Tensil

Tensil is a free and open source platform providing customizable machine learning inference accelerators designed for edge and embedded FPGA devices. It includes an RTL generator, model compiler, and drivers that enable users to create tailored accelerators optimized for their ML workloads without the need for model quantization or compression. Focused initially on convolutional neural networks on small edge FPGAs, Tensil aims to offer flexibility, improved performance per watt, and accessibility to ML hardware acceleration for small to mid-sized companies, academics, and hobbyists. The core technology is open source, with plans for enterprise "pro" features and a cloud service for architecture search and deployment.

platform:web platform:fpga pricing:free pricing:freemium form:open-source form:cli-tool form:library form:api feature:customizable-hardware feature:ml-inference feature:fpga-support feature:systolic-array feature:rtl-generator feature:model-compiler feature:drivers feature:edge-computing use-case:machine-learning use-case:inference-acceleration target:developers target:academics target:researchers target:engineers target:hobbyists

Features

Customizable Hardware

Ml Inference

Fpga Support

Systolic Array

Rtl Generator

Model Compiler

Drivers

Edge Computing

Testimonies

No testimonies available for this tool yet.

Basic Info

Category Artificial Intelligence

Website GitHub Doc Discord Other

Availability & Pricing

Code Access Open Source
Pricing Model
Free Freemium

AI Curation

Curator Agent updated description, category, subcategory, and 3 more fields for this tool

9 months ago