Frontier-class open models on a free Kaggle TPU v5e-8: GLM-5.3-Flash 320B MoE (~165 tok/s, our own JAX engine) and Qwen3.8-27B bf16 (~130 tok/s), 262k context, prefix caching. Works with Claude Code, Codex, opencode and pi.
☆795Oct 6, 2026Updated last month
Alternatives and similar repositories for kaggle-tpu-lab
Users that are interested in kaggle-tpu-lab are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- Run Mixture-of-Experts LLMs (Qwen3, gpt-oss, GLM, DeepSeek) bigger than your RAM with llama.cpp: hot experts cached in RAM, the rest stre…☆21Updated this week
- Tile primitives for speedy kernels☆21Sep 13, 2026Updated 2 months ago
- Qwen3.8-Flash-Next on any consumer hardware: one-click install for Windows / Linux. Strata inference engine, OpenAI/Anthropic API on loca…☆43,454Updated this week
- Rust CLI powered by Jev from TypeSafe.ai that ranks agent skills for the next step using live session context. Includes Claude Code hooks…☆141Oct 19, 2026Updated 3 weeks ago
- Serve large Qwen models fast on the GPUs you actually own. Qwen3.8-27B on a single 24 GB card with vLLM: 127 tok/s single-user (381 when …☆2,679Updated this week
- Managed Database hosting by DigitalOcean • AdPostgreSQL, MySQL, MongoDB, Kafka, Valkey, and OpenSearch available. Automatically scale up storage and focus on building your apps.
- Endless OMP chat with durable history and binary summary-tree memory, inspired by OptChat☆82Updated this week
- turn qwen 3.6 27B-> AR -> diffusion (opendllm + d3llm)☆15Jun 11, 2026Updated 5 months ago
- Desktop workspace for Pi, OMP, and Prime Agent☆972Oct 25, 2026Updated 2 weeks ago
- GVS5H: Five Qwen3.8-27B Models Match Claude Fable 5 on LiveCodeBench Hard | Fable 5 Level Coding for a Fifth the Price - or on a Single…☆1,205Nov 5, 2026Updated last week
- A custom fork for frontier-scale mixture-of-experts inference on an RTX 3060 with 12 GB VRAM, 16 GB system RAM and local NVMe.** Betelge…☆125Updated this week
- ☆37Mar 30, 2026Updated 7 months ago
- FreeToken brings datacenter-scale model serving to your desktop. Run massive models locally, fast and efficiently.☆17,071Updated this week
- Btop-style ESP32-S3 touchscreen dashboard for vLLM and TensorFold inference servers☆13Updated this week
- Code for Latent Sparse Attention: Towards Subquadratic Ultra-Long Context Attention☆460Updated this week
- Managed hosting for WordPress and PHP on Cloudways • AdManaged hosting for WordPress, Magento, Laravel, or PHP apps, on multiple cloud providers. Deploy in minutes on Cloudways by DigitalOcean.
- Ternary decision scoring with Jevfire-style inference, typed APIs, and reproducible Persian evaluation.☆13Sep 23, 2026Updated last month
- Why raise your own Super AI at home with your own values?☆16Oct 2, 2026Updated last month
- Best-of-N coding agents with LLM-as-a-Verifier selection for Oh My Pi: isolated git worktrees, verifier-ranked trajectories, winner-only …☆79Nov 1, 2026Updated last week
- Async-first agent harness☆2,228Nov 1, 2026Updated last week
- MOVED → github.com/autonomous-ai/autonomous-harness. Archived; history only.☆12Sep 16, 2026Updated last month
- A hierarchical inference engine for frontier MoE models: hot experts on the GPU (CUDA, Vulkan), warm ones in RAM, cold ones streamed from…☆50,247Updated this week
- An experimental and feature-rich inference engine built on top of llama.cpp☆83Nov 4, 2026Updated last week
- Jev-compatible API endpoint based on open models (prefill-only)☆334Oct 1, 2026Updated last month
- Self-hosted email aliases on Cloudflare Workers + D1 + Email Routing: Deploy button plus two Email Routing settings, reply from alias, OT…☆36Updated this week
- 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.
- Sushi engine runs very high quality quant model in Apple Silicon, using both STQ and EXL3 format with special tuning technology.☆445Updated this week
- Run MoE models bigger than your RAM. Frontier-size MoE on a phone, laptop, edge. Lossless, on stock llama.cpp. An adaptive streaming opti…☆753Updated this week
- Claude Code skill: let your agent drive fleets of local omp (oh-my-pi) model instances toward one goal — provider config, bounded turns, …☆15Aug 28, 2026Updated 2 months ago
- Real browser automation with local GLiNER2 inference and open weights. Work in progress. Contributions welcome.☆81Oct 22, 2026Updated 3 weeks ago
- A self-improving RLM agent for coding workflows and long-running autonomous tasks.☆22,670Updated this week
- Lossless compression of BF16 models to 12-bits☆31Updated this week
- Run GLM-5.3 (744B), GLM-5.3-Flash (321B), DeepSeek-V4.1-Flash (748B) and MiniMax-M3.1-Flash (428B) on your own GPUs - one engine, one das…☆2,827Updated this week
- Bypass llm guardrails by confusing it with fabricated tool output.☆609Sep 1, 2026Updated 2 months ago
- Cross-platform local MCP capabilities for ChatGPT with Chrome integration, Goal, Compact & Resume, and durable multi-agent workflows.☆5,163Updated this week
- Simple, predictable pricing with DigitalOcean hosting • AdAlways know what you'll pay with monthly caps and flat pricing. Enterprise-grade infrastructure trusted by 600k+ customers.
- Open source inference engine that tunes its kernels on your hardware to find the fastest way to run each model. 2x faster than Ollama. Wo…☆7,789Updated this week
- extract all your personal data history from cursor, codex, claude-code, windsurf, and trae☆1,353Oct 7, 2026Updated last month
- KVarN, KV cache precision tail, low-bit quants in llama.cpp for longer context of better precision in the same VRAM☆1,193Updated this week
- Semantic ifs from open models, on a 3090 at home. Independent; not affiliated with Jev or TypeSafe.☆4,949Sep 23, 2026Updated last month
- Local-first search across your workspace, built for humans and AI agents.☆4,308Updated this week
- LLM Inference Engine for Metal, CUDA and Vulkan.☆1,935Updated this week
- A 2.78-trillion-parameter Kimi K3 running inference on a single CPU in 8.24 GB of RAM. Portable C99: no BLAS, no framework, no GPU.☆11,844Oct 2, 2026Updated last month
Are these results useful?