LoreonLabsPlatform
DocsHome
  • Overview

Intelligence

  • Markets
  • Builders
  • Research
  • Ecosystems
  • Launchpads
  • Search
Ecosystems

Python

Megatron-Bridge

Training library for Megatron-based models with bidirectional Hugging Face conversion capability

PythonEmerging
GitHubWebsite
Stars
—
Forks
—
Contributors
8
Last push
2mo ago

Recent commits

Latest commits.

  • [doc] chore: add Bailing MoE V2 news and fix chronological order (#3260)
    d581ba9Yu Yao2mo ago
  • [model] feat: Model Support for Ling MoE V2 Model (#2028)
    9197938Changlong Yu2mo ago
  • Add Hybrid CP arg to initialize_model_parallel (#3259)
    2bb907aParth Mannan2mo ago
  • perf: Set micro_batch_size=2 for qwen3 235B B300 V2 configs (#3252)
    bb767c9Raghav Hrishikeshan Mukundan2mo ago
  • build: drop rc0 pre-release tag and add dynamic git versioning (#3219)
    f5391dboliver könig2mo ago
  • [docs] feat: Add AGENTS.md for AI coding agent guidance (#3256)
    3ee19b2Yu Yao2mo ago
  • [model] fix: Use adaptive transpose_on_export for GPT-OSS expert weights (#3250)
    b139a4bYu Yao2mo ago
  • [model] fix: preserve dtype in Kimi-K2.5 INT4 roundtrip quantization (#3220)
    4cb379fYu Yao2mo ago

    • Top contributors

    Builders behind this project.

    ko3n1g
    330 commits
    ananthsub
    272 commits
    yaoyu-33
    170 commits
    malay-nagda
    102 commits
    chtruong814
    64 commits
    maanug-nv
    62 commits
    cuichenx
    58 commits
    dimapihtar
    57 commits