datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
llama-cpp-wheelsIf you like this please consider liking and donating (https://buymeacoffee.com/aiencoder)
🏭 llama-cpp-python Mega-Factory Wheels
"Stop waiting for pip to compile. Just install and run."
The most complete collection of pre-built llama-cpp-python wheels in existence — 8,333 wheels across every platform, Python version, backend, and CPU optimization level.
No more cmake, gcc, or compilation hell. No more waiting 10 minutes for a build that might fail. Just find your wheel and… See the full description on the dataset page: https://huggingface.co/datasets/AIencoder/llama-cpp-wheels.whisperjav-wheels
WhisperJAV Pre-built Wheels
Pre-built Python wheels for WhisperJAV dependencies that are difficult to compile from source.
Repository Structure
whisperjav-wheels/
├── llama-cpp-python/
│ ├── cu124/ # CUDA 12.4 wheels
│ ├── cu121/ # CUDA 12.1 wheels (legacy)
│ └── metal/ # Apple Silicon wheels
└── README.md
Available Wheels
llama-cpp-python
For local LLM translation support (whisperjav-translate --provider local).… See the full description on the dataset page: https://huggingface.co/datasets/mei986/whisperjav-wheels.qwentts-cpp-python-wheels
qwentts-cpp-python wheels
Optional backend-specific wheel variants for qwentts-cpp-python.
The default PyPI package is CUDA 12.8:
pip install qwentts-cpp-python
Install a backend-specific wheel from this repository with --find-links:
pip install "qwentts-cpp-python==0.3.1+cpu" -f https://huggingface.co/datasets/andito/qwentts-cpp-python-wheels/tree/main/whl/cpu
pip install "qwentts-cpp-python==0.3.1+cu124" -f… See the full description on the dataset page: https://huggingface.co/datasets/andito/qwentts-cpp-python-wheels.vllm-0.28.0-wheels-py312zerogpu-blackwell-wheels
ZeroGPU Blackwell (sm_120) wheels
CUDA-extension wheels built for Blackwell (sm_120). PyPI doesn't currently publish sm_120 wheels for most of these libs, and ZeroGPU Spaces can't compile at boot, so Spaces should pin URLs from this dataset directly in requirements.txt.
Usage
URL template:
https://huggingface.co/datasets/multimodalart/zerogpu-blackwell-wheels/resolve/main/wheels/<TAG>/<wheel-filename>
Example, for a Space on torch 2.12 / cu130 / py3.13 that needs… See the full description on the dataset page: https://huggingface.co/datasets/multimodalart/zerogpu-blackwell-wheels.wheelstorch-wheelsWindows_wheelsprebuild-wheels-utilanzaro-python-wheelswan2gp-wheelsTurboCpp_Wheels
Llama.cpp Spinoff - Python Binary Wheels
Pre-built Python whl binaries for [mention your specific fork name here] to allow easy installation without compiling C++ code.
Installation
Install the latest version via pip:
pip install turbocpp
Features
Pre-compiled: No need to install compilers (CMake, GCC/Visual Studio).
Optimized: Built with specific performance optimizations (AVX, AVX2, cuBLAS).
GGUF Supported: Supports the latest GGUF format.
slime-v0.3.0-wheels
🧩 slime v0.3.0 cu129 Wheelhouse
Prebuilt wheels and installation notes for reproducing a slime v0.3.0 + SGLang + Megatron-LM CUDA 12.9 environment.
✨ What is this?
This repository is a reproducibility wheelhouse for installing slime v0.3.0 with SGLang and Megatron-LM on a Linux x86_64 CUDA 12.9 / cu129 stack.
It is not a model checkpoint repository and it is not a training or evaluation dataset. It stores prebuilt Python wheel artifacts, pinned… See the full description on the dataset page: https://huggingface.co/datasets/zhangdw/slime-v0.3.0-wheels.mtr-bus-2026-1MTR Bus's ETA data for late 2025/early 2026, archived by Wheels Atlas.
llama-cpp-wheelsduckjam-wheels
Duck Jam — the kit
The wheels a Duck Jam template repository resolves: duckjam — the SDK, the trusted
runner and the CLI — and one distribution per round. index.html is a find-links page,
and a generated template already names it, so a clone of one needs nothing but uv sync.
