Catter58/CASELLM-26b-a4b-evaluation

CASELLM-26b-a4b-evaluation (GGUF, Q4KM)

Q4KM quantization of Catter58/CASELLM-26b-a4b-evaluation-full.

• —Architecture: Gemma4 (MoE, 26B total / 4B active)
• —Quantization: Q4KM (~16 GB)
• —Converted with llama.cpp convert_hf_to_gguf.py

Usage (llama.cpp)

llama-cli -m casellm-26b-a4b-Q4_K_M.gguf -p "Hello"

Usage (Ollama)

ollama run reinhardbit/casellm-26b-a4b-evaluation