Gemma-3-4B-IT fine-tuned for uncensored creative writing and roleplay. Made by RACER IS OP.
Google’s Gemma 3 4B, as Ollama builds it, with the picture-reading part taken out. The text weights are the same bytes. It writes text. It cannot look at an image.
Contact & Chat: https://etherpad.wikimedia.org/p/ollama
https://huggingface.co/nidum/Nidum-Gemma-3-4B-it-Uncensored-GGUF
Google Gemma-3-4B-it llama.cpp quantization by Fabrício Carraro
專為中華民國台灣語境打造的繁體中文指令微調模型
Gemma 3 4B Instruct fine-tuned on Occitan (Lengadocian) via RS-LoRA (r=32). Good balance of quality and speed. Q2_K to f16 quants.
Gemma 3 4B Instruct fine-tuned on Occitan via DoRA (r=16). Lightweight model for resource-constrained setups. Q2_K, Q4_K_M, Q5_K_M, Q8_0, f16.
nidum-gemma-3-4B-it-uncensored
A 4B-parameter Persian-specialized language model built on Google's Gemma 3 architecture, fine-tuned on high-quality Persian text data while preserving multimodal capabilities for native-quality responses.
gemma3:4b but Emu Otori (info taken from https://projectsekai.fandom.com/wiki/Otori_Emu)
Bilingual (IT/EN) offline AI assistant optimized for Raspberry Pi. Runs completely offline with 3.67 tokens/s on Pi 4.
gemma3:4b but KAngel (info taken from https://needy-streamer-overload.fandom.com/wiki/Ame-chan)
gemma3:4b but mai from bunny girl senpai [made for a schoolfriend] (info taken from https://aobuta.fandom.com/wiki/Mai_Sakurajima)
gemma3:4b but rui kamishiro (info taken from https://projectsekai.fandom.com/wiki/Kamishiro_Rui)
gemma3:4b but raffina (info taken from https://puyonexus.com/wiki/Raffina)
gemma3:4b but niko (info taken from https://oneshot.fandom.com/wiki/Niko)
gemma3:4b but rafisol (info taken from https://puyonexus.com/wiki/Rafisol)
gemma3:4b but agar agar cookie (info taken from https://cookierunkingdom.fandom.com/wiki/Agar_Agar_Cookie)
llmfan46/gemma-4-31B-it-uncensored-heretic-GGU with Vision