Somewhere Systems

centience

Somewhere Systems' Autonomous Agent Engine

dataclysm

Pull high-quality, efficient embeddings for PubMed, arXiv and Wikipedia from Huggingface and use for local LLM inference/Retrieval Augmented Generation (RAG)

llama2mlx

Karpathy's llama2.c transpiled to MLX for Apple Silicon

sentinel

Sentinel is a system which uses Voice Agents to send outbound dialers (telemarketers) to the shadow realm

surreal-engine

Procedural art interface with a neural rendering suite

agent-examples

PoCs for Agents

ggml-k8

Run GGML models with Kubernetes.

setup-scripts

Setup scripts for various VMs and tasks

arxiv-tools

Tools to bulk download arxiv data

ggml-k8s

Run GGML models with Kubernetes.

megablocks-public

metavoice

AI for human-level speech intelligence

multishot

Embed datasets from Kaggle like Cornell/arXiv and others

S2PC-server

Node/Express server for networked user support with the S2PC VR/Web/Mobile Player Controller

snippets

Apache 2.0 code snippets from Somewhere Systems

uniword

The world's first pronounceable programming notation.

llm

Large Language Model: 160kg of parameters, fueled by burgers & fries. Exploring emergent language through extreme scale & questionable dietary choices. Research purposes only.

nextwave-current

.TS/Next-based static microblog template