<!-- SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved. -->
<!-- SPDX-License-Identifier: Apache-2.0 -->

<a id="model-training-with-gr00t"></a>

# Model Training with GR00T

With a LeRobot dataset collected on real hardware ([Data Collection in Real](data_collection_real.md)) or in simulation
([Data Collection in Sim](data_collection_sim.md)), the next step is to fine-tune a GR00T N1.7 manipulation policy on it
and evaluate the trained policy on the SO-101.

#### 🚧 Work in progress

This training guide is still being written. In the meantime, the
[Sim-to-Real SO-101 learning path](https://docs.nvidia.com/learning/physical-ai/sim-to-real-so-101/latest/index.html) walks through GR00T N1.6 training and evaluation for the
SO-101 end to end, while the GR00T N1.7 model is still being finalized.

<!-- References -->
