mirror of
https://github.com/mudler/LocalAI.git
synced 2026-09-13 06:45:26 -04:00
Offer AtomicChat IQ4_XS and Q4_K_M builds alongside the existing Qwen3.8 Flash Next variants. Include every shard and the F16 projector. Use the publisher's required mmap and parameter fitting settings. Pin downloads to a verified revision and document direct installation. Assisted-by: Codex:GPT-6