# synchopate/llama-modchip

> Tool for running large MoE models (DeepSeek-V4-Flash 284B, Nemotron-30B) on pure CPU with 64 GB RAM and no GPU, using a custom Fractal-64 compression format and speculative decoding.

- Magnitude: 6.8 out of 10 — Early signal
- Stars: 507 total · +36 stars measured 2026-11-09, 00:02–02:08 UTC, ≈ 414 by evening
- Star trust: star growth looks organic
- Category: Language models · Language: Python · License: Apache-2.0 · Created: 2026-10-29 · Last push: 2026-11-08
- GitHub: https://github.com/synchopate/llama-modchip · Homepage: https://github.com/synchopate/llama-modchip#readme · Page: https://gitnova.dev/en/r/synchopate/llama-modchip

## Useful for

- Run DeepSeek-V4-Flash 284B locally on 64 GB RAM without a GPU
- Wire Claude Code CLI to a local model via ./modchip claude
- Run ./modchip bench to measure tok/s and compare your CPU

## Why it’s here

- Star-counter measurements on 2026-11-09 (UTC), 00:02–02:08: 471 → 507 stars (+36). This is the change over that interval.
- Estimated end-of-day forecast: about +414 stars, using observed gains and the previous day.
- Over the last two days the pace is 46× that of the previous 9 days.
- The repository is 11 days old and already has 507 stars. With less than two weeks of history, there's no usual pace to compare the spike against yet.
- Hacker News: “Llama-modchip: 284B DeepSeek on 64GB RAM, no GPU” — 366 points, 15 h ago.
- About 11 forks a day — people are taking the code.

## Star trust

Star growth looks organic.

Star-trust labels are heuristics based on the repository’s behavior, not a check of every stargazer.

## Numbers

- Forks: 26
- Issues and pull requests: 13
- Watchers: 8
- Average over the last week: 122 per day
- Usual pace: too little history (under two weeks)
- Stars in the last hour (measured): 14
- Latest release: v2.4.0 (2026-11-04)

## Stars per day, last 12 days (oldest → newest, today is partial)

2026-10-29 … 2026-11-09: 3, 43, 21, 2, 2, 1, 0, 1, 9, 7, 423, 36

## Magnitude by day, last 2 days

11-08 5.5, 11-09 6.8

## Hacker News

- Llama-modchip: 284B DeepSeek on 64GB RAM, no GPU — 366 points, 159 comments: https://news.ycombinator.com/item?id=50411961

## Spotted in now

- Spotted on Hacker News

## More in this category

1. **NVIDIA/auto-ontology** — 5.4 · Peaking · Language models · Python · +86 stars measured 2026-11-09, 00:02–02:09 UTC, ≈ 633 by evening
   NVIDIA's Python framework that builds a semantic ontology graph over enterprise metadata, giving LLM agents trusted metric and entity definitions for reliable text-to-SQL.
   Full card: https://gitnova.dev/en/r/NVIDIA/auto-ontology.md
2. **neuml/txtai** — 4.7 · Breakout · Language models · Python · +8 stars measured 2026-11-09, 00:03–02:09 UTC, ≈ 63 by evening
   Python framework for semantic search, LLM orchestration and language model workflows. Combines vector indexes, graphs and relational databases into an embeddings database for RAG, agents and multimodal search.
   Full card: https://gitnova.dev/en/r/neuml/txtai.md
3. **JustVugg/colibri** — 4.0 · Steady · Language models · C · +30 stars measured 2026-11-09, 00:03–02:09 UTC, ≈ 306 by evening
   A pure-C, zero-dependency inference engine for running large MoE models (744B–2.8T parameters) on consumer hardware by treating VRAM, RAM, and storage as a single multitier hierarchy and streaming experts from disk.
   Full card: https://gitnova.dev/en/r/JustVugg/colibri.md
4. **Niko1221/Strata** — 3.8 · Steady · Language models · C++ · +37 stars measured 2026-11-09, 00:03–02:10 UTC, ≈ 401 by evening
   A C++ local inference engine that runs the 125B MoE model Qwen3.8-Flash-Next on a regular PC with one NVIDIA GPU (12-24 GB) and 64 GB RAM, exposing an OpenAI/Anthropic-compatible API on localhost.
   Full card: https://gitnova.dev/en/r/Niko1221/Strata.md
5. **togg53192-cmd/jailbreaks** — 3.5 · Cooling · Language models · +42 stars measured 2026-11-09, 00:03–02:10 UTC, ≈ 300 by evening
   A collection of 31 system prompts, persona locks, and AGENTS.md files for 25+ models (Claude, Gemini, Grok, GLM, DeepSeek, Kimi, Qwen, Gemma, GPT-OSS). Intended for research on how instruction hierarchies and policy overrides affect model…
   Full card: https://gitnova.dev/en/r/togg53192-cmd/jailbreaks.md

---

Magnitude (0–10) measures how fast and how unusually interest in a repository is growing right now. It is not a quality score.
Days are UTC. “So far today” is a fact; “expected by the end of the day” is a forecast.
Summaries and use cases are written by an LLM (DeepSeek V4.1 Flash) from the README and may be inaccurate: verify specific claims (benchmarks, speed, hardware) in the repository itself.
Data as of 2026-11-09 02:18 UTC, updated every 30 minutes.