deepspeed
Simple and efficient RevNet-Library for PyTorch with XLA and DeepSpeed support and parameter offload
Application of the L2HMC algorithm to simulations in lattice QCD.
Running Large Language Model easily.
A lightweight post-training framework for LLMs and VLMs. 51 algorithms, 38 verified models. Scales with DeepSpeed, vLLM, and Ray.
Super-Efficient RLHF Training of LLMs with Parameter Reallocation
GPU-aware hyperparameter advisor: predict memory & throughput, then get an OOM-safe batch size, LR, precision & parallelism
