Kicking off GPU Mode [D]
Hey ! I’m starting a series to document my work on GPU infrastructure, LLMs, and CV. Stop #1 is up: A brief look at why GPUs…
Latest
I silently break training codes or configs so I made pybench [P]
It is like pytest but for statistical tests: it ensures no regression of your metrics at a statistical level. It manages tedious things such that seeds,…
Late Submission of NeurIPS Review [R]
I submitted one of my NeurIPS review ~6 hrs later than the official deadline. Will this still affect my own submission? Asking because I’m a first…
A debugger for RL reward functions that detects reward hacking during training [P]
While experimenting with GRPO training, I kept running this shit that when reward increases, it becomes difficult to tell whether the policy is genuinely improving or…
Live Continual Learning in Machine Learning [D]
My question on live continual learning use cases was removed by moderators here because they think i asked basic level question about live continual learning which…
How’re you deploying LLMs in production now-a-days? What’s the best and most affordable way? [D]
I've been developing an AI product using LLM APIs (from OpenRouter) but want to deploy an open-source LLM in my own Prod env. which I can…
Showcase: geolocating a dashcam video without GPS, only from the footage [P]
Sharing a project I have been working on called Third Eye. It does visual geolocation. Given a video, it figures out where it was filmed using…
For ECCV, Springer Metor. How are we supposed to upload the files? [D]
source files + final paper pdf. ZIP containing the source files and final paper.pdf. Where does the supplemental materiel get uploaded? Because in that email it…
Kuma: compiling PyTorch models into self-contained WebGPU executables [P]
I've been experimenting with a compiler/runtime project that I'm not entirely sure is a good idea, so I'd love some feedback from people who've worked on…
Dev Log on Steam Recommender[P]
Since the steam sale is live I wanted to post a Dev log on my personal project https://nextsteamgame.com/ sharing some outcomes from the web traffic and…