zheliuyu/Liger-Kernel-dev
Efficient Triton Kernels for LLM Training
Interested in @flash-linear-attention+NPU @Liger-Kernel+NPU @Ascend/docs @verl/HybridFlow+NPU
Repositories
zheliuyu/MoonEP-Triton-Distributed
MoonEP-Triton-Distributed: MoonEP reimplemented on Triton-distributed with NVSHMEM symmetric memory.
zheliuyu/zheliuyu
zheliuyu/fla-dev
🚀 Efficient implementations for emerging model architectures
zheliuyu/triton-ext-lk
A collection of out-of-tree extensions for the Triton language and compiler
zheliuyu/ms-swift-dev
Use PEFT or Full-parameter to CPT/SFT/DPO/GRPO 600+ LLMs (Qwen3.6, DeepSeek-R1, GLM-5.1, InternLM3, Llama4, ...) and 300+ MLLMs (Qwen3-VL, Qwen3-Omni, InternVL3.5, Ovis2.5, GLM4.5v, Gemma4, Llava, Phi4, ...) (AAAI 2025).
zheliuyu/docs-dev
The repository provides docs & api & tutorials & FAQ and all things like that.
zheliuyu/DeepSpec-triton
DeepSpec: a full-stack codebase for training and evaluating speculative decoding algorithms
zheliuyu/verl
verl/HybridFlow: A Flexible and Efficient RL Post-Training Framework
zheliuyu/LLaMA-Factory-Doc-dev
LLaMA Factory Document
zheliuyu/LLaMA-Factory-dev
Unified Efficient Fine-Tuning of 100+ LLMs & VLMs (ACL 2024)
zheliuyu/kernels-npu-dev
zheliuyu/Ascend-CI-dev
CI system for Ascend
zheliuyu/tracking-map-dev
tracking-map is used to track the progress of adapting some popular ecosystem tools for the Ascend NPU.
zheliuyu/VeOmni-dev
VeOmni: Scaling Any Modality Model Training with Model-Centric Distributed Recipe Zoo