amd-wsung102/mori
Modular RDMA Interface
Repositories
amd-wsung102/MSLK
MSLK (Meta Superintelligence Labs Kernels) is a collection of PyTorch GPU operator libraries that are designed and optimized for GenAI training and inference, such as FP8 row-wise quantization and collective communications.
amd-wsung102/sglang
SGLang is a high-performance serving framework for large language models and multimodal models.
amd-wsung102/ATOM
AiTer Optimized Model
amd-wsung102/FlyDSL
FlyDSL is the Python front‑end of the project: Flexible LaYout DSL.
amd-wsung102/aiter
AI Tensor Engine for ROCm
amd-wsung102/FBGEMM
FB (Facebook) + GEMM (General Matrix-Matrix Multiplication) - https://code.fb.com/ml-applications/fbgemm/