OOM killed at epoch 99 of 100 Test data leaked into training set The data is too noisy to use Restart kernel, run all, pray Who annotated this?! Works perfectly on my machine Loss went to NaN on step 1 Server crashed during the live demo CUDA out of memory “Quick 5- min sync” = 45 minutes The server is suspiciously slow today Just add ONE more feature before launch Forgot to shuffle the dataset Wait… which cell did I run last? The labels are completely wrong I’ll document this later Can we make it real-time? Server rebooted overnight. RIP my results. Stakeholder ghosted after week 2 Forgot to kill my process… sorry team! Dataset has 47 samples. Total. Pet photobombed the video call Extreme class imbalance strikes again Actually, let’s pivot the entire approach Free! SSH connection died mid- experiment Internet died at demo time We need more data This should be simple, right? Nice pajama top on the call This notebook is 2000 cells long Who left training running ALL weekend? A meeting could have been an email Requirements changed… again Someone hogged all GPUs The business need has… evolved Jupyter kernel died mysteriously As per my last email… Ground truth isn’t actually true Can we get 99.9% accuracy? When will the model be production- ready? OOM killed at epoch 99 of 100 Test data leaked into training set The data is too noisy to use Restart kernel, run all, pray Who annotated this?! Works perfectly on my machine Loss went to NaN on step 1 Server crashed during the live demo CUDA out of memory “Quick 5- min sync” = 45 minutes The server is suspiciously slow today Just add ONE more feature before launch Forgot to shuffle the dataset Wait… which cell did I run last? The labels are completely wrong I’ll document this later Can we make it real-time? Server rebooted overnight. RIP my results. Stakeholder ghosted after week 2 Forgot to kill my process… sorry team! Dataset has 47 samples. Total. Pet photobombed the video call Extreme class imbalance strikes again Actually, let’s pivot the entire approach Free! SSH connection died mid- experiment Internet died at demo time We need more data This should be simple, right? Nice pajama top on the call This notebook is 2000 cells long Who left training running ALL weekend? A meeting could have been an email Requirements changed… again Someone hogged all GPUs The business need has… evolved Jupyter kernel died mysteriously As per my last email… Ground truth isn’t actually true Can we get 99.9% accuracy? When will the model be production- ready?
(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.
OOM killed at epoch 99 of 100
Test data leaked into training set
The data is too noisy to use
Restart kernel, run all, pray
Who annotated this?!
Works perfectly on my machine
Loss went to NaN on step 1
Server crashed during the live demo
CUDA out of memory
“Quick 5-min sync” = 45 minutes
The server is suspiciously slow today
Just add ONE more feature before launch
Forgot to shuffle the dataset
Wait… which cell did I run last?
The labels are completely wrong
I’ll document this later
Can we make it real-time?
Server rebooted overnight. RIP my results.
Stakeholder ghosted after week 2
Forgot to kill my process… sorry team!
Dataset has 47 samples. Total.
Pet photobombed the video call
Extreme class imbalance strikes again
Actually, let’s pivot the entire approach
Free!
SSH connection died mid-experiment
Internet died at demo time
We need more data
This should be simple, right?
Nice pajama top on the call
This notebook is 2000 cells long
Who left training running ALL weekend?
A meeting could have been an email
Requirements changed… again
Someone hogged all GPUs
The business need has… evolved
Jupyter kernel died mysteriously
As per my last email…
Ground truth isn’t actually true
Can we get 99.9% accuracy?
When will the model be production-ready?