt]

model should be identical but conditioned on different time steps that will be added to the self attention layers in the unet im pretty sure

would be quite impractical to train T differnet unets lmao

Im not sure about stable diffusion but the way a traditional diffusion model trains is it samples uniformly up to time step T and learns the noise to get rid of at that time step

[indigo current]

For the different time steps , it's possible because they uses a concept of sinusoidal position embedded for time related constraints

So this unet is actually time dependent unet

[eternal bolt]

yes thats how they condition the time step

[eternal bolt]

this unet is dependent on time step and im pretty sure stable diffusion also conditions on text embeddings and such in order to achieve text to image generation

I suggest you read up on the original DDPM paper before u try to read stable diffusion because it goes over the general concept of diffusion models before it conditions on anything like clip representations. In a nutshell a traditional diffusion model is a model that predicts p(x_t-1 | x_t) starting from random noise, so it goes without saying that its dependent on time

[indigo current]

Btw I already read the ddpm paper as well

And read the whole stable diffusion process as well

According to me both of these denoising unet and denoising step are identical steps

Because the formulation on the paper also shows the same thing

But actually if both of these are identical than why the mentioned it

[eternal bolt]

???

[eager prism]

How can i install libraries with pip during submission with no internet?

[tawny lagoon]

🚀 Dive deep into the world of Transformers with my comprehensive guide! 📚✨ Whether you're new to the field or looking to level up your understanding, this post has got you covered. From the basics to advanced concepts, I break it all down in an engaging and accessible way.

Check it out here: Understanding Transformers: A Comprehensive Guide

Feel free to share with your fellow AI enthusiasts and learners. Let's unlock the power of State-of-the-Art Transformers together! 💡🤖 #Transformers #AI #DeepLearning #nlp #KaggleKnowledge

[indigo current]

Okay next question

What is latent space

[eternal bolt]

latent space is usually a lower dimensional more abstract representation of the space. for example instead of generating an image in the high dimensional pixel space you would generate it in a lower dimensional latent space to save computation and then project it back to the pixel space with an autoencoder. diffusion models are pretty difficult to understand you should really start with like autoencoders then vaes then diffusion models

[compact osprey]

Hi, my graduating project is to do segmentation for satellite images using Segment anything model, so i am trying to fine tune the model on my dataset, first i do freezing to the image encoder block and tune the propmt encoder and mask decoder blocks and for the propmt i use point, when it run the ruslts not good at all, the dice score is terrible, so i want someone to help me figure what the problem is to finish the project.

[ornate saddle]

is there a dataset consists of the texts about laws posted by lawyers

[proven lagoon]

Hey, as in you want court judgement documents? And also for which country?

[ornate saddle]

yeah, as long as it's about laws of the US

[soft ferry]

Hello, I'm learning to code and beign useful

[arctic plume]

wat are you learning to code

[slim cloak]

Hello everyone, I have been in Deep Learning space for about some time now..Often when I watch new tutorials (I was watching lora implementations in pytorch from umar jamil's lec)..I just can't keep up with many of things in code...most of things feels new and gets overwhelming to understand
How do I bridge this gap effectively?

[harsh edge]

#❗┊link-kaggle-account I am not able verify my phone number. It is showing "Too many requests" error.

[soft ferry]

I'm learning python, while trying to understand and create demos of genai use cases

[fringe hedge]

I'm no expert in ML or DL, but when it comes to learning code, first look at what it's doing in the context of the code; what two parts is it linking? Then look up its documentation, whether that's python or sklearn or javadocs and try to understand, and then ask chat GPT "What is X doing in this bit of code?" and copy paste in the bit of code you're curious about.

[uncut folio]

Guys does anyone have free resources to learn deep learning and GenarativeAi?

[arctic plume]

https://github.com/microsoft/ML-For-Beginners

[wicked berry]

hello, I'm learning to code of great competitions

[uncut folio]

Is Any course available?

[compact osprey]

is anyone here try to fine tune Meta Segment Anything model on satellite images ??

[uneven flare]

raviharitasa

[grim thistle]

Hello, I ma new here. any welcome ?

[covert saddle]

hello, i started working on ml like a month ago, in which i completed andrew ngs supervised learning course and am almost done with the advanced algo course. besides that i am using "hands on ml with sci kit" book

[arctic plume]

welcome

[fluid delta]

hello, I am kellia, I am a finance Student and I am looking forward to more about kaggle

[slender radish]

Hello i would like to ask about something

If i want to have an app that matches a resume against a job posting and assign to each resume score from0-100

Could anyone please worked on a similar task guide me how to approach it?

I don't want to use APIs

[calm fiber]

Hello my name is Nicolas and I`m a Data Science student at the PUC in Chile. My main interest are topics related to machine learning algorithms

[rich halo]

horse1218 has been warned

Reason: Bad word usage

[viscid spade]

Great opportunity for everyone to learn new skills and techniques. Complete the Cloud Skills Challenge: Fundamentals of Machine Learning!

Challenge Link:
https://learn.microsoft.com/training/challenges?id=35934283-b24e-4684-94d3-f1f2567865a2&WT.mc_id=cloudskillschallenge_35934283-b24e-4684-94d3-f1f2567865a2&wt.mc_id=studentamb_291122

[vital gyro]

hey

[fading urchin]

hey

[rich halo]

cryptoninjaok has been warned

Reason: Bad word usage

cryptoninjaok has been banned

Reason: Too many infractions

[wicked berry]

does anybode have experience with zarr or other libs to store datasets for fast loading?
Because I want to save a giant dataset into a zip file but I don't understand if zar will try to load whole zip into RAM
its 150 gb the dataset and its of mri images that I use 2d slices of and even those are painfully slow to load

[rich halo]

abdullahawann has been warned

Reason: Bad word usage

[vale bough]

Is Kaggle stuff here?

[fading blade]

Hi! I'm a 20 y/o coding student who loves math and AI that would like to team up with someone in the Math competition!

[shut saddle]

hey, I see websites saying there are 17m users on kaggle but the numbers on the leaderboard only add up to 200k including noobs, where is the reason of the difference?

[dapper pivot]

Just want to double check, this has stopwords removed right?
https://www.kaggle.com/datasets/andrewmvd/trip-advisor-hotel-reviews/data

[short wedge]

The 17M number refers to the total number of users who have signed up to Kaggle. The 200k number you are seeing is likely the number of users who are ranked for competitions - which means people who have submitted to point-awarding competitions (eg. not titanic). That's much lower since many users on Kaggle only enter tutorial competitions, download datasets, or do courses.

[short wedge]

I'd recommend asking on the discussion forum for that dataset, the author is probably the best person to answer your question.

[quaint cave]

Hey, I created a nootebook trying to do a customer segmenatition, and I'd like to know how good it is. The link is: https://www.kaggle.com/code/pytoleo/eda-customer-segmentation
Does anybody can tell me how I should improve?

And also, does anybody recommend a dataset for starting with tensorflow?

[magic loom]

hey

I am creating a sentiment analysis model using https://www.kaggle.com/code/prakharrathi25/sentiment-analysis-using-bert

I need some help

[shut saddle]

oh... I see, thank you! :)

wait, since submitting a competition entry automatically makes one a contributor regardless of the result, why is not everyone who's ranked in competitions automatically a contributor or higher?

[short wedge]

You need to do a few others things to reach contributor beyond submit to a competition. And submitting to tutorial competitions counts for reaching contributor and doesn't count towards the ranking number.

[shut saddle]

I see, that's understandable then!

[reef tiger]

Can anyone help me in implementing this paper "Robust and accurate object detection through adversarial learning" https://arxiv.org/abs/2103.13886

[dim ferry]

Hey, I just created a video talking about the rise of AI (For a college assignment). And I need to show the engagement on the video uploaded, hopefully you guys are willing to give it a like and give me a feedback on the comment section.
Here's the youtube link: https://youtube.com/watch?v=SfKGHKzkm-o

Few things I definitely lacking on:

• The materials are so shallow (Since I'm only given one week to make and get engagement on the video)
• I'm not a native English speaker, so excuse my English

Thank you and have a good day 😄

[earnest lodge]

Good job!

[tawny lagoon]

https://www.kaggle.com/code/saurabhbadole/decoding-the-attention-black-box-using-bertviz

https://www.kaggle.com/discussions/general/496770
BertViz is a tool designed to visualize the inner workings of a specific part of large language models (LLMs) called the attention mechanism. I am sure you will love this blog and find it insightful 🙂

[wicked berry]

Hello, I would like to help you with this. Can I dm you?

[reef tiger]

Yeah sure please do

[fringe birch]

i was wondering, will google cloud next always be in the usa, or will it also be in the eu in the coming years

[dim ferry]

Thank you!

[hardy zodiac]

https://learn.microsoft.com/en-us/training/modules/get-started-ai-fundamentals/?wt.mc_id=studentamb_328079

[tight mesa]

I am seeking a remote position with the role of machine learning and computer vision developer.
Please DM me if you are interested

[fair glen]

Hello,

I’m seeking interview participants with experience working in the data science sector as part of my Computer Science MSc dissertation project.

