#❓┊ask-a-question

1 messages · Page 7 of 1

patent thicket
#

hey guys do i have to do the assignment in kagle copulsarily?

tiny kiln
#

Hello I am unable to verify my Malaysian phone number. Did anyone else same issue as I did? If so, how did you solve it? Thanks

livid eagle
#

@everyone "In Long-Running Operations, the ToolContext maintains state between pause/resume cycles. What happens if multiple users trigger the same resumable tool simultaneously? Does ADK provide built-in concurrency handling for tool_confirmation states, or do developers need to implement their own locking mechanisms to prevent state corruption?"

obsidian pecan
#

Authentication error: pls make sure you have added 'GOOGLE_API_KEY' to your kaggle secrets. Details: unexpected response from the service .....

#

How to fix

west wren
obsidian pecan
hidden rain
west wren
#

make sure the check box is selected once you add the google api key and name is exactly GOOGLE_API_KEY or if you named it something else change the name of GOOGLE_API_KEY to whatever you named the secret key

arctic loom
#

When designing an AI agent that uses multiple tools through the MCP interface, how do you ensure the agent selects the most appropriate tool in real time, especially when tool choice impacts latency, cost or reliability? or did I miss the point?

obsidian pecan
west wren
obsidian pecan
#

Adding photo is disabled here

#

I got it that tick mark is missing, 👍

hidden rain
hidden rain
obsidian pecan
#

But save version is getting failed

hidden rain
obsidian pecan
west wren
# obsidian pecan But save version is getting failed

Are you referring to the notebook versions? You don't necessarily need to save versions, unless you want to update it later on or maybe you have the save and run all selected while saving versions and it could be failing for that given a lot of people might be trying at the same time. I just create a new notebook and do the projects there following the steps, saves me time

west wren
# obsidian pecan How

it is the button between the power and the menu bar in the notebook near top right corner

hidden rain
obsidian pecan
hidden rain
#

@obsidian pecan reset and clear every cell

obsidian pecan
#

Ok

obsidian pecan
hidden rain
west wren
#

maybe try a quick save instead of save and run all after you click the save version or you can create a new separate notebook and redo the steps, will help understanding it as well, that is what I have been doing

obsidian pecan
#

Ok

hidden rain
#

@obsidian pecan dm me the error screenshot

obsidian pecan
#

I am trying again

patent night
#

Do we need to submit the assignments? If yes, how to ?

west wren
lilac solar
obsidian pecan
#

Not able to complete section 3 try the ask web interface

#

Still running almost 1 hr

obsidian pecan
#

Non empty folder already exist: /kaggle/working/sample-agent override existing content [y/n]:

golden ore
#

Helloo

lilac solar
lean drum
#

api_key project name?

fluid spade
#

Hi, where can I find whether I registered correctly for the capstone project participation?

violet mist
#

anyone here know the error of phone number verification

#

on kaggle

lilac solar
lilac solar
rocky oxide
#

Where is the assignments

cyan helm
#

Hello everyone, I’d be grateful if someone could help me understand this question better. what happens if an AI agent kinda loses track of what it was doing midway through a multi-step process?
Can the orchestration layer actually fix that kind of context drift or confusion?

#

How might context drift or state misalignment occur in multi-step tool interactions, and what role does the orchestration layer play in correcting it?

obsidian pecan
lilac solar
shadow elk
#

Hi all, good to connect with everyone here. Had a quick question- I keep getting a 429 resource_exhausted error when running the final enhanced currency agent, is anyone else facing this issue

civic ore
#

Hi there

#

NameError Traceback (most recent call last)
/tmp/ipykernel_48/2763458470.py in <cell line: 0>()
----> 1 runner = InMemoryRunner(agent=root_agent)
2 response = await runner.run_debug("What are the latest advancements in quantum computing and what do they mean for AI?")

NameError: name 'InMemoryRunner' is not defined

lilac solar
#

Trun on internet toggle

civic ore
#

Already on

lilac solar
#

@everyone Phone number verification now working

lofty cliff
#

name is not defined
problem
halp me

civic ore
#

Bro can you clarify, so that we can help you.

crystal basin
#

why skip the section 2?

civic ore
#

Do not skip any section.

crystal basin
#

why?

#

but my agent worked." i skip the mcp section 2.0 og all this this is the right or not?

lofty cliff
#

NameError Traceback (most recent call last)
/tmp/ipykernel_48/1012248524.py in <cell line: 0>()
----> 1 retry_config=types.HttpRetryOptions(
2 attempts=5, # Maximum retry attempts
3 exp_base=7, # Delay multiplier
4 initial_delay=1, # Initial delay before first retry (in seconds)
5 http_status_codes=[429, 500, 503, 504] # Retry on these HTTP errors

NameError: name 'types' is not defined

floral crescent
#

Hi

crystal basin
#

what?

bitter remnant
tulip gull
#

hy im getting error when i run the day one

#

??

crystal basin
#

iam all ready done on this."

loud jewel
#

Hii I'm getting error to link kaggal account

livid eagle
#

When using MCP tools like the Everything Server, what's the recommended fault-tolerance pattern if the external MCP server becomes unavailable mid-execution? Does ADK provide retry mechanisms specifically for MCP connections, and how should agents handle partial tool execution where some MCP tools succeed but others fail?

crystal basin
#

iam all ready have done this." without error

burnt sand
#

Authentication Error: Please make sure you have added 'GOOGLE_API_KEY' to your Kaggle secrets. Details: Unexpected response from the service. Response: {'errors': ['No user secrets exist for kernel id 99284506 and label GOOGLE_API_KEY.'], 'error': {'code': 5}, 'wasSuccessful': False}.
THIS ERROR IS STILL NOT GOING FROM DAY 1 AFTER ADDING NEW SECRET KEY ANYONE PLEASE HELP
@everyone

limpid arrow
#

Hi everyone i have a question related to day 1 assignment 1a

#

In that assignment we are creating a simple agent that can use google_search as a tool to search the web which makes it a simple agent

narrow shuttle
#

Do we have to code the agents and stuff?

#

Like, use a coding language?

#

I'm not sure about it.

livid eagle
#

The notebooks show using AgentTool for calculations, but mention sub-agents as a different pattern. What specific architectural criteria should we use to decide between wrapping an agent as a tool versus implementing a true sub-agent handoff? Are there performance or memory implications for each approach in high-volume systems?

clear copper
#

Hello guys i have ran into an error can anyone help? ```py
import pandas as pd
import kagglehub
from kagglehub import KaggleDatasetAdapter
import os

folder_path = os.path.join(os.getcwd(), 'kaggle', 'intro_to_ml', 'datasets')
os.makedirs(folder_path, exist_ok=True)

df = kagglehub.load_dataset(
KaggleDatasetAdapter.PANDAS,
"dansbecker/melbourne-housing-snapshot",
folder_path
)

print("First 5 records:", df.head())
bash
(venv) PS C:\Users\davit\Desktop\fromzerotohero> & C:/Users/davit/Desktop/fromzerotohero/venv/Scripts/python.exe c:/Users/davit/Desktop/fromzerotohero/kaggle/intro_to_ml/01.py
Traceback (most recent call last):
File "c:\Users\davit\Desktop\fromzerotohero\kaggle\intro_to_ml\01.py", line 11, in <module>
df = kagglehub.dataset_load(
File "C:\Users\davit\Desktop\fromzerotohero\venv\lib\site-packages\kagglehub\datasets.py", line 141, in dataset_load
return kagglehub.pandas_datasets.load_pandas_dataset(
File "C:\Users\davit\Desktop\fromzerotohero\venv\lib\site-packages\kagglehub\pandas_datasets.py", line 86, in load_pandas_dataset
read_function = _validate_read_function(file_extension, sql_query)
File "C:\Users\davit\Desktop\fromzerotohero\venv\lib\site-packages\kagglehub\pandas_datasets.py", line 108, in _validate_read_function
raise ValueError(extension_error_message) from None
ValueError: Unsupported file extension: ''. Supported file extensions are: .csv, .tsv, .json, .jsonl, .xml, .parquet, .feather, .sqlite, .sqlite3, .db, .db3, .s3db, .dl3, .xls, .xlsx, .xlsm, .xlsb, .odf, .ods, .odt

teal smelt
# clear copper Hello guys i have ran into an error can anyone help? ```py import pandas as pd i... 
import pandas as pd
import kagglehub

# Download dataset (this returns a local path)
path = kagglehub.dataset_download("dansbecker/melbourne-housing-snapshot")

# Check the files inside
import os
print("Files in dataset:", os.listdir(path))

# Usually, the main file is 'melb_data.csv'
csv_path = os.path.join(path, "melb_data.csv")

# Load it with pandas
df = pd.read_csv(csv_path)

print("✅ Dataset loaded successfully!")
print(df.head())
#

Try this code

#

🐣🐣🐣

clear copper
#

lol((

nova matrix
proper cairn
#

Hey Everyone

#

Is there anyone

upper scroll
livid eagle
#

The resumable App uses InMemorySessionService in the demo. For enterprise deployment, what persistent session storage backends does ADK support, and are there specific considerations for maintaining session state across server restarts or in load-balanced environments?

chilly cove
#

i have been trying to install the google adk and it hasnt been working on my kaggle notebook or my jupyter notebook what am i doing wrong 😫

#

please someone help me

fluid bridge
#

somone know if yastrday i create agent and if need also multi agent in anhther nootbook or some

hearty agate
#

In the first notebook, when I get to the part with running the ADK web interface, it runs the code and gives me an IP address to access it ("For local testing, access at http://xxx.x.x.x:xxxx".) But when I go there, I see "This site can’t be reached, XYZ refused to connect."

Any help would be appreciated 🙏🏻

limpid garnet
#

I'm in france, but i can't get the google api key

lilac solar
lilac solar
normal wadi
#

“I wanted to confirm whether this course is dependent on the Capstone Project or if it’s assessed separately.

lilac solar
#

@everyone please make sure Your phone number is verified

hot latch
#

Hey Everyone, i need help understanding which version of gemini ai i can use ? 1.5 pro not available to use ? @everyone

neon rain
#

I send any massesge in live

#

i can't send any massage in during live

lilac solar
hot latch
#

which is ideal bossman ?

#

which is better*

lilac solar
#

@hot latch 2.5

#

2.5 Flash lite

hot latch
#

thank you for your guidance !

#

@lilac solar

lilac solar
#

2.5 Flash lite

It's optimized to be extremely fast and more cost-efficient, making it ideal for high volume or "lite" tasks that still need to be smart.

hot latch
#

yes sir 🫡

hearty agate
#

Anyone else seeing this error a lot? WARNING:google_genai.types:Warning: there are non-text parts in the response: ['function_call', 'function_call'], returning concatenated text result from text parts. Check the full candidates.content.parts accessor to get the full model response.

warm hill
hearty agate
warm hill
#

yes

nimble rampart
lilac solar
digital shadow
#

can someone else me with this
File "<frozen runpy>", line 198, in run_module_as_main
File "<frozen runpy>", line 88, in run_code
File "C:\Users\idkmy.venv\Scripts\adk.exe__main.py", line 7, in <module>
sys.exit(main())
~~~~^^
File "C:\Users\idkmy.venv\Lib\site-packages\click\core.py", line 1462, in call
return self.main(args, kwargs)
~~~^^^^^^^^^^^^^^^^^
File "C:\Users\idkmy.venv\Lib\site-packages\click\core.py", line 1383, in main
rv = self.invoke(ctx)
File "C:\Users\idkmy.venv\Lib\site-packages\click\core.py", line 1850, in invoke
return _process_result(sub_ctx.command.invoke(sub_ctx))
~~~~^^^^^^^^^
File "C:\Users\idkmy.venv\Lib\site-packages\click\core.py", line 1246, in invoke
return ctx.invoke(self.callback, ctx.params)
~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\idkmy.venv\Lib\site-packages\click\core.py", line 814, in invoke
return callback(args, **kwargs)
File "C:\Users\idkmy.venv\Lib\site-packages\google\adk\cli\cli_tools_click.py", line 421, in cli_run
logs.log_to_tmp_folder()
~~~~^^
File "C:\Users\idkmy.venv\Lib\site-packages\google\adk\cli\utils\logs.py", line 72, in log_to_tmp_folder
os.symlink(log_filepath, latest_log_link)
~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
OSError: [WinError 1314] A required privilege is not held by the client: 'C:\Users\idkmy\AppData\Local\Temp\agents_log\agent.20251111_172502.log' -> 'C:\Users\idkmy\AppData\Local\Temp\agents_log\agent.latest.log'

can someone help me??

lilac solar
#

Dm

worldly iron
#

Hey guys. I have a doubt that if someone doesn't do the daily assignments will he get the certificate ?

stable herald
#

Does ur python code is running in day2a

gusty flint
# worldly iron Hey guys. I have a doubt that if someone doesn't do the daily assignments will h...

1. Will there be any certificates?

  • There is no official certificate from Google for completing the 5-day course.
  • What you do get is a Kaggle badge for completing the optional capstone project.
  • A badge isn’t a diploma, but it’s still a nice credential (proof of skill) on your profile.

2. Will there be free goodies for everyone?

  • No.
  • This isn’t a merch festival. Google sometimes highlights top projects or offers small perks, but there are no guaranteed goodies for all participants.

3. When does the course start?

  • The course runs from November 10–14, 2025.
  • Daily livestreams start at:
    11 AM PT / 8 PM CET / 12:30 AM IST.

4. Where will we get the course material?

  • Everything will be posted here:
    👉 https://www.kaggle.com/learn-guide/5-day-agents

  • You’ll also get reminders through:

  • Kaggle Discord

  • Follow-up email

  • Links shared right after each day’s material drops

  • This is your central repository (main place to store things) for all codelabs, whitepapers and assignments.

5. Is Google giving anything for completing this course?

  • Aside from knowledge and trauma from debugging…
    You get:

  • A Kaggle badge if you complete the optional capstone

  • Recordings of all livestreams

  • Access to all materials

  • Direct interaction with Google engineers during the week

  • But no certificates, no guaranteed swag, and no prize money for simply finishing.

6. What is the capstone project?

  • It’s an optional final project you submit on Day 5.

  • You build your own AI agent using everything you learned.

