42 yesterday

Optimized 4.5B Gemma 4 model with Unsloth Dynamic QAT & MTP (4.3GB). Features near-lossless multimodal reasoning, an active 4-token speculative window, and an expanded 64K context size. Best for edge execution on systems with 8GB to 16GB RAM.

tools thinking
ollama run 4skl/gemma4-e4b-mtp

Details

yesterday

42cc992a8b53 · 4.3GB ·

gemma4
·
7.46B
·
Q4_0
GGUF���1�������+��������������general.architecture����������gemma4-assistant �������general.type
{ "draft_num_predict": 4, "num_ctx": 65536, "stop": [ "<turn|>" ], "temp

Readme

No readme