The purpose of the study is to understand the importance of human domain knowledge when creating explainable artificial intelligence (XAI) models. The study involves creating a repeatable XAI model development framework that incorporates domain knowledge. The framework will consider each stage in the model development life cycle and outline how domain knowledge can be optimally implemented at each stage, from data collection to quality assurance.

To make this research possible, I need your help. I’m sure there are plenty of experts in the Kaggle community with exactly the knowledge I am looking for.

All interview participants will be anonymised in the research, there are a total of 11 interview questions, and the expected interview duration is 15-20 minutes.

If you are interested in helping my research, please message me and I’ll share further details.

Many thanks,

Phillip

[lapis bay]

This is Muhammad Taqi Raza a researcher in the field of computer vision and deep learning. I have done several projects related to machine learning, one of which is poverty detection using satellite imagery. Moreover, I have hands on practice in training deep learning models. If some one is interested in working with me, I could be available.

[copper locust]

hy everyone, I completed some courses like intro to ML and deep learning and I did some competitions like titanic problem. Now I feel a little lost. Can anyone give me kind to roadmap on how to do kaggle now or what should be my approach? just looking for good advices.

[hallow pewter]

kaggle down?

[night sail]

Kaggle is down? What's going on?

[barren ermine]

Kaggle is down for me too !

[earnest isle]

is competition ending soon?

[night sail]

It's up again

[short wedge]

Which competition?

[barren ermine]

First, Id say you should try to learn Advanced Python and do Exploratory Data Analysis instead of ML Competitions. Doing the ML intro in kaggle is really not enough, you have 2 choices: find a course like ZTM to have a solid foundation on ML and DL or you can always check people notebook (the ones that has the most markdown and informations) so you can understand why for exemple we use a Regressor or a classifier. After that, take notes and try to replicate this on your own notebooks.

[patent plume]

Hey! Can anyone suggest me any platform where I can load and train my ML model of the given dataset that is of around 136 GB

[hallow pewter]

dask is helpful with dealing large datasets

[wicked berry]

how is it helpful?

rather if I have a single laptop, can dask help me?

I am using zarr and deeplake currently

[hallow pewter]

Instead of loading the whole dataset and frying your RAM's soul begging you to have a mercy on it. You load the dataset in chunks or rather as partitions. It also provides tools with multiprocessing and inference on multiple devices. I myself am new to using the framework. And it really helped me a lot when dealing with huge datasets

[wicked berry]

interesting, sounds like what zarr, h5py and deeplake do, but I haven't seen anyone mention dask and compare it to them

[hallow pewter]

I didn't use the libraries you mentioned so can't compare dask to them. But if they do the job for you then it's good.

[inland flicker]

not the F1 point distribution in AutoML Grand prix 💀

[sharp dirge]

Hello Everyone, I'm working on ML project I have used PIMA dataset https://www.kaggle.com/datasets/uciml/pima-indians-diabetes-database?resource=download for Diabetes Detection I want to use another similar Dataset that I can use for comparison. Can anyone share me any publicly available such dataset that I can use. Thanks!

[still viper]

Hello everyone, I am learning and working on ML projects. My bio is available at https://abhijit-ai-folio.netlify.app

In my current role as a Python Test Automation Developer, I thrive on exploring data and domain before translating it into actionable item. I have spent around 11+ years in the BFSI domain exploring mutiple businesses and veriticals within. I have pursued certification courses on Data Science, Machine Learning and simulatenously working on developing End-to-End Machine learning internship projects. My goal is to step-in into the Machine Learning ecosystem of Data Science as an ML Engineer.

[misty mango]

Hi, guys! I'm a student participating in my first Kaggle Code Competition, so I'd be greatful if you could help me, a beginner, understand the requirements for an elligible submission.

From what I've read in the Kaggle documentation, it is necessary for the submission notebook to be ran "top to buttom" in less than 9 hours of CPU / GPU runtime. That means that in my submission I should train the model on the training dataset and also predict on the test dataset in less than 9 hours, right? So, if I'm using an ensemble solution, I should manage to train all my component models in the time limit. Or I'm getting it all wrong?!

I'm asking this question because I've noticed in this and other competitions' code tabs that there are public inferrence-only notebooks that import model(s) trained elsewhere (uploaded as Kaggle datasets) and use them directly to predict on the test dataset. This shortens the total runtime of those notebooks. Is this kind of notebook allowed to be a final submission? Or is this just a way to avoid exhausting the GPU weekly quota while also allowing one to see how well their predictions perform on the public leaderboard and also making certain notebooks public for the community without revealing too much of the training process used.

If this isn't allowed, then how are my submissions supposed to compete with these sped-up notebooks, with high public scores, especially in the efficiency section of the contest?

Thank you in advance!

[earnest isle]

kaggle down?

[spice seal]

I just submitted something for the titanic competition 10 monutes ago and it went fine

I think Kaggle is working

[earnest isle]

yeap already on the work baby

[bleak salmon]

Hey everyone!! Super excited to join and begin my developer journey

[rotund umbra]

Hey y'all! Super stoked to be part of this crew and learn from all the awesome data scientists here.

[shut saddle]

hey, what is the state of the art model for sentiment analysis? (just binary, positive/negative)

[honest ibex]

how you guys been running the llms? do you pay for the vertex ai subscription or use private hardware?

[honest ibex]

i would say some spinnoff of llms with fine tunning + prompting, or remove the language modelling head and use the logits from the encoder/decoder to train a classifier

[shut saddle]

can this work for classifying massive amounts of text messages?

[honest ibex]

yes. if i wanted to do sentiment analysis that is what i would do. tho for simpler implementations you can use document embeddings and have something that kind of works. Do you mean massive in the sense of a lot of texts, or that the texts itself are massive. If the latter, it would get tricky with llms as you would have to look into models that can have longer context window and probably more expensive computation.

[marsh void]

Hi everyone, I am very excited and I hope I can join any team available here

[shut saddle]

oh thanks!

[earnest isle]

kaggle down i repeat kaggle down

[buoyant sparrow]

Hi! I'm not familiar with software programming. But I think I see things before even happen from my Qualitative analytical skills, on that way, I can see the Qualitative state of events.

Can someone help? I'm looking to intertwine.

[elfin stag]

can someone help me

related to stock prediciton model

[floral tangle]

Would love to hear your reviews, this is my first such writing
https://www.kaggle.com/discussions/general/499806

[bold trench]

[bold quail]

does anyone know what the percentage % sign means in colab? or notebooks?
I know ! is used to run shell commands, but what about %?

[jagged chasm]

anyone looking to start a study group to read machine learning by rashka?

[azure lynx]

hi

[azure lynx]

in

[still hatch]

! is used to execute shell commands, % is used for magic commands which are a special command provided by IPython. Magic commands don't execute anything in the shell, they are used to provide special functionality from IPython such as timing code execution.

[rich halo]

asadali_ has been warned

Reason: Bad word usage

[river pagoda]

Hi everyone, are there someone here?

I got a question about how can I start with my data sciences degree in online formation because there aren’t careers of data sciences or computer science in general. I’m studying accountability and I wanna make a models for the auditoriums in finance institutions

Sometimes I’m feel so alone and in kaggle I didn’t can to speak with someone

And also my English is some worse, cause I’m from Central America, I hope you will be able understand. Best regards my friends 🙌🏻

[bronze swift]

can anyone help with this error? I can run it on jupyter but when i try doing it on kaggle this pops up

[short wedge]

Kaggle is a very self-study focussed platform. Many people learn best with the structure that comes from a class with a dedicated teacher and fellow students. You may find that is the best place to start, then consider Kaggle as an extra source of projects to hone your skills.

[river pagoda]

Oh, good
Just I didn’t know where I can to start

[marsh void]

@river pagodaso you know where to start?

[river pagoda]

So so

Idk where, I don’t wanna only study on Kaggle

[marsh void]

https://www.kaggle.com/learn

@river pagoda

here

[river pagoda]

Yes I got it

[lethal token]

hi all, just join the discord here 🙂

[fair mural]

hello everyone

[quartz topaz]

Hi everyone

[crystal elbow]

Hello

[visual locust]

Hi everyone

[vivid moss]

Hello Everyone I just joined 5 min. before messaging

[marsh void]

is there any team here that I can volunteer for ?

if so why don't I create the team?

[bold trench]

Hi, my kaggle notebook just shows Editor Loading...
Can't do anything. Any fixes plz

[pseudo wren]

Hello Everyone! Here to learn stuff about ML and DS.

[tiny wadi]

Hi, Very nice to be here,

[fair mural]

if i have a note book and i need help and someone rate it and tell me my mistakes

[vestal lagoon]

hello guys, any chance someone could teach me about for loops? i can't seem to understand them... im a beginner but want a programming job maybe if you have a job you can give me as well thanks my name is Gupta

[urban gyro]

Y’all are welcome.

