Test data leaked into training set Requirements changed… again Who left training running ALL weekend? Just add ONE more feature before launch Jupyter kernel died mysteriously OOM killed at epoch 99 of 100 Can we get 99.9% accuracy? Forgot to shuffle the dataset The labels are completely wrong SSH connection died mid- experiment Forgot to kill my process… sorry team! We need more data Restart kernel, run all, pray As per my last email… Works perfectly on my machine The business need has… evolved Actually, let’s pivot the entire approach Server rebooted overnight. RIP my results. A meeting could have been an email Internet died at demo time This notebook is 2000 cells long This should be simple, right? Who annotated this?! “Quick 5- min sync” = 45 minutes The server is suspiciously slow today Someone hogged all GPUs CUDA out of memory Pet photobombed the video call Nice pajama top on the call Extreme class imbalance strikes again Stakeholder ghosted after week 2 Server crashed during the live demo I’ll document this later Wait… which cell did I run last? Loss went to NaN on step 1 Ground truth isn’t actually true Can we make it real-time? When will the model be production- ready? Dataset has 47 samples. Total. The data is too noisy to use Free! Test data leaked into training set Requirements changed… again Who left training running ALL weekend? Just add ONE more feature before launch Jupyter kernel died mysteriously OOM killed at epoch 99 of 100 Can we get 99.9% accuracy? Forgot to shuffle the dataset The labels are completely wrong SSH connection died mid- experiment Forgot to kill my process… sorry team! We need more data Restart kernel, run all, pray As per my last email… Works perfectly on my machine The business need has… evolved Actually, let’s pivot the entire approach Server rebooted overnight. RIP my results. A meeting could have been an email Internet died at demo time This notebook is 2000 cells long This should be simple, right? Who annotated this?! “Quick 5- min sync” = 45 minutes The server is suspiciously slow today Someone hogged all GPUs CUDA out of memory Pet photobombed the video call Nice pajama top on the call Extreme class imbalance strikes again Stakeholder ghosted after week 2 Server crashed during the live demo I’ll document this later Wait… which cell did I run last? Loss went to NaN on step 1 Ground truth isn’t actually true Can we make it real-time? When will the model be production- ready? Dataset has 47 samples. Total. The data is too noisy to use Free!
(Print) Use this randomly generated list as your call list when playing the game. There is no need to say the BINGO column name. Place some kind of mark (like an X, a checkmark, a dot, tally mark, etc) on each cell as you announce it, to keep track. You can also cut out each item, place them in a bag and pull words from the bag.
Test data leaked into training set
Requirements changed… again
Who left training running ALL weekend?
Just add ONE more feature before launch
Jupyter kernel died mysteriously
OOM killed at epoch 99 of 100
Can we get 99.9% accuracy?
Forgot to shuffle the dataset
The labels are completely wrong
SSH connection died mid-experiment
Forgot to kill my process… sorry team!
We need more data
Restart kernel, run all, pray
As per my last email…
Works perfectly on my machine
The business need has… evolved
Actually, let’s pivot the entire approach
Server rebooted overnight. RIP my results.
A meeting could have been an email
Internet died at demo time
This notebook is 2000 cells long
This should be simple, right?
Who annotated this?!
“Quick 5-min sync” = 45 minutes
The server is suspiciously slow today
Someone hogged all GPUs
CUDA out of memory
Pet photobombed the video call
Nice pajama top on the call
Extreme class imbalance strikes again
Stakeholder ghosted after week 2
Server crashed during the live demo
I’ll document this later
Wait… which cell did I run last?
Loss went to NaN on step 1
Ground truth isn’t actually true
Can we make it real-time?
When will the model be production-ready?
Dataset has 47 samples. Total.
The data is too noisy to use
Free!