denizyuret/Knet.jl

Health v2 · maintenance only

Flags: no_license

Adoption not part of the score

1432 stars · 223 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Knet (Koç University deep learning framework) is a deep learning library implemented in Julia with GPU support and automatic differentiation via dynamic computational graphs. Models are defined in plain Julia code without a separate modeling language.

Use cases

• train neural networks in Julia
• build deep learning models with GPU acceleration
• define models with automatic differentiation in plain Julia
• learn deep learning with a minimal framework
• compare deep learning framework performance benchmarks
• implement CNNs and RNNs in Julia

When to choose

• you are already working in the Julia ecosystem
• you want models defined in plain Julia with dynamic graphs
• you need GPU-backed training with automatic differentiation
• you want a lightweight educational deep learning framework

When to avoid

• you need a large ecosystem of pretrained models and tooling
• you prefer Python-based frameworks like PyTorch or TensorFlow
• you need production-grade support and frequent updates
• you want a framework with a large active community

Facets

library · maturity maintenance

deep-learning machine-learning gpu-computing deep-learning machine-learning data-science windows automatic-differentiation neural-networks julia dynamic-computational-graphs gpu linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="denizyuret/Knet.jl")