[velvet kestrel]

Hello everyone. Good to be here. Looking forward for a lot of fruitful discussions and learning in this journey to master DS.

[ivory ermine]

Hey I want to build a ai system for scientific and Technological discoveries

[verbal radish]

I have a question for submission.csv files:

57543,0.1
57544,0.9
57545,0.5
etc.

in a csv file, is this in one column or in many? If that doesn't make sense, given a pandas dataframe, how do I turn it into this? Thank you!

[knotty ledge]

Hi everyone, I am a green hand of kaggle

I am an AI research engineer majoring in AI for Science. Recently we held a very interesting competition and many teams and people joined us. Here we provide a playground:https://aistudio.baidu.com/projectdetail/7459168?channelType=0&channel=0. And here is the sign-up page:https://competition.atomgit.com/competitionInfo?id=7f3f276465e9e845fd3a811d2d6925b5

Chance of co-author to publish a paper in IJCAI and bonus 35000$ ~

[sudden niche]

hi

[hazy mica]

hello

[green crystal]

hey

[past brook]

hello guys ! I have a question: can we use deep learning for forcasting ?

[bold horizon]

Has anyone without a university degree or masters applied for a internship that requires a degree?

[scenic jacinth]

Hello guys! Do any one of you need to team up to study machine learning together?

[knotty ledge]

Good Afternoon(+08:00 Beijing) everyone, I am here again
I am an AI research engineer majoring in AI for Science. Recently we held an exciting competition and many teams and people joined us. Here we provide a playground:https://aistudio.baidu.com/projectdetail/7459168?channelType=0&channel=0. And here is the sign-up page:https://competition.atomgit.com/competitionInfo?id=7f3f276465e9e845fd3a811d2d6925b5

It has been forked 94 times, come here and join us

[knotty ledge]

https://tenor.com/view/meme-gif-8470516175995282617

[azure lynx]

Hi everyone

[slate hamlet]

Thank you so much!

[fair mural]

hello how you guys doing

[slow whale]

Hi

[tidal pagoda]

how do i update my bio on kaggle's server?

[winter osprey]

Hi everyone

[wind charm]

he, yes me

[tacit vine]

Hello everyone, I need help with coming up with a data product for the monetization team in the Bank I work

Incorporation AI/Ml would be a big plus but not compulsory

If there is any project or paper you can point me to, please do 🙏

[civic void]

Hi. I am computing enginery student, and I want to know more about ia because I have some projects in my school

[torn ledge]

#❓┊ask-a-question message

[wraith garnet]

Heyyy Im a data science enthusiast. Hope you guys are doing good 🙂

[elfin stag]

I have a customer churn model

I did SMOTE to handle the class imbalance

then did ensemble

I manage to get 88 accuracy

what else to do?

[arctic crown]

Hi guys!
I have a dataset that I've been working with recently on a ML project that I'd like to upload to kaggle.
Any tips on what I should look out to make sure it is clear enough for anyone to understand? :>

[nimble iron]

@echo spoke hey! you need something?

[keen badger]

guys suggest me some data sets for clustering

[knotty ledge]

Good noon everybody, IJCAI 2024 Contest: Rapid Aerodynamic Drag Prediction for Arbitrary Vehicles in 3D Space, Sign Up Now! :https://aistudio.baidu.com/projectdetail/7459168?channelType=0&channel=0 baseLine code: https://aistudio.baidu.com/projectdetail/7459168?channelType=0&channel=0

[viscid spade]

Complete the Cloud Skills Challenge: Natural Language Processing (NLP)!

What you will get:

1. Badge from Microsoft
2. ⁠E-Certificate

Challenge Link:
https://learn.microsoft.com/en-us/training/challenges?id=aabad9dd-e4b7-4357-b237-54bd2a91f71e&WT.mc_id=cloudskillschallenge_aabad9dd-e4b7-4357-b237-54bd2a91f71e&wt.mc_id=studentamb_291122

Fill out the form here after completing all modules for Certificates:
https://forms.office.com/r/e3Wn065scn?origin=lprLink

[viscid spade]

Great resource for everyone to learn about Natural language processing in just 30 to 40 minutes. This module will end in 1 day. I have shared it because I learned pretty much from it

[viral locust]

hello, I am a beginner in computer science and my project is: Identification of texts generated by artificial intelligence or by human beings.

I really don't have the qualities to reproduce it I need a mentor or someone to help me
please i really need help please

[reef tiger]

Does anyone have any resources to get into algo trading?

[rich halo]

.paitom has been warned

Reason: Bad word usage

[wicked berry]

Hi everyone
Can you help me please select csv dataset for my study project
It must contain uncleared data, that will be used to create regression or classification model

[stiff musk]

Hello all, I’m a program manager for AI

[chrome pike]

anyone can help with the flood prediction competition?

[deep carbon]

What help do you need ?

[chrome pike]

the data is too hard to do a heat map on it.
do you think i should normalize the train data before doing anything?

[chrome pike]

the data is too hard to do a heat map on it.
do you think i should normalize the train data before doing anything?

[deep carbon]

I also tried heatmap but it was very wierd 😅

It's always good to perform standardization

[chrome pike]

i need to do this project, for a final class.
do you have any recommendation to do with the data before i start? because the data seems really weird to me right now and i cant work with it.

[deep carbon]

I also checked pca but there's no sense, for this perticular data.... So I directly jumped for model building 😅

Lr and xgboost gave me best r2 score

[vestal belfry]

learner

[wary night]

hi, can somebody tell me any limits, restrictions on kaggle

cause i want to start using it but im scared i might get a ban

[short wedge]

Each competition has it's own rules you should read. In general, don't create multiple accounts to get more submissions, that is always against the rules and will result in banning. But generally if you follow the rules you have nothing to worry about!

[wary night]

Yeah sure, but for example if I host Stable Diffusion(finetuned) I get banned

[arctic plume]

is that what you were intending on doing?

[rich halo]

bechir00 has been warned

Reason: Bad word usage

[worn cave]

hello is any one there

[devout slate]

Hello guys! Ghulam Hussain here, I am passionate about data analysis and data science. Currently developing skills in Python, data manipulation, and statistical analysis. Eager to contribute to the Kaggle community by participating in discussions, exploring datasets, and learning from experienced Kagglers.

Areas of Interest: Data Analytics -Data Science -Machine learning

[sand vault]

How's everyone doing?

[elfin stag]

overfit?

[sudden gorge]

hello my fellow kagglers

[worldly knot]

Hi all, I am new to Kaggle and here to learn more.

[worn cave]

is any one this server participating in RSNA???

[acoustic mica]

Hi everyone

im new to ML but ive spent quite a while learning Tensorflow, and the basics, my plan now is to learn NLP and create a model, before moving onto something more intense

[sand vault]

What is an "Over fit" or are you looking Correlation better variables using I think it's called the coefficient matrix?

[acoustic mica]

overfit is basically when

the model learns the data too well

and starts finding patterns that arent actually there

[fringe vigil]

Hi, I am new to ML. Currently understanding concepts and soon would be doing some hands on

[thorny knot]

For those who are using (or interested in) Lightning Studio, I have discovered a way to install LaTeX Compiler (similar to Overleaf, but free and faster than the free version of Overleaf) on Lightning Studio. If many people are interested in this installation method, please react, and I will publish a notebook on Kaggle about it.

[sand vault]

Oh wow I didn't know that was a thing. Thank you for taking the time to explain it to me. I appreciate it.

[fleet zodiac]

Hi, I am Alin Castillo. I would like to improve my machine learning techniques.

[dusk tree]

hi

[tribal granite]

hello everyone. I am Grish Rana. Straight outta Tanahun, Nepal.

[shell frost]

I'm manager of a start-up agency. I want to meet english natives for improving english level of me and our developers. I can help you to solve your financial problems, if you're due to it. 🙂
Feel free to DM me and let's discuss in details.
Preferred

• Familiar software engineering
• 25+ years old
• US citizen

[dusk tree]

i am looking for freelance oppurtunities in field of llms

[tender drift]

URGENT!!! Please Help!
hello @everyone, I get access to meta's Llama3-8b. But how do I get my access token? Any idea..Please help its urgent!!

[thorny knot]

Do you know any websites or sources that share short harmful videos? Or are there any tricks to find a lot of harmful videos on TikTok? I need this data to train a harmful video detection model. If anyone knows, please share some sources with me.

[obtuse creek]

nah

[obtuse creek]

like gore type videos ?

[thin wyvern]

there are some subreddits where you can find them

[light plank]

I am looking for some good in AR / VR ... want to create a team for a hackathon , if someone you knew or any one let me know

[livid barn]

too many garbage LLM competitions these days

[elfin hatch]

Some examples ?

[empty current]

Hello!

I'm new to this server, and also quite new to machine learning. If I have a computer-vision related question about a chemistry project, which I think needs both segmentation and classification, where do I ask - in #❓┊ask-a-question or in #computer-vision ? Thanks in advance 😊

[short wedge]

@empty current Either is fine, more people are hanging out in ask-a-question so that might increase your chances of getting a response. Typically the more specific and clear your question the better.

