monkeyfu 

Seq2SeqSharp is a tensor based fast & flexible encoder-decoder deep neural network framework written by .NET (C#). It can be used for sequence-to-sequence task, sequence-labeling task and... More information

CRFSharp core algorithm APIs

The APIs wrapper for CRFSharp core algorithm

A tensor libaray based on CPU. It includes many operations and can be used for neural networks

A tensor libaray based on CUDA. It includes many operations and can be used for neural networks

The console tool for CRFSharp encoder and decoder

A utility for common alogrithms

Core tensor primitives, CPU operations, memory management, storage, and device abstractions for TensorSharp.

Runtime services and contracts for TensorSharp GGUF inference: parsing, tokenization, prompt rendering, sampling, and output protocols.

Direct CUDA backend for TensorSharp using the CUDA Driver API, cuBLAS, PTX kernels, and quantized CUDA ops.

Native MLX backend for TensorSharp inference on Apple Silicon / Metal.

GGML backend integration for TensorSharp, including native interop for CPU, Metal, and GGML CUDA execution.

Model architecture implementations, multimodal encoders, and model execution helpers for TensorSharp.

Command-line host for TensorSharp local inference, diagnostics, prompt inspection, and JSONL batch processing.

ASP.NET Core server package for TensorSharp with embedding and chat inference, web UI, and Ollama/OpenAI-compatible APIs.

Agentic layer for TensorSharp: Agent Skills (SKILL.md bundles, progressive disclosure, script execution) and code execution — a sandboxed shell tool plus a deterministic apply_patch editor, in the style of... More information

Logging services for TensorSharp.

Host-neutral chat pipeline for TensorSharp: model service, sessions, generation, skills loop and the Web UI request/stream contract, independent of ASP.NET Core

Peer-to-peer TCP distributed tensor parallelism for TensorSharp, enabling multi-node inference without external coordination services.

ASP.NET Core server host package for TensorSharp with embedding and chat inference, web UI, and Ollama/OpenAI-compatible APIs.