  • Completing it earns you:

  • A Kaggle badge

  • A chance for your project to be highlighted by Google/Kaggle

  • This is more about applying skills than “competing.”

  • Submission deadline is Nov 30th. (Thanks to @dusty timber for correction)

A Short Informational Video About The Course :

https://youtu.be/fWgURAgn_30

worldly iron
gusty flint
worldly iron
gusty flint
worldly iron
gusty flint
cursive egret
#

I get an error for the 1b lab. This is the block of code i'm trying to run:

root_agent = SequentialAgent(
name="BlogPipeline",
sub_agents=[outline_agent, writer_agent, editor_agent],
)

The error I get:
ValidationError Traceback (most recent call last)
/tmp/ipykernel_48/3912492297.py in <cell line: 0>()
----> 1 root_agent = SequentialAgent(
2 name="BlogPipeline",
3 sub_agents=[outline_agent],
4 )
5

/usr/local/lib/python3.11/dist-packages/pydantic/main.py in init(self, **data)
248 # __tracebackhide__ tells pytest and some other tools to omit this function from tracebacks
249 tracebackhide = True
--> 250 validated_self = self.pydantic_validator.validate_python(data, self_instance=self)
251 if self is not validated_self:
252 warnings.warn(

ValidationError: 1 validation error for SequentialAgent
Value error, Agent OutlineAgent already has a parent agent, current parent: BlogPipeline, trying to add: BlogPipeline [type=value_error, input_value={'name': 'BlogPipeline', ...l_error_callback=None)]}, input_type=dict]
For further information visit https://errors.pydantic.dev/2.12/v/value_error

craggy coral
#

How can we use AI agents for use of internal organisation sensitive data to predict or summarise.

tardy otter
#

if u know python u can move on to it after completing these course

craggy coral
tardy otter
tidal iron
#

Hi
I'm trying to verify my phone number in kaggle’s web site but it gives me ' permission denied '.
I've registered the course before,but I can't verify my phone number.
Can someone help me?

lilac solar
finite leaf
#

Hi, Is the 429 error in day 1 assignment solved?

#

Anyone completed the task?

limpid arrow
hidden rain
limpid arrow
#

running all at once causes too many requests and the api closes your connection

finite leaf
#

Yes, i moved to the next only after getting the output of one by one. But still stuck with the error

limpid arrow
finite leaf
scenic barn
#

Hi everyone! 👋
I’m from India, and the livestream timings (12:30 AM IST) aren’t convenient for me as I have college classes during the day. Could someone please let me know how I can mark my attendance in this situation?

Thanks in advance! 🙏

hollow flame
#

Hello Everyone 👋🏻

NameError Traceback (most recent call last)
/tmp/ipykernel_48/3849244353.py in <cell line: 0>()
----> 1 enhanced_currency_agent = LmAgent(
2 name="enhanced_currency_agent",
3 model=Gemini(model="gemini-2.5-flash-lite", retry_options=retry_config),
4 # Updated instruction
5 instruction="""You are a smart currency conversion assistant. You must strictly follow these steps and use the available tools.

NameError: name 'LmAgent' is not defined

What is this error??

slender ferry
#

have anyone finished day 1a ?

slate cedar
#

what is the benefit of getting the kaggle badge without any digital certificates

stable sentinel
slender ferry
worldly iron
slender ferry
worldly iron
slender ferry
worldly iron
slender ferry
finite leaf
calm flint
#

Guys do we need to sumbit day2 ? if yes then how should we do it ?

slender ferry
calm flint
#

how?

slender ferry
calm flint
#

thats it?

#

so no submission buttons?

pale seal
#

hello everone

#

i have quic

#

k

#

question

#

how to completed the day 2

#

please tell me

slender ferry
slender ferry
pale seal
pale seal
gilded galleon
#

how to get the HW DAY2? I send Stop by mistake!

#

I need Day2 instruction

celest wing
#

സാധനം കയ്യിലുണ്ടോ? Looking for teammates from kerala. Anyone struggling to complete day 1 task, please DM

slender ferry
slender ferry
limpid arrow
pale seal
slender ferry
pallid briar
#

Hello fellow Kagglers... I am a paying AI Studio Pro user escalating a critical platform failure. For over a month, my chat sessions have been systematically failing to load, and several are now completely inaccessible.
This is an environment-agnostic system failure. I have confirmed it occurs across all my devices (iMac, MacBook, iPhone) and browsers (Chrome, Safari), on both the web and native iOS apps.
The failure is directly correlated with chat history length. Performance degradation begins around 100k tokens. For my main project chats (500k-800k tokens), the failure is total. Weeks of critical MLOps work are now inaccessible.
The previous "scroll-to-top" workaround is no longer viable for these large sessions.
This platform instability is a mission-critical bottleneck that has already forced me to architect a custom terminal-based API solution as a disaster-recovery measure.
My question: Is this a known issue with the AI Studio platform's handling of long-session data for Pro accounts? What is the formal channel to escalate this high-severity service failure to your engineering team?
Thanks in advance. 🙂

wanton fiber
wicked sleet
#

I was getting error when i executed the day 1 assignment which was caused due to exhaustion of the free tokens of the google API. do we need to have the paid account for that or we'll get the credits ?

trail ivy
#

Where can we deeply exercise for all the transferred knowledge and information to practice, consolidate and expand Them. Thanks

undone seal
timber kettle
#

Hey, while I am registering I opted-out for capstone project , now I have edited and opted it. Will it be considered?

azure abyss
#

How does standardizing the agent-tool interface help make AI systems more flexible, modular, and easier to update or maintain in the future?”

celest wing
#

Beyond enabling tool use, how does MCP support better data handling and interpretation for AI models through its use of content types and schemas?

celest wing
#

Describe the minimum required metadata for a tool exposed via MCP, specifically focusing on the fields that define the expected input and output structure.

quasi trellis
#

Getting "🔑 Authentication Error: Please make sure you have added 'GOOGLE_API_KEY' to your Kaggle secrets. Details: Connection error trying to communicate with service." even though my google api key is connected to the notebook. Any thoughts?

blissful musk
quasi trellis
#

already done

blissful musk
#

The same issue occured with me, I factory reset all, created a new secret by removing the previous one, it worked for me

quasi trellis
#

Ill try it, thank you

robust cedar
#

Under which channel should I put my question for Tomorrow's Live Stream (5-Days of AI Agents Intensive Course?

blissful musk
azure abyss
#

What is a session in context engineering, and how does it help an AI agent maintain context and coherence during a conversation?

celest wing
#

Dinu, From Kerala, India. Finally my own agent is ready. Try now : https://kkb-production.jupyter-proxy.kaggle.net/k/276269928/eyJhbGciOiJkaXIiLCJlbmMiOiJBMTI4Q0JDLUhTMjU2IiwidHlwIjoiSldUIn0..IczX7JGovhJpeIS6vA4PAA.kUJuvYgyPivALU94ICYzP83frW_0dEzdbvoEDCQjhK0ph3OD6879a-p3f6WdWq7Jyj3Ydgc1Uz1uCHELTjG4HKsTExyn5zJRUVZJNbKLHlovUfZiY_8U-VFA7O6M4jTniYWqGmaPQHF7O-_avIUy1ZoWDnx9HRIRurGTfYZcjVg_80T0OXagKjZRx8nRH1iMW4__2niB-O5-5eRp8pYKHG3PB8ym8EUtwrxIFLyBJQRKgoj6KkkuAeo08zXifQ3f.DnRSfNpgS6nx1ZcYq2xUqA/proxy/proxy/8000/dev-ui/

nova matrix
idle nexus
#

Anyone completed the exercise of day 2b? tiny image retrieves max 1 image and was wondering how you guys solved it

sonic rain
#

Hi everyone! 👋
I just started and completed the Day 1 task — everything worked fine, but I don’t really understand the Python code and workflow. Could someone please guide me?

steep iris
#

Hi! Was wondering if it was possible to acquire data from previous, closed competitions (Specifically NFL Big Data Bowl 2025) for a school project.

glacial carbon
#

Finished completing the day 2 assignments. Thanks to all the organizer!

hot saffron
#

Hi can someone help me to connect to the agent ? it keep showing " Authentication Error: Please make sure you have added 'GOOGLE_API_KEY' to your Kaggle secrets. Details: Connection error trying to communicate with service."

gaunt goblet
hot saffron
gaunt goblet
#

Hi All,

I have a small doubt. I'm ticking the checkboxes in the kaggle webpage for each day. However, when I refresh the page they are dissappearing. Is this happening to everyone or only me?

gaunt goblet
hot saffron
gaunt goblet
hot saffron
gaunt goblet
hot saffron
#

OK . Thankyou for your help. May be I first go to find out why i can't receive the SMS

wanton fiber
chrome marsh
#

Finished task of Day 3! Today was quite easy, everything was just fine.

sleek patio
#

hey guys! anyone here to help?

#

i couldn't setup my agent. Due to that i am unable to complete the day 1 task.

boreal hamlet
#

Hi

civic ore
#

@everyone If anyone wants to join a group where we complete assignments quickly and help each other out, feel free to DM me.

native aurora
#

@civic ore yes please

umbral heath
silk crown
#

Hii

dreamy sapphire
#

hey hey. Did any1 mange to expose the tiny image from exercise 1b on adk web ? I see the base64 encoded data in the event, but the UI doesn't render it.

forest pecan
#

hello everyone

forest pecan
forest pecan
dreamy sapphire
weak karma
#

Hi, I’ve completed Day 2 Kaggle assignment but can’t access #day2-assignments.
Can someone please unlock it?

verbal bloom
#

i can't able to save my files

#

if i try to save it, it displays failed

hexed jewel
#

Hi, I've a busy academic schedules as the sems are nearing. Actually, when should I have to submit capstone project. Can I complete YouTube playlist later...???

lyric glen
#

i have also not complete day 1 assignment

#

can anyone help me

sour grotto
#

hii

#

can i do prompt enginner course also

#

?

tardy otter
#

its not for tht

#

this course is basically related to LLM, I am not wrong

sour grotto
#

yes

tardy otter
sour grotto
#

after completing it can i do prompt enginner course also

tardy otter
#

If u want u can

sour grotto
#

good to here

#

thank you

tardy otter
#

prompt engineering is part of tht

late seal
#

Guys. I have not started at all because unfortunately I was unable to verify my phone by Kaggle support team until now. Does anyone know if i should save the material because I want to still be able to obtain the certificate at my own pace. If anyone was able to reach out to Kaggle support let me know.

jagged osprey
#

Day 3a 5.4 user and country not stored not sure why. Is this common issue for this lab step? I'll investigate but give 5.3 output below, it had a warning not sure if relevant.

Session: state-demo-session

User > Hi there, how are you doing today? What is my name?
gemini-2.5-flash-lite > Hello! I'm doing well, thank you for asking. I can't recall your name right now. If you tell me, I can remember it for you.

User > My name is Sam. I'm from Poland.
gemini-2.5-flash-lite > It is a pleasure to meet you, Sam! I will remember that you are from Poland.

User > What is my name? Which country am I from?
WARNING:google_genai.types:Warning: there are non-text parts in the response: ['function_call'], returning concatenated text result from text parts. Check the full candidates.content.parts accessor to get the full model response.
gemini-2.5-flash-lite > I see that your name is Sam and you are from Poland.

#

still have not figured out why Day 3a 5.4 output is

Session State Contents:
{}

🔍 Notice the 'user:name' and 'user:country' keys storing our data!

mild furnace
#

I am trying to create the Gemini API key(1.2). It gives option to import a project or create a project. My page to create API key does not match the instuctions in the setup in #5dgai-announcements . What am I missing?

nova matrix
normal hedge
#

Help plssss.
Response = await runner.run_debug("...")
Not working

silk crown
#

Hi

normal stream
#

Hi

slender ferry
#

Day 1a - From Prompt to Action

Unable to commit this
The commit fails everytimes when i do it
any solution for this ?

hexed wave
#

can you help me to configure api key

#

day 1- from prompy to action

#

unable to commit

#

everytime i try it fails

slender ferry
hexed wave
#

okay but api key possible to configure?

slender ferry
hexed wave
#

how

#

and how to restart kernal in codelab console

craggy flame
#

@everyone If anyone wants to join a group where we complete assignments quickly and help each other out, feel free to DM me.

celest wing
#

Does anyone know how to download a score sheet and rank script in kaggle ? in Discord?

covert crane
#

certificate available after course....?

celest wing
#

!rank

molten radish
#

How is the day 3 going ?

nova matrix
elfin gorge
#

What are the fundamental questions one should always ask when crafting the context and prompt of an LLM interaction, to ensure the model has the right intent, scope, constraints, and reasoning path embedded from the start?

glacial carbon
#

Can we upload the file on github after we finish working on it?

thorn beacon
#

hey guys, i have tried to run excersises on google colab aswell as kaggle, but I encountered an issue, does anybody know how to solve it ?

response = await runner.run_debug(
"What is Agent Development Kit from Google? What languages is the SDK available in?"
)

AttributeError Traceback (most recent call last)
/tmp/ipython-input-3364943494.py in <cell line: 1>()
----> 1 response = await runner.run_debug(
2 "What is Agent Development Kit from Google? What languages is the SDK available in?"
3 )

AttributeError: 'InMemoryRunner' object has no attribute 'run_debug'

honest cape
#

can anyone solve ?

wintry apex
#

Do I have access to the Kaggle resources after the 5 days workshop?

prime otter
#

Authentication Error: Please make sure you have added 'GOOGLE_API_KEY' to your Kaggle secrets. Details: Connection error trying to communicate with service. ......guys am facing an issue in day 1 starting because of this am not able to move forward ...I did everything

small tartan
#

In secrete key name the key as "GOOGLE_API_KEY" AND Add the secrete key you generated in google AI Studio

oak magnet
#

Is That Important To Attend Live Class Can We only Read Whitepapers

twilit depot
#

It’s showing my work here in the “All of Your Work” section 7, but how will I know if this is my complete work?

barren crater
#

Hey guys while saving the version in today's assignment, it's saying failed in the bottom left section where it shows Active Events (it usually says successful while saving regular assignments )

#

Any idea anyone

vivid oasis
#

Under the first AI Agent with ADK in section 2.3 B please how can I Created New Session: debug_session_id

#

Under the first AI Agent with ADK in section 2.3 B please how can I Created New Session: debug_session_id

slender ferry
#

same for me
anyone pls tell how to finish this

like this 1a, 1b, 4a, 4b cant able to commit it

#

who have finished 1 and 4 ?
if you guys finished pls dm me and tell me how to solve it

oak magnet
heady sandal
#

Is there certificate??

civic ore
dreamy thicket
#

hi my name is parth and i looking for a team?? can i be a part of ur teamm??

primal garden
#

Api key kaha melige

next shale
#

do we get the certificate

gentle mirage
#

How is evaluation different than workflow testing or chain testing?

junior halo
junior halo
junior halo
#

@everyone I am trying to execute notebook 2b - Day 2b - Agent Tools Best Practices and getting below error. any help would be useful!

section 4.4 - ### 4.4: The Workflow Function - Let's tie it all together!

NameError Traceback (most recent call last)
/tmp/ipykernel_48/3733309264.py in <cell line: 0>()
----> 1 approval_info

NameError: name 'approval_info' is not defined

update:

there was readonly error while running previous section which leads to this error. i had to wait for sometime and then try running, it works!

nova matrix
mystic yarrow
#

Hi everyone!
Unfortunately, I couldnt make it to go all of 5day agents intensive.. Should I forhet about sleep today and try to catch all missing days, or it doesnt matter so much and i can finish it on weekend for example? Thanks!

And sorry for my English, if I wrote smth wrong

thorn beacon
#

hey guys, do you know how late we have time to upload capstone ?

pure merlin
#

construct@[native code]
r@https://www.kaggle.com/static/assets/app.js:2:970558
construct@[native code]
i@https://www.kaggle.com/static/assets/app.js:2:969329
t@https://www.kaggle.com/static/assets/app.js:2:972364
g@https://www.kaggle.com/static/assets/app.js:2:971002
@https://www.kaggle.com/static/assets/app.js:2:988779
@https://www.kaggle.com/static/assets/app.js:2:986453
ce@https://www.kaggle.com/static/assets/app.js:2:985093
i@https://www.kaggle.com/static/assets/app.js:2:985299. Whts this error

modern surge
#

hi guys. I am a bit late to the AI course. just starting on it today. Is there any deadlines for this course i should know before hand?

brisk forge
#

can anyone suggest how to delete orevious agents and multiagent and start from scratch in code lab?can anyone help

vivid oasis
wanton fiber
vivid oasis
vivid oasis
wanton fiber
vivid oasis
#

Ok thanks 🙏

upbeat moth
#

i don't have credit card to log in google cloud, can someone tell me any alternative way?

#

for 5b assignment

clear crest
#

is the project given let me know?

steady grail
#

Facing a problem to execute day 4 evaluate baseline performance line 45 showing client error.
It says 404 not found gemini-1.5-glash-001 is not found for api version v1beta. What to do?

analog yoke
#

Is it necessary to do day 5b

#

Actually I don't have any. Cloud SDK credentials

#

And it was given there optional

#

@upbeat moth same problem bro

lyric horizon
#

hey guys after finishing all 5 day assignments

#

what to do ??

#

like do we get teh cirtificate or we finish the capstone project , if we have to do the project where can we get details abt that

chrome marsh
graceful copper
#

Will we get any swag ?? If yes will all of us get or some and how will be that decided??

warped lotus
#

What happens if we do not sign up to Google Cloud?
Are we eligible for certification?

wet wigeon
#

@mortal slate @scenic crater This is my question for today's session
How do autonomous agents exchange information and coordinate tasks through agent-to-agent communication, and what protocols are typically used to ensure reliable message passing?”

nova matrix
turbid cairn
#

Can anyone clarify me that will we be getting any certificate?? As we don't have to submit our codelabs anywhere then how will we get??

sterile minnow
#

Can you design a self improving Agent model updating its A2A protocols based on performance?

full siren
quartz wasp
#

for the Day -5b Agent Deployment lab, i have integrated Cloud SDK of my other Free Tier gmail account into Kaggle using my other primary gmail account, and trying to execute 3.3 step (deploy the agent).. it throw below error INFO:vertexai_genai.agentengines:Creating in-memory tarfile of source_packages INFO:vertexai_genai.agentengines:Using agent framework: google-adk Cleaning up the temp folder: sample_agent_tmp20251114_195700 Deploy failed: 403 PERMISSION_DENIED. {'error': {'code': 403, 'message': 'Request had insufficient authentication scopes.', 'status': 'PERMISSION_DENIED', 'details': [{'@type': 'type.googleapis.com/google.rpc.ErrorInfo', 'reason': 'ACCESS_TOKEN_SCOPE_INSUFFICIENT', 'metadata': {'method': 'google.cloud.aiplatform.v1beta1.ReasoningEngineService.CreateReasoningEngine', 'service': 'aiplatform.googleapis.com'}}]}} add Codeadd Markdown

#

all necessary APIs are enabled, i believe it should allow to link any google account in the Secrets ?

hollow elk
#

Sorry, I'm starting late due to work. where the Whitepapers and all the materials to start working on all assignments and Capstone? thanks

fervent raven
#

Is it possible to create an agent that logs into a bank account (I can save the credentials in a secure location) and gets the statement and converts it into google sheets format. Every successive periodic login of the agent appends to the sheet

wraith sparrow
# fervent raven Is it possible to create an agent that logs into a bank account (I can save the ...

Yes, this is technically possible! You can build an AI agent that automates this workflow. Here's how you'd approach it:

Technical Approach

1. Authentication & Web Automation

  • Use browser automation tools like Playwright or Selenium to handle the login process
  • Store credentials securely using environment variables or a secrets manager (never hardcode them)
  • Handle 2FA if your bank requires it (this adds complexity)

2. Statement Extraction

  • After logging in, navigate to the statements section
  • Download the statement (usually PDF or CSV)
  • Parse the data using libraries like:
    • pdfplumber or PyPDF2 for PDFs
    • pandas for CSV files

3. Google Sheets Integration

  • Use the Google Sheets API with the gspread library
  • Authenticate with a service account
  • Append new transactions while avoiding duplicates (check dates/amounts)

4. Scheduling

  • Run periodically using cron jobs (Linux/Mac) or Task Scheduler (Windows)
  • Or use cloud services like AWS Lambda, Google Cloud Functions, or GitHub Actions

Important Considerations

Security Risks:

  • Bank websites often have anti-bot measures that may block automation
  • Storing credentials anywhere increases risk - even "secure" locations can be compromised
  • Many banks explicitly prohibit automated access in their terms of service

Better Alternatives:

  1. Use your bank's API if they offer one (many modern banks do)
  2. Open Banking APIs (Plaid, Yodlee, TrueLayer) - these are designed for this exact use case and are much more secure
  3. Official bank integrations - some banks directly integrate with accounting software
  4. Manual CSV exports - tedious but safest for occasional use
wraith sparrow
static mural
#

Hello @everyone 👋🏽 Any update on the certificate and next steps?

slender ferry
#

how to finish 4a and 4b ?

#

i am unable to finish it

severe lance
severe lance
slender ferry
#

cuz they use google cloud sdk for that

#

so u no need to do it

#

how to finish the agents intensive - capstone project ?

hollow spear
#

Will all of them get swag rewards

#

We had done

#

Class

#

And assignment

tidal bloom
#

Did kaggle just stop giving the l4 gpus , I try to late submit on arc-2025 but it doesn't give me l4 gpu option, previously it was available , but not now?

fervent gust
#

Is webapp / API okay for submission of capstone project ? or notebook only ? since that would be easy to deploy.

hollow moss
#

Hey

nova goblet
#

How to start Capstone Project Which Topic I need to choose , Plz Suggest @lone basalt

finite leaf
#

Hi, I would like to join a team for Capstone Project. Is there any slot available?

junior halo
#

@everyone - anyone looking for a member who can join you to do captstone project? i am interested!

sly drift
#

@lone basalt Is the capstone project mandatory to be submitted as a team or can it be individual? cbf dealing with delays due to ppl not putting in effort and time.

dusk zealot
#

Prizes

The top 3 teams in each track (including teams of 1) will receive Kaggle swag, and their projects will be featured on Kaggle’s social media platforms. All Capstone participants will receive a badge and a certificate of participation

Which Social media? I cannot find Kaggle social medias. Can anyone help

burnt gazelle
#

Is anyone looking for a member for the Capstone Project? I am interested

dreamy thicket
analog steeple
#

Do we get a certificate for attending 5-day intensive course?

severe lance
analog steeple
#

It's mentioned that only who participate and complete in capstone project will only receive a certificate

junior halo
dreamy thicket
burnt turtle
#

I too wanted to join a team

teal smelt
# 

# Import necessary modules
import json
import requests
import subprocess
import time
import uuid
import os
import re 

# Components
from google.adk.agents import LlmAgent
from google.adk.a2a.utils.agent_to_a2a import to_a2a
from google.adk.models.google_llm import Gemini
from google.adk.runners import Runner
from google.adk.sessions import InMemorySessionService
from google.genai import types 

# Handling Warnings
import warnings
warnings.filterwarnings("ignore")

# YAYYAYAYAYAY!!!!!
print("✅ ADK components and libraries imported successfully.")

ImportError: cannot import name 'DeploymentResourcePoolServiceClient' from partially initialized module 'google.cloud.aiplatform_v1.services.deployment_resource_pool_service' (most likely due to a circular import) (/usr/local/lib/python3.11/dist-packages/google/cloud/aiplatform_v1/services/deployment_resource_pool_service/__init__.py)

@ancient valley @fervent totem I get this error while trying to work on an ai agent on kaggle
I understand what circular imports are, but didn't these same packages work during the course? why not in my notebook

heavy quartz
#

can anyone guide me for capstone project i didn't attended the live streams because i am a student

wraith sparrow
dim forum
#

ValueError Traceback (most recent call last)
/tmp/ipykernel_48/2266491298.py in <cell line: 0>()
4
5 if PROJECT_ID == "gen-lang-client-0614975654" or not PROJECT_ID:
----> 6 raise ValueError("⚠️ Please replace 'your-project-id' with your actual Google Cloud Project ID.")
7
8 print(f"✅ Project ID set to: {PROJECT_ID}")

ValueError: ⚠️ Please replace 'your-project-id' with your actual Google Cloud Project ID.

dim forum
slender ferry
crude umbra
#

I completed every 5 days' assignments alongwith capstone project of kaggle and google 5 day intensive course that took place in March and April 2025 but could not yet obtain certificate of it....do let me know how to obtain its certificate

merry pulsar
#

Hi
I know this is going to prolly sound a bit beginner to me since I am but I need to know when doing the codelabs where do we insert the code into to get started? I am just starting out and have no idea where I should be putting the code in the codelabs into ??

severe lance
slender ferry
#

ik 5b is optional

feral canyon
#

Day 3a last cell of notebook (cross-session state sharing) mentions session-specific state and user-specific state. from where user-specific state became available after starting a new session?

ivory smelt
#

I'm on 4b and when I want to analyze a failure, it does not show me the pencil/edit or the comparison as shown in the screen video in the notebook.
Does anyone know how to do it?

Where I'm having this issue:
Day 4b - Agent Evaluation
3.3: Run the Evaluation
👉 Do: Analyze a Failure

pale cedar
sonic rain
#

hello where can i find the instruction guide?

elder seal
#

how can i upload my capston project from Github repo ?

crude meteor
tardy copper
# crude meteor

can you share the screenshot of your file structure in vs code along with code

brazen shore
#

Me too

#

Can i use mobile for work? If yes, how to use mobile

frosty plinth
#

Last date of Prject Submission?

brisk dust
#

I wonder whether kaggle has the tutorial of Pytorch,cause I see that most courses are using Tensorflow

brazen shore
#

I have no credit or debit card so how i can get access Google cloud service

sleek orchid
#

Hi everyone,
I’m working on a Master’s thesis about the practical engineering differences between Spiking Neural Networks and traditional ML models, especially around integration, deployment, and maintainability.

If you’ve worked with SNNs or ML systems, I’d be very grateful for your input in this 4-minute anonymous survey:

https://forms.gle/tJFJoysHhH7oG5mm7

I’ll share a summary of the insights with the community once the study is done.
Thanks a lot for your help!

fluid flicker
#

Does anyone know how to generate an image using prompts given to the Agent?

shadow elk
#

Hi all, having an issue running the notebook that @spiral bison shared - Demo 2 Notebook: Data Science Agent Report Generator. When running the explorer cell, the cli_data_explorer module is not available and is not a pip install either. Any suggestions. thanks.

outer grove
#

Hey 👋

We’re Siray.AI – we just opened our API completely free for the entire kaggle community (no credits, no limits during launch phase).
Available right now via API:

Nano Banana 2 – current #1 image model on every blind arena
SeeDream & SeeDance – state-of-the-art video generation
Gemini 3 Pro (full 2M context + image/video understanding)

We’d LOVE an official Siray.AI node so the n8n/ComfyUI community can drag-and-drop these into workflows without custom HTTP requests.

Docs: https://docs.siray.ai
API Playground & free key: https://siray.ai

please let me know if anyone would like to make the node. Or who is interested in trying our API service and give feedbacks.

dense wolf
#

Heyy everyone do provide suggestions for our uni project that we can implement it is a

Community Engagement Project for Data Science Students

Mental Health Awareness
SDG 3

Expected Output
Dashboard of mental health trends among youth

burnt gazelle
#

Does anyone have an issue with submitting the writeup

cedar lion
#

I want to show the "Instructions" as "system Prompt". If I ask the Agent "Who Are You",it should show "instructions" as output. How to do it .

polar pollen
#

Hey guys how do you make model architecture diagrams??

viral zenith
#

hello everyone is there any course on kaggle to learn Scikit-learn. or what would be the best way to learn it.

knotty imp
slow topaz
#

Hello, please can someone tell me how and where we should send the project?

wanton fiber
flint ocean
#

#❓┊ask-a-question guys can some one say how you guys started learning about ML and how you guys started doing hckathons in kaggle and where did you lean ML related stuffs can you give me how do i participate and learn to tune models and increase my scores and stuffs??

burnt gazelle
#

I've been trying to submit my project for 3 days. When I paste my GitHub project link into the Project Links field and click Save Draft, the page refreshes, and the link disappears (it is not being saved). When I click Submit, the modal opens but stays stuck on “Submitting…” indefinitely.

viral dagger
#

Im preparing for my first ever data science interview
Any tips or>

earnest cliff
#

Which dataset should I choose? Hi, I'm about to create my first portfolio, but looking at datasets on Kaggle makes me hesitant to choose one that's already been "done." I don't want to pick one that's been used a lot and that recruiters might see as basic.