[civic frost]

Hello i am a beginner and wants to be a data scientist i needed some guidance for a career path and some questions related to it

[thin wyvern]

I mean it depends on what types of harmful videos they're looking for

[empty current]

Thank you very much! I will try to make a graph to illustrate my problem and post in #❓┊ask-a-question later : )

[tidal ibex]

I am looking for a partner with whom I can learn deep learning. Ideally, this person is as passionate about artificial intelligence and data science as I am. Together, we can explore neural networks, dive into complex algorithms, and work on innovative projects that push the boundaries of what's possible. If you are enthusiastic about mastering deep learning and eager to embark on this educational journey with a dedicated study buddy, let's connect and start this exciting venture.

[fickle shuttle]

Hello everyone

[lost sail]

Hi, my name is Mário. I'm Brazilian and I want to train and improve my skills in Artificial Intelligence.

[shut saddle]

hey, I'm doing text sentiment classification (just positive and negative). so far I'm just using svm and rf on tfidf, would someone have a suggestion of a model that works well? thanks in advance

[placid quartz]

So, while using hf trainer I set eval accumulation steps to 1, the evaluation speed is ridiculously low, is there any workaround for this?

[dry sigil]

hello

[wicked berry]

I’ve applied to Social Summer Of Code Season 3 and want to get more good builders to participate!

Use this link: https://ssocseason3.devfolio.co/?ref=7a4f1d3ac4 to apply, if you're in for a building spree.

[fading grove]

i keep getting this a lot, any help?

[wild holly]

how to upload models from huggingface to kaggle？

[teal glacier]

I am searching for a team partner for Kaggle competitions. Anyone interested, please DM me

[teal glacier]

Hi Mário! Nice to meet you. It's great that you're looking to improve your skills in Artificial Intelligence. If you need any help or want to collaborate on projects, feel free to reach out

[wicked berry]

Hello everyone. I'm Sanjula and I'm starting my journey towards becoming proficient in AI/ML. Nice to meet you.

[lost sail]

Hi MVikram! Thanks!

[teal glacier]

@wicked berry welcome! It's great to have you here

[eternal shuttle]

hi everyone new to kaggle AI and DS student

[wicked berry]

Hello everyone

[fickle berry]

oh hello.

is anyone there?

[upbeat violet]

Hi everyone, I'm trying to submit a file for the flood regression competition but it keeps giving me the submission file not found error. I've tried making it submit an exact copy of the sample submission file but I still get the error. any help at all would be appreciated!

[upper glen]

Hello everyone. I'm Chris and I'm beginning with AI/ML skills. Nice to meet you 🙂

[south rain]

Hi everyone I am Priyansh Singh. Currently I am learning AI/ML.

I am new in this feild and I am starting with
IBM AI Engineering Professional Certificate course on Coursera(https://www.coursera.org/professional-certificates/ai-engineer)
any suggestion?

[plain galleon]

Hi guys, New in the field and trying to get in the society with after completing Simplilearn Master' Data Scientist course. Currently experimenting with deep learning models and make my base stabler.

[fair cairn]

yo

[wicked helm]

hello @plain galleon where r u from ??

[upper sequoia]

Have you any one doing the Titanic prediction

[edgy finch]

Hi everybod, I'm new in this group. I'm full stack developer and trying to change a data scientist

[hallow jetty]

Hey, Priyansh. I'm thinking about taking an AI/ML course too. You think it's worth it?

[south rain]

I am not that sure about the course but rating is good and it will give basic idea about AI/ML.
I am following this course

[hallow jetty]

I only ask because I'm not the type to pay $49 per month on coursera lol

[exotic stratus]

hi

[south rain]

@hallow jetty bro you don't have to purchase that course you can simply audit that course and learn for free.
EVEN I can't spend $49/month 😂

[wicked helm]

yuss...

[snow jetty]

Hii everyone

I am ruchita and currently gaining experience in AI /ML

[upbeat violet]

wait I'm trying to learn and the option to audit is gone????

[south rain]

Check for individual courses
Most of them have audit option

[thorny knot]

I am training a Video Classification Model (from transformers import VideoMAEImageProcessor, VideoMAEForVideoClassification). It is working correctly in training phrase, but in validation phrase, it only use CPU for validation, no GPU is used. Below is the link to notebook. Thanks for helping!
https://www.kaggle.com/code/an1001/tiktokvideoclassification?scriptVersionId=180260409

[placid quartz]

So, while using hf trainer I set eval accumulation steps to 1, the evaluation speed is ridiculously low, is there any workaround for this?

[fluid rose]

Hi, im new to kaggle and ML, and ive just started learning these. I found a few udemy courses on these and i thought of buying them for the certificates

i just wante to know as to how valuable it is for me to buy the course just for those certificates, like does having these certificates really help me in my shortlisting process in interviews? im currently in 2nd year of my engineering in computer science so im asking this with respect to my college placements

[light gulch]

does anyone knows about liquid neural nets available to deploy ?

[stray nest]

guys! I had been trying to motivate myself to work on Kaggle competitions a lot unfortunately being single ain't giving me that.

if someone is looking for a seasoned guy in Computer Vision and Natural language processing, I am happy to be part of your team

[wicked berry]

Hi

[empty current]

I can give you a real-life task on Computer Vision to motivate you 😂 interested in training a net that can determine the size and shape of objects in black-and-white images? 🔬
Like automatic thresholding, just better because of dealing with background noise etc.

[stray nest]

alright LOL can you be more specific?

[empty current]

#❓┊ask-a-question message

^it's basically this

[stray nest]

@empty current I think I have an answer for you

You require contour algorithms with morph filter with erosion from OpenCV to remove those minute background noise

I'm meaning how you don't want the particles to stick in the segmentation

then a better approach is to use median blur and then use KMeans clustering with higher numbers of clusters to get those segmentations

your problem is basically a background substraction problem and watershed like algorithms won't be effective in this case

[wet violet]

can anyone explain why this is happening?

I am trying to do this one: https://www.kaggle.com/competitions/spaceship-titanic/data

[regal pier]

How could anyone do that from what you shown?

[civic valley]

is it because:

X_test = torch.ones(1, n_values)

this looks to be like there is something wrong in your 'svm.predict' function. possible returning the bias, or weights, instead of the calculated result.

[wet violet]

im not sure how I can send code through here

should I just copy paste it

[empty current]

Thank you very much for your suggestions! Can I DM you later to discuss this more in detail? I don't want to pollute the chat because maybe it doesn't interest the other participants here

[robust lintel]

Hi

I am a beginner in ML

I just submitted my first ML model. Ofcourse it was the titanic one. And it's a disaster

Sorry I just realised there's a titanic channel. Imma go there.

[stray nest]

Sure! I don't have time to actual write the code for you. But, I am happy to guide you

[runic owl]

KaggleX fellow program application shows coming soon !

[tight quail]

Hi @runic owl the first step of the application is to participate in the skill assessment challenge (beginner Kaggle competition). We encourage you those who are interested to get started today. 🙂

[runic owl]

Okay, thanks !

[tight quail]

Here is the link, KaggleX Skill Assessment Challenge: https://www.kaggle.com/t/005ae82d6fda45a9a3048d900fd2eb17

[halcyon fern]

Hi what preparation do you recommend before attempting the challenge?

[tight quail]

Hi Jayendra,

We encourage you to do your best. Please note that your score will not reflect your application status, so we ask that you have at least one submission.

Here are some links to references before you get started:
Take a look at a few videos to learn a bit more about site etiquette, Kaggle lingo, and how to enter a competition using Kaggle Notebooks.

Check out the Titanic - Machine Learning from Disaster beginner competition.

[halcyon fern]

Thank you so much!

[real rover]

Can i apply in KaggleX Fellowship from Bangladesh?

[tight quail]

Hi Asib, we're only able to open the program to fellows who are from a limited set of countries due to legal requirements outside of our control.

Here is the current list of eligible countries: Argentina, Australia, Austria, Brazil, Canada, Colombia, Croatia, Denmark, Finland, Ghana, Greece, India, Italy, Japan, Kenya, Mexico, Netherlands, Nigeria, Peru, Philippines, Poland, Portugal, Rwanda, South Africa, Switzerland, Turkey, UAE, Uganda, Ukraine, United Kingdom or United States

I've noted Bangladesh for the request list. We hope to expand the list in the future.

[real rover]

Understood. Thank you

[mild trellis]

Bro how often is the kaggle x program released?

[wicked berry]

Hello everyone!

[jagged chasm]

anyone want to study ml

[crystal ermine]

Hello everyone. I published several datasets on Kaggle, everything helping in building machine learning systems for Ecommerce platforms. Can you guys please take a look ? I wanted feedback and possible improvements. https://www.kaggle.com/kartikeybartwal/datasets
Thank You 😊

[timber dew]

Will there be a channel for the RSNA comp?

