249M-param MoE transformer built from scratch in PyTorch. GQA, RoPE, SwiGLU, sparse MoE with 3 aux losses, AMP training loop no Trainer abstractions. Architecture mirrors LLaMA/Mistral/Mixtral decisions, fully inspectable.
☆19May 22, 2026Updated 5 months ago
Alternatives and similar repositories for H64LM
Users that are interested in H64LM are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- ☆16Oct 25, 2026Updated 2 weeks ago
- Implementation of a fast semantic chunker in C++, installable in python 3.7+ projects.☆22Sep 20, 2025Updated last year
- The open source workflow agent.☆11Oct 10, 2026Updated last month
- Privacy-first iOS app with on-device LLM inference via llama.cpp. Chat with local models, get AI health insights from HealthKit, and trac…☆33Aug 3, 2026Updated 3 months ago
- Making MoE fine-tuning accessible to anyone with consumer hardware.☆18Oct 4, 2026Updated last month
- Managed Kubernetes at scale on DigitalOcean • AdDigitalOcean Kubernetes includes the control plane, bandwidth allowance, container registry, automatic updates, and more for free.
- A lightweight transformer language model built from scratch in PyTorch, trained on a single consumer GPU with a full pipeline for data pr…☆43Sep 30, 2026Updated last month
- A zero-dependency Python module for inspecting and converting coding-agent session files (.jsonl) — Claude Code, Codex, and Pi — into the…☆49Jul 10, 2026Updated 4 months ago
- [COLM '24] Source-Aware Training Enables Knowledge Attribution in Language Models☆23Apr 1, 2025Updated last year
- Glow-TTS with Stochastic Duration Predictor and Stochastic Pitch Predictor☆20Jun 5, 2023Updated 3 years ago
- This repo is a alternative solution for standard self-attention. It uses tree routing with learnable functions for cheaper self-attention…☆43Oct 11, 2026Updated last month
- Docling → Jev vs Docling → Gemini 3.8 Flash vs Gemini reading the PDF: same documents, one clock, scored against arXiv's own metadata☆14Sep 22, 2026Updated last month
- Demo of fine-tuning QA models for answering FAQ of cloud providers documentation☆11Jun 20, 2026Updated 4 months ago
- LLM Context Manager for inference optimization☆25Jul 28, 2025Updated last year
- Chew is a Go library for processing various content types into markdown/plaintext.☆42Feb 16, 2025Updated last year
- 1-Click AI Models by DigitalOcean Gradient • AdDeploy popular AI models on DigitalOcean Gradient GPU virtual machines with just a single click. Zero configuration with optimized deployments.
- Lightweight OpenAI wrapper using FastAPI. Add rate limits to OpenAI usage, optionally log and store all API calls, and share regulated Op…☆15Feb 15, 2024Updated 2 years ago
- Examples for the Activate conference☆11Sep 11, 2019Updated 7 years ago
- Groq-powered MAD: The first work to explore Multi-Agent Debate with Large Language Models :D☆12Jul 5, 2024Updated 2 years ago
- A small, local-first AI chat workspace for any OpenAI-compatible model. Cache-efficient conversations, MCP tools, web search and persiste…☆39Updated this week
- Open-source AI agent that reasons by blending many AI models, does real work on its own, and keeps learning — safe, self-hostable, and fu…☆29Updated this week
- Nerves system for Raspberry Pi 5 with support for the Hailo AI accelerator☆10Dec 16, 2024Updated last year
- Build and operate production-safe agents that run in your own environment.☆17Oct 28, 2026Updated 2 weeks ago
- Fork of github.com/UCSBarchlab/OpenTPU for the TGPTPU project☆14Jun 1, 2025Updated last year
- Experimental validation: Agentic Search (context trees) vs traditional RAG for code retrieval. Agentic wins with 99% fewer tokens and 2× …☆13Jan 6, 2026Updated 10 months ago
- Virtual machines for every use case on DigitalOcean • AdGet dependable uptime with 99.99% SLA, simple security tools, and predictable monthly pricing with DigitalOcean's virtual machines, called Droplets.
- Precision Knowledge Editing (PKE): A novel method to reduce toxicity in LLMs while preserving performance, with robust evaluations and ha…☆14Nov 26, 2024Updated last year
- A PyTorch-based optimizer wrapper for continual learning via selective fine-tuning, guided by the condition number κ (kappa) of model ten…☆13Mar 13, 2026Updated 7 months ago
- 3D Spiking neural network simulation exploring Spike Timing Dependent Plasticity (STDP)☆11Apr 1, 2026Updated 7 months ago
- A Rust binding for LibHailoRS☆11Jan 18, 2025Updated last year
- A 21M-parameter LM with a 16.8M-row product-key memory: as good as a 114M dense model and runs with the table on an SSD. Triton kernels f…☆16Oct 23, 2026Updated 2 weeks ago
- ☆13Apr 22, 2026Updated 6 months ago
- ☆16Jul 23, 2026Updated 3 months ago
- Live structural code graph for AI agents — one-curl install wires MCP + an agent skill into Claude Code, OpenCode, Codex and Cursor. Inde…☆13Oct 24, 2026Updated 2 weeks ago
- lets build claude code from scratch☆36May 13, 2026Updated 5 months ago
- 1-Click AI Models by DigitalOcean Gradient • AdDeploy popular AI models on DigitalOcean Gradient GPU virtual machines with just a single click. Zero configuration with optimized deployments.
- Abliterates models quickly and efficently, Qwen2.5-7B: 95%→2% refusal at 0.095 KL☆44Nov 2, 2026Updated last week
- TalkingData AdTracking Fraud Detection Challenge☆10May 8, 2018Updated 8 years ago
- ☆11Mar 1, 2026Updated 8 months ago
- MIT-licensed code knowledge graph: index any source repo into a queryable graph of symbols, call chains and blast radius — with wiki gene…☆17Oct 3, 2026Updated last month
- A self-hosted, AI-native personal knowledge base. Notes, tasks, projects and contacts, with an AI layer that reads what you already wrote…☆20Updated this week
- ☆16Dec 16, 2024Updated last year
- The News Landscape Toolkit (NELA)☆16Oct 14, 2020Updated 6 years ago
Are these results useful?