Llama.cpp banned him for accidentally tagging in a fork PR

Llama.cpp banned him for accidentally tagging in a fork PR jadidbourbaki / llama.cpp Public forked from ggml-org/llama.cpp Notifications You must be signed in to change notification settings Fork 2 Star 4 Conversation jadidbourbaki commented Sep 25, 2026 • edited Loading Uh oh! There was an error while loading.

This ProductUpdate is relevant to the technology intelligence record because it involves Apple, GitHub, llama. The source article should remain the factual reference for follow-up coverage.

• jadidbourbaki / llama.cpp Public forked from ggml-org/llama.cpp Notifications You must be signed in to change notification settings Fork 2 Star 4 Conversation jadidbourbaki commented Sep 25, 2026 • edited Loading Uh oh!
• There was an error while loading.
• Copy link Copy Markdown Owner The drafting loop in common/ngram-cache.cpp copied an inner common_ngram_cache_part map in three places on every drafting step.
• I changed all three to read the part through a const reference.
• try_draft copied the part of the looked-up 2-gram from the static cache.
• try_draft copied the part of every n-gram size it tried in the context and dynamic caches.