Wave optical models and inverse algorithms for label-agnostic imaging of density & orientation.

Specification for the bioimage.io model description file.

Deep-learning library for image restoration.

Fork of tlambert03/microsim — light microscopy simulation in python. Development fork for testing and submitting fixes upstream.

A set of tools to simulate expansion microscopy expriments under different labeling and imaging conditions. Python 3 port, updated from Python 2.

GB10/DGX Spark random-access memory characterization using a GUPS-derived workload — cpu_gups, cuda_gups, managed_gups. Pascal-verified, GB10 hardware run pending.

OpenMP Offloading Validation & Verification Suite; Official repository. We have migrated from bitbucket!! For documentation, results, publication and presentations, please check out our website ->

GB10/DGX Spark hardware-aware fork: UMA-safe expert eviction, /proc/meminfo EMV memory tracking, PSI-based swap pressure detection for out-of-core MoE on unified memory platforms

Code for "Language Models as Causal Effect Generators" (SD-SCMs) — adds GB10/DGX Spark unified memory support

Implementing transformers from scratch

Optimized SGLang MoE kernel configs for NVIDIA DGX Spark (GB10) - GLM-4.7-FP8 and other MoE models

Nemotron-3-Super-120B on DGX Spark - vLLM configuration, hardened agentic stack, and benchmark methodology

An interactive NVIDIA-GPU process viewer and beyond, the one-stop solution for GPU process management.

NVML unified memory shim for NVIDIA DGX Spark Grace Blackwell GB10 - enables MAX Engine, PyTorch, and GPU monitoring

A unified tool for collecting system logs and other debug information

Enhanced GPU throttle diagnostic for DGX Spark (GB10): NVML direct telemetry, throttle cause decoder, PCIe link monitoring, baseline drift detection, timeline capture.