Sep 25 24 days ago

1. 23:39

   Switch to 35B i1-Q4_K_M (same compression, faster load); size GPU reservations from measurements

   markjouh
   Verified signature

   2bc79b6
2. 23:03

   Refine UI: typography, in-panel progress, inline errors, token tooltips

   markjouh
   Verified signature

   4257d42
3. 21:58

   Simplify UI, show token coding costs, and skip automatic decode verification

   markjouh
   Verified signature

   ca5fce0
4. 21:42

   Increase input limit to 8192 tokens and 64 KiB

   markjouh
   Verified signature

   5273448
5. 21:35

   Use GGML BF16 cache enum supported by the binding

   markjouh
   Verified signature

   b9a2a13
6. 21:34

   Add published Qwen3.5 Base Q8_0 arithmetic compressor

   markjouh
   Verified signature

   a81897d