Recently, I saw a list of the 10,000 largest companies in my country on my country's open website.

But I wanted to hear your recommendations on which dataset to choose for my portfolio.

neat anchor
solid wedge
#

Hello everyone! 👋
I’m excited to share my capstone project:

🛡️ SENTINELS – Multimodal Disaster Intelligence System
An AI-powered system for real-time disaster detection, severity analysis, risk prediction & interactive mapping.

🔗 Kaggle Notebook: https://www.kaggle.com/code/mukthanjalibonala/sentinels-multimodal-disaster-intelligence-agent

Connect with me on LinkedIn 👉 https://www.linkedin.com/in/mukthanjalibonala/

Would love feedback, suggestions, and support 🙏

Thank you! 💙

tepid edge
#

I enrolled in Google's 5-day AI agent course. but due to some issue, I cannot continue this course. Now i want to complete this course.Can anyone please help me? How can i complete this?

wraith ledge
#

is deployment related for capstone projects, deployment of agents or deployment of application? Im confused with requirements of capstone project

fluid yew
#

I want to know about this part on the bonus point:

Agent Deployment
(5 points) If you either have code or otherwise show evidence (e.g. in your code or write up) of having deployed your agent using Agent Engine or a similar Cloud-based runtime (e.g. Cloud Run).

For the deployment of Agent Engine, does it mean using ADK in the script?

craggy goblet
#

Hello , I am a student in currently doing bachelor's in Artificial intelligence , I have a few questions
I wish to get into NLP , I have already learned basics of pytorch and know theory of most of Machine learning , though I do not know how much depth in classical MLis required for going into deep learning, specially when it comes to the maths and derivations but I am familiar with the concepts of neural nets , representation learning , ML models , ensemble learning , gradient descent etc though online sources and courses in collage . Could someone give me a rough guide or idea for how I should proceed

dense sail
#

Can I submit the capstone project after 1st dec. Didn't take much time to finish it due to my other engagements. Can I get a week more to submit it.

wraith sparrow
#

submitting submission.csv file in a comp its committing from a long time each time i try to submit the file nothing seems happening and saving version option it only shows committing...

severe lance
#

Hello everyone, I’ve a huge announcement for you today. Did anyone ever try to use power law into machine learning project, I think that it would be very interesting and helpful to suppress any errors and predict perfectly any event !

sage parcel
#

I'm facing challenges to submit my project, specifically it's taking long time but still not get submitted. Someone help please. @everyone

austere kraken
#

Hi. Excuse me, I'd like to participate in the contest, but I'm new to the platform. How can I do this with a team? Is a team considered a group? Do I need to specify this somewhere, or is there a separate application?

wraith sparrow
#

U js need to join the contest as individual n u will be team with one person u u can add others by merging with their teams like urs

rich jacinth
#

i am receiving this error on last step of Kaggle notebook submission. I added a .csv output file to notebook, which runs successfully. Any help troubleshoot this? : "Cannot submit
An Evaluation system has not been configured for this competition."

wet wigeon
#

I am unable to submit the capstone project..it is just buffering for a long time

#

any suggestions?

glacial carbon
glacial carbon
near flame
#

hi guys, I am relatively new to programming compared to the seasoned ones. At my current state, I want to learn about git and github. what do u guys recommend for learning git and github. I like to go a bit deep into things, but not too much and I would love to have some free resources recommended. Thank you all.

brisk seal
#

Hi,
Shahid here,
I'M Data science and Business analyst student, in final year,
Can anyone please suggest me, on what project i can work on?

noble path
#

||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​||||​|| _ _ _ _ _ _ https://imgur.com/TC6h8P4 https://imgur.com/iiKXKB5 https://imgur.com/JAkE28j https://imgur.com/keASgw9

honest thorn
fresh forge
#

hi everyone there are anyone work with image caption , i need some help

oblique gale
#

Hello Kaggle community,

I have a question that you might find trivial. Given a tabular task (so basically we use classic ML algo namely LR, RF, XGBoost, ...), do we add/remove features (assume all features are numerical)? When and how and why? I would like to make things clear in my head, because when trying to understand from chatgpt or similar AI LLMs it doesn't give a consistent answer and it looks biaised based on how I ask the question, sometimes it is good to make many new features from the original ones namely using sin function, product of two features, ... and fix a threshold based on their correlation with the target feature to remove "uninformative" features, sometimes it says if you have three too correlated features keep only one of them since information is redundant ( so in the correlation matrix we can find a small square matrix along the diagonal where corr coeff are too high like >0.8), ... So in short, my understanding of "good" practices in this regard is pretty much blurry and I would like to make it clear, I would really appreciate it if you can give me somehow clear logic (math based) answer or a suggestion where I could find such an answer!

keen pewter
#

Hi all! I'm trying to visualize my dataset in 3D and the Kaggle runtime keeps crashing on me when it gets large. Is there a way to run my visualization stuff on my local computer and just run the ML side remotely on Kaggle servers?

old mist
#

In kaggle notebook, I got the error when I run "from transformers import Trainer"

ImportError Traceback (most recent call last)
/tmp/ipykernel_77/3090685106.py in <cell line: 0>()
----> 1 from transformers import Trainer

/usr/local/lib/python3.11/dist-packages/transformers/utils/import_utils.py in getattr(self, name)
2152 isdummy = True
2153
-> 2154 def getattribute(cls, key):
2155 if (key.startswith("") and key != "_from_config") or key == "is_dummy" or key == "mro" or key == "call":
2156 return super().getattribute(key)

/usr/local/lib/python3.11/dist-packages/transformers/utils/import_utils.py in _get_module(self, module_name)
2182 module_file: str,
2183 import_structure: IMPORT_STRUCTURE_T,
-> 2184 module_spec: Optional[importlib.machinery.ModuleSpec] = None,
2185 extra_objects: Optional[dict[str, object]] = None,
2186 explicit_import_shortcut: Optional[dict[str, list[str]]] = None,

127 
128

/usr/local/lib/python3.11/dist-packages/transformers/trainer.py in <module>
136 )
137 from .training_args import OptimizerNames, ParallelMode, TrainingArguments
--> 138 from .utils import (
139 ADAPTER_CONFIG_NAME,
140 ADAPTER_SAFE_WEIGHTS_NAME,

ImportError: cannot import name 'is_torch_optimi_available' from 'transformers.utils' (/usr/local/lib/python3.11/dist-packages/transformers/utils/init.py)

How to fix that?

wraith bone
#

Hello. In playground competitions, are you allowed to use an external dataset and it's features to train a model?

stiff grove
#

hello everyone, i have some questions about competitions in kaggle,
Submissions to this competition must be made through Notebooks. In order for the "Submit" button to be active after a commit, the following conditions must be met:

CPU Notebook <= 9 hours run-time
GPU Notebook <= 9 hours run-time
Internet access disabled
Freely & publicly available external data is allowed, including pre-trained models
Submission file must be named submission.csv

will these be assesed based on the notebook you use to train your (pre-trained) models? what if i use another source like colab pro i and just save the best models the upload to kaggle to inference, does that count as cheating or anything?

muted spindle
stiff grove
fierce basalt
#

hey guys i just want to express that i want to learn DSA for interview and logic building for better code efficiency and memory management i know python very well and libraries for Ml. but when i started DSA i can see my level very well
leave DSA i can't even solve hackerrank's python problems
am i cooked pls save me

maiden cedar
#

I have an error when using API Gemini 3 Pro on Google Ai Studio, can't "read & understand" the youtube link. The result is just google search index. How to configure on Google Ai Studio to use the power of API Gemini 3 pro?! Best regard all

grizzled sable
dense sentinel
# maiden cedar I have an error when using API Gemini 3 Pro on Google Ai Studio, can't "read & u...

Gemini 3 Pro cannot directly read or understand a YouTube link.
Go to:
In Google AI Studio - Project Settings - Tools - Turn ON Web Content Retrieval

Then paste this JSON in your book:

{
"tools": [{ "google_search": {} }],
"contents": [
{
"role": "user",
"parts": [{ "text": "Summarize this video: <YOUTUBE_LINK>" }]
}
]
}

Replace <YOUTUBE_LINK> with your actual YouTube link.
It will only work well if the video’s transcript is publicly available.

stark phoenix
#

web dev or cybersecurity? Which one you would choose on 2026?

maiden cedar
tawny linden
#

The judges requested a video that is only 2 minutes long. However, including the speech explanation, creative elements, and other requirements might stretch the length to a maximum of 4 minutes. Will this be acceptable?

harsh yarrow
#

Hey guys if there is any course for students going on pls make sure to update me

oblique sluice
severe lance
#

Hello guys, I hope everyone is good. I’d like to know if anyone here finally received a Google kaggle certificate for the previous Google’s 5Day intensive course ? I didn’t receive anything !

wraith ledge
#

Is it possible to get my certificate with real name? Cause I was using random username for Kaggle and got certificate with it

strange aurora
verbal crest
static summit
#

Well why account it's not showing

spark knoll
#

Hi yall does anyone have experience with Amazon Personalize?

wraith ledge
strange aurora
crude meteor
#

I recieved a badge from kaggle is that what they you recieve at end of the December anyone if know pls inform me

lucid sparrow
#

Hi friends! I want to create self hosted competition with prizes. Is it possible to add thread for it on kaggle discord?

pallid swan
wraith sparrow
#

guys feedback needed

wraith sparrow
#

i tried asking pplx comet to find the relevant forum in kaggle discussion to post about a google survey form in hope of quick responses but i got quick flagged and ban by kaggle bot in return

past plume
#

how i can download whitepaper

wraith sparrow
#

scroll down to find the whtiepaper link

wraith sparrow
#

official kaggle api sucks anyone knows a better alt ❔

wraith sparrow
#

@worthy yoke bro

willow dust
#

I’m new to this field, and I’m working with a dataset (the current playground competition) that contains many extreme values. I don’t yet have experience handling them, so I’d appreciate learning what approaches you usually take in such cases.

barren phoenix
#

Does this exist anywhere/ would this be usefull for people im planning to build this ???
A lot of kagglers iterate on kaggle notebooks completely online. (Using kaggle runtime...)
I could use a chrome extension which acts like cursor/ claude code for kaggle

#

(please ping me if you reply to me )

hardy lagoon
#

ExceptionGroup: unhandled errors in a TaskGroup (1 sub-exception)

ocean portal
#

I'm stuck on section day 5b Agent deployment 1.4. Every time I enter my google project ID it gets an error message. I've tried multiple times but doesn't seem to work. I don't think I can progress past this point. Any suggestions?

tribal stag
#

hi

undone holly
#

Hi, how did you guys learn to do EDA?

charred patio
dusty timber
# undone holly Hi, how did you guys learn to do EDA?

I learn to do EDA (or anything) by doing. 🫡

Coming up with a project that interest (you'll remain curious in order to slog through) you to do something end-to-end is the best start. It would highlight the gaps in your knowledge that you need to fill & learn. Especially if you cannot starting coding anything without asking in Cursor/Antigravity/NotebookLM/GPT chat tools/etc. This is still the way I learn, today. These tools speed up the reading & comprehension, compared to 3-5 years ago where we need to read documentation ourselves. The horror! 🤣

undone holly
#

Yes of course by doing, but when i wanted to do something more complex than iris data set i get lost like what should i do first, so i look at kaggle at other peoples edas and am like how do they know they should do that and that. So i started looking up tutorials and found some books. Just was wondering if other had that problem and where they informed them selfs.

plucky sparrow
#

Hello team,
I couldn’t attend the live sessions due to my semester exams, but I have joined via the Discord link and I am completing all the assigments
Could you please confirm if I’ll receive the certificate after successful completion?
Thank you.

dusty timber
# undone holly Yes of course by doing, but when i wanted to do something more complex than iris...

Yes, there's a term for that - "tutorial hell". I overcome it by doing a separate project from scratch, after following a tutorial. To use the concepts covered in the tutorial.

Depends how you're oriented UI first or code first. Then start from there. Some projects I know what the UI should look like, so I use FastAPI or Streamlit to scaffold some simple UI elements and then wire up with code. Some projects I know what the backend code entails and I start coding the scaffolding of my features that I want the project to have.

That feeling of not knowing how to start? It does not go away I'm afraid. I just overcome it with starting a new repo, ask AI chat some questions, copy some scaffolding code and then just start.

With Google Studio and Cursor and now Antigravity, it is so much easier to start now. The challenge for me now is to stop using this AI tools for everything as they have become a crutch - I feel incomplete and insecure when I cannot ask it questions that I feel like I know the answers to. Imposter syndrome rears its head again. I feel like I need their confirmation that my ideas is sound and correct.

dusty timber
dense wolf
#

Is it worth learning hadoop?

glass sage
#

I am trying the Kaggle 5 day AI agent, following the steps, but got error when trying to authentication in notebook complaining the GOOGLE_API_KEY, which I did, but it does not work. Authentication Error: Please make sure you have added 'GOOGLE_API_KEY' to your Kaggle secrets. Details: Connection error trying to communicate with service.

#

Any suggestions? Thanks.

fast saffron
#

Hello, I need a team mate experienced in deep learning and llms with sophisticated approaches. Must have GPU machine or cloud access. Interested individuals dm me

sage sandal
past plume
#

how to download the whitepaper

dull terrace
#

Hello i have this question, i don't know if this is the right place to ask.
But can anyone tell me how the batching in llm api call is cheap and why it is the case ??

wraith sparrow
coral aurora
#

I've been playing a game called Happy Factory recently. In the game, you need to set up an automated production line to make various items, and each item may have multiple recipes. There are simple formulas like aA=bB and compound formulas like aA+bB=cC+dD. All items can eventually be generated from several basic items through different recipe path diagrams. There may be rings on A single branch of the generation tree of A target item, such as A<-bB+cC, B+dD<-eE+fF, E<-wD+gG. Here, on the recipe path for generating A, a ring appears at node D. It is known that substances passing through a loop will only be consumed and not increased. I need to calculate the formula path diagram that consumes the least basic resources to generate a certain quantity of target products, where different basic resources have different importance weights. How should an algorithm model be designed?

coral aurora
# coral aurora I've been playing a game called Happy Factory recently. In the game, you need to...

