Model Intelligence Sheet
cmndcntrlcyber/gemma4-26b-a4b-code-trainer-gguf overview
Code Trainer — Phase 5 Gemma GGUF LoRA adapter from cmndcntrlcyber/gemma4 26b a4b code trainer v10 dpo https://hf.co/cmndcntrlcyber/gemma4 26b a4b code trainer…
Runs locally from ~13.10 GB disk (16 GB VRAM class GPUs with llama.cpp / guIDE).
Repository Files & Downloads
Model Details
| Model ID | cmndcntrlcyber/gemma4-26b-a4b-code-trainer-gguf |
|---|---|
| Author | cmndcntrlcyber |
| Pipeline | — |
| License | — |
| Base model | google/gemma-4-26B-A4B-it |
| Last modified | 2026-09-12T10:36:40.000Z |
Model README
---
base_model: google/gemma-4-26B-A4B-it
tags:
- gguf
- llama-cpp
- quantized
- code-generation
- gemma-4-26b-a4b
---
Code-Trainer — Phase 5 Gemma GGUF
LoRA adapter from cmndcntrlcyber/gemma4-26b-a4b-code-trainer-v10-dpo merged into
google/gemma-4-26B-A4B-it and quantized via llama.cpp. Use with llama-cli, llama-server,
Ollama, LM Studio, vLLM (experimental GGUF), or text-generation-webui.
Files
gemma-4-26B-A4B-it-Q4_K_M.ggufgemma-4-26B-A4B-it-IQ4_XS.gguf
Quick start (llama-server)
llama-server -m <file>.gguf --host 0.0.0.0 --port 8080 --ctx-size 4096Run cmndcntrlcyber/gemma4-26b-a4b-code-trainer-gguf with guIDE
Download guIDE — the AI-native code editor with local LLM inference and 69 built-in tools.
Source: Hugging Face · Compare models