Built and uploaded by .github/workflows/publish-wheels.yml on every tag.
lit-gpt-wheelsrandom-wheelssageattention220-wheelsanzaro-wheels-persistentvllm-wheelsadarft-build-wheelsdependency-wheels
MAAP3D patched dependency wheels
This is a versioned artifact repository, not a dataset or a Python package
index. It publishes pinned downstream wheels used to beta-test specific
MAAP3D deployment profiles. These artifacts are not official releases of the
upstream projects.
The wheels contain third-party code and retain their applicable upstream
licenses. They are not relicensed under MAAP3D's AGPL license. See
LICENSE, the individual files under licenses/,
and… See the full description on the dataset page: https://huggingface.co/datasets/MAAP3D/dependency-wheels.casual-conv1d-pre-built-wheels
Pre-built CUDA Wheels: Causal-Conv1d & Flash-Linear-Attention
This dataset is a collection of pre-built Python wheels for ninja, flash-linear-attention, and causal-conv1d specifically compiled for modern NVIDIA GPUs. It is organized to help developers bypass the highly resource-intensive C++ compilation times typically encountered in CI/CD pipelines (like GitHub Actions) and Docker builds.
Dataset Details
Uses
This dataset is open-sourced and… See the full description on the dataset page: https://huggingface.co/datasets/Feargal/casual-conv1d-pre-built-wheels.Prebuilt-wheelsxlerobot-final-v1-xlerobot_mobile_straight_wheelstate_v1
XLeRobot Final v1 whole-body wheel-feedback integration record
A public, real-hardware LeRobot Dataset v3 integration record for an
XLeRobot-compatible 0.4.0 differential two-wheel configuration. This is a
transparent engineering dataset, not a household-task benchmark or an
autonomy-performance claim.
Recorded system
Two Seeed Studio Pro SO-101 arms under leader/follower teleoperation
Differential two-wheel base
Pan/tilt head
Three RGB cameras: left wrist… See the full description on the dataset page: https://huggingface.co/datasets/NOJIMA21/xlerobot-final-v1-xlerobot_mobile_straight_wheelstate_v1.openmayhem-runtime-wheels
OpenMayhem Runtime Wheels
vLLM 0.24.0 downstream runtime with the speculative FlashInfer host-metadata
lifetime fix. The original upstream native binaries, dependencies, licenses,
and platform metadata are preserved. This is not an official vLLM release.
Each wheel has a SHA-256 provenance record. The source patch, regression test,
and offline reproducible packager are included. Source base:
https://github.com/vllm-project/vllm/commit/ee0da84ab9e04ac7610e28580af62c365e898389
The… See the full description on the dataset page: https://huggingface.co/datasets/TracNetwork/openmayhem-runtime-wheels.stack-lego-wheelsThis dataset was created using LeRobot.
Dataset Structure
meta/info.json:
{
"codebase_version": "v3.0",
"robot_type": "so101_follower",
"total_episodes": 34,
"total_frames": 24088,
"total_tasks": 1,
"chunks_size": 1000,
"data_files_size_in_mb": 100,
"video_files_size_in_mb": 500,
"fps": 30,
"splits": {
"train": "0:34"
},
"data_path": "data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet",
"video_path":… See the full description on the dataset page: https://huggingface.co/datasets/EvanMath/stack-lego-wheels.trex_wheelsgpu-wheels
gpu-wheels
Personal cache of prebuilt Python wheels for packages that are slow to compile from source (flash-attn, etc.), so rented GPU instances (vast.ai and similar) don't have to recompile from scratch every time.
Compiling flash-attn from source can take 45-90 minutes. If a new instance has the exact same torch version, CUDA version, Python version, and C++ ABI as a wheel already in this repo, installing from the cached wheel takes seconds instead.
Repo layout… See the full description on the dataset page: https://huggingface.co/datasets/DanielTobi0/gpu-wheels.