HackerTrans
TopNewTrendsCommentsPastAskShowJobs

smaddrellmander

90 karmajoined 3 個月前
idlemachines.co.uk

Submissions

The annotated PyTorch training loop

idlemachines.co.uk
81 points·by smaddrellmander·20 天前·19 comments

Mae vs. MSE: more than just the mean vs. median debate

idlemachines.co.uk
1 points·by smaddrellmander·24 天前·0 comments

DiffusionGemma: Discrete diffusion in a large language model

idlemachines.co.uk
3 points·by smaddrellmander·上個月·1 comments

Heaven knows I'm perplexed now

idlemachines.co.uk
2 points·by smaddrellmander·上個月·0 comments

Reading MAI's efficiency gain. How to pick architectures like serious people

idlemachines.co.uk
9 points·by smaddrellmander·上個月·0 comments

MAI-Thinking-1: Building a Hill-Climbing Machine [pdf]

microsoft.ai
3 points·by smaddrellmander·上個月·0 comments

Are contrastive losses just cross entropy all along?

idlemachines.co.uk
2 points·by smaddrellmander·上個月·0 comments

Every token, everywhere, all at once

idlemachines.co.uk
2 points·by smaddrellmander·2 個月前·0 comments

The cut in the Mixture of Experts compute graph

idlemachines.co.uk
1 points·by smaddrellmander·2 個月前·0 comments

DeepSeek V4 from the Inside

idlemachines.co.uk
2 points·by smaddrellmander·2 個月前·0 comments

Softmax, can you derive the Jacobian? And should you care?

idlemachines.co.uk
134 points·by smaddrellmander·3 個月前·47 comments

Gemma 4 is not your standard transformer

idlemachines.co.uk
2 points·by smaddrellmander·3 個月前·0 comments

comments

smaddrellmander
·4 天前·discuss
Love this! Cute graphics and the way the balls combine is really fun. Great job.
smaddrellmander
·2 個月前·discuss
A similar vein, asking why in PRs. Most people don’t do the above, and asking why when it doesn’t make sense really helps understanding what your colleagues are actually thinking. Helps create a unified perspective on the codebase.