Shared Ultralytics logos, media, sample images, pretrained model weights, dataset artifacts, and release assets.

Training with FP16 weights in PyTorch

🎓 2020 Undergraduate Graduation Project in Jiangnan University ALL codes including Data-convert, keras-Train, model-Evaluate and Web-App

The high-performance distributed tensor layer — load once, share everywhere.

ClimWIP allows to calculate & apply performance and independence weights to CMIP models

Utility tool for Intel(r) OpenVINO(tm) IR models. The tool can display detailed model information, layer information and can check compatibility. The tool also can extract weight data and feature map from the IR model.

Verified multi-source distribution for local LLM weights

Pre-compiled ASR model weights for the VoxRT on-device runtime. Encrypted .vxrt v2 format. streaming-medium-pc: FastConformer 32M, CTC + RNN-T, CC-BY-4.0 (NVIDIA NeMo).

Wake-phrase model weights (.vxrt, ~100 KB, AES-GCM encrypted) for the VoxRT custom on-device inference runtime. Paired with voxrt-wake-word-{android,ios,linux}. Custom phrases at voxrt.com.

Flow-driven recursive skill evolution for agentic LLM orchestration.

14-way keyword-spotting model weights (.vxrt v2, AES-256-GCM encrypted) for the VoxRT on-device runtime. Streaming Conformer-Medium (636 K params, 1.28 MB). Paired with voxrt-kws-{linux,browser}.

The Brain Collector: Find ML model weights inside Android

Computer Vision Basics to advanced. My journey through this subfield of AI

Quantization Aware Training

Template for Cog models with built-in model weights caching and CDN integration

Pinned chess neural network weights and minimal model documentation.

This GitHub repository provides an implementation of a Genetic Algorithm (GA) for finding optimal weights of a Neural Network (NN).

A deterministic, random-access archive format and toolkit for compressing, storing, diffing and patching large language model weights.

Implementation of 3D U-Net for Brain MRI skull-stripping along with pre-trained Keras model checkpoints (.h5) for medical image segmentation & grading.

A pure Standard ML reader for the safetensors weight format — dequantizes f16/bf16/f32/f64 to reals