msaroufim
Follow
Lists (1)
Stars
A terminal UI dashboard for monitoring multiple Claude Code sessions in tmux
FlexAttention based, minimal vllm-style inference engine for fast Gemma 2 inference.
PostTrainBench measures how well CLI agents like Claude Code or Codex CLI can post-train base LLMs on a single H100 GPU in 10 hours
The Chrome OS Virtual Machine Monitor - Mirror of https://chromium.googlesource.com/crosvm/crosvm/
Daytona is a Secure and Elastic Infrastructure for Running AI-Generated Code
A curriculum for learning about gpu performance engineering, from scratch to what the frontier AI labs do
Post processing library used to analyze memory snapshots
Bolt is a C++ template library optimized for GPUs. Bolt provides high-performance library implementations for common algorithms such as scan, reduce, transform, and sort.
General plug-and-play inference library for Recursive Language Models (RLMs), supporting various sandboxes.
CUDA Tile IR is an MLIR-based intermediate representation and compiler infrastructure for CUDA kernel optimization, focusing on tile-based computation patterns and optimizations targeting NVIDIA te…
Convert nvprof profiles into about:tracing compatible JSON files
A Python script to convert the output of NVIDIA Nsight Systems (in SQLite format) to JSON in Google Chrome Trace Event Format.
Effortlessly migrate GitHub repositories to Codeberg! Seamlessly transfer project. Powered by Bash, curl, and jq.
Code for the paper “Four Over Six: More Accurate NVFP4 Quantization with Adaptive Block Scaling”
A good looking terminal emulator which mimics the old cathode display...
slime is an LLM post-training framework for RL Scaling.
Miles is an enterprise-facing reinforcement learning framework for LLM and VLM post-training, forked from and co-evolving with slime.
VLA-0: Building State-of-the-Art VLAs with Zero Modification
An early research stage expert-parallel load balancer for MoE models based on linear programming.
A fast communication-overlapping library for tensor/expert parallelism on GPUs.
Private voice keyboard, AI chat, images, webcam, recordings, voice control with >= 4 GiB of VRAM.
https://github.com/eunomia-bpf homepage, documents and blogs