Skip to content
Categorías de proyectos

Aprendizaje automático · GitHub

Explora herramientas para entrenar, evaluar y ejecutar modelos, y compara el interés de la comunidad.

Proyectos útiles de GitHub

12,806 proyectos

Rose

MatthewK78

79estrellas
Rose preview

🌹 Rose: Range-Of-Slice Equilibration PyTorch optimizer. Stateless optimization through range-normalized gradient updates.

  • Stateless - No momentum, variance estimates, or even step counters. Memory footprint: parameters + gradients + working memory, nothing more. Rose uses less memory than even 8-bit versions of other optimizers, while still retaining full accuracy.
  • Improved generalization - Training loss runs higher while validation loss runs lower, consistent with implicit regularization rather than memorization.
231estrellas

This repository provides the code used to implement the framework to provide deep learning models with total uncertainty estimates as described in "A General Framework for Uncertainty Estimation in Deep Learning" (Loquercio, Segù, Scaramuzza. RA-L 2020).

1,517estrellas
gemma-tuner-multimodal preview

Fine-tune Gemma 4 and 3n with audio, images and text on Apple Silicon, using PyTorch and Metal Performance Shaders.

  • Fine-tune Gemma (text-only CSV) - ✅ - ✅ - ✅ - ✅
  • Fine-tune Gemma image + text (caption / VQA CSV) - ✅ - ⚠️ varies - ⚠️ varies - ⚠️ varies

tensai

mattn

120estrellas

A tiny neural-network framework in pure Go with AVX2 SIMD kernels (GOEXPERIMENT=simd)

  • Matrix operations - Matrix plus basic operations such as Dot, Add, T, and AddBias. Tensors are float32 (tensai.Float)
  • Low-allocation training - layers reuse their forward/backward scratch buffers across training steps (a full MLP step runs in ~29 allocations), so GC stays out of the training loop; Predict always returns freshly allocated results

KANX

Mattral

38estrellas
KANX preview

One library, four surfaces. Production-grade Kolmogorov-Arnold Networks || TensorFlow + PyTorch + ONNX. || A small KAN beats a 10× larger MLP on smooth, separable target. One library. Two backends. Real ONNX export. Docker + Kubernetes ready.

42estrellas

60+ ML/AI/DL/RL/Bayesian algorithms implemented from scratch in NumPy -- plus mlscratch, a pip-installable package (pip install scratchkit) with a consistent, scikit-learn-style API and 1,100+ tests.

ds

matyushkin

233estrellas

👨‍🔬 In Russian: Обновляемая структурированная подборка бесплатных ресурсов по тематикам Data Science: курсы, книги, открытые данные, блоги и готовые решения.

RTM3D

maudzung

300estrellas
RTM3D preview

Unofficial PyTorch implementation of "RTM3D: Real-time Monocular 3D Detection from Object Keypoints for Autonomous Driving" (ECCV 2020)

  • Realtime 3D object detection based on a monocular RGB image
  • Support distributed data parallel training

SFA3D

maudzung

1,129estrellas
SFA3D preview

Super Fast and Accurate 3D Object Detection based on 3D LiDAR Point Clouds (The PyTorch implementation)

  • Super fast and accurate 3D object detection based on LiDAR
  • Fast training, fast inference