Skip to content
Robotics

Geoffrey19/cei2-groot-n17-base-drawer200-40k

fine-tuned model for robotic arm control

2 min readupdated

This model is a fully fine-tuned version of the NVIDIA GR00T N1.7 architecture (3.14 billion parameters) designed to control a complex robotic system. It was trained on real-world drawer manipulation demonstrations using a 7-axis Franka FR3 arm and a 16-joint LEAP hand. The model predicts target joint positions based on external RGB camera data and the robot's current state. This release is intended for robotics and autonomous agent researchers.

Key features

  • Based on NVIDIA's base GR00T N1.7 model with 3.14B parameters.
  • Trained for 40,000 steps on real teleoperation demonstrations.
  • Supports joint arm and hand control (23 absolute joints).
  • Uses 16-step relative action chunks for smoother control.

What you can do

  • Run open-loop evaluation of the model on a prepared dataset in GR00T format.
  • Use the model as a starting point for further fine-tuning on your own robotic tasks.
  • Analyze modality configurations and joint order (`joint_order.json`) for hardware integration.

Example

Example command to run open-loop model evaluation on the dataset:

python gr00t/eval/open_loop_eval.py \
    --dataset-path  \
    --embodiment-tag NEW_EMBODIMENT \
    --model-path  \
    --traj-ids 0 1 2 --execution-horizon 16
Read moreDetails
Benchmarks: final training loss dropped to 0.0167 at step 40,000.
Hardware requirements: float32 format weights, single GPU required.
Limitations: no held-out evaluation or real-robot tests, training data only.
Training details: bfloat16 mixed precision over float32 weights, peak learning rate 2e-5 with cosine decay.
Discuss Telegram 0 7

Will it run?

Runs in F32

Fits without quantisation.

  • F32weights 12 GB · ready files13 GB
  • BF16weights 5.9 GB6.9 GB
  • FP8 / INT8recommendedweights 2.9 GB3.7 GB
  • 4-bitweights 1.6 GB2.3 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: FP8 / INT8 — from RTX 2060 · F32 — from RTX 5080

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # Geoffrey19/cei2-groot-n17-base-drawer200-40k
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download Geoffrey19/cei2-groot-n17-base-drawer200-40k --local-dir ~/models/cei2-groot-n17-base-drawer200-40k
    # The model is in ~/models. How to run it is in the model card (link above).
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

New posts about versions of GR00T-N1.7-3B (quants, finetunes, LoRAs) will show up under the bell.

Quantised versions 3All on Hugging Face
Fine-tunes 12All on Hugging Face
LoRAs and adapters 1All on Hugging Face

About the model

Geoffrey19/cei2-groot-n17-base-drawer200-40k
Author
Geoffrey19
Category
Robotics
Released
Posted

On NeyroLearn

Hugging Face

70

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments