now     waiting · proxy idle 8m/10m
next    idle 8m/10m → train Qwen3.5 4B
paused  no

  Model             Learning       Pending    Last update
  Qwen3.5 4B        new only       38         today 14:02  ✓ swapped
  Granite 4.2 3B    off            —          —
  LFM3 1.7B         unsupported    —          unsupported

learning   Qwen3.5 4B · 14 conversations · since 11-10 09:12
  tool_error → fixed 5 · user_correction 4 · retry → success 3 · topics pytest, git
last       11-10 14:02 · repeat_errors −18% · drift 0.002 · ✓ swapped
