Patrick Devine
|
4cc3be3035
Move envconfig and consolidate env vars (#4608)
|
11 miesięcy temu |
Sang Park
|
4434d7f447
Correct typo in error message (#4535)
|
11 miesięcy temu |
Daniel Hiltgen
|
ec231a7923
Remove VRAM convergence check for windows
|
11 miesięcy temu |
Patrick Devine
|
6845988807
Ollama `ps` command for showing currently loaded models (#4327)
|
11 miesięcy temu |
Daniel Hiltgen
|
4142c3ef7c
Always use the sorted list of GPUs
|
11 miesięcy temu |
Jeffrey Morgan
|
bb6fd02298
Don't clamp ctx size in `PredictServerFit` (#4317)
|
11 miesięcy temu |
Daniel Hiltgen
|
354ad9254e
Wait for GPU free memory reporting to converge
|
11 miesięcy temu |
Jeffrey Morgan
|
c9f98622b1
Skip scheduling cancelled requests, always reload unloaded runners (#4189)
|
1 rok temu |
Jeffrey Morgan
|
dfa2f32ca0
unload in critical section (#4187)
|
1 rok temu |
Daniel Hiltgen
|
f56aa20014
Centralize server config handling
|
1 rok temu |
Jeffrey Morgan
|
942c979232
allocate a large enough kv cache for all parallel requests (#4162)
|
1 rok temu |
Daniel Hiltgen
|
20f6c06569
Make maximum pending request configurable
|
1 rok temu |
Mark Ward
|
63c763685f
log when the waiting for the process to stop to help debug when other tasks execute during this wait.
|
1 rok temu |
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 rok temu |
Daniel Hiltgen
|
d6e3b64582
Fix concurrency for CPU mode
|
1 rok temu |
Daniel Hiltgen
|
9b5a3c5991
Merge pull request #3914 from dhiltgen/mac_perf
|
1 rok temu |
Jeffrey Morgan
|
00b0699c75
Reload model if `num_gpu` changes (#3920)
|
1 rok temu |
Daniel Hiltgen
|
b123be5b71
Adjust context size for parallelism
|
1 rok temu |
Bryce Reitano
|
36a6daccab
Restructure loading conditional chain
|
1 rok temu |
Bryce Reitano
|
284e02bed0
Move ggml loading to when we attempt fitting
|
1 rok temu |
Daniel Hiltgen
|
34b9db5afc
Request and model concurrency
|
1 rok temu |