Thanks for suggesting. But since this is a very different model, this would not be possible. Adding 5 chapters on implementing this, training this, evaluating this, etc would be a whole other book.
@rasbt
-
Challenges in Model Evaluation and Elo Rating Leaderboards
By
–
Good question. I think the challenge here is that the evaluation is more challenging because of a) it's relative to all the other models
b) it's expensive to run the other models PS: Fun fact, I just wrote a tutorial on Elo ratings and constructing a leaderboard last week: -
TRM motivation and specialized model design considerations
By
–
Thanks for sharing. Points a 1 and 2 are actually what motivated the work on TRM as mentioned in the paper. I think point 3 makes sense as this is a small, specialized model, not a large generalist model.
Point number 4 also makes sense, this has always been true for ML and DL. -
Neural Network Implementation in NumPy with Manual Backpropagation
By
–
Been there, done that . But how about a a neural net in NumPy with manual backprop? https://
github.com/rasbt/python-m
achine-learning-book/blob/master/code/ch12/neuralnet.py
… That was 10 years ago, makes me feel so old now -
Super Mario Kart AI racers lacked random items due to compute limits
By
–
And at the same time In Super Mario Kart (SNES) the AI racers couldn’t get random items because of compute limitations (implementing individual random generators would have been to costly)
-
Early Video Game AI: If-Else Statements and Public Introduction
By
–
Has to be if-else statements. This was how video game AIs were originally implemented (starting with Pong, Space Space Invaders, PAC-Man).
Probably the first touch point of the general public with the term AI, where the term was deliberately used in a product for the general -
ARC Primary Focus with Sudoku as Retroactive Addition Example
By
–
Yes, very well said. Although, I do think this was primarily built with ARC in mind (for which there is no search algo) and Sudoku was added retroactively to show it is not specific to ARC.
The other thing about Sudoku is is that it is kind of easily self-verifiable. -
Tool-calling LLM with specialized units as tools
By
–
That would basically come back to the idea of having a tool-calling LLM where each of the units would be a specialized tool, no?
-
Lightning AI Cloud Compute Platform Offers Competitive GPU Pricing
By
–
I love this upgrade. Lightning AI is my go-to cloud compute platform due to its user-friendliness (great UI, persistent environment, multi-GPU and multi-node support, etc), and now the prices are also really great. An A100 for $1.55/hour through Lambda Labs or an H100 for $2.70 x.com/_willfalcon/st…
-
AI boom increases work expectations instead of reducing hours
By
–
Just talked about that last week. Since the post-2022 AI boom, companies expect workers to work more, not less.
I think we are closer to a 6-day than a 4-day work week.