iarai / concurrent-dataloader
Profiling and Improving the PyTorch Dataloader for high-latency Storage
☆18Updated last year
Related projects ⓘ
Alternatives and complementary repositories for concurrent-dataloader
- Inference framework for MoE layers based on TensorRT with Python binding☆41Updated 3 years ago
- DL Dataloader Benchmarks☆18Updated 2 months ago
- ☆22Updated 11 months ago
- A Python library transfers PyTorch tensors between CPU and NVMe☆98Updated last week
- 📑 Dive into Big Model Training☆110Updated last year
- Fairring (FAIR + Herring) is a plug-in for PyTorch that provides a process group for distributed training that outperforms NCCL at large …☆63Updated 2 years ago
- Distributed DataLoader For Pytorch Based On Ray☆24Updated 3 years ago
- A performant, memory-efficient checkpointing library for PyTorch applications, designed with large, complex distributed workloads in mind…☆146Updated 2 weeks ago
- Research and development for optimizing transformers☆125Updated 3 years ago
- FTPipe and related pipeline model parallelism research.☆41Updated last year
- MLPerf™ logging library☆30Updated this week
- ☆35Updated 3 years ago
- Odysseus: Playground of LLM Sequence Parallelism☆57Updated 5 months ago
- Memory Optimizations for Deep Learning (ICML 2023)☆60Updated 8 months ago
- An I/O benchmark for deep Learning applications☆69Updated 3 weeks ago
- Sequence-level 1F1B schedule for LLMs.☆17Updated 5 months ago
- NAACL '24 (Best Demo Paper RunnerUp) / MlSys @ NeurIPS '23 - RedCoast: A Lightweight Tool to Automate Distributed Training and Inference☆61Updated last month
- High performance RDMA-based distributed feature collection component for training GNN model on EXTREMELY large graph☆48Updated 2 years ago
- ☆88Updated 2 months ago
- A resilient distributed training framework☆85Updated 7 months ago
- CUDA 12.2 HMM demos☆17Updated 3 months ago
- PyTorch bindings for CUTLASS grouped GEMM.☆68Updated 4 months ago
- ☆70Updated 2 years ago
- pytorch-profiler☆50Updated last year
- ☆79Updated 2 months ago
- PSTensor provides a way to hack the memory management of tensors in TensorFlow and PyTorch by defining your own C++ Tensor Class.☆10Updated 2 years ago
- Dynamic Memory Management for Serving LLMs without PagedAttention☆238Updated last week
- ☆11Updated last year
- Official implementation of ICML 2024 paper "ExCP: Extreme LLM Checkpoint Compression via Weight-Momentum Joint Shrinking".☆41Updated 4 months ago