DachengLi1 / AMP
(NeurIPS 2022) Automatically finding good model-parallel strategies, especially for complex models and clusters.
☆34Updated 2 years ago
Related projects ⓘ
Alternatives and complementary repositories for AMP
- ☆65Updated 3 years ago
- ☆46Updated 5 months ago
- Automated Parallelization System and Infrastructure for Multiple Ecosystems☆75Updated this week
- nnScaler: Compiling DNN models for Parallel Training☆74Updated 3 weeks ago
- ☆70Updated 2 years ago
- PyTorch library for cost-effective, fast and easy serving of MoE models.☆103Updated 3 months ago
- ☆88Updated 2 months ago
- FTPipe and related pipeline model parallelism research.☆41Updated last year
- ☆131Updated 3 months ago
- PyTorch bindings for CUTLASS grouped GEMM.☆53Updated 3 weeks ago
- ☆44Updated last year
- An experimental parallel training platform☆52Updated 7 months ago
- PipeTransformer: Automated Elastic Pipelining for Distributed Training of Large-scale Models. ICML 2021☆55Updated 3 years ago
- A Python library transfers PyTorch tensors between CPU and NVMe☆98Updated last week
- PET: Optimizing Tensor Programs with Partially Equivalent Transformations and Automated Corrections☆114Updated 2 years ago
- An Attention Superoptimizer☆20Updated 6 months ago
- ☆8Updated last year
- ☆73Updated last year
- SparseTIR: Sparse Tensor Compiler for Deep Learning☆131Updated last year
- ☆167Updated 4 months ago
- ☆16Updated last year
- Stateful LLM Serving☆38Updated 3 months ago
- Quantized Attention on GPU☆30Updated 2 weeks ago
- PyTorch implementation of paper "Response Length Perception and Sequence Scheduling: An LLM-Empowered LLM Inference Pipeline".☆74Updated last year
- ☆55Updated 5 months ago
- A resilient distributed training framework☆85Updated 7 months ago
- ☆23Updated last year
- ☆52Updated last week
- ☆26Updated 3 years ago
- ☆21Updated last year