NeMo-RL

license

Apache-2.0(OSI)

source

public

maintainer

NVIDIA

core features withheld

no

NeMo-RL is Apache-2.0, an OSI license, with the full source public and no enterprise or paid directory anywhere in the repository. NVIDIA's separate NeMo Customizer microservice withholds nothing from the published code, and the project's orientation to NVIDIA GPUs is a hardware practicality rather than a license restriction.

• https://developer.nvidia.com/nemo-customizer recorded 2026-09-09

  The URL redirects to docs.nvidia.com/nemo/microservices/latest/customizer/index.html, NVIDIA's documentation for "NeMo Customizer" under the separate NeMo Microservices platform; its own banner points to a distinct GitHub repo, "NVIDIA-NeMo/nemo-platform", not NVIDIA-NeMo/RL. A case-insensitive search of the saved 85,344-byte body for 'nemo-rl' or 'nemo rl' returns 0 matches for each. Nothing on the page describes a feature withheld from the nemo-rl repository; Customizer reads as a separately branded managed microservice rather than a paid tier of nemo-rl.

• https://github.com/NVIDIA-NeMo/RL recorded 2026-09-09

  README states "NeMo RL is licensed under the Apache License 2.0" and describes the library as the re-architected repo of NeMo Aligner. Embedded repo JSON reads "isPrivate":false and "spdxId":"Apache-2.0". The Bare-Metal Quick Start runs `uv run python examples/run_grpo.py` with no key, account or license step. The opening sentence recorded on 2026-08-09 is no longer on the page; the README was rewritten.

• https://raw.githubusercontent.com/NVIDIA-NeMo/RL/main/LICENSE recorded 2026-09-09

  The fetched body is the unmodified Apache License, Version 2.0 text end to end - Sections 1 through 9 plus the appendix template with its placeholder "Copyright [yyyy] [name of copyright owner]" line still unfilled - an OSI-approved grant of a perpetual, worldwide, royalty-free copyright and patent license with no restriction on who may use the software or at what scale, only the Section 4 attribution/change-notice/NOTICE obligations.

GitHub stars are the only usable adoption signal for NeMo-RL: the 'Used by' and contributor counts are not exposed on the page, and the nemo-rl package on PyPI is a non-functional placeholder that points back to GitHub, so no download figure exists at all. That leaves the reading on a weaker signal than most entries in this category.

• https://github.com/NVIDIA-NeMo/RL recorded 2026-08-09

  Sidebar records 1,890 stars (displayed '1.9k', up from the ~1.7k previously recorded), 501 forks, 45 watchers, and a commit count of 1,301. The 'Used by' and contributor-count sidebar sections are unloaded skeleton placeholders in the static page (populated by client-side JavaScript), so no usable count is available from either.

• https://pypi.org/project/nemo-rl/ recorded 2026-08-09

  The project page for the 'nemo-rl' PyPI listing reads: "WARNING: This package is not functional and a placeholder from NVIDIA. Please refer to https://github.com/NVIDIA/NeMo-RL for instructions on installing and using NeMo-RL." It is version 0.0.0, released Jun 3, 2025, a 6.0 kB source distribution with no functional code, so PyPI download counts cannot serve as an adoption signal for this product.

NeMo-RL covers GRPO, GSPO, DAPO, PPO, DPO, SFT, on-policy distillation and reward modeling, and trains through PyTorch FSDP2 or Megatron Core with tensor, pipeline, context, sequence and expert parallelism. Its Megatron Core backend gives it the same large-scale parallelism as Megatron-LM, along with SGLang rollout and speculative decoding for faster generation.

• https://raw.githubusercontent.com/NVIDIA-NeMo/RL/main/README.md recorded 2026-09-28

  The README Features list includes "GRPO/GSPO/DAPO, SFT(with LoRA), DPO, and On-policy distillation.", "PyTorch FSDP2, TP, CP, and SP for efficient training", "Performant parallelisms with Megatron Core (TP/PP/CP/SP/EP/FSDP)", "SGLang rollout support for optimized inference.", "Speculative Decoding support for rollout acceleration and training", a "GDPO Support" news entry and an RM (reward modeling) section in its Table of Contents. The README also newly names PPO, MOPD and X-Token off-policy distillation, which the recorded value omits.