Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
A lightweight suite of motion imitation methods for training controllers.
| Date | Stars |
|---|---|
| 2026-07-24 | 2182 |
| 2026-07-25 | 2185 |
| 2026-07-28 | 2185 |
| 2026-07-30 | 2185 |
| 2026-08-10 | 2215 |
| 2026-08-18 | 2234 |
| 2026-08-19 | 2236 |
| 2026-08-20 | 2238 |
| 2026-08-21 | 2241 |
| 2026-08-22 | 2242 |
| 2026-08-24 | 2244 |
| 2026-08-25 | 2251 |
| 2026-08-26 | 2253 |
| 2026-08-27 | 2257 |
| 2026-08-28 | 2261 |
| 2026-08-29 | 2264 |
| 2026-08-30 | 2265 |
| 2026-08-31 | 2270 |
| 2026-09-01 | 2273 |
| 2026-09-02 | 2280 |
| 2026-09-03 | 2279 |
| 2026-09-04 | 2283 |
| 2026-09-05 | 2286 |
| 2026-09-06 | 2289 |
| 2026-09-07 | 2290 |
| 2026-09-08 | 2293 |
| 2026-09-09 | 2300 |
| 2026-09-10 | 2304 |
| 2026-09-11 | 2306 |
| 2026-09-12 | 2307 |
| 2026-09-13 | 2308 |
| 2026-09-14 | 2315 |
| 2026-09-15 | 2319 |
| 2026-09-16 | 2322 |
| 2026-09-17 | 2323 |
| 2026-09-18 | 2328 |
| 2026-09-19 | 2328 |
| 2026-09-20 | 2329 |
Today
+1 stars today
This week
+21 stars this week
This month
+88 stars this month
Momentum
10.0
growth rate 0.91%/day
# MimicKit <p align="center"> <img src="images/teaser_humanoid.gif" width="32%" style="margin-right: 1px;"> <img src="images/teaser_g1.gif" width="32%" style="margin-right: 1px;"> <img src="images/teaser_go2.gif" width="32%"> </p> This framework provides a suite of motion imitation methods for training motion controllers. This codebase is designed to be clean and lightweight, with minimal dependencies. A more detailed overview of MimicKit is available in the [Starter Guide](https://arxiv.org/abs/2510.13794). For a more feature-rich and modular motion imitation framework, checkout [ProtoMotions](https://github.com/NVlabs/ProtoMotions/). Instructions for each method are available here: - [DeepMimic](docs/README_DeepMimic.md) - [AMP - Adversarial Motion Priors](docs/README_AMP.md) - [AWR - Advantage-Weighted Regression](docs/README_AWR.md) - [ASE - Adversarial Skill Embeddings](docs/README_ASE.md) - [LCP - Lipschitz-Constrained Policies](docs/README_LCP.md) - [ADD - Adversarial Differential Discriminator](docs/README_ADD.md) - [SMP - Score-Matching Motion Priors](docs/README_SMP.md) --- ## Installation This framework supports different simulator backends (referred to as `Engines`). We highly recommend using a package manager, like [Conda](https://docs.conda.io/projects/conda/en/stable/user-guide/install/index.html), to create dedicated Python environments for each simulator. 1. Install the simulator of your choice. <details> <summary>Isaac Gym</summary> Install [Isaac Gym](https://developer.nvidia.com/isaac-gym). To use Isaac Gym, specify the argument `--engine_config data/engines/isaac_gym_engine.yaml` when running the code. </details> <details> <summary>Isaac Lab</summary> Install [Isaac Lab](https://isaac-sim.github.io/IsaacLab/main/source/setup/installation/index.html). This framework has been tested with `2ed331acfcbb1b96c47b190564476511836c3754`. To use Isaac Lab, specify the argument `--engine_config data/engines/isaac_lab_engine.yaml` when running the code. </details> <details> <summary>Newton</summary> Install [Newton](https://newton-physics.github.io/newton/guide/installation.html). This framework has been tested with `v1.0.0`. To use Newton, specify the argument `--engine_config data/engines/newton_engine.yaml` when running the code. </details> 2. Install the requirements. ``` pip install -r requirements.txt ``` 3. Download assets and motion data from [here](https://1sfu-my.sharepoint.com/:u:/g/personal/xbpeng_sfu_ca/EclKq9pwdOBAl-17SogfMW0Bved4sodZBQ_5eZCiz9O--w?e=bqXBaa), then extract the contents into [`data/`](data/). --- ## Training To train a model, run the following command: ``` python mimickit/run.py --mode train --num_envs 4096 --engine_config data/engines/isaac_gym_engine.yaml --env_config data/envs/deepmimic_humanoid_env.yaml --agent_config data/agents/deepmimic_humanoid_ppo_agent.yaml --visualize true --out_dir output/ ``` - `--mode` selects either `train` or `test` mode. - `--num_envs` the number of parallel environments used for simulation. Not all environments support parallel envs, this is mainly used for Isaac Gym envs and other environments, like DeepMind Control Suite does not support this feature and should therefore use 1 for the number of envs. - `--engine_config` configuration file for the engine to select between different simulator backends. - `--env_config` configuration file for the environment. - `--agent_config` configuration file for the agent. - `--visualize` enables visualization. Rendering should be disabled for faster training. - `--out_dir` the output directory where the models and logs will be saved. - `--logger` the logger used to record training stats. The options are text `txt`, TensorBoard `tb`, or `wandb`. - `--video` either `true` or `false` to enable headless video recording, which are then recorded by the logger. Instead of specifying all arguments through the command line, arguments can also be loaded from an `arg_file`: ``` python mimickit/run.
Excerpt of 8,149 characters
Read on GitHub
Simon Fraser University, NVIDIA
52
26
11
5
4
3
2
1
1
1
Sense_wang
1
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:278181fdd9cb448e, topic:robotics, readme:humanoid
matched fp:278181fdd9cb448e, topic:animation
matched fp:278181fdd9cb448e, topic:reinforcement-learning