gpu-memory-management

Here are 3 public repositories matching this topic...

basnetlachu / memopt

MEMOPT: Universal Memory Fabric for AI Infrastructure. Open-sourced by Sophisticates.

ai cuda apache2 vmm kv-cache llm-inference ai-infrastructure gpu-memory-management ai-memory-system sophisticates memopt

Updated Sep 24, 2026
Python

Serving engine for tabular foundation models (TabICL, TabPFN): fit a table once, keep its encoded context on the GPU, and serve fast repeat predictions to many tenants, with memory-aware admission and multi-GPU placement.

python multi-tenant gpu tabular-data inference-server model-serving kv-cache foundation-models in-context-learning ray-serve tabpfn gpu-memory-management tabicl

Updated Oct 21, 2026
Python

arponbiswasanik / optzero

Star

A custom distributed training framework implementing ZeRO-1 and ZeRO-2 memory partitioning from scratch in PyTorch. Features async gradient hooks, sharded AdamW, and distributed checkpointing without relying on FSDP or HuggingFace Trainer.

pytorch autograd distributed-training backend-infrastructure gpu-memory-management zero-redundancy-optimizer

Updated Aug 4, 2026
Python

Add this topic to your repo

To associate your repository with the gpu-memory-management topic, visit your repo's landing page and select "manage topics."

Learn more

Navigation Menu

gpu-memory-management

Here are 3 public repositories matching this topic...

basnetlachu / memopt

VectorInstitute / tabctx

arponbiswasanik / optzero

Add this topic to your repo