The model I am currently designing is as follows: the formula for the fixed formula is fixed_weight_matrix, the weights of the training formula are formula_weights, and the formula_weights are converted to non-negative numbers through the non_neg_activation function. Then calculate the change in the product, namely resource_change. In resource_change, I will pick out base_idxs to calculate the consumption of basic resources, target_idxs to calculate the output of target resources, and non_base_idxs to calculate the increase or decrease of non-basic resources (used for automatically balancing formula formulas). But I failed. It seems that the loss terms resist each other and cannot converge.

sweet topaz
#

I have uploaded a model to Kaggle, but it does not show in my models section, although I can see it using the URL. Can anyone tell me what the reason is?

wary orbit
#

Hi guys. How can one get the output of an agent outside the workflow? Like for logging and external processing

dapper bay
#

Hi Guys I have completed the 5-Day AI Agents Intensive Course with Google but I don't see an option to download my completion certificate. I know there is one because people post it on linkedin

wild hawk
#

Can anyone help me with Housing Prices Competition? I have some questions

wary orbit
craggy wagon
mellow vortex
#

Hi
is there any kaggle discord bit by which I can add my and friends handle and see their and mine ranking with score in competetions??

gritty mantle
sick wagon
dusty timber
sick wagon
#

okayyy thanks buddy

mossy edge
#

hey guys new here, are there anybody good at stata, who knows how to best compare the iv effect of on a binary dv and an index dv ?

stiff grove
#

hello guys, im having a trouble with dependencies download for offline inference, can sb help me with that

icy copper
#

Anyone know what the new AI quotas for daily and monthly are for? Just noticed them when I clicked my profile to check my current GPU quota.

dusty timber
umbral crow
#

Has anyone been using Antigravity? It's utterly insanely powerful. Anyone know of any other similar platforms they prefer more?

wild hawk
#

Can anyone help me with advice in my Classification task?

simple frost
wraith sparrow
lapis garden
#

Hi everyone, I am not sure where to post this so I decided to post it here. I graduated with a degree in data analytics and maths 3 years ago but after graduation I found it really hard to stay motivated and focused on working through datasets and making my own projects.

During my undergraduate courses I was given datasets by my teacher and some questions on what I need to do. I believe that leaving this structured environment and transitioning to a independent learning was my main challenge since I was really confused on how to approach each problem, how to keep myself motivated, etc

So now it has been 3 years and I haven't really looked at data analytics or tried working on improving my skill but I want to give it another shot and hopefully obtain a data science-related job. Does anyone have any suggestions on how I can get started? and what tips do you have to stay motivated when learning?

simple frost
wraith sparrow
lapis garnet
#

🚀 MedGemma Challenge Project Live!

Built PneumoScan AI for Google's MedGemma Impact Challenge - chest X-ray analysis with AI.

🔗 Live Demo: https://lnkd.in/dRKxmGPP
💻 Code: https://lnkd.in/dzmduVnV

Features:
• Upload chest X-rays → AI analysis
• Clinical findings & recommendations
• Medical term explanations
• Edge-deployment ready

Using: MedGemma models + Gradio + Kaggle chest X-ray dataset

Looking for feedback! Anyone else working on this challenge?

#medgemma #kaggle #healthcareAI

viscid acorn
#

Does anyone know Google Search Optimization?

wraith sparrow
viscid acorn
balmy mortar
#

Hi anyone knows how can I use kaggle or colab in vsc as i used the colab extension it's unstable especially when I'm adding the dataset

dusty timber
#

it's unstable

What does "unstable" mean? Crash? Errored out?
Have you looked at / read the Kaggle API documentation?
You meant local vsc? or codespaces?

tender flume
#

Can anyone tell me where can I find tumor segmentation mask dataset for 2D image segmentation using UNet

rugged juniper
balmy mortar
dusty timber
#

Give a concrete example, of your command(s) that is giving problems and the extension name that you use? And the dataset URL and the command you use to download.

I can't help with abstract scenarios.

lapis garnet
slender crater
#

Is anyone aware about the Microsoft Student Ambassador Program?? Or even participating in it ???

weary oxide
slender crater
#

yep

weary oxide
balmy mortar
#

What kind of a program is that when most of the knowledge is open source available

rugged juniper
pale lantern
#

Hello, I hope you're doing well. Is there anyone who is experienced in fine-tuning Hugging Face models who could help me with a project?
Please dm

wraith sparrow
#

how do i fix this errorhttps://media.discordapp.net/attachments/1454044110712340606/1463147235456974888/image.png?ex=6970c58c&is=696f740c&hm=1d77d7a45aa6d8653f05417434b1583847976cb0630b8db0b4bf0ae6495609a3&=&format=webp&quality=lossless&width=1228&height=653

polar stirrup
desert sail
rough fjord
#

Hi, how to add a file to a dataset of mine? I forgot to add the file of the labels when I downscaled another dataset. I click "new version" but there is no option at all to add or modify files.

#

Do I have to do the whole thing from scratch? It took quite a while to do it

civic ore
# rough fjord Hi, how to add a file to a dataset of mine? I forgot to add the file of the labe...

No, you do not have to do the whole thing from scratch.

  1. Go to your Dataset page.
  2. Click New Version.
  3. A window will pop up. Look at the top or left sidebar of this popup. You will see your existing files listed.
  4. To Add a File: clear/delete the specific file references if you are replacing them, or simply look for the "Upload" button (cloud icon) or the "Drag and drop files" area usually located above the file list.
  5. Drag your missing labels.csv (or equivalent) into this upload zone.
  6. Add a version note (e.g., "Added missing label file") and hit Create.
rough fjord
tiny egret
#

Hey, I’m trying to predict the next tick in a BTC price stream that updates once per second (i.e., every 1 second there is a new price update): https://data.chain.link/streams/btc-usd-cexprice-streams.

I found/reconstructed several other BTC price streams that are highly correlated with this Chainlink stream and update many times per second. Using the fact that they’re correlated, my goal is to accurately predict all upcoming Chainlink price moves with |Δ| ≥ 10 (up or down) ~500ms before the next Chainlink price update using signal from the faster streams early enough to act before the next Chainlink tick prints.

ML work so far: I aligned features from the faster streams to the “next Chainlink tick” target and trained a simple two-stage pipeline (1) a classifier to predict whether the next |Δ| exceeds a threshold, and (2) a regressor to predict the magnitude |Δ|. Decent results.

If someone could give me advice on how to better model this problem, I would be super grateful. I would pay for your time. Please lmk.

eternal hedge
#

is there any one to help me

granite compass
#

Set your email ID as public on Kaggle so how to achieve this

#

Set your email ID as public on Kaggle so how to achieve this@civic ore

mental cosmos
#

hi
i have a doubt like im running a big data and traing model using random forest but its loading with no output tried to change and restart it no effects
using kaggle notebook any clues?

pastel night
mental cosmos
timid jungle
#

i need to buy a copy of O Reily's hands on machine learning book, but its a bit costly, anybody aware of any online sites where second hand options might be available?

polar crescent
#

Just get an online pdf and get it printed from a local print shop and get it binded if you want. Will cost you half or less than actual cost

brittle mesa
wraith sparrow
#

hey i was wondering if theres any existing tools for like maintaining the network/connection of all the arxiv papers (where two papers are neighbors based on their similarities)

stiff grove
#

hello, I see that kaggle has quota for AI, can anyone show me how to use it or give me an article to read about that

thin gull
#

Hii guys can anybody help . I am new in kaggle competition and unable to make submission . I am getting this red flag

Cannot submit
Your Notebook cannot use internet access in this competition. Please disable internet in the Notebook editor and save a new version.

west palm
#

Hello everyone! Could you please suggest some strong AI/ML, deep learning, or reinforcement learning project ideas that would add significant weight to my resume? This will be my final-year project, and I can dedicate about a year to working on it.

twilit hearth
thin gull
twilit hearth
#

You can use Optuna library

thin gull
cold crest
#

Hi guys I need help!!! Have you seen architectural drawing? There if you check wall section drawings: you will see arrows pointing towards particular sections. I want to identify the labels for those arrows. If we move from left to right on the drawings: we need to traceback the arrows to the label text and give the final output. How to do it?

dense wolf
#

Guyzz I need opinion on how do I get started as I took part in this competition and how do I make myself stand out it's the round where our models will be selected so we will have a national opportunity

We choose the following problem set:

Agentic Honey-Pot for Scam Detection & Intelligence Extraction
Participants must design anautonomous AI honeypot systemthat detects scam messages and actively engages scammers using a believable persona. Once a scam is detected, the AI agent must continue the conversation to extractbank account details, UPI IDs, and phishing links. Interactions will be simulated using aMock Scammer API, and outputs must be returned in a structured JSON format.

Level Objective:
Help participants clearly understand both problem statements
Ensure focus on a single problem domain
Prepare participants for API and system implementation in the next level

stoic mesa
#

Hey is there one of those getting started competitions for just python and SQL? I want to get started but a lot of those are in R, which I do plan on learning, but later

wraith sparrow
hollow garden
#

I am a graduate student from a non-science background. I do not yet have a strong foundation in mathematics or statistics, but I want to develop my data science skills. I would like to use generative AI to learn efficiently, so could you give me some ideas on how to study?

wraith sparrow
hollow garden
#

I will review

dense wolf
wraith sparrow
half owl
#

Yo guys, I have a DS question.

Is it important/helpful to understand the math behind each algorithm?

If you respond, please select my message so Discord notifies me, thank you

marble moat
#

hi i have few doubts about rag can any one expalin
i have one form multiples users can sbmint eg: loan request form there are nearly 1L submisson in that form record . i want to get some info form that reconds how the rag works here or i am going in right path or i need to fine tune the llm for this or what i need to do.

stoic mesa
#

hey um when I do the Intro to SQL excersise 'Select, From & Where' I get hit with this error when setting up the feedback system /usr/local/lib/python3.11/dist-packages/google/cloud/bigquery/table.py:1727: UserWarning: BigQuery Storage module not found, fetch data with the REST endpoint instead. warnings.warn(

#

the feedback code:

# Set up feedback system
from learntools.core import binder
binder.bind(globals())
from learntools.sql.ex2 import *
print("Setup Complete")```
pine heath
karmic quartz
#

Hey, guys! how are you? Quick question about competitions: is there a way to host a competition without automated scoring? I want to analyze submissions individually and manually

rain shoal
#

Hey I am a highschool student and want to know if getting the kaggle certificates for ML and Data science worth it for my resume. Has anyone had kaggle or ML and data science projects as a key part of their college resume and got into their dream school?

limpid garnet
#

where can i ask a question about a kaggle course while linking an image ?

wraith sparrow
wraith sparrow
brittle mesa
#

Hi,I'm a newbie to Machine Learning.I'm learning about ML since 3 months and I do know well about cleaning, visualizing,imputing,few supervised learning algorithms and I'm intending to continue learning with others which I think is best way to learn faster.If anyone really want to learn with me please DM me so that we can team up for further competitions in kaggle

inland marsh
#

im running a submission file and i have tried submitting it 5 times on different days with multiple tweaks but it is running into an error that says notebook threw an exception and there is no specific error output to work on , any help?

gloomy cairn
lavish arrow
#

Hello there, I have a question.
in the google ia studio i deleted my first project by mistake. How do I get it back or create a new one for the AI agent?

ebon hearth
grand sapphire
#

Hi
I'm looking for alternatives to gemini-2.5-flash as Google has slashed down RPD from 1000 to 20.
Generous free tier alternatives only.

wraith sparrow
wraith sparrow
#

anyways anyone tried a prompt attack: like roleplaying sonnet 4.5 to behave like opus 4.6 it strictly said no to act like that
ig anthropic very strict for such roleplays

opaque rapids
#

Does anyone know how to increase the titanic ML score on kaggle? I'm still learning

snow turtle
#

How to fix qqplot error I mean I get yeo johnson transformation error which raise axis=0 ya 1 ,I tried to solve more than twice but I am entered that loop cycle so if you have any method then tell me method to fix!!

golden wadi
#

Hey everyone! 👋 I’m just starting my journey into Data Science and am looking to dive into my first few projects. Does anyone have recommendations on which Kaggle datasets or 'Getting Started' competitions are best for a total beginner? Also, if anyone is open to offering some mentorship or guidance, I’d love to connect. Thanks in advance!

mental cosmos
#

Hii, I have a question I'm in 3rd CS AIML from India haven't been able to get an internship. Was thinking of learning flask for deployment of my projects. Would it be wise to do so or should I Focus on something else??

wraith sparrow
mental cosmos
winged rapids
#

Hello, can I still get a Kaggle badge and certificate if I finish the AI Agents Intensive Course?

verbal crest
earnest plinth
#

hi guys, can anyone recommend a good book for beginners in machine learning?

humble shoal
#

(you can defo find it for free online too (for legal reasons not advice he he))

wraith sparrow
wraith sparrow
sinful root
#

Hey, if it's not legal don't do it.

humble shoal
#

ok nothing to see

west tusk
#

Hello guys this is Sharan ,Do you guys have any crazy ideas for building let's Collab and do the project , let's do it together ,Network is networth 🤝

clever creek
#

Hi crew, I am new to Kaggle, computer science & AI Agents. My question is if we connect our Google_API_Key to our Kaggle Notebooks will it use up my google data or charge me for going over data usage on my google account?

clever creek
#

Also can we copy the ADK and use it to build our own agent in a notebook outside of Kaggle?

rugged juniper
#

Yes, absolutely. Google ADK is a Python library, not a Kaggle-exclusive tool

pip install google-adk
you can install it on your own computer or in a different environment
let me know anytime if you need help

little tundra
#

Hi guys, I'm a new data engineering student. I have good fundamentals in Python and SQL. About a month ago, I started building my first project, and I've faced some knowledge gaps, such as how to use important tools like Docker, Airflow, and PostgreSQL.

