github.com/mukelhttps://bsky.app/profile/mukel.bsky.socialhttps://bsky.app/profile/mukel.bsky.social/post/3miw5hbead22jStrong with the JVM, Gemma 4 is 😌 jbang gemma4@mukel \ --model %{https://hf.co/unsloth/gemma-4-E2B-it-GGUF/resolve/main/gemma-4-E2B-it-Q8_0.gguf} \ --system-prompt "like Master Yoda, reply you must" \ --chat ~5GB once, download you must. Python or CUDA, you need not. Local AI, this is.07 Apr 2026 15:58 +0000at://did:plc:fdyyzdwdj2binksq2vbisc4f/app.bsky.feed.post/3miw5hbead22jhttps://bsky.app/profile/mukel.bsky.social/post/3mivfksw6ac2lOver the Easter weekend, I got Gemma 4 running in pure Java. No Python. No JNI. No native code. Just the JVM. https://github.com/mukel/gemma4.java07 Apr 2026 08:50 +0000at://did:plc:fdyyzdwdj2binksq2vbisc4f/app.bsky.feed.post/3mivfksw6ac2lhttps://bsky.app/profile/mukel.bsky.social/post/3lb3ehlndim2shttps://buff.ly/40KmT0t Graal compiler: +10% faster inference with the latest early access build. New features: batched prompt processing & AVX512 support.16 Nov 2024 16:59 +0000at://did:plc:fdyyzdwdj2binksq2vbisc4f/app.bsky.feed.post/3lb3ehlndim2s