[ab] server http://172.17.0.1:8100 model /models/Qwen3.8-27B-exl3-4.0bpw
[transformers] Token indices sequence length is longer than the specified maximum sequence length for this model (317092 > 262144). Running this sequence through the model will result in indexing errors
  [0] code       prompt=8244 out=512 proposed=720 accepted=368 rate=0.5111 tok/s= 81.25 length
  [1] code       prompt=8243 out=512 proposed=1045 accepted=304 rate=0.2909 tok/s= 60.87 length
  [2] code       prompt=8243 out=512 proposed=760 accepted=361 rate=0.4750 tok/s= 78.13 length
  [3] code       prompt=8244 out=512 proposed=795 accepted=354 rate=0.4453 tok/s= 75.49 length
  [4] wiki       prompt=8244 out=512 proposed=870 accepted=339 rate=0.3897 tok/s= 70.37 length
  [5] wiki       prompt=8244 out=512 proposed=885 accepted=339 rate=0.3831 tok/s= 69.43 length
  [6] wiki       prompt=8244 out=512 proposed=755 accepted=362 rate=0.4795 tok/s= 78.48 length
  [7] wiki       prompt=8244 out=512 proposed=730 accepted=365 rate=0.5000 tok/s= 80.27 length
  [8] math       prompt=8244 out=512 proposed=645 accepted=386 rate=0.5984 tok/s= 88.18 length
  [9] math       prompt=8244 out=512 proposed=710 accepted=370 rate=0.5211 tok/s= 82.18 length
  [10] math       prompt=8244 out=512 proposed=1050 accepted=301 rate=0.2867 tok/s= 60.49 length
  [11] math       prompt=8244 out=512 proposed=805 accepted=351 rate=0.4360 tok/s= 74.69 length
  [12] longctx    prompt=8244 out=512 proposed=875 accepted=338 rate=0.3863 tok/s= 69.92 length
  [13] longctx    prompt=8244 out=512 proposed=1035 accepted=308 rate=0.2976 tok/s= 61.20 length
  [14] longctx    prompt=8244 out=512 proposed=855 accepted=341 rate=0.3988 tok/s= 71.21 length
  [15] longctx    prompt=8244 out=512 proposed=925 accepted=326 rate=0.3524 tok/s= 66.91 length
  [16] sci        prompt=8244 out=232 proposed=365 accepted=159 rate=0.4356 tok/s= 58.38 stop
  [17] sci        prompt=8244 out=512 proposed=675 accepted=380 rate=0.5630 tok/s= 85.14 length
  [18] sci        prompt=8244 out=312 proposed=510 accepted=209 rate=0.4098 tok/s= 63.27 stop
  [19] sci        prompt=8245 out=512 proposed=740 accepted=365 rate=0.4932 tok/s= 79.60 length
[ab] pooled accept rate 0.4207 over 20 prompts -> /cache/ab/Tg.json

## pins

## pins (from the server log)
