R_REDDYX.XYZ
AIPython★ 1,290⑂ 137 forksApache-2.0

NVIDIA-NeMo/labs-molt

LABS-MOLT: AGENTIC RL FRAMEWORK FOR RESEARCH

Scalable framework for agentic RL built on HuggingFace and supporting context‑ and expert‑parallelism with FSDP. Contains 9 000 lines of Python and already has 1 164 stars on GitHub.

// KEY FEATURES

  • Trains RL agents on large language models via HuggingFace Transformers.
  • Implements context‑ and expert‑parallelism for distributed training across multiple GPUs.
  • Uses FSDP for model sharding and reducing GPU memory consumption.
  • Provides CLI and Python‑API for launching experiments with logging in Weights & Biases.
#agentic-rl#context-parallelism#expert-parallelism#fsdp#huggingface#llm#rl-framework#rlhf#rlvr#tensor-parallelism
Open on GitHub →

New repositories every 30 minutes

REDDYX AI scans GitHub 24/7 and ships the best AI/ML/Web3 projects to Telegram.

Join on Telegram
← Full catalog·Full index

// SIMILAR REPOSITORIES

Full-Stack AI Template★ 1289PHONECLAW: On-device AI Agent★ 1291KUBEAI – Integral Inference Operator★ 1291LLM AUTOMATICALLY COLLECTS WIKI★ 1292CHUNKHOUND: Local Code Intelligence★ 1288SEARCH2AI - Smart Search for LLM★ 1288

← FULL CATALOG

NVIDIA-NeMo/labs-molt — Labs-Molt: Agentic RL Framework For…