[blazing urchin]

hey kaggle team, nothing important but there is small type on website
link: https://www.kaggle.com/kagglex/#prospective-fellows

[thorn gyro]

typo?

[thorn gyro]

Hi, its an individual assessment challenge and not to be done with a team, right?

[green mortar]

Hello, how many hours a week is a Fellow expected to work for 15 weeks in the KaggleX Fellowship Program?

[south oracle]

why am i so slow at tensorflow, etc? ;-;

how to get faster at debugging when there isnt a clear error message

eg program crashes

[sharp ferry]

best thing to do when using TensorFlow is , just start using pytorch

who has infinite time , to resolve issue which make no sense , always broken TF

[south oracle]

is tf just worse?

[sharp ferry]

yes it is

its allergic to windows as well

[rich halo]

tribino85 has been warned

Reason: Posted an invite

[earnest delta]

Can I apply for KaggleX Fellowship program from India as a college student?

[plucky vector]

The application are not open yet right for kagglex fellowship?

[ornate saddle]

what's the best way to learn DSA. I am a high school, student, and i like computer science, but recently i feel tired and confused while learning DSA

[hot plover]

what is kaggleX

[empty current]

#announcements

[wicked helm]

follow a cource and code

[karmic badge]

Uhh...what is dsa

Oh

I found most ppl recommend python

For dsa,llm,data analyst/science/engineering

Learn basic..you can learn basic from MIT operncourse on yt

Like learn object and how does each statement functioning and structured

And in data science you need to learn string manipulation

As you gonna need to use them most of the time ig..and things like tuple,list, dictionary

Then you can go advance like data cleansing, pipeline,web scrapping

Then algorithm

Also python modules

[misty oxide]

Hi everyone! I'm a data scientist who works in forecasting but looking to expand my skillset and do some competitions! Looking forward to learning from everyone 🙂

[raw mica]

hi

[sharp ferry]

hello

[south rain]

hi

[distant reef]

Hi everyone! My name is Ana Paula, I'm a Data Analyst from Brazil, and I'm studying to become a Data Scientist.

[sharp ferry]

good luck

[tight quail]

Hi @thorn gyro the skill assessment will be completed individually.

[tight quail]

Hi @plucky vector, there are two parts of the application.

Step 1: Review and verify that you meet all of the program eligibility requirements
Step 2: Complete to skill assessment challenge: https://www.kaggle.com/t/005ae82d6fda45a9a3048d900fd2eb17
Step 3: Complete the application form.

We're currently experiencing issues with the application form link. We highly encourage you to get started on the skill assessment in the meantime. We will reach out once the application form is ready. Thank you for your patience and understanding.

[plucky vector]

Thank you for informing!
😊

[fluid rose]

Hi, im new to kaggle and ML, and ive just started learning these. I found a few udemy courses on these and i thought of buying them for the certificates
i just wante to know as to how valuable it is for me to buy the course just for those certificates, like does having these certificates really help me in my shortlisting process in interviews? im currently in 2nd year of my engineering in computer science so im asking this with respect to my college placements

[carmine hearth]

Hello everyone!! I'm Harshit just another explorer of Data Science

[small mirage]

Hello ^^

[south fossil]

Hello everyone!

[wispy harness]

I can’t edit certain cells in my kaggle notebook, is anyone experiencing the same issue?

[night sail]

Hey guys. I want to rank up on Kaggle. In order to become "Kaggle Expert", do I have to become "expert" in every category (competitions, datasets, notebooks, discussions)? Or just becoming "expert" in one of those categories is enough?

[night sail]

Hello. The certificates like Coursera and Udemy have small to zero effect on your interview. It is not relevant in the job market. Some of them can give you great learning start into the field. But I would not invest big money into that. There is a lot of free content. Good luck on your learning journey!

[tacit token]

Just One

[night sail]

Super. Thanks

[empty current]

I had this problem once, too. I copied some code from a website and when I wanted to paste it, no cell was editable anymore. I pasted the clipboard content to a plaintext editor and copied it from there again, and after that it worked again 🤷
I guess the usual things like refreshing the browser window you tried already?

[valid musk]

Can we check out old projects done on KaggleX old cohorts?

[short wedge]

@valid musk Check out: https://www.kaggle.com/bipoc2021-project-showcase and https://www.kaggle.com/cohort2-2023-project-showcase

[valid musk]

Thanks @short wedge !

[woeful willow]

Hello friends how does one actually get a kernel to start using the gpu from kaggle?:

[wet nacelle]

is anyone experienced willing to help mentor me, ive done the basic competitions of titanic and house prices and i also did some predictive maintenance project but im kinda lost

[tacit token]

The GPU will start automatically when tasks that require it are initiated, such as deep learning (especially image classification). If your GPU isn't working during image classification tasks, be aware of potential bottlenecks between the CPU and GPU. You can Google this for more information.

[south oracle]

is it normal to feel frustrated in this field or am i doing something wrong?

[sharp ferry]

its not about this field particularly but everything related to code is frustrating

unless you are using TF for ML

then its your choice to live frustrating life

[cursive heath]

Hi, I'm CallyI3D an architect exploring AI/ML. I just got on to Kaggle and that Pepper🌶️ emoji is giving me a hard time. any ideas? i hope to get used to this platform to learn and upskill.

[south oracle]

i heard that tf is used more in industry. Is that false?

[sharp ferry]

may be too old news

transformers are based on torch

[south oracle]

Can you, in theory, get good scores on a competition by just using ensembles on others' submissions without understanding the problem at all?

And does this have carryover to job skills?

[sharp ferry]

if you get something to work , then it has value
no matter what or how you did

but interviews need more theoretical knowledge even if the jobs won't require that much

in the end you just gonna copy paste most of code anyways

[fair glen]

Hey everyone,

As part of a University project, I'm conducting a survey to understand how incorporating domain specialists into species distribution model development would affect user trust and decision-making. The survey has 5 multiple choice questions and should only take a couple of minutes. All respondents are anonymous.

I'd greatly appreciate your time taken to complete this survey, your responses will be valuable for my project. Thank you.

https://www.surveymonkey.com/r/9WRDCBH

[flat depot]

Hi @tight quail , I am not sure if the ranking on KaggleX skill assessment challenge can affect my application process? Thank you.

[shadow cove]

hey Can you guys suggest me best youtube for learning and developing ai ml

and also discord channels where voice channels are active for ai ml developers

[tight quail]

Hi Quang, thank you for your question. Your score for the KaggleX skill assessment challenge itself will not reflect your application status, but please do your best. We also consider your responses in the application form.

[flat depot]

Thank you so much for your information.

[solemn kayak]

wsp

[wispy harness]

I’ve tried refreshing it many times, but it wouldn’t work. This issue just occurred out of the blue. But I’ve solved it just by downloading the notebook and importing it again

[wicked berry]

feeling silly NGL

[inner minnow]

Hi everyone! I'm new here, I hope to learn a lot about machine learning from you guys!!

[south oracle]

Why do you need to write all loops by yourself in pytorch?

seems like i just need to save my own function for the basic default calls

[sharp ferry]

because pytorch wants to give you full control

not like TF where you just model.fit()

btw there are multiple pipelines from huggingface which you can use to save the work

[rapid timber]

Hello people

[rapid timber]

Found you 😂

[solemn kayak]

codeyrun

Im out

jkjk

[karmic badge]

found you too

[wicked pivot]

Hi Kaggle team I am a latino from Colombia, I would like to know if I can apply for the KaggleX Fellowship Program, thanks in advance

[shell delta]

Hello everyone.

[fading blaze]

Hello

[undone gyro]

hello

[south oracle]

yea but if youre making a relatively basic model, it seems overkill

[sharp ferry]

yes , it is

but its a good practice , as you mentioned before you can save that code in a file and import train function which will take all neccessary arguments

def train(model, train_dataloder, test_dataloader, lr, epochs, path_to_save):
  # training block 

[plucky rock]

hello

[spice creek]

Hello 👋🏻

[hot horizon]

Hello

[green glade]

hello everyone i am a 2nd year electrical engineering student
I have good understanding of non-convolutional neural networks and I have also made a project on it (image classification).
I had a doubt if I should go for generative ai or continue learning computer vision
i have interest in both the domain but generative ai seems fancier but ig it's harder
please help!!

[south oracle]

do both, youll find the answer lol

[rapid timber]

Imagine not using java for every kaggle problem 😋

[sharp ferry]

tbh, i hate java

[rapid timber]

lmao

[rapid timber]

guys I got a 0.765 on the titanic thing 🔥

@sharp ferry and I used java 😂

[sharp ferry]

why are you using java for ML, don't you like to be happy in life

[rapid timber]

Never.

[south oracle]

iirc gender submission gives you 75

cant we make a virtual machine version of python or cpp so that we can get rid of java for good?

[rapid timber]

and???

