GR00T-N1.7-3B-p150 / LICENSE-NOTICE.md
changh95's picture
Stage-2 release: DiT denoise as one persistent megakernel (snapshot 009de892, image 4cfc5e1ffe29)
1edaf57 verified
|
Raw
History Blame
5.9 kB

Licence notice — changh95/GR00T-N1.7-3B-p150

This repository is a tt-model container package: a Docker image (image/), the code inside it (code/), a manifest and a card. It contains no model weights. Five licences apply to what is here and to what the package fetches.

component where licence
Model weights nvidia/GR00T-N1.7-3B @ 2fc962b973bccdd5d8ce4f67cc63b264d6886495 not in this repott-model serve / scripts/download_weights.sh fetch exactly the files listed in tt-model.yaml (weights.allow_patterns) into your Hugging Face cache NVIDIA License (the LICENSE file of the weights repository, byte-identical to the N1.5 / N1.6 one; a verbatim copy is LICENSE in this repo). §3.3: use is limited to non-commercial purposes — research or evaluation only. §3.1: redistribution must carry the same licence and a complete copy of it. §3.2: derivative works must carry the same use limitation. This is not the NVIDIA Open Model License and not Apache-2.0. The N1.7 model card's sentence about commercial use is not what the LICENSE file in the weights repository says; this package follows the LICENSE file.
Vendored tokenizer / image-processor files code/gr00t_p150/assets/tokenizer/n17/{vocab.json, merges.txt, tokenizer_config.json, preprocessor_config.json} in this repo and in the image (/opt/tt-metal/gr00t_p150/assets/tokenizer/n17) byte-identical (sha256 in code/gr00t_p150/assets/tokenizer/n17/PROVENANCE.md) to the same-named files of Qwen/Qwen3-VL-2B-Instruct @ 89644892e4d85e24eaac8bacfd4f463576704203, Apache License 2.0. GR00T N1.7's backbone (nvidia/Cosmos-Reason2-2B) is a fine-tune of that model and ships the same tokenizer / processor files; they are not part of the weights repo, which is why they ship here. Verified: loading them reproduces the golden N1.7 input_ids, pixel_values and state bit-exactly.
Demo observation media/{exterior_image_1_left,wrist_image_left}_{tm15,t0}.png, media/demo_frames.png, media/demo_observation.json, code/gr00t_p150/demo/n17/ (four frames, raw state, instruction, the reference initial noise and the fp32 / bf16 reference actions) in this repo and in the image one step of the DROID sample episode distributed with the Isaac-GR00T repository (demo_data/droid_sample, episode 1, step 40; frames at t-15 and t of both 320×180 cameras); DROID itself is released by its authors under CC-BY-4.0. The reference actions were computed with NVIDIA's unmodified code on the pinned weights and are therefore derivative of the weights (NVIDIA License §3.2 applies to their use).
Port and serving code: code/models/experimental/gr00t/ (the tt-nn port of GR00T N1.5 / N1.6 / N1.7: common/, reference/, tt/ incl. the tuned Qwen3-VL tower tt/qwen3vl_vision.py and tt/megakernel/, tests/, benchmarks/), code/gr00t_p150/ (FastAPI server, schemas, smoke test, demo builder), code/scripts/ in this repo and in the image (/opt/tt-metal/...) Apache-2.0 (SPDX-License-Identifier: Apache-2.0, © 2026 Tenstorrent USA, Inc. in every source file). Because the port only runs the NVIDIA-licensed weights and is distributed together with the reference actions, the package as a whole is offered under the NVIDIA License's non-commercial use limitation; the Apache-2.0 grant on the code itself is unaffected.
tt-metal / tt-nn: code/models/common/utility_functions.py, code/models/tt_dit/** (incl. the verbatim Qwen3-VL tower encoders/qwen3vl/vision_qwen3vl.py, the oracle the tuned tower is checked against), code/models/demos/deepseek_v3_b1/unified_kernel_descriptor.py and unified_kernels/{kernel_op_api,kernel_utils,dataflow_utils,mcast}.hpp (the megakernel's generic_op descriptor and the device-kernel headers it #includes) and the whole tt-metal build inside the image (/opt/tt-metal, commit 668c290757550588d0ce46b180c344a462a2aaf5) in this repo (the 20 files) and in the image Apache-2.0 (tenstorrent/tt-metal).
Third-party Python packages inside the image (requirements.lock: torch 2.11.0+cpu, torchvision 0.26.0+cpu, transformers 5.12.1, opencv-python-headless 4.8.1.78, numpy 1.26.4, fastapi, uvicorn, pydantic, pillow, safetensors, huggingface_hub, loguru, ...) and the Ubuntu 22.04 base in the image their respective licences (BSD-3 / Apache-2.0 / MIT / HPND; Ubuntu package licences).

Practical reading

  • You may download the weights and run this package for research or evaluation. You may not use the weights, the reference actions, or the actions this server produces for a commercial purpose (NVIDIA License §3.3), and you must keep the NVIDIA License with any copy you redistribute (§3.1).
  • The Apache-2.0 code (code/models/experimental/gr00t, code/gr00t_p150, code/scripts) and the Apache-2.0 tokenizer / processor files can be reused under Apache-2.0 on their own; running them against GR00T weights puts you back under the NVIDIA License.
  • The upstream paper is the GR00T N1 report (arXiv:2503.14734); GR00T N1.7 has no separate paper — its upstream is the main / n1.7-release line of Isaac-GR00T and the nvidia/GR00T-N1.7-3B model card. The backbone is nvidia/Cosmos-Reason2-2B (Qwen3-VL-2B family).
  • Hugging Face metadata of this repo: license: other, license_name: nvidia-license, license_link: https://huggingface.co/nvidia/GR00T-N1.7-3B/blob/main/LICENSE, base_model: nvidia/GR00T-N1.7-3B.