Machine Learning & Deep Learning Packages
Machine-learning and deep-learning frameworks and model libraries for training, fine-tuning, and serving models (PyTorch, scikit-learn, Transformers).
Packages in Machine Learning & Deep Learning
Kind
Sort:Used by Most Apps
Triton
A Python language and compiler for writing highly efficient custom GPU kernels for deep learning.
LightGBM
A fast, distributed, high-performance gradient boosting framework based on decision tree algorithms, used for ranking, classification, and other machine learning tasks.
llama-cpp-python
Python bindings for llama.cpp to run local LLMs with an OpenAI-compatible API
ModelScope
Alibaba's open model-hub library for NLP, CV, speech, and multi-modal AI models.
fastembed
Lightweight, ONNX-powered Python library for fast text, image, and sparse embedding generation.
gse
A Go library for multilingual text segmentation with dictionary, DAG, and HMM cutting modes for Chinese, Japanese, and English.
tiktoken-rs
Fast Rust library for tokenizing text and counting tokens for OpenAI GPT models with tiktoken BPE.
DeepSpeed
Distributed deep learning optimization library for training and serving massive models
PyTorch Lightning
The deep learning framework that separates PyTorch research code from engineering boilerplate, scaling from CPU to multi-node GPU clusters without code changes.
EasyOCR
Ready-to-use Python OCR supporting 80+ languages and all popular writing scripts.
Candle
A minimalist machine learning framework for Rust with a PyTorch-like API and GPU support.
Gymnasium
The standard Python API for reinforcement learning environments, maintained as the community successor to OpenAI Gym.
kornia
A differentiable computer vision library for PyTorch, with 500+ GPU-ready ops for image processing, augmentation, and geometry.
xFormers
Hackable, memory-efficient Transformer building blocks with custom CUDA kernels
Kokoro
Python inference library for the open-weight Kokoro-82M text-to-speech model
CatBoost
A fast, high-performance gradient boosting library with best-in-class native categorical feature support
SWE-bench
Benchmark harness that evaluates whether a language model can actually resolve real-world GitHub issues.
MLX-VLM
Run and fine-tune Vision Language Models on Apple Silicon with MLX.
NVIDIA NCCL
Optimized primitives for collective multi-GPU and multi-node communication
FastEmbed-rs
Rust library for generating vector embeddings and reranking documents locally via ONNX
whisper-rs
Safe Rust bindings to whisper.cpp for running OpenAI's Whisper speech-to-text model locally.
skops
Securely persist and share scikit-learn models without pickle, and auto-generate model cards for them.
conformer
PyTorch implementation of the Conformer convolutional module, adding local convolutional context to Transformer self-attention.
PaddleOCR
Multilingual OCR and document-parsing toolkit that turns PDFs and images into structured data.