Don't be supportive of people doing it for the first thing or anything nooo just kill whatever joy they get from it 😁

[south oracle]

sorry, congratulations
i didnt mean it ike that, i just tried submitting the gender submission to see if mine was actually that much of an improvement so i thought itd be an interesting thing to note if just starting out

[rapid timber]

Now I feel bad 😭

It's fine

[rapid timber]

Yeah but this is gonna piss my friend off watch 💀

[south oracle]

apparently there are a few notebooks with over 80 percent accuracy but 77 was the max i could manage lol

with a neural network

https://www.kaggle.com/code/ldfreeman3/a-data-science-framework-to-achieve-99-accuracy/ somehow this improves accuracy, didn't read it yet
but i wonder if it's just overfitting

[rapid timber]

How are people getting 1 on the titanic thing then 💀

[rapid timber]

i used support vector machine

[south oracle]

the test dataset is publically available

so they just read directly from it

[south oracle]

i see

88 is the maximum legit accuracy it seems

achieving an accuracy of 1 legitimately makes no sense anyway, because it kind of implies that the passengers who died on the titanic were predetermined, which is absurd since there's obviously a luck aspect

[rapid timber]

Lets just find information on the people who were on the titanic to see if that was influenced 💀

The only possible thing could be some conspiracy where all the people on it were pushed to get on it

[south oracle]

from what i remember, class 1 (whichever the richer class was?) passengers, women and children were more likely to survive, which is something you can get from the EDA, and those factors are actually the basis for probabilistic prediction

in every knowledge competition that doesnt have an enddate, the top accuracies are always 1 because the test datasets are publically available

[rapid timber]

Why don't they ban 100 % then 😭

[empty current]

There is an FAQ or something where they explain why they decided not to remove them

[rapid timber]

Definitely answers my question 🔥

What FAQ where

[coarse onyx]

I want some upvotes 👉👈 😁

[rapid timber]

https://tenor.com/view/my-goals-are-beyond-your-understanding-reverse-flash-eobard-thawne-the-flash-gif-16637212

[south oracle]

i did sentiment analysis in java because i had to do something for my java course but

isn't python what's used in every ml job?

[silent linden]

would be really great if anyone could upload a whisperx whl to kaggle

i have trouble download it

[unique fjord]

was running a notebook and noticed it has issue importing tensorflow ?

[wicked berry]

Damn that sucks

Wish I could help

[wicked berry]

Hi,
Anyone worked before on wind speed forecasting using deep learning ?

[plucky vector]

Why do the error score change after running the same code for training again?

[rapid timber]

I use java for everything, do you need help?

[rapid timber]

Not exactly, it's just very popular for it

[rapid timber]

Random seeds or algorithims that don't produce the same outcome every time

[south oracle]

no, i already did it

i just dont like java because for some reason it makes my low end pc perform bad

the virtual runtime environment and all

[rapid timber]

The JVM?

I'm not entirely sure if python is going to be faster

End goal is usually accuracy for kaggle though not speed

I'm sure the fastest as usualy would be rust but I don't know how much ML support it has

You know what, I know a guy lemme ask him.

[south oracle]

it just makes my pc lag for some reason

im not talking about the actual speed

[rapid timber]

Ohh

[south oracle]

i dont see any reason to switch to java anyway

[rapid timber]

You don't have a dedicated like

ML Set up

[south oracle]

ik python for high level stuff and cpp for low level stuff so

[rapid timber]

Have you tried kaggles GPU's?

[south oracle]

yea no i wasnt talking about ml actually

intellij in general makes it lag for some reason

[rapid timber]

😢

[south oracle]

maybe some bug or something

[rapid timber]

That's just heavy jet brains ide's for you 💀

What are your specs

[south oracle]

no gpu for one

8 gb ram

[rapid timber]

Is this on a laptop or?

[south oracle]

and this processor lmao

laptop

[rapid timber]

Hmmm

[south oracle]

im surprised i can run basic games like valo somehow

maybe cuz it's new gen?

[rapid timber]

Because valo supports low end laptops

Uhhh

You don't care about speed

[south oracle]

i guess

[rapid timber]

But intellij eats ram

[south oracle]

i only ran it for my college assignment

[rapid timber]

I have a random idea

[south oracle]

im staying away from java for now

i see no reason to use it

[rapid timber]

WHY I LOVE JAVA

Theres really not unless you're like me and wanna master EVERYTHING in java for no reason

I do have a real reason for ML in java though

FTC/FRC

[south oracle]

it's just a waste of time to learn a new language when i already know ones i can do all my stuff on

[rapid timber]

Which is java or?

[south oracle]

just python and cpp

that combo takes care of most of it for me

[rapid timber]

So yeah, no real reason to go learn java just for ML

[south oracle]

i did learn basic C# for unity but i dont use it anymore
which ig is just microsoft java

[rapid timber]

Java has some advantages for some situations, pretty sure ML is not one of them right now.

[rapid timber]

Same, don't really like it

We had a game design class that forced me to use it

Which fun fact you can use java code in unity if you try hard enough 💀

[south oracle]

languages are tools ig

i dont even follow tutorials basically these days, i just use docs

and maybe one reference code or something

[rapid timber]

SAME

I had to escape tutorial hell forever ago

[south oracle]

i also dislike java because it requires you to define everything unlike python, but then also doesn't give you full low level control unlike cpp

[rapid timber]

I hate dynamic typing

And it can give you low level control with some classes

It can even use cpp code

[wicked berry]

any one please help me in 1 computer vision task

[atomic cove]

Hi!! I'm Ashani, a second-year undergraduate of BSc (Hons) in Artificial Intelligence. Looking forward to learning from everyone..

[wicked berry]

Absolute masochist💀💀💀💀

[trim mulch]

Hi there this is Eriel!

[sharp ferry]

and it will never be ,

[sharp ferry]

hello

[rapid timber]

I've been using DL4J, Weka, and Apache MLlib with no problem at all

[sharp ferry]

start using python and you will get to see problems with java

why make things complicated

[rapid timber]

I mean

Maybe concurrency and performance?

[rapid timber]

FTC 💀

[rapid timber]

I have used both, pythons way easier for everything as usual but I need to use java

[sharp ferry]

why

[rapid timber]

I told you I'm gonna use it in FTC like we have to every year

Robotics competition that requires java since the robot runs on Android and they want us to use it for some reason

Yk I should ask

[rapid timber]

Possibly train model with one language and use it in another?

[sharp ferry]

yes you can use model in any language once you got the weights

what is FTC though ?

[sharp ferry]

ok

[rapid timber]

All I know how to do is reuse the model in the same language with serialization right now 😭

[sharp ferry]

don't you use c or embedding hardware for that purpose

[rapid timber]

Not sure how to export it into something I can use in another language

[rapid timber]

The SDK provided is in java and Android uses java

[sharp ferry]

you have to build whole model to use it , simple as it is

[sharp ferry]

ohhk ,

[rapid timber]

But like we've had to use models before that were literally in the form of a tar.gz for computer vision

Which Ik is just compressed but

[sharp ferry]

don't know about computer vision , CNNs but llms are just matix multiplications , once you have weights

[rapid timber]

what's cnn again

I forgor

[sharp ferry]

used to train images classifications models

or anything related to images

[rapid timber]

Does it stand for something

[sharp ferry]

convolution NN

[rapid timber]

Ohhh

I also wanna make

I think it's it's called a large action model?

[sharp ferry]

never heard of anything like that

[rapid timber]

Instead of generating stuff or whatever it's given a set of things it can do it just has to decide which to do

Im basically replacing the robot driver entirely 💀

[sharp ferry]

more like Reinforcement learning

[rapid timber]

Yeah

I'm gonna have to make most of the data myself

And then do the reward bs with the AI

Right now I'm making a stacked ensemble learning program with 10 fold cross validation for kaggle

Funsies

I plan to use outside of kaggle one day hopefully

Gonna add some more fancy stuff for the decision tree algorithm because right now it is absolutely useless 🔥

[sharp ferry]

best of luck ☘️

[rapid timber]

ty

[rapid timber]

Can I show you the stats it's outputting ? 💀

[sharp ferry]

why not

[pale dune]

hii

[hazy hedge]

Can anyone recommend some large datasets available on the internet for linear regression and predicting something?

[karmic badge]

You lookup for that on kaggle

[karmic badge]

Steam games data ....it has 24x42k

Somewhere 900k cells

[sharp ferry]

what's that data about ?

[karmic badge]

Oh linear regresion right

Mb

[sharp ferry]

i mean how does that data looks like ?

features ?

[karmic badge]

Games review%....price,price after discount, game genres

[sharp ferry]

ohhk 👍

[sharp ferry]

but isn't it hard to find relation ?

[karmic badge]

I do not think so

[south oracle]

I had this really weird idea the other data for making a model that predicts a person's BMI or weight from an image

[karmic badge]

Like game genre=sold%

[sharp ferry]

their no pattern for price and reviews

[south oracle]

Not sure where to find data for that though

[sharp ferry]

weird idea , be weird enough to create data

[south oracle]

probably famous people

can scrape their images and weights

[sharp ferry]

that can be a feature to consider

[karmic badge]

Or you can corelate overall review% with sold%

Idk ...i didn do machine learning

Didn even know knn algorithm

[sharp ferry]

won't be a good distribution of data ,

[sharp ferry]

Know your Near Neighbours

[south oracle]

Do you think it would have any practical uses, though?