My question is: Do you think I should stop my project and improve my foundation, or just keep going and learn these tools to finish the project and, after that, build a solid foundation?

wraith sparrow
flint bramble
#

Hi guys,
I'm facing severe overfitting across multiple models like XGBoost, LightGBM, and others. It's a multiclass classification problem (5 classes) for comment label prediction. Even after applying parameter tuning and PCA, the validation performance remains significantly lower than the training score. Any suggestions on how to properly diagnose and fix this?

rugged juniper
flint bramble
flint bramble
ashen mango
#

Same for every interactive notebook, the issue is that, when something is ruinning in interactive notebook then if internet disconnectes or page is refreshed while its running then it shows that it is connecting yet it never connectes or shows the ourput on jupyter or console untill we restart the session or stop it and start a new one, even doing so the output still never printed. This is problematic because as a researcher I need and many people need to run experiments and if this happnes then we need to wait same amount of time to do it again and get the logs. Please fix this issue!

#

is there any solution? Because I lost some of my very important expriments because of this!

#

Please help me guys!

sinful root
ashen mango
sinful root
#
  1. Not a bro 2) I've passed along your feedback
#

FYI: For long running experiments, we recommend using "Save Version" to kick off a non-interactive session.

solid valve
#

Hi guys I want to learn machine learning in the area of sports analytics, so what courses on coursera would give me a good introduction to that?

quick sonnet
#

Hello everyone! New guy here looking to make some friends and further my Data Analytics skills!

deep hill
#

hiii! does anyone know where i can find game-arena-related stuff? does it have its own discord?

sinful root
sinful root
coral lava
sinful basin
#

Hi everybody. I have a few question about the metodology in a machine learning project. First of all, what are the first steeps u folow when you receive a full schema in CSV?. I mean, i want to know which steeps you folow since you receive the raw data in a schema until u start the feature engineering. Cuz lately i'm feeling i understand the concepts and Technics but no the process.

brazen chasm
#

Is it possible to get a certificate for doing the guide: 5-Day AI Agents Intensive Course with Google?

stable tusk
#

supposing you're in an offline AI programming competition, what would you do first right after you've read the instructions and problems of the ML, NLP and CV problems (given that you're allowed to use LLM when coding)

wraith sparrow
verbal crest
swift rune
#

Hello, i am a host for a community competition, i have uploaded a new solution.csv and would like to request a rescore for the public leaderboard scores. How can i request it? i have sent an email to kaggle support and haven't got a response

vapid pelican
gritty current
#

does anyone has idea how to get into gsoc as a begginer

wraith sparrow
gritty current
errant grove
#

do any of u have trouble downloading the api json in kaggle. I generated a new api token but it does not download anything

#

They said it should automatically download ur api key and info in json

#

I have none

pallid oasis
#

Hi guys do any of you know how to fix Out Of Memory issues in kaggle? I fine-tuned a model just fine with a small context size dataset using pyTorch but now I changed the dataset to one with a larger context size and I also changed the code a bit. Is this the place to be asking this?

pallid oasis
#

Guys I'm trying to use the 2 kaggle GPUs when fine-tuning on PyTorch but it keeps failing. Would like some help pls

eternal hedge
wet wagon
#

help

#

Persona couldn't verify your identity
Your identity couldn't be verified by Persona. If you feel this is in error, contact support.

inner laurel
#

🚀 We’re Hiring: AI Engineer (Full-Time | Remote – India)

NavGurukul’s AI Learning Labs is building AI tools to impact 1 crore underserved students through technology-enabled learning.

📍 Location: Remote-first (with ~30% travel to campuses or Igatpuri HQ)

What you’ll work on:
• Build and experiment with AI solutions for learning at scale
• Design and deploy LLMs (GPT, BERT, open-source models)
• Work on RAG pipelines, vector search, prompt engineering
• Build evaluation frameworks, A/B testing, and model benchmarks
• Develop APIs (FastAPI/Flask) and production-ready AI systems
• Mentor engineers and contribute to AI strategy

What we’re looking for:
• Strong Python + AI/ML frameworks (PyTorch, TensorFlow, Hugging Face)
• Experience with LLMs, NLP, vector databases, prompt engineering
• Knowledge of MLOps, model deployment, monitoring
• Passion for building AI with real social impact

✨ Experience: 2+ years (all levels welcome)

🔗 Apply here: https://recruiterflow.com/navgurukul/jobs/295

If this excites you or someone you know — please share!
Let’s build technology that creates opportunities for millions of learners ❤️

wraith sparrow
#

how is literature review done or anyone of u do it im new to the term so anyone pls ans 😓

safe laurel
#

Hello everyone 👋

I recently reached Expert level in Kaggle Notebooks, and now my next goal is to reach Master level.

I mostly work with Deep Learning using PyTorch, and I enjoy building models and experimenting with architectures. I’m looking for project ideas or research-style notebooks that could help me create high-quality work for Kaggle and improve my DL skills.

I’m especially interested in:
• Deep Learning projects
• Computer Vision or NLP
• PyTorch-based implementations
• Interesting ML ideas are also welcome

If anyone has suggestions for projects, datasets, or Kaggle-style experiments that could help reach the Master level, I would really appreciate it.

torpid ruin
#

Can you tell me how you reached level expert ? @safe laurel

safe laurel
torpid ruin
#

Ok

#

What was the duration that took to become an expert?

safe laurel
torpid ruin
#

Ok

#

So it's notebooks?

#

Or anything else

errant grove
safe laurel
#

More than one year

safe laurel
# torpid ruin So it's notebooks?

Mostly notebooks, but competitions also helped a lot. Good explanations and useful implementations usually get more attention and upvotes.

hidden orbit
#

Hello guys, who's in the titanic competition?
I could use some help, I'm stuck

viscid marsh
viscid marsh
hidden orbit
hidden orbit
viscid marsh
#

yep

spare canopy
#

Hey guys I actually need someone to study with like those who know more about the ece branch and is passionate about vlsi , chip designing or embedded systems (lwky I'm in my 2nd sem) if anyone is interested dm mee (⁠θ⁠‿⁠θ⁠)(⁠θ⁠‿⁠θ⁠)

short haven
#

i am looking for unclean messy data in kaggle for practicing data cleaning .
how i can find them ?

hearty stag
#

Hi everyone,

I’m encountering a very frustrating technical glitch with one of my notebooks and I’m hoping someone from the Kaggle team or the community can help.

The Issue:
My notebook appears completely empty (no code cells) when viewed from the public link or through the "Viewer" mode. However, when I click "Edit," all the code and cells are perfectly visible and intact in the Editor.

What I have tried so far:

  • Save & Run All (Commit): The version finished successfully (green status) with no errors.

  • Permissions: The notebook is set to "Public."

  • Privacy Check: I’ve checked the notebook using different browsers and different account; the code remains invisible.

Clearing Outputs: I’ve tried clearing all cell outputs before saving to ensure it’s not a metadata/rendering size issue.

Specific Notebook Link:
https://www.kaggle.com/code/furqonaryadana/tri-survival-stack-distance-stratified-blending

Has anyone experienced this "Ghost Notebook" behavior recently? Is there a specific metadata setting or a backend sync issue that causes the Viewer to fail while the Editor works fine?

I’ve spent a lot of time debugging the system rather than the code itself, and I would appreciate any insights or a fix from the Kaggle team.

Thanks in advance!

sinful root
# hearty stag Hi everyone, I’m encountering a very frustrating technical glitch with one of m...

That sounds super frustrating - sorry to hear it Furqon! The best way to get this bug addressed by our engineering team is by reporting it here: https://www.kaggle.com/contact#/other/issue . We do actively manage and respond to issues we get there (and it helps us keep track of them and make sure we're taking care of it.) Hopefully we can get a resolution for you!

Contact Kaggle Support

A lot of the common inquiries we receive are listed below. Please click on the one that applies to you to learn more.

dusk elm
#

Hi all , I have been working on extensive doj parser and research pipeline in nativelab for parsing Epstein case files and segregate them as per timing event sequence missing sequence and hypothesis generation , for this currently m choosing , multi context pdf chunk reasoning with back context support , based on continuation keywords and event consolidation , any other technique for retrieval of large files without loading all in context with maximum back context would be helpful. Also needed help on image identification models for low quality images , if any .

limpid garnet
#

`from kaggle_environments import make, evaluate

Create the game environment

Set debug=True to see the errors if your agent refuses to run

env = make("connectx", debug=True)

List of available default agents

print(list(env.agents))

env.run(["random", "random"])

Show the game

env.render(mode="ipython")`

#

returns:
`['random', 'negamax']

FileNotFoundError Traceback (most recent call last)
/tmp/ipykernel_55/3927067547.py in <cell line: 0>()
11
12 # Show the game
---> 13 env.render(mode="ipython")

/usr/local/lib/python3.12/dist-packages/kaggle_environments/core.py in render(self, **kwargs)
371 args = [self]
372 player_html = get_player(
--> 373 window_kaggle, self.html_renderer(*args[: self.html_renderer.code.co_argcount])
374 )
375 if mode == "html":

/usr/local/lib/python3.12/dist-packages/kaggle_environments/envs/connectx/connectx.py in html_renderer()
192 def html_renderer():
193 jspath = path.abspath(path.join(dirpath, "connectx.js"))
--> 194 with open(jspath, encoding="utf-8") as f:
195 return f.read()

FileNotFoundError: [Errno 2] No such file or directory: '/usr/local/lib/python3.12/dist-packages/kaggle_environments/envs/connectx/connectx.js'`

#

Why ?

sour barn
#

@limpid garnet what is this???

limpid garnet
real rain
#

hello, i studied finance in undergrad at the uorochester and looking to transition into data science. i have beginner level coding experience. i'm starting out on kaggle and would appreciate advice or suggestions to improve in the data science path. thank you:)

thick drift
#

Question for the community: what's the best way to handle the "synthetic data credibility" problem on Kaggle?

I just built a Starbucks recommendation engine, but since there's no public POS data, I generated 100K synthetic transactions constrained by real data (FRED CPI/wages, Open-Meteo weather, actual menu
nutrition).

My approach:

  • Every assumption explicitly documented
  • Benchmark comparison against known Starbucks metrics (avg ticket $5-6, Grande preference ~45%, etc.)
  • Perturbation test: vary each assumption ±30% and measure downstream impact
  • Bootstrap stability test across 50 resamples

Is this enough to make synthetic data credible? Or will reviewers always dismiss it regardless of validation rigor?

Project: https://www.kaggle.com/datasets/shiratoriseto/starbucks-recommendation-engine
Validation notebook: https://www.kaggle.com/code/shiratoriseto/starbucks-validation-stress-test

supple basalt
#

Does anyone know how to utilize the GPU resources associated with a Competition? It keeps taking from my base quota, and I get: Accelerator Quota Exceeded
You've exceeded the number of hours allowed to use GPU. Your quota will reset to 30 hours in 2 days. You can continue working in this session until you've reached your 9 hour limit. You will not be able to commit unless you turn off GPU.

Please let me know how to avoid this issue in the future.

green shell
left kite
#

Hello! There was an announcement not too long ago that there is going to be another AI agents intensive that goes into much deeper details than the 5 days one. It was supposed to be at the beginning of this year. Any news about this?

echo scaffold
#

How do I access the Feedback channel?

sinful root
echo scaffold
sinful root
left kite
lavish prairie
#

Hey Kaggle community

First hackathon ever. Solo. No team. Built my submission for the Google DeepMind AGI Hackathon in a few hours and called it done.

I built an original AI benchmark measuring metacognition — not what AI knows, but whether it knows what it knows. Three tasks: knowledge boundary awareness, confidence calibration, and sycophancy resistance. Output is a classification system I'm calling the Kardashev Trust Classification. Type III = genuine epistemic integrity. Type I = surface mimicry.

Results on one frontier model were interesting. Perfect on knowledge boundary awareness. Significant miscalibration on confidence in contested domains. That gap is exactly what I care about.

I'm AuDHD — autistic and ADHD 😅 The hyperfocus and pattern recognition is literally where this came from. I see where things don't fit. That's the whole point.

All my code was written with AI assistance. I have a problem with syntax 😬 The ideas are mine. The execution is... we'll see.

About an hour of research. Few hours to build. First hackathon. No team because I genuinely struggle to collaborate 😭 I still showed up though.

I'm not sure spending just a few hours on a final submission is normal. I may have generated full chaos. Actually I think I just generated chaos 💀

If anyone has worked on metacognition benchmarking or confidence calibration I'd genuinely welcome your perspective. Not looking for validation. Looking for what I'm missing.

Results June 1st. We'll see 🐟

— GiGi

sinful root
lavish prairie
#

Idk what the benchmark scores actually mean ... I'm still learning idk everything and never will.

finite bridge
#

Hi all, I have a question which hopefully isn't too naive to ask: what's the point of late submissions in competitions? is there a penalty for late submissions?

dapper crag
#

Hi everyone 👋,

I’m using XGBoost, LightGBM, and CatBoost for this churn prediction task, and I noticed something interesting:

👉 CatBoost is taking significantly more time to train (~45+ minutes)
while:

  • XGBoost finishes much faster
  • LightGBM is extremely fast

⚙️ My CatBoost Setup

  • iterations = 2000
  • depth = 8
  • learning_rate = 0.03
  • eval_set = (X, y)
  • early_stopping_rounds = 200

❓ My Questions

  1. Why is CatBoost much slower compared to XGBoost and LightGBM?

  2. Is this due to:

    • handling categorical features internally?
    • ordered boosting?
    • GPU vs CPU differences?

  3. Am I using it incorrectly (e.g., eval_set issue)?

  4. How can I reduce training time without hurting AUC?

Would really appreciate insights from experienced users 🙏

Here is my Notebook link: https://www.kaggle.com/code/mahmudulmadu/s6e3-adv-fe-xgboost-lightgbm-catboost-ensemble

vital hazel
lavish arrow
#

Hello there, i have a question.
The first day and the first task, after i create API and everything, i press restart and run all cells, but it just turnrning after the restart, it's not running. What should i do?

uneven marsh
#

Hi there, I have a quick question.. I have an upcoming datathon this weekend (was a last minute decision), and I wanted to ask what general steps should I take to solve the problem cases? Although I may have been in this discord for some time, I haven't found time to catch on AI and ML due to the immense work load from my school. All I (think) know is that I need to clean the dataset, run it through two or three base models, put the intermediate predictions into a stacking model and get the final predictions, where my team and I will give our solution statement. Is there anything else that I should/need to know? #❓┊ask-a-question

gloomy cairn
#

Hey, any 💡 idea HOW & WHERE can I get an API key 🔑 for a OpenSource 🆓 LLM

Which are freemium alternatives?

silk citrus
# gloomy cairn Hey, any 💡 idea HOW & WHERE can I get an API key 🔑 for a OpenSource 🆓 LLM W...

