kwai / Megatron-Kwai
[USENIX ATC '24] Accelerating the Training of Large Language Models using Efficient Activation Rematerialization and Optimal Hybrid Parallelism
☆45Updated 3 months ago
Related projects ⓘ
Alternatives and complementary repositories for Megatron-Kwai
- ☆64Updated 3 months ago
- PyTorch bindings for CUTLASS grouped GEMM.☆68Updated 4 months ago
- ☆52Updated last week
- nnScaler: Compiling DNN models for Parallel Training☆74Updated 3 weeks ago
- ☆65Updated 3 years ago
- A baseline repository of Auto-Parallelism in Training Neural Networks☆142Updated 2 years ago
- ☆79Updated 2 months ago
- Dynamic Memory Management for Serving LLMs without PagedAttention☆238Updated last week
- Since the emergence of chatGPT in 2022, the acceleration of Large Language Model has become increasingly important. Here is a list of pap…☆175Updated 2 weeks ago
- A collection of memory efficient attention operators implemented in the Triton language.☆219Updated 5 months ago
- Puzzles for learning Triton, play it with minimal environment configuration!☆121Updated last week
- Zero Bubble Pipeline Parallelism☆281Updated last week
- ☆140Updated 6 months ago
- ☆70Updated 2 years ago
- A fast communication-overlapping library for tensor parallelism on GPUs.☆224Updated 3 weeks ago
- High performance Transformer implementation in C++.☆82Updated 2 months ago
- ☆74Updated last month
- Summary of some awesome work for optimizing LLM inference☆37Updated 2 weeks ago
- USP: Unified (a.k.a. Hybrid, 2D) Sequence Parallel Attention for Long Context Transformers Model Training and Inference☆357Updated this week
- flash attention tutorial written in python, triton, cuda, cutlass☆202Updated 5 months ago
- ATC23 AE☆43Updated last year
- A tiny yet powerful LLM inference system tailored for researching purpose. vLLM-equivalent performance with only 2k lines of code (2% of …☆104Updated 4 months ago
- Flash-LLM: Enabling Cost-Effective and Highly-Efficient Large Generative Model Inference with Unstructured Sparsity☆180Updated last year
- A low-latency & high-throughput serving engine for LLMs☆245Updated 2 months ago
- ☆138Updated 2 weeks ago
- ☆42Updated 7 months ago
- [ICML 2024] Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference☆202Updated 2 weeks ago
- Disaggregated serving system for Large Language Models (LLMs).☆359Updated 3 months ago
- Automated Parallelization System and Infrastructure for Multiple Ecosystems☆75Updated this week
- Performance of the C++ interface of flash attention and flash attention v2 in large language model (LLM) inference scenarios.☆29Updated 2 months ago