VOOZH about

URL: https://www.nuget.org/packages/Shorokoo.Modules/

⇱ NuGet Gallery | Shorokoo.Modules 0.0.1



Shorokoo.Modules 0.0.1 

dotnet add package Shorokoo.Modules --version 0.0.1
 
 
NuGet\Install-Package Shorokoo.Modules -Version 0.0.1
This command is intended to be used within the Package Manager Console in Visual Studio, as it uses the NuGet module's version of Install-Package. 
<PackageReference Include="Shorokoo.Modules" Version="0.0.1" />
For projects that support PackageReference, copy this XML node into the project file to reference the package. 
<PackageVersion Include="Shorokoo.Modules" Version="0.0.1" />
 


 Directory.Packages.props
 

 

 
<PackageReference Include="Shorokoo.Modules" />
 


 Project file
For projects that support Central Package Management (CPM), copy this XML node into the solution Directory.Packages.props file to version the package. 
paket add Shorokoo.Modules --version 0.0.1
The NuGet Team does not provide support for this client. Please contact its maintainers for support. 
#r "nuget: Shorokoo.Modules, 0.0.1"
#r directive can be used in F# Interactive and Polyglot Notebooks. Copy this into the interactive tool or source code of the script to reference the package. 
#:package Shorokoo.Modules@0.0.1
#:package directive can be used in C# file-based apps starting in .NET 10 preview 4. Copy this into a .cs file before any lines of code to reference the package. 
#addin nuget:?package=Shorokoo.Modules&version=0.0.1
 


 Install as a Cake Addin
 

 

 
#tool nuget:?package=Shorokoo.Modules&version=0.0.1
 


 Install as a Cake Tool
The NuGet Team does not provide support for this client. Please contact its maintainers for support.

Shorokoo.Modules

Baseline neural-network library for Shorokoo: ready-made layers, loss functions, optimizers, and initializers built from Shorokoo modules.

  • Initializers (Shorokoo.Modules.Initializers) — Zeros, Ones, Uniform, Normal, XavierUniform, XavierNormal, KaimingUniform, KaimingNormal, TruncatedNormal, LeCunNormal. All shape-only [TrainableParamInitializer]s; the random ones are seeded (deterministic), and Xavier/Kaiming/LeCun compute fan-in/fan-out in-graph from the shape vector.
  • Layers (Shorokoo.Modules.Layers) — Linear, Conv1d, Conv2d, Conv3d (hyperparameter-driven geometry via the dynamic Conv lowering), ConvTranspose2d (default geometry, kernel inferred from the weight), BatchNorm2d/BatchNorm1d (training/eval flag, running stats via StateUpdate), LayerNorm, RMSNorm, GroupNorm, InstanceNorm2d, Dropout (training flag), Embedding, MultiHeadAttention / TransformerEncoderLayer (+ the Attention.ScaledDotProductAttention helper), LeakyReLU/ELU (hyper alpha), PReLU (learnable slope), and the Pooling / GatedLinear.GLU helpers (MaxPool2d, AvgPool2d, GlobalAvgPool2d, GlobalMaxPool2d, Flatten). Plain activations are tensor one-liners — x.Relu(), x.Gelu(), x.Sigmoid(), x.Tanh(), x.Softmax(axis) — and need no modules.
  • Losses (Shorokoo.Modules.Losses) — L2Loss (MSE), L1Loss, HuberLoss(delta) / SmoothL1Loss, CrossEntropyLoss (logits + int64 class indices), NLLLoss, BCELoss, BCEWithLogitsLoss, KLDivLoss (log-probs + probs). All map (predictions, targets) → scalar loss.
  • Optimizers (Shorokoo.Modules.Optimizers) — SGDOptimizer, SGDMomentumOptimizer, AdamOptimizer (with bias correction), AdamWOptimizer, RMSpropOptimizer, AdagradOptimizer, with strongly typed hyperparameter sets and learning-rate schedules (Schedules.*). Optimizer state (moments, velocity, accumulators) is created inside each module via optimizer-owned [StateInitializer]s — OptimizerStateZeros (param-shaped) and OptimizerScalarZeros (a rank-0 scalar, e.g. Adam's timestep) — and threaded with StateUpdate; never declared in the Inline signature.
dotnet add package Shorokoo.Modules

using Shorokoo.Modules.Optimizers;
using Shorokoo.Modules.Losses;

var rig = TrainingRig.FromScratch(
 MyModel.ComputationGraph,
 CrossEntropyLoss.ComputationGraph,
 AdamOptimizer.ComputationGraph,
 sampleInputs,
 new AdamOptimizerHyperparameters { LearningRate = 1e-3f });

Documentation: https://github.com/Shorokoo/Shorokoo

Product Versions Compatible and additional computed target framework versions.
.NET net10.0 net10.0 is compatible.  net10.0-android net10.0-android was computed.  net10.0-browser net10.0-browser was computed.  net10.0-ios net10.0-ios was computed.  net10.0-maccatalyst net10.0-maccatalyst was computed.  net10.0-macos net10.0-macos was computed.  net10.0-tvos net10.0-tvos was computed.  net10.0-windows net10.0-windows was computed. 
Compatible target framework(s)
Included target framework(s) (in package)
Learn more about Target Frameworks and .NET Standard.

NuGet packages (1)

Showing the top 1 NuGet packages that depend on Shorokoo.Modules:

Package Downloads
Shorokoo

Define, train, and run neural networks in pure C#. Meta-package that brings the Shorokoo runtime (Shorokoo.Core), ready-made layers (Shorokoo.Modules), and the [Module] source generator (Shorokoo.CodeGen). Add exactly one backend: Shorokoo.LinuxCPU, Shorokoo.LinuxGPU, Shorokoo.WinCPU, or Shorokoo.WinGPU.

GitHub repositories

This package is not used by any popular GitHub repositories.

Version Downloads Last Updated
0.0.1 62 6/25/2026
0.0.1-preview 56 6/25/2026