Jeffrey Morgan
|
dfa2f32ca0
unload in critical section (#4187)
|
1 vuosi sitten |
Daniel Hiltgen
|
f56aa20014
Centralize server config handling
|
1 vuosi sitten |
Jeffrey Morgan
|
942c979232
allocate a large enough kv cache for all parallel requests (#4162)
|
1 vuosi sitten |
Daniel Hiltgen
|
20f6c06569
Make maximum pending request configurable
|
1 vuosi sitten |
Mark Ward
|
63c763685f
log when the waiting for the process to stop to help debug when other tasks execute during this wait.
|
1 vuosi sitten |
Mark Ward
|
f4a73d57a4
fix runner expire during active use. Clearing the expire timer as it is used. Allowing the finish to assign an expire timer so that the runner will expire after no use.
|
1 vuosi sitten |
Daniel Hiltgen
|
d6e3b64582
Fix concurrency for CPU mode
|
1 vuosi sitten |
Daniel Hiltgen
|
9b5a3c5991
Merge pull request #3914 from dhiltgen/mac_perf
|
1 vuosi sitten |
Jeffrey Morgan
|
00b0699c75
Reload model if `num_gpu` changes (#3920)
|
1 vuosi sitten |
Daniel Hiltgen
|
b123be5b71
Adjust context size for parallelism
|
1 vuosi sitten |
Bryce Reitano
|
36a6daccab
Restructure loading conditional chain
|
1 vuosi sitten |
Bryce Reitano
|
284e02bed0
Move ggml loading to when we attempt fitting
|
1 vuosi sitten |
Daniel Hiltgen
|
34b9db5afc
Request and model concurrency
|
1 vuosi sitten |