royjhan
|
a5f23d766e
Merge branch 'main' into royh-batchembed
|
10 months ago |
Roy Han
|
512e0a7bde
Clean up
|
10 months ago |
Roy Han
|
aee25acb5b
move normalization to go
|
10 months ago |
Jeffrey Morgan
|
717f7229eb
Do not shift context for sliding window models (#5368)
|
10 months ago |
Roy Han
|
80c1a3f812
playing around with truncate stuff
|
10 months ago |
Roy Han
|
c111d8bb51
normalization
|
10 months ago |
Roy Han
|
5213c12354
clean up
|
10 months ago |
Roy Han
|
49e341147d
add server function
|
10 months ago |
Roy Han
|
c406fa7a4c
api/embed draft
|
10 months ago |
Roy Han
|
ff191d7cba
Initial Draft
|
10 months ago |
Michael Yang
|
9d91e5e587
remove confusing log message
|
10 months ago |
Daniel Hiltgen
|
fb9cdfa723
Fix server.cpp for the new cuda build macros
|
11 months ago |
Jeffrey Morgan
|
ead259d877
llm: fix seed value not being applied to requests (#4986)
|
10 months ago |
Jeffrey Morgan
|
34f142797a
llm: always add bos token to prompt (#4941)
|
10 months ago |
Michael Yang
|
829ff87bd1
revert tokenize ffi (#4761)
|
11 months ago |
Michael Yang
|
de781b37c8
rm unused infill
|
11 months ago |
Michael Yang
|
3e21799377
rm unused system prompt
|
11 months ago |
Michael Yang
|
26a00a0410
use ffi for tokenizing/detokenizing
|
11 months ago |
Michael Yang
|
714adb8bd1
bump (#4597)
|
11 months ago |
Daniel Hiltgen
|
b37b496a12
Wire up load progress
|
11 months ago |
Sam
|
e15307fdf4
feat: add support for flash_attn (#4120)
|
11 months ago |
Michael Yang
|
58876091f7
log clean up
|
11 months ago |
Daniel Hiltgen
|
920a4b0794
Merge remote-tracking branch 'upstream/main' into pr3702
|
1 year ago |
Michael Yang
|
44869c59d6
omit prompt and generate settings from final response
|
1 year ago |
jmorganca
|
fcf4d60eee
llm: add back check for empty token cache
|
1 year ago |
Jeffrey Morgan
|
18d9a7e1f1
update llama.cpp submodule to `f364eb6` (#4060)
|
1 year ago |
Daniel Hiltgen
|
23d23409a0
Update llama.cpp (#4036)
|
1 year ago |
ManniX-ITA
|
c942e4a07b
Fixed startup sequence to report model loading
|
1 year ago |
Jeffrey Morgan
|
7c9792a6e0
Support unicode characters in model path (#3681)
|
1 year ago |
Daniel Hiltgen
|
0a0e9f3e0f
Apply 01-cache.diff
|
1 year ago |