[sharp ferry]

not at all

why predict from photo

[karmic badge]

Hmmm...ive tried to use sql to extract the steam games data

But its didn come to fruition

[south oracle]

will it good look on resume if i scrape everything from scratch? I could demonstrate skills for a lot of ml related jobs, from data collection and analytics, feature engineering to model design

because everything else that i put on resume, someone else will have done better than me

[sharp ferry]

for what purpose

[karmic badge]

My project for my portfolio

[sharp ferry]

yup it is

[sharp ferry]

i don't think it will have any value

if you gonna invest time , do something worth or something to learn even if copy

[south oracle]

what should you even put on your resume? good models already exist for every good dataset out there

[sharp ferry]

CS student ?

[sharp ferry]

well just build transformers from scratch , train a model using only numpy

try different techinques to improve loss on pretrained model

[south oracle]

whats the point of that if youre never gonna do it on the job though?

[sharp ferry]

you can't build something extra ordinary that easily , reaserchers wasting their life still coming up with hopeless papers which will go to dust in few. months

[sharp ferry]

if you know how it works , then you can do most of the things you will get asked on job

like you will have an idea of limitations and improvements

[south oracle]

how would you stand out from all other applicants though?

[karmic badge]

My major is computer networking...also just recently get into data science realm bcuz of i considered it to be the forefront in the industrial environment+ im hired to work with team of data science bcuz the HR thot i wan the engineer role itself and they didn really any kind of project that need for network engineer

[sharp ferry]

have good projects on github , use latest tech , try using/implementing latest papers

or write good blogs on medium or HF

[sharp ferry]

lol ,

you want to pursue career in networking or in ML then ?

[karmic badge]

And not to mention that networking company have started to do network automation, involving ci/cd pipeline

[south oracle]

latest tech as in newer frameworks? or just the standard pytorch

[sharp ferry]

even in networking field you can not avoid this much of automation and python

[south oracle]

can apply ml to computer networks (botnet detection)
that was my prof's research topic and he put it on the exam

[sharp ferry]

read research papers , JAX is new ML framework which is promising for future works

[sharp ferry]

as long as your data has some meaning , but really i don't have any idea on this

[karmic badge]

I can go for both..but yh i need to learn math like calculus ,etc

[sharp ferry]

you can not focus on both after a certain point , ML is just getting broader

[boreal dust]

ah is there a go-to place for ML research papers? I've mostly just been studying the bread and butters of ML so I don't have a lot of exposure to new technologies

[sharp ferry]

https://arxiv.org/abs/2402.19155

[karmic badge]

Data science/network

[sharp ferry]

arxiv website is home for all research papers

[boreal dust]

ah sweet

[sharp ferry]

networking looks cool though ,
i like watching the guy networkchuck he's awesome in that field

[karmic badge]

Like this ```python
import os
from dotenv import load_dotenv
import requests
import click
from pprint import pprint

load_dotenv()

class Trello:
def init(self):
self.trello_api_url = "https://api.trello.com/1"
self.trello_api_key = os.getenv("trello_api_key")
self.trello_api_token = os.getenv("trello_api_token")

def get_all_boards(self, boardname=""):


    url = f"{self.trello_api_url}/members/me/boards"

    headers = {"Accept": "application/json"}
    query = {"key": self.trello_api_key, "token": self.trello_api_token}
    response = requests.request("GET", url, headers=headers, params=query).json()

    for board in response:
        if boardname == board['name']:
            print(board['id'])

[sharp ferry]

just a lazy devloper , don't wanna bother opening trello for looking at tasks

[south oracle]

do they have to be useful papers or just any papers?

[sharp ferry]

unless you have too much time ,

[south oracle]

like, revolutionary papers

[sharp ferry]

just search for popular papers and you might get some recommendations

[south oracle]

alright, thanks

[karmic badge]

my current data

is consist of this :

"Store Number": Which likely serves as a unique identifier
"AreaStore": Denoting the store's physical area in square meters
"Property": Specifying whether the store is owned or rented
"Type": Classifying the store's category
"Old/New": Indicating whether the store is a new or established one
"Checkout Number": Potentially representing the number of cash registers
"Revenue": Revealing the financial performance or earnings of each store

what you might suggest i can do with this?

for data science project?

@sharp ferry

im srry with the ping if you didn feel comfortable with it

[sharp ferry]

its ok , i won't mind : )

[sharp ferry]

regression problem , find out revenue depending on those features ?

[south oracle]

Can someone help me understand why my code is running out of RAM? I was doing the facial keypoints recognition competition

[timid crypt]

https://www.kaggle.com/code/abhishek0032/marathon-analysis-indian-atheletes hello everyone i have created a project check it out and if you have any suggestions let me know if you like it please upvote

[sharp ferry]

don't know if there is way for it to reduce ram usuage like reducing batch size while training to prevent OOM

[south oracle]

I posted in #❓┊ask-a-question
Does using dataloaders use less ram? I was just iterating over the entire dataset

Im more concerned with how the RAM's going that high because the dataset is only like, 200 MB?

[sharp ferry]

set low batch size

[south oracle]

Wait, wouldn't RAM already be used when I read it into a pandas dataset though?

[sharp ferry]

batch size = 16 or 8

no

[south oracle]

So it is stored in disk until it's used?

[sharp ferry]

yes

[south oracle]

same with the pytorch dataset?

Because like I said I just iterated over the dataset without using a dataloader. Does that load the entire dataset into RAM?

[sharp ferry]

just set batch size to lower number

[sharp ferry]

you are using data size of 64

which is too big

[south oracle]

that was for training
didn't have any problems there so i commented out that part
i'm not even using batches for submission, which is where the RAM is running out

I'm not using that variable anywhere in the submission code

[sharp ferry]

ram only keep track of current usage other stuff gets into heap , or swap memory

[south oracle]

But point stands, I'm not using that data size variable for submission so it's gotta be something else
had no issues in training or validation

[sharp ferry]

don't have idea what you doing

[sharp ferry]

don't know about kaggle but colab sometimes don't free up GPU vram , so needs to restart

[south oracle]

that might be it

[karmic badge]

i got this

---------------------------------------------------------------------------
ModuleNotFoundError                       Traceback (most recent call last)
Cell In[3], line 1
----> 1 import pyspark
      2 from pyspark.sql import SparkSession
      3 from pyspark.sql.functions import col, expr

ModuleNotFoundError: No module named 'pyspark'

[sharp ferry]

pip install it then

[karmic badge]

this is on kaggle

you can do that on kaggle?

[sharp ferry]

yes , you should be

wow , you believe in safari

[karmic badge]

how does kaggle notebook interact ?...like i alr have it installed on my mac

or its smt like container/virtual

[sharp ferry]

it should be normal jupyter notebook then

[karmic badge]

where the installation doesnt efct?

[sharp ferry]

you installed what ? you mean downloaded notebook right ?

[karmic badge]

okay like i alr hv the modules installed on my mac ...that i can simply import the module whenever i open a python file

[sharp ferry]

yes

[karmic badge]

so why does kaggle notebook asked to install

like does the environment in kaggle is different?

[sharp ferry]

you are using kaggle notebook on kaggle website or in your local vs code

[sharp ferry]

are you there ?

[karmic badge]

yh

so i think the environment is differen

im using kaggle's notebook

bcuz whenever i start a fresh new session its aksed me to install it

[sharp ferry]

well ofcourse it will if you are using kaggle notebook online

whenever you start a session you get new server assigned to you which will get cleaned after you stop using it

so you have to install dependencies everytime you start a session

[thorny knot]

I am training a Binary Object Detection Model, but I don't know how to set the bounding box for the background class. It means if there is no cardiac area in the image, the model shouldn't create any bounding box for this image. Thanks for helping.

Below is the related source code:

self.detection_model = FasterRCNN(
            self.backbone,
            num_classes=2,  # Binary detection: presence or absence of cardiac area
            #rpn_anchor_generator=self.anchor_generator,
            #box_roi_pool=self.roi_pooler,
            min_size=256,
            max_size=256
        )

if len(unique_labels) == 2 and 0 in unique_labels and 4 in unique_labels: 
            # If only background, return default bounding box and label
            box = torch.tensor([[0, 0, 0, 0]], dtype=torch.float32)
            box_label = torch.tensor([0], dtype=torch.int64)
        else:
            x,y = label.shape
            label_temp = zoom(label, (256 / x, 256 / y), order=0)
            _, bbox = find_and_draw_contours(label_temp, 0)
            box = torch.tensor(bbox, dtype=torch.float32)
            box_label = torch.tensor([1], dtype=torch.int64)

when training model, I get the error below:

AssertionError: All bounding boxes should have positive height and width. Found invalid box [0.0, 0.0, 0.0, 0.0] for target at index 1.

