Build, train, and export Torch-backed NeuralFn models via the Python API.
npx skills add https://github.com/adamdroberts/NeuralFn --skill neuralfn-torch
Please help me install this Agent Skill. Skill: neuralfn-torch Source: https://github.com/adamdroberts/NeuralFn/tree/main/.cursor/skills/neuralfn-torch Command: npx skills add https://github.com/adamdroberts/NeuralFn --skill neuralfn-torch
Build, train, and export Torch-backed NeuralFn models using the Python API to enable rapid development and deployment of transformer-based systems.
Install NeuralFn, import the NeuralFn Python API, and run the provided end-to-end example to build, train, and export a Torch-backed model.
High-intent search queries and answers about installing and using this skill.
To train and export Torch models, use the NeuralFn Python API to configure ModelSpec and BlockSpec with presets like nanogpt, llama, or jamba, compile the graph with TorchTrainer, and export the trained weights to a .pt file for inference or deployment in Python projects.
NeuralFn supports presets including nanogpt, llama, and jamba for configuring transformer-based model graphs. You select a preset when setting up your ModelSpec and BlockSpec before compiling the graph with the Torch runtime.
Yes, building, training, and exporting Torch-backed NeuralFn models requires the NeuralFn Python SDK and the Torch runtime. Optional MCP tooling is also available for integrating the training workflow with model automation pipelines.
Yes, the NeuralFn workflow generates fully runnable Torch modules that support autoregressive outputs. Once training is complete via TorchTrainer, you can export the weights and run inference directly within your Python projects.
The best way to deploy Torch-backed NeuralFn models is to train them using TorchTrainer and export the resulting weights to a .pt file. This produces an inference-ready, fully runnable Torch module that integrates directly into Python deployment environments.
Configuring ModelSpec and BlockSpec involves defining the architecture layers and parameters for your chosen NeuralFn preset. Once set, you compile the graph with the Torch runtime to prepare the model for training with TorchTrainer and subsequent weight export.
adamdrobertsConstruct, manipulate, and execute NeuralFn neuron graphs with the neuralfn Python package.
zechenzhangAGIReduce LLM inference latency using draft models, Medusa heads, and Lookahead Decoding.
AngelP17Automate end-to-end ML pipelines from data ingestion to deployment.
aayushsoamTrain scalable PyTorch models with automatic distributed training across hardware configurations.
