AI๊ฐ ์์ง์ด๊ณ ๋ฏธ๋๊ฐ ์
๋ฐ์ดํธ๋ฉ๋๋ค.gemma-4-12B-coder-fable5-composer2.5-v1-GGUF
> **No matter your GPU. No matter your RAM.** If you've got **~4.5 GB** of VRAM *or* unified memory free, > you can run your own private, offline coding assistant right now. ๐ > This is the **v1 / code edition** โ distilled from **real chain-of-thought** so it *thinks through* a problem > before writing the solution. ๐ง ๐ป All local, all yours, no API, no cloud. A focused fine-tune of Gemma 4 12B on **verifiable Python coding** data โ every training example's reasoning leads to code that **actually passed its tests**. The result reasons in the open (edge cases, complexity, approach) and then emits a clean, runnable solution. ๐ **๐๐ฅ IT'S HERE โ v2 is OUT NOW!** v2 has shipped โ the **GGUF quants are live and ready to run** โ **grab v2 here**. ๐ The **full `safetensors` master** (build / fine-tune on top) goes up **tomorrow**. v2 is **agentic + coding** focused โ the piece v1 was missing.
๋ฒ์ญ ์ค๋น ์ค์ ๋๋ค. ์๋ฌธ์ ํ์ํฉ๋๋ค.
> **No matter your GPU. No matter your RAM.** If you've got **~4.5 GB** of VRAM *or* unified memory free, > you can run your own private, offline coding assistant right now. ๐ > This is the **v1 / code edition** โ distilled from **real chain-of-thought** so it *thinks through* a problem > before writing the solution. ๐ง ๐ป All local, all yours, no API, no cloud. A focused fine-tune of Gemma 4 12B on **verifiable Python coding** data โ every training example's reasoning leads to code that **actually passed its tests**. The result reasons in the open (edge cases, complexity, approach) and then emits a clean, runnable solution. ๐ **๐๐ฅ IT'S HERE โ v2 is OUT NOW!** v2 has shipped โ the **GGUF quants are live and ready to run** โ **grab v2 here**. ๐ The **full `safetensors` master** (build / fine-tune on top) goes up **tomorrow**. v2 is **agentic + coding** focused โ the piece v1 was missing.