Hello! Well...If you know how to make wrappers, you can make one similar to VS Code. The first one i used was Code Studio( made by Syncfusion) on July 2025. Was the fisrt " VS code wrapers" with all these models, agent mode etc.

You can also try Openrouter , check " free" or " Experimental-->they are calling people to free for all if you send feedback. Check Experimental models on Openrouter--> apply to be tester.

teal smelt
teal smelt
uneven marsh
uneven marsh
teal smelt
teal smelt
short haven
#

I am struggling to find a messy dataset to practice data cleaning. But i am tired of find a perfect dataset.
Could u help me to get that?

sinful jasper
short haven
sinful jasper
#

So that's what make you for cleaning you can use "pandas_profiling" library for complete EDA then you can do cleaning after that

short haven
#

I have plan to clean data in pandas and numpy

sinful jasper
#

Then try to use pandas

fluid cave
#

I’m currently a 4th semester student majoring in Knowledge Discovery / Data Mining, and I’m starting to plan my thesis. but idk what domain, what theme and title should i use :c
I know i still have time but if i get some and create it earlier i feel saver.. if anyone have suggestion im open to it, i still have a lot of time to study about it like 1-2 year.

real frigate
#

Guys, the minor consent form not loading (no input fields)

#

Anyone who can help me?

analog raft
rigid phoenix
#

Is there any GSOC channel

stoic valve
#

Is there any leaderboard of LLM for construction for Knowledge graph

lofty halo
#

can anyone help me understand how do they rank participants??

#

i quite literally got 11 submissions ALL above 0.500!!

old jolt
#

Where can I find more about creating Kaggle multi-turn benchmarks on the website? Asking for the measuring AGI challenge.

#

Also, how do I DM a mod? I would like for my real name to not be used.

sinful root
old jolt
old jolt
#

😭 username is not editable

sinful root
#

On Kaggle, your display name is editable. You would go into your profile, and click "Edit Your Public Profile". You can then change the text that is displayed for your name.

solar stirrup
#

hello

#

@violet crow @twin stream

#

In the kaggle benchmark, Especially the featured ones, There are more than one collums are shown. if these collums are from the one task which a differents task still won't be able to show multiple collum like that with names

#

example: #
Model
Average Score
AssetOpsBench Score
ITBench Score
1
Gemini 3 Pro Preview
2
o3
3
GPT-5
4
Claude Opus 4.5
5
GPT-5.1
6
Gemini 2.5 Flash
7
Claude Sonnet 4.5
8
Gemini 2.5 Pro
9
Qwen 3 235B A22B Instruct
10
Claude Haiku 4.5
11
, This has multiple collums.

#

I want to show, multiple collums for my kaggle benchmarks

#

Any way to show multiple collums for kaggle bench, <@&1130936848743682170> staff @sinful root

tame sigil
vocal thicket
vocal thicket
tame sigil
#

8 Votes that count for medals and tiers

vocal thicket
turbid latch
#

Hi , can I apply for late submission for a competition

#

??

vocal thicket
turbid latch
#

Okay Thank you ..

hearty harness
#

any one from ai developer ???

true needle
#

Hey guys, just wanted to start everything again, what do you suggest? btw want to get in tech role.

rustic dock
#

I wanna discuss and ask questions about the latest playground competition

#

Predicting irrigation need

#

If anyone is also taking part, i would be grateful if you answer some of my dumb doubts.

#

pong on reply please

finite hawk
#

hi, i have started the gemma4 hackaton and have already finetuned the model. i look for a team who can join in, possibly with experience in Kaggle best practices as i am a noob. - how do people proceed normally building a team for kaggle ? Any advice is welcome (within the terms)

fast saffron
azure sinew
#

Hi everyone, I'm currently working on a crypto trading ML project using XGBoost and Optuna for hyperparameter tuning.

I'm running into some difficulties with Optuna — it often gives me models that are either too regularized (low variance predictions) or too noisy and unstable.

I'm not sure how to properly balance the search space or objective to get more consistent results.

Has anyone experienced something similar or can suggest best practices for tuning with Optuna in time series problems?

Any advice would be really appreciated!

carmine jasper
#

Happy Weekend!

Hello Everyone!
If you know someone who have good skills in Python and Machine Learning, Please invite me!

Our Company is open to hire Python and Software Engineer.

Requirements:
2+ years of Software Engineering Experience
C1 or Native English Level
Good vision of Software Trent

Benefits:
Competitive Income
Supporting Several roles and chances
Multiple Role Working is enable

Important:
Our company is designed for Capability Person.

Questions:
For Junior Persons?
Do not give up, strong enthusiasm is also big point and our company also focus on the person's enthusiasm.

Thanks again.
Sophia

sinful root
languid coral
#

Hi For the Gemma 4 good hackathon,

Quick clarification on prize eligibility — the rules state that projects are eligible to win both a Main Track Prize and a Special Technology Prize. Does this mean a single entry can only win one prize from each track? Or is it possible to win multiple prizes within the same track (e.g., both the Ollama and Cactus prizes from the Special Technology Track, or both the Future of Education and Digital Equity prizes from the Impact Track)?

Thanks in advance!

tepid kayak
#

Hi - I'm new to Kaggle and getting a 'Content not viewable in your region' on all images on Alexis Cook's Titanic Tutorial. Anyone know why this could be? Thanks

karmic peak
#

Hi everyone! I’m looking to get more involved in events, hackathons, and meetups focused on Agentic AI. Since local opportunities are a bit limited in my current area, I’m eager to connect with the broader community online. Does anyone have recommendations for active Discord servers, virtual hackathons, or global online meetups where I can contribute and learn?
I am looking for a way to get more involved.

slender crater
#

Hey guys ! I've got Coursera plus and edX premuim at cheaper rates . 2000 rupees for 1 year. Dm to know more .

grand fable
#

Hey! Quick question, did you use CPU or GPU to train your models?

granite sinew
#

I am new to Data Science
Can anyone guide me how I can start my journey?
Like I know python quite well solved more than 300 problems

snow quartz
#

Hello everyone,
Is there any data set available in ಕನ್ನಡ regarding Electronic Health Record?

frosty remnant
#

Hello I wanna is there anyone uses claude with libraries like
Caveman ,claude-mem,obsidian
Or with olama i wanna know how to utilise with maximum efficiency

arctic raven
#

Hi guys rate my website from my profile, any improvement, recommendation etc?

carmine jasper
#

Hello Everyone!
If you know someone who have good skills in Python and Machine Learning, Please invite me!

Our Company is open to hire Python and Software Engineer.

Requirements:
2+ years of Software Engineering Experience
C1 or Native English Level
Good vision of Software Trent

Benefits:
Competitive Income
Supporting Several roles and chances
Multiple Role Working is enable

Important:
Our company is designed for Capability Person.

Questions:
For Junior Persons?
Do not give up, strong enthusiasm is also big point and our company also focus on the person's enthusiasm.

Thanks again.
Sophia

cunning pike
#

I need a fraud detection model that is not card based

carmine jasper
#

Hello Everyone!
If you know someone who have good skills in Python and Machine Learning, Please invite me!

Our Company is open to hire Python and Software Engineer.

Requirements:
2+ years of Software Engineering Experience
C1 or Native English Level
Good vision of Software Trent

Benefits:
Competitive Income
Supporting Several roles and chances
Multiple Role Working is enable

Important:
Our company is designed for Capability Person.

Questions:
For Junior Persons?
Do not give up, strong enthusiasm is also big point and our company also focus on the person's enthusiasm.

How to apply?
DM with resume and 1min's record of your English Speaking

Thanks again.
Sophia

green bluff
#

trying to do the 5-day course for AI and am encountering the 429 error - how do I solve?

sinful root
green bluff
sinful root
#

It's variable - so hard to predict

worn pollen
#

Hello everyone. I've recently landed an internship for ML engineer role at a startup. By background, I'm a fullstack + genAI dev (javascript + python). I will have to give some time to revise python, some for sql, as I know them but haven't used them for a bit.

I prepared basics like linear, logistic regression, bias variance tradeoff, for interview. What libraries can I learn, algorithms to explore that can help me prepare for my internship??

halcyon summit
#

hello everyone, i am actually trying to make a ensemble with two models, one predicting the std of signal and one predicting the signal. i want to use the first to filter what i feed to the second, my question is can the second be trained on same dataset as the first or do i need separate dataset? because the first have been trained on one dataset, if i wanna use his predictions and use the same dataset, result will be optimistic(i think) because the first model will predict on the training set he already seen and the second will use those optimistic predictions as feature. i do not know if its acceptable or not (i an a newbie)

halcyon summit
worn pollen
serene nebula
#

Hope all is well, I was just wondering if there was a mod or high staff I could speak with about offering a partnership with the server?

sinful root
fast basalt
#

Hello, can i have a Txt To Speech from HuggingSpaces that works perfectly? Cuz Huggingspaces models is having tons of errors

If anyone can help me please mention it because i need Txt to speech with voice cloning

fast basalt
#

@verbal crest

Sorry to bother you bud but may i ask about Huggingspaces models having errors when running it inside Kaggle, is that models error or kaggle needs to download something?

teal smelt
tame sigil
#

my frind got his account hacked and got banned but now he got his account back and need to be unbanned
can anyone help?

sinful root
tame sigil
#

here on discord

sinful root
#

You can still file an appeal there

violet vault
#

hello

worn pollen
#

Hello everyone. I'm a beginner in Data Science + ML, and want to explore how to improve my skills in this domain. I've heard that Kaggle competitions can help in this, but how do i get started with this? In short, how can i improve my Data cleaning/ML thinking skills by giving Kaggle competitions.
By the way, i'm heavy into DSA, so i came with a preconception that kaggle competitions are similar to a problem solving DSA contest. Can anyone help me highlight the differences bw these two fields?

potent field
#

why i cant verif my account

abstract tinsel
#

I m trying to train a model but after half an hour kaggle starts to like freezing my instance i tried 2-3 times happening with me again and again

final delta
#

Check my bio 😁

fiery perch
#

My script on a T4 is taking long to start. Way longer than usual. Is something up with the T4s right now?

fiery perch
#

Ok they seem to be running now but but it looks like the time that it was waiting in queue has also eaten up my GPU quota. Anyone here having the same problem? Or anyone who can throw light on this?

autumn umbra
fiery perch
#

Ok. Is P100 capacity any better?

autumn umbra
fiery perch
#

I just need to build stuff with the right cuda and Torch for the RTX

autumn umbra
fiery perch
#

Oh thanks for the heads up. Let me see if I can cross compile on the cpu.

tepid socket
#

RTX ^)))

onyx storm
#

@flint wing

#

Hi

#

I send you a friend request

#

@flint wing

worldly musk
#

Hi @wraith cliff my teammate and I placed 4th in the Tunix Hackathon (Team SandSpec). Although we submitted all the required documentation we have yet to hear back from Kaggle regarding our prize. Since you were the competition host I thought I’d ask you here directly. Do you know if it is being processed? Only because we haven’t received any confirmation from anyone. Any guidance would be appreciated! Tagging @sinful root in case you know of any action items I can take.

verbal owl
#

I was not active on Kaggle in the pass few months, but I find something strange recently. In my "Your Work" page, there are many notebooks and datasets created by other users. I could not delete them from this page. What is happening to my account? Has my account been hacked?

sweet locust
#

I've been rate-limited for downloading orbit-wars games more than 36 hours ago. would that ever clear up? I cant even see games on the browser now.

tender flume
#

hello guys.
I added tqdm bar in my project.
But when I ran something like this: !python train.py --model vit_b_16 --epochs 5 the bar is not loading properly instead it is printing every steps
How can I solve this issue?

gloomy ledge
twin stream
sweet locust
twin stream
sweet locust
twin stream
#

👍 I'll take a look and see what I can find out for you

twin stream
sweet locust
sweet locust
twin stream
tough phoenix
#

Hello everyone, I have been working on the NeuroGolf 2026 championship and was wondering if anyone would be able to answer some questions I had on the terms and conditions for submitting work and what is allowed to be used (for instance, is it acceptable for me to use Perplexity and Claude to accomplish the 400 neural networks in tandem with my own application?)

warm mango
#

​Hi everyone,
​I’m an independent researcher and I’ve just finished a paper on AI Governance Architecture (Assumption-Gated Runtime).

​I’m planning to submit it to arXiv under the cs.AI category, but I currently need an endorsement. I was wondering if anyone here with the necessary permissions would be willing to help me out?

​You can check out my work/preprint here: https://github.com/zhangze1007/Assumption-Gated-Biodesign-Runtime-Artifact

​Endorsement Code: OEOMPS

​Thanks in advance for your support!

tough phoenix
# warm mango ​Hi everyone, ​I’m an independent researcher and I’ve just finished a paper on A...

hi zhangze, I am drafting a position paper that articulates how discrete typed structure beats continuous shared substrate when precision matters in AI and computational verification work. I would like to cite your repository and your accompanying paper as a cross-domain example of this pattern emerging independently in synthetic biology. Would that be okay with you? It would be published on Zenodo

warm mango
# tough phoenix hi zhangze, I am drafting a position paper that articulates how discrete typed ...

Hi Bilal, thanks for your interest. That means a lot to me as an independent researcher.

Quick heads-up: the paper is currently under moderation at engRxiv and should have a DOI shortly. In the meantime, the GitHub artifact is the most stable public reference. Once the preprint is live, I'll update the repository with the DOI link, and you're welcome to use either for citation.

If you want to discuss the conceptual connection between your position paper and my work, I'd be happy to chat more. Thanks again for reaching out.

tough phoenix
# warm mango Hi Bilal, thanks for your interest. That means a lot to me as an independent res...

Thanks for the quick response and the citation guidance, that's very helpful. I'll cite the GitHub artifact per CITATION.cff for now and swap in the engRxiv DOI once it's live.

