Stories by jxmorris12

Colophon

Training a 70B Model from Scratch

A Meticulous Guide to Advances in Deep Learning Efficiency

I think Yann Lecun was right about LLMs (but perhaps only by accident)

Please Commit More Blatant Academic Fraud (2021)

Demystifying Noise Contrastive Estimation

It's time to become an ML engineer (2022)

Approximating KL Divergence (2020)

The Ultra-Scale Playbook: Training LLMs on GPU Clusters

Implementing LLaMA3 in 100 Lines of Pure Jax

Outperforming cuBLAS on H100: A Worklog

Gravel Map

Stochastic Integration for Poets

NASA writes space-proof code [video]

Large Lambda Model

Softmax forever, or why I like softmax

Diffusion Without Tears

AGI Safety Course Workbook

7
8
9
10
11
12
13
14
15
16