PrototypeCommon
yanring/Megatron-MoE-ModelZoo
Best practices for training DeepSeek, Mixtral, Qwen and other MoE models using Megatron Core.
- Owner
- yanring (personal account)
- Stars
- Stars: 206
- Forks
- Forks: 35
- Language
- Python
- License
- MIT
- Created
- 2025
- Topics
- llm-trainingmegatron-lmmixture-of-experts
- Tier
- Prototype (100 to 249 ★) What this tier means
More repos by yanring
All 2 repos by yanring- yanring/jianzhi-Offer-LeetcodeStars: 400Alpha
Nearby in Python
All Python repos- per-simmons/unreal-agent-harnessStars: 206PythonPrototype
- Carlehyy/OpenOntologyStars: 206PythonPrototype
- sandbaseai/sandbase-skillsStars: 206PythonPrototype
- peppermintos/iceStars: 206PythonPrototype
- PirateBox-Dev/PirateBoxScripts_WebserverStars: 206PythonPrototype
- kennethreitz-archive/argsStars: 206PythonPrototype