[sharp ferry]

never worked on such project but try using None instead of zeros then ?

any reference for above implementation ?

[thorny knot]

I have read a post in the PyTorch forum where someone set the [[1,2,3,4]] or just random for the box in case there is no object in the image. I have only trained for 10 epochs, and it seems to be working quite accurately.

[karmic badge]

What else math i need to know for data science

?

[wicked pivot]

Hi Kaggle team I am a latino from Colombia, I would like to know if I can apply for the KaggleX Fellowship Program, thanks in advance cc @tight quail

[mint yoke]

I'm working on an unsupervised anomaly detection on fast furior transformed data, trained it on autoencoder but not luck , no expected real life results. So I'm now embedded those frequency using sinusoidal embedding and used transformer for training now problem is reconstruction error in training and test data almost same what to do ?
Do need to scale each fft individually to retain the fft structure or keep scaling globally?

[severe cove]

Torch or Tensor?

[sharp ferry]

Torch anyday

[dense bear]

Hi! Can I run stable diffusion and kobold ai on kaggle without be banned?
I'm too recording a course about stable diffusion, can I use kaggle to teach and the students use without be banned? How it works?
Can I train checkpoints to stable diffusion on keaggle without be banned?

[wicked berry]

Coral.

[wicked berry]

I think yeah

[dense bear]

Will be great

[wicked berry]

Nice

[dense bear]

@wicked berry But both can generate nsfw content, this don't will banned?

[empty current]

there are for sure kaggle policies on nsfw in the internet out there, did you search already?

[dense bear]

Yes, it's only for research

I'm asking because I have to notice the students

And kobold aí, for example

You can generate texts with violence, it will banned? Idk

And stable diffusion for example, some models are biased to nsfw, if I generate a image like this accidentally I will banned? Only asking to me know the better way

I found

It's prohibited nsfw for txt and images

I can run a content moderator to test the models to know if it have nsfw or not, must be sufficiently 🙂

[tight quail]

Hi @wicked pivot we're glad to hear that you are interested in the KaggleX program.

Colombia is an eligible country for the program. You can get started with the application process by first participating in the skill assessment challenge. Please visit kaggle.com/kagglex for more details.

[timid crypt]

I have created a notebook in You tube Dataset to find the best channels and content creators in the field of data science ,check it out and if you have any suggest then comment and if you like it please upvote https://www.kaggle.com/code/abhishek0032/youtube-channels-analysis-with-advance-visuals

[outer mountain]

What happened to the KaggleX Skill Assessment Challenge?
https://www.kaggle.com/competitions/kagglex-cohort4

Its a dead link

I am worried I will not be able to work on the coding challenge because the database is taken down

[tight quail]

Hi Ishan, the competition is still live, can you please try clearing your cache then use this link: https://www.kaggle.com/t/005ae82d6fda45a9a3048d900fd2eb17

If you continue to experience issues, please email us at [e-mail]. Thanks you!

[cosmic olive]

Hello, can I ask what should happen after sending the submission? I have send it but I didn´t receive an email and no message popped up after sending it. Thanks

[tacit token]

Hello guys, i see that Discord isn't very active, if you wanna joint a small group of active people let DM me please, Thanks

[grim linden]

Hi @outer mountain , Are you able to access the KaggleX webpage now? I did delete the cache and access the link but still not able to access it.

I just wanted to access the dataset for the competition. Could anyone share the dataset if they downloaded it?

[pulsar crag]

Same

Not able to access the challenge webpage

[outer mountain]

There are several people who are unable to access. And yes, I have personally asked my friends and family to see if they can open that page from their own system. They were equally unable to open.

[slate birch]

I am also getting the same error. I just sent a mail to [e-mail] as suggested by @tight quail

[pulsar crag]

Hopefully they will fix it soon

[short wedge]

@outer mountain @cosmic olive @grim linden @pulsar crag @slate birch I can confirm the competition is currently inaccessible, we'll get it back online tomorrow once we understand what happened. Sorry for the inconvenience.

[pulsar crag]

Thanks @short wedge

[outer mountain]

Thank you for the confirmation. I was one of the affected who could not download the file. Someone on the discussions board responded to me with a link of a copy. I hope it should be okay to post for others who are yet to download.

https://www.kaggle.com/datasets/stephanievelezrph/kagglexfellowship

[grim linden]

Hey @short wedge
Thanks for the the quick reply.

Just wanted the datasets for the assignment

Thanks @outer mountain for sharing the data. Will check it out

[tight quail]

@Ishan Mitra @Alonso Alarcón Aguilar @pkn_2018 @madarshbb @Raunak Kaushik

Hi all,
Thank you again for making us aware of the access issue. We've resolved the error and you should now be able to access the skill assessment, using this link: https://www.kaggle.com/t/005ae82d6fda45a9a3048d900fd2eb17. Please let us know if you still experience an error.

Again we apologize for the inconvenience.

[empty current]

Do I understand it right that I cannot surpass the novice stage without registering my phone number? Everytime I want to leave a comment somewhere, it tells me I need to add a phone number first.

[rich halo]

archangel7431 has been warned

Reason: Bad word usage

[unreal wave]

Hi I am Imran. I am new to data science and want to start with you.

[empty current]

Hallo Imran!

[south oracle]

is doing ml supposed to feel this uncomfortable or am i just burned out? lol

[outer mountain]

Yes thank you so much!

I can now see the site is up and the deadline is updated to reflect 23rd as the deadline now. Thanks!

[high canyon]

@mellow sentinel Hi, I sent a message about teaming up for the RSNA competition but I think we have to be friended to dm

[short wedge]

@high canyon We paused DMs on the server to deal with a discord spam attack, you should be able to send DMs again in a couple of hours.

[wicked berry]

hello

[dense bear]

hi guys

I'm tryng run stabel diffusion on kaggle

when I open the local URL open a page and show me a message that the page refused the conection

I put the port on firewall windows and yet don't work

what can I do to corret this?

[worthy portal]

Hello guys I am Aaron! I am a college student who is interested in Data Science and AI. Currently I am looking into the uses of KANs in medical images

[raven badger]

Hello there! I am Sai, an ML engineer working on the arc-agi challenge. Excited to be a part of this discord community 🙂

[karmic badge]

Does feature engineering involves library like sklearn?

[dense bear]

hi

I'm trynt to run the stable diffusion

but when give me the link

the page instead of run the program give me this message

anyone knows how can I fix it?

[wicked berry]

New to the club.

[obtuse creek]

i had some issues saving the model in .h5 file type on kaggle , earlier i used to save it in .pkl , .keras type format but now its like we have to save only in the format .weights.h5

@coral hedge

@fair cliff @old kayak , sorry for the ping , its actually urgent for me

sorry again

[old kayak]

Yeah h5 has depreciated

Faced issues recently

.keras should work

Could you share a screenshot

Worst case, download your notebook, run it locally as a Jupyter notebook or on Google collab and save your model

[obtuse creek]

.keras also doesnt work

earlier it used to work

[old kayak]

Oh cause you’re using save_weights(), this expects you to save the weights of the model alone, not the whole model

That needs to be saved into a weights.h5

Try model.save(“name.h5”)

Or model.save(“name.keras”)

These should solve your issue

[fair cliff]

@obtuse creek thanks for sharing this question here, one more question are you pinning the version of the packages you are using in your notebooks, this is a way to safe guard the notebook from breaking in case of changes in versions of underlying libraries used.
@old kayak good spot with the usage issue

[fair cliff]

model.save(“name.h5”)
Or model.save(“name.keras”)
Saves the whole model (meta data, weights, and other data that is model specific)

save_weights()
Saves only the weights so that it can be applied on top of any other compatible model(s)

Is this what the above APIs are doing @old kayak

[civic geyser]

Hi everyone, I'm looking to learn more about data science and develop the skills for the profession. What would you guys reccomend me to follow? Should I try a specific online course, stick to youtube, or try a boot camp?

[wicked berry]

Just try everything. Participate in kaggle competitions too

Books are great too

[civic geyser]

thank you, any books that you would reccomend?

[dense bear]

my stable diffusion on kaggle is pinging but dont open the page, any suggestion?

[normal sonnet]

same as here bro . i didn't know about this

[dense bear]

I did it, what notebook are you using?

[exotic stratus]

Ian Goodfellow has good ones but nowadays people prefer video lectures

Kaggle has a good hands on series to get you familliar with machine learning

besides that, deeplearning.ai has a good series by Andrew Ng who teaches machine learning from the ground up, building mathematical concepts and ideas that are cruical for ML.

i'd suggest you start off with video lectures , try applying what you learnt on real world datasets on kaggle and try competitions. you might not do well initially, but keep learning and practising and you'll be an expert sooner than you'll know.

[hasty prairie]

Hello, may I know when will KaggleX Fellowship Program be available for Vietnamese applicators?

[civic geyser]

Thank you