it's time to update your Gemma 4 GGUFs
New GGUF files from bartowski and unsloth fix broken chat templates for Gemma 4
Deep Dive
Chat Template was fixed a few days ago. Choose your favorite dealer: links to bartowski and unsloth GGUF files for Google Gemma 4 models (31B, 26B-A4B, E4B, E2B) on Hugging Face.
Key Points
- Chat template bug fixed for all four Gemma 4 variants (31B, 26B-A4B, E4B, E2B)
- Updated GGUF files released by bartowski and unsloth on Hugging Face
- Fix ensures proper token formatting for local inference tools like llama.cpp and Ollama
Why It Matters
Fixes multi-turn conversation and prompt formatting, essential for reliable local LLM deployment.