I'd very much like to take you up on the discussion offer. The conceptual connection feels load-bearing to me, not incidental. What's the easiest format for you, discord message or threaded email? Please let me know

warm mango
tough phoenix
hidden pulsar
#

I am looking developers who can speak english well

proud beacon
#

I'm hosting a hackathon in Georgia would @mild geode be at all interested in sponsoring :)) We've hosted one before @ https://wavehack.org

ivory vigil
#

torch environment problem on P100?

import torch
torch.zeros(10).cuda()
"""
/usr/local/lib/python3.12/dist-packages/torch/cuda/__init__.py:435: UserWarning: 
    Found GPU0 Tesla P100-PCIE-16GB which is of cuda capability 6.0.
    Minimum and Maximum cuda capability supported by this version of PyTorch is
    (7.0) - (12.0)
    
  queued_call()
/usr/local/lib/python3.12/dist-packages/torch/cuda/__init__.py:435: UserWarning: 
    Please install PyTorch with a following CUDA
    configurations:  12.6 following instructions at
    https://pytorch.org/get-started/locally/
    
  queued_call()
/usr/local/lib/python3.12/dist-packages/torch/cuda/__init__.py:435: UserWarning: 
Tesla P100-PCIE-16GB with CUDA capability sm_60 is not compatible with the current PyTorch installation.
The current PyTorch install supports CUDA capabilities sm_70 sm_75 sm_80 sm_86 sm_90 sm_100 sm_120.
If you want to use the Tesla P100-PCIE-16GB GPU with PyTorch, please check the instructions at https://pytorch.org/get-started/locally/

  queued_call()
"""
teal smelt
trail meteor
#

Hi guys, I am looking for a practical work which gives some certificates, I am an automation expert and want to learn about AI, LLM models, everything about AI, can any one help me

worn crater
#

WARNING: Skipping jupyterlab as it is not installed.
WARNING: Retrying (Retry(total=4, connect=None, read=None, redirect=None, status=None)) after connection broken by 'NewConnectionError('<pip._vendor.urllib3.connection.HTTPSConnection object at 0x7c14fe17a8a0>: Failed to establish a new connection: [Errno -3] Temporary failure in name resolution')': /simple/google-genai/
how to fix this error

vapid tundra
#

hello, i cant verify my phone number in the kaggle website. its sayin too many requests even tho i didnt do many, ive waited 15+ mins, i need help.

jagged ember
keen pecan
#

I verified my number

sinful root
midnight goblet
#

guys

#

for Exercise: Conditions and Conditional Statements, Question 5 what are the answers?

teal smelt
#

Oh shoot i cant share images

# 
# mozzarella
get_labels('solid', 91, 330, 6, 0, 750, 1)

# cookies
get_labels('solid', 34, 170, 3.5, 0, 220, 14)

# soda
get_labels('liquid', 240, 170, 0, 0, 85, 44)

(You can find these answers by visiting the mentioned site in the notebook or by viewing the images..)

teal smelt
teal smelt
midnight goblet
rapid void
#

I'm a little new to data science in general. I was wondering if someone could let me know what is the best way for me to learn about data analysis and cleaning specifically so I can then use it for training models

oblique grove
#

guys please help
GPU P100 doesn't work with pytorch cuda 12.8 now on kaggle
any fast fix i need to do a project real quick

oblique grove
lean knoll
#

i wanna do web dev and ai ml .what can i do as in carreer?

#

like is there a job for such a proflie

#

??

ember crypt
#

hi kaggle 目前有哪些AI竞赛可以参加？

ember crypt
rustic garden
#

My name is Ayaan and i want to pursue data science and my question is for all data scientist for becoming a data scientist is certificate and projects is enough or a specific degree is required if yes which degree is required

lavish basalt
#

How many of you love Python more than JavaScript and also learnt Agentic AI and API calling to automate the workflow?

hexed echo
#

when will the new kaggle course of vibe codding starts ?

dusty timber
#

Isn't the date of the duration stated on that page?

hexed echo
dusty timber
#

What? Course is not scheduled till Jun 15th.
What live are you expecting today?

modern pelican
#

Hi, could you please let me know when the training session will start?

midnight crown
#

hi everyone this might be a repeat question but the 5 dday vibecoding course was supposed to start today, the communication said i will receive a setup mail by yesterday , did not get it , also the learning material or a live video on kaggle youtube channel is not available. Did i do anything wrong during the registeration? , help.

crude tartan
#

I don't know same is with me.

dusty timber
dusty timber
#

Where is everyone getting this information from? That the course starts today?

#

If from an influencer, kindly someone inform that person to update with the correct information.

slow ingot
#

The 5-Day AI Agents: Intensive Vibe Coding Course With Google will start in June 15 - 19, 2026

slow ingot
pine acorn
#

also are there any other even i can join right now or will begin soon other than gen ai agents

teal smelt
teal smelt
pine acorn
#

what kind of competition, do i have submit my code?

teal smelt
pine acorn
#

thank for details but you last text made it confusing. can you brief what exactly i as a participant has to do between 15 jun to 20 jun?

teal smelt
# pine acorn thank for details but you last text made it confusing. can you brief what exactl...

During the 5 day period, they'll do a Live stream(each day, usually around 12:30am IST) and will teach you a topic everyday...
After each livestream you'll get a few assignments that you have to do (take it as homework after the Daily lecture)

Then, by the 5th day if you follow the materials properly, you'll be ready to participate in the competition 🐣...

The competition is a way to showcase the skills you've learnt along the 5day period

shadow rapids
#

where can i watch the previous recorded video

dusty timber
# pine acorn thank for details but you last text made it confusing. can you brief what exactl...

If I may ask: how did you come to be at this Discord server? Via an influencer post? Did they not link to the course/competition website? If they have, have you not read the course/competition website?

Or they did not link to the website at all? And just to this server? Now, this would be problematic. Which explains a lot on why so many were expecting the 5Day AI course to commenced yesterday. And how so many are here without knowing details about the course/competition.

crystal rose
#

Hi,I am a 12th pass student and I am entering college so what course or technology should i learn to match to industrial requirements

coral valley
minor meadow
#

how you guys think this platform can help me land on a career , i'm a bsc mathematics fresh graduate , done google data analytics professional certificate . can you point me out how should i work on this platform to actually become valuable , or should i join a data analytics offline course where they teach from square 1

sinful root
keen pecan
#

@sinful root what skills must a cse student master keeping in mind the industrial requirements ?can u share your experience?

sinful root
#

The environment is changing very fast. When I was in college, the hot skill was Java2 and HTML. Honestly, the biggest advice I'd give to this community during the 5 Days course is to read the materials we provide carefully, look for answers before asking questions, and try to imagine what a good answer to your question might look like. If you have no idea, probably no one else will either.

primal needle
#

im new to machine leanring and i want to advance fast to deep learning ,RAGS and stuff i wanted to ask the profesionals how should i approch it in what sequence should i learn it that itll give me 80% of results from 20% effort

#

im looking for clever ways that will help me advance fast .like what are the algorithems that actually matter in machine leanring that i should learn and move on to advance feilds, and so on

sterile parrot
# primal needle im new to machine leanring and i want to advance fast to deep learning ,RAGS and...

Sorry to be blunt but "80% results from 20% effort" is impossible. It takes effort to learn. If you are just after money, go into business then lots of money there. Computer Science/ML has been ruined by people who want a quick buck and they have had a devastating impact on society by all the people getting fired and every product getting worse, eg Github's uptime or Windows 11 constantly breaking and all the AI and Crypto scams. Please stay out of the field if that is your goal. Remember the lunchbox quote from Jurassic Park:
"I'll tell you the problem with the scientific power that you're using here: it didn't require any discipline to attain it. You read what others had done and you took the next step. You didn't earn the knowledge for yourselves, so you don't take any responsibility for it. You stood on the shoulders of geniuses to accomplish something as fast as you could and before you even knew what you had you patented it and packaged it and slapped it on a plastic lunchbox, and now you're selling it, you want to sell it!"
If you are actually interested in the field, I suggest Deep Learning with Python, Third Edition by François Chollet and Machine Learning Bootcamp both from Manning Publications.

frosty remnant
#

Hey anyone is working on agentic project or something should i check the repo bcz in my insta algo there are the creator who make it sounds easy to create an Jarvis personal assistant and central intelligence layer and brain like obsidian with claude code
Anyone check these type of repo?

teal smelt
#

@sinful root @twin stream someone sent a malicious link

hearty harness
#

Hey everyone!
Are you interested in AI/ML? If yes, let me know!
I’m currently working on a personal AI assistant project inspired by the idea of Jarvis — but it’s not a copy of Jarvis. I’m trying to build something unique and different with my own vision and features.
My goal is to turn this project into a real startup in the future.

hallow sierra
#

so umm

mellow merlin
#

Hi everyone,
I am planning to build a local desktop data tool, and I need your expert advice on both the technical side and the business/pricing strategy.
My main idea is to create a local app that provides a massive, high-performance leap compared to Excel. It will allow users to process, clean, and analyze giant datasets (millions of rows) completely offline on their machines, with a dead-simple, one-click user interface. I plan to offer this tool as a monthly subscription.
Alternatively, I am considering a web-scraping product price tracker (also on a monthly subscription).
From your experience:
Which of these two ideas has higher profit potential and would be easier to sell as a monthly subscription?
For the Excel alternative, what would be a reasonable monthly price to charge users or small businesses for a tool like this?
What Python GUI frameworks and data engines (like Polars) do you recommend to keep the app fast, local, and incredibly user-friendly?
Thanks in advance for your insights!

autumn pawn
#

Is it allowed to use jupyter notebook or colab for Titanic?

thick musk
#

The World Has a Data Problem. We Fix It.
Every AI team hits the same wall eventually.
You have the model. You have the architecture. You have the engineers. But you don't have the data, and everything stops.
Maybe your dataset is too small to train on. Maybe it carries sensitive patient records, financial transactions, or personal identifiers that legal won't let you touch. Maybe you've been waiting months for a vendor to deliver labeled data that still isn't ready. Maybe your edge cases are so rare in real life that your model keeps failing exactly where it matters most.
This is not a skill problem. This is a data problem. And it is quietly killing more AI projects than any other single reason.
We generate synthetic data.
Not as a workaround. Not as a compromise. As a legitimate, statistically rigorous alternative that lets your team move again. We produce tabular, text, image, and time-series synthetic datasets that mirror the distributions, correlations, and behavioral patterns of real-world data without exposing a single real record.
We have solved this for teams in healthcare who couldn't share patient data across departments. For fintech companies building fraud detection models with almost no real fraud examples to train on. For startups that needed 10x their dataset size before a funding deadline. For enterprises blocked by GDPR, HIPAA, and compliance teams that said no to everything.
The problem you are sitting with right now, whether it is a privacy blocker, a data scarcity issue, a class imbalance, a regulatory wall, or a timeline that real data collection simply cannot meet, has a solution. We will tell you exactly what it is within 24 hours of hearing from you.
No long sales cycles. No vague proposals. You describe your data problem in plain language, and we come back with a concrete plan.
Send us your situation: [synthox.ai@gmail.com]
The only thing worse than a data problem is spending another month pretending it will resolve itself.

heavy gust
#

hello ,
i am new to kaggle and want to try the titanic competition, can anyone tell me how to start it ?

merry trail
#

hello @everyone does anyone work with ONNX runtime engines, and optimizing pytorch models using this, and accelerating models on diff hardware acceletors ?

#

request to please dm if you do.

karmic kelp
teal smelt
#

You know, it's just a typical youtube livestream that will be later available as a video 🐣🐣🐣

karmic kelp
loud willow
#

what should i do first if I wanna increase my accuracy?

dusty timber
#

Understanding why Accuracy is the right metrics?
Or, is there another way to determine 'best'? What is the aim - best model - what determines a 'best' model? lowest training time? AUC vs Accuracy?

It depends on context doesn't it?

loud willow
#

Oh....so good accuracy doesn't necessarily mean it's a good model either?

dusty timber
#

No, it really depends on the problem you're solving. Thus, context.

#

Also, is it for a competition/project or an app you have to deliver? Then it means, the metric is whatever the competition is measuring to have a Leaderboard so that ranking is possible. Or the objectives of the project/capstone.

quasi breach
#

Hello is there anyone who has worked in space debris

urban flame
#

🚀 Hey @everyone!

We’re building PromptGram and aiming to reach 150 GitHub stars ⭐

If you like AI, FastAPI, microservices, or developer tools, please check it out and support the project by starring the repo 🙌

GitHub Repo: https://github.com/dewangsahuji/promptgram

Every star really helps and motivates us to improve the project further 💙

vapid tundra
#

hi, ive verified my account with the face verification, but its still not letting me do the phone verification it says ive tried to many times (Which i didnt) im trying to use gpu

delicate yarrow
#

Hello everyone, what is the difference from a data scientist to a data analyst in practical terms?

teal smelt
teal smelt
vapid tundra
teal smelt
#

yea, the ticket method usually works, you'll just have to wait a lil

vapid tundra
#

problem is*

teal smelt
#

you can try it on google colab if that helps?

vapid tundra
teal smelt
#

oh damn ;-;

vapid tundra
#

its fine tho its not end of the world, i can still manage

teal smelt
#

we can just wait now i guess

#

yep

vapid tundra
#

thanks anyway, ill wait for the ticket

tranquil solar
#

Yo!! I am a DS and AI enthusiast. I am new here and I joined so that I could learn how to build projects and add it to my profile. Can someone explain how this all works or is there a channel guide where I can get all the info?

Also, if anyone wants to team up with me to create these projects then ping me. I would most likely be on the learning and help as much as possible.

valid jay
old gull
#

Where can I share my tiny model golf?

#

It's a huggingface competition. The winner takes home 50 runpod credits

dusty timber
old gull
#

ty

onyx storm
#

What is the best data science course?

narrow stag
# onyx storm What is the best data science course?

I don't think there's a single "best" data science course because it depends on your background and goals.

For beginners, I'd focus less on finding the perfect course and more on building a strong foundation in Python, SQL, statistics, and data analysis. In my experience, people often spend too much time comparing courses and not enough time working on real datasets.

A good course can help you learn concepts, but real progress usually comes from applying those concepts through projects. If I had to choose, I'd rather complete a decent course and build three solid projects than finish the most popular course without any hands-on practice.

What's your current background? Programming, analytics, or something completely different?