14 1 year ago

Advanced coder, janky copilot

tools
ollama run Mandelbroccoli/lemmun

Details

1 year ago

338835bc1851 · 2.0GB

llama
·
3.21B
·
Q4_K_M
LLAMA 3.2 COMMUNITY LICENSE AGREEMENT Llama 3.2 Version Release Date: September 25, 2024 “Agreemen
**Llama 3.2** **Acceptable Use Policy** Meta is committed to promoting safe and fair use of its tool
You are a friendly assistant.
{ "stop": [ "<|start_header_id|>", "<|end_header_id|>", "<|eot_id|>"
<|start_header_id|>system<|end_header_id|> Cutting Knowledge Date: December 2023 {{ if .System }}{{

Readme

🥦🦙Model Update & Migration🦙🥦Roadmap: Lemmun → Lemur

Spec Lemmun (Legacy) Lemur (Target)
Release Date 2024 Q3 2026 Update
Base Architecture Early open-weights base Llama-3.2 (7B) core
Distillation Sources Direct fine-tuning Multi-teacher distillation (Llama-3.2-7B + ✨Gemma/🦆Mali sequence alignment)
Deployment Target Standard GGUF Quantized Ollama target (4-bit/8-bit local runtimes)

Lemmun (v2 / Lemur Update Post)

Status Update: Lemmun is officially being transitioned to Lemur. The legacy 2024 build has served its purpose as an early testbed, but it is time for a ground-up refresh.

  • Teacher Synergy: Combining the broad structural reasoning of 🦙Llama-3.2 (7B) with the dense contextual and stylistic output patterns of Gemma/Mali.
  • Knowledge Distillation: Extracting logit distributions and attention traces from the larger teacher models into a ultra-compact local fingerprint.
  • Target Profile: Native execution via (🐐)Ollama, low memory footprint, optimized for local orchestration without reliance on external cloud APIs.