Models

Gemma4GR v3: Gemma 4 E4B fine-tuned for cleaner Greek (stray foreign characters in ~1% of answers vs ~40% for stock). v2 tags were the base model (export bug), now fixed. Not a source of facts.

Uncensored Gemma 4 E4B & E2B (0/465 refusals) with MTP drafter. 32k context, up to 80 TPS on RTX 3050 8GB.

Stronger Gemma 4 E4B coding agent for Claude Code, 64K context, native tool-calling, runs fully local on 16GB Apple Silicon.

Gemma 4 E4B & E2B (Q4_K_M) with MTP drafter. 32k context, up to 80 TPS on RTX 3050 8GB.

Gemma 4 E4B admin and devops agent for Claude Code, 32K context fully on GPU for stable throughput, native tool-calling, runs local on 16GB Apple Silicon.

This is a coding model derived from the model Gemma E4B, optimized specifically for mac., use the model along with the adaptive runtime, look at the adaptive runtime section of readme

Gemma 4 E4B (Google DeepMind) with thinking mode disabled. Compact multimodal model — 4.5B effective / 8B total parameters. Supports text, image and audio input. Designed for edge devices and local deployment. Knowledge cutoff: January 2025.

Fine-tune of Gemma 4 E4B (4B parameters) specialized in offensive and defensive cybersecurity. Trained with LoRA (r=16) on a dataset covering vulnerability analysis, secure code review, and Java insecure code remediation.

Gemma 4 E4B (Google DeepMind) with thinking mode enabled. Compact multimodal model — 4.5B effective / 8B total parameters. Supports text, image and audio input. Designed for edge devices and local deployment. Knowledge cutoff: January 2025.

A modality-stripped variant of google/gemma-4-E4B-it for text-only classification, entity extraction, and structured-memory extraction. The vision encoder (150M params) and audio encoder (300M params) are removed; the text path is unchanged.

Gargantua 2.5: the main openTARS assistant on Gemma 4 E4B. Uses your computer’s tools, sees the screen and understands audio. Runs locally.

https://huggingface.co/HauhauCS/Gemma-4-E4B-Uncensored-HauhauCS-Aggressive

😈 Nieocenzurowana Gemma 4 E4B (Q6_ Abliterix) dla Ollamy. ⚡ Ultralekki agent rozumowania i kodowania, 🛠️ Natywne wywoływanie narzędzi, 🧠 Deep Thinking Rust 1.99.1, Linux/Windows CLI, Claude Code, OpenCode i Edge AI. 🚀 Matryca #16HEX

Google Gemma 4 E4B QAT checkpoint converted to GGUF by TurinTech. Text only; size tags map to the Hugging Face files. Measured Q4_0: Mean KLD 0.0014, Same top p 98.4% vs BF16 on 20 fixed documents (the selection tier); protocol ryzen9950x-2026-10.

😈 Uncensored Gemma 4 E4B (i1-Q4_K_M Abliterix) for Ollama. ⚡ Ultralight Reasoning & Coding Agent, 🛠️ Native Tool Calling, 🧠 Deep Thinking Rust 1.99.1 , Linux/Windows CLI, Claude Code, OpenCode & Edge AI. 🚀 #16HEX Matrix.

SwayAlgo Finance Gemma 4 E4B-it 1 is a Q4_K_M GGUF finance assistant for Indian business, accounting, reasoning, and finance MCQ tasks.

BibleAI is a production-ready Gemma 4 E4B model specialized for Bible study, theology, church history, and faith-centered Q&A.

Fine-tuned Gemma 4 E4B for quality-checking Bethesda/Starfield game string translations. Detects 16 issue codes: tag mismatches, source-language leak, English leak, AI artifacts, repetition, truncation, and more.

A fine-tuned Gemma 4 E4B model for Bible study. Quotes Scripture precisely from the BSB, provides Greek/Hebrew word studies, and presents Protestant, Catholic, and Orthodox perspectives.

LoRA adapter on Gemma 4 E4B for 3D print job planning. The first successful fine-tune after v1 (which parroted a template). This is the non-QAT sibling of the recommended v3 QAT model.