NVlabs/motionbricks

Source
GitHub
First trending
Category
Model inference
GitHub stars
1,108
Main language
Python

This page introduces an external open-source repository. It is not an HDATF product.

NVlabs/motionbricks

What it does

MotionBricks generates short human or humanoid reference motions between recent poses and selected target poses. It serves animation and robotics simulation developers who need transitions that can be chained into longer movement. The model takes four recent poses and one to four target poses, then returns generated pose sequences and predicted foot contacts at a 30 Hz motion sampling rate. Its documented architecture separates root path and duration prediction, masked motion token prediction and decoding of those tokens into continuous motion. The inspected generation code checks pose counts and output buffers, while a headless example feeds the last generated poses into the next call. Pretrained SOMA and Unitree G1 models, motion and scene editors, and a Python runtime using Warp, Newton and TensorRT are included. The bundled runtime requires an NVIDIA GPU, and the README reports demonstration testing on Windows and Linux with an RTX 4090 rather than establishing compatibility with every machine. Text and scene geometry are not model inputs, physical robot safety requires separate control and validation, and code, documentation, model weights and sample motion clips have different licenses.

License

Custom license Custom or non-standard license. Read the license file before any reuse.

More in this category

  • Tejas-TA/predikit
    predikit is a Python library that exposes trained machine-learning models as typed tools for language-model agents. A developer supplies a fitted model, a Pydantic input definition, a tool description and an output name. The library validates incoming values, aligns them with model features, runs prediction and returns a named result dictionary. It exports schemas for OpenAI-style tool calls, LangChain tools and ordinary Python functions, and a registry groups several tools. Ensembles support collecting results, averaging and voting, including weighted variants. For classifiers with prediction probabilities, confidence thresholds can warn, raise an error or call a fallback tool, while asynchronous invocation runs the blocking prediction in a thread pool. Optional integrations load models from MLflow or Snowflake and expose registries through an MCP server, while core model support targets scikit-learn-compatible estimators and XGBoost. Feature names must match training columns, unnamed arrays depend on field order, and the library does not train the model or establish that its probabilities are calibrated or its predictions valid for a new business setting.
  • affifuddin-hazam/ghost-infer
    ghost-infer is a Rust language model runtime that takes model files in the GGUF format and conversation input and generates replies locally. It focuses on an individual using an existing computer rather than a server batching requests for many users. It provides command-line and chat interfaces, and an OpenAI-format chat API that streams replies or returns tool-call information. Tool-call information describes an execution request, and an external Harness using this API must supply its own execution loop and tools. For supported mixture-of-experts models, which select only some experts for each calculation, it keeps shared computation on the GPU and distributes expert weights between RAM and a GPU cache while the CPU computes the rest. It supports Windows 10 or 11, GPU execution needs a supported NVIDIA card and driver, and source builds additionally need NVRTC, a GPU code compilation library. The installer includes that compiler library and can start with existing model files or downloaded models, so core inference does not require an external paid model API. Requests run one at a time, supported architectures and file formats are limited, and image input and multiple generated choices per request are not supported.
  • wici-ai/inferssd
    InferSSD is a local inference engine that streams large MoE models from SSD when they cannot fit entirely in RAM or GPU memory. It extends llama.cpp rather than supplying a model, taking supported GGUF weights and user requests to produce generated responses. It reads the experts selected for each token from SSD, caches them in RAM and uses CPU and GPU paths according to the environment. Browser chat and an OpenAI-compatible API allow external clients to connect. Weights must be downloaded separately and can require substantial SSD space plus runtime working memory. Official documentation describes Apple Silicon Metal and Linux CUDA environments; source builds also need Python and C/C++ tooling. The README header calls Windows work in progress while its body describes Windows 11 packages, so the actual release's supported backend needs checking. Neither universal model support nor hardware-independent speed is established; SSD, cache and request conditions matter. The network API lacks authentication and needs access controls, while model licenses remain separate from the engine. Published throughput does not guarantee answer quality or performance on our hardware.
  • canivel/dynolab
    Dyno Lab is an Apple Silicon research app for observing local model behavior and testing hypotheses. Researchers supply questions, prompts, models and experiment settings to obtain responses, activation measurements and saved artifacts. It provides request inspection, research notebooks, comparative experiments, linear probes and small sparse-autoencoder workflows. The app bundles Python and MLX, but model weights must be downloaded and started separately. It requires Apple Silicon and macOS 14 or later; pairing a Windows NVIDIA worker for larger GGUF models is an experimental GPU-pool feature. Capturing an already-loaded model and experiments that load another copy have different memory requirements. Python SDK, HTTP API and a local stdio MCP bridge are available, but starting MCP alone does not launch a model or experiment. Results are stored locally, yet an external assistant may send results it reads to its provider, so exposure needs review. Activation magnitude and model-emitted thinking are not safety certification or a faithful explanation of internal computation. Official material distinguishes exploratory experiments from validated safety monitors.
  • huawei-csl/MACaQ
    MACaQ is a calibrated quantization toolkit for reducing the numerical precision of supported model weights and, on selected backends, activations to make local inference more practical. Its inputs include matching original model checkpoints and representative calibration text, prompts or conditioning images, while its outputs include compact GGUF or safetensors checkpoints and recorded settings. In the inspected dense text path, activation statistics guide weight balancing and a block Hadamard transform before low-bit quantization and codebook refinement. Separate adapters cover documented text, diffusion and speech model families, so model agnostic should not be read as automatic compatibility with every architecture. The llama.cpp path offers two-, three- and four-bit weight formats, whereas other adapters offer W4A4, meaning four-bit weights and activations, with some components intentionally left at higher precision. Full transformed GGUF workflows use the bundled runtime, while MACaQ-R refines selected native MoE weight codes without changing their formats and is documented to work with standard llama.cpp. Preparation requires backend-specific dependencies, matching source weights, calibration data and working storage; the guides specify Python 3.12, Linux for vLLM and a CPU alternative for llama.cpp. Quantization is not repeated when saved checkpoints are loaded, but runtime compatibility and original model licensing still apply, and the diffusion guide explicitly notes unclear licensing for one community audio checkpoint. The README reports memory, speed and model-quality results, but these were not reproduced here and cannot establish the same trade-off for another model, device or application.

Only repositories in the ranked Trendshift lists are included, and the lists are used only to find candidates. We do not copy their ranks. Descriptions, licenses and star counts come from each GitHub repository. The notes are our own reading. We have not tested these projects, and a place on a trending list does not prove quality.

View on GitHub (opens in a new tab)