 DannyDaemonic
|
3498588e0f
Add --simple-io option for subprocesses and break out console.h and cpp (#1558)
|
пре 2 година |
 Johannes Gäßler
|
0728c5a8b9
CUDA: mmq CLI option, fixed mmq build issues (#2453)
|
пре 2 година |
 klosax
|
8a88e5855c
perplexity : add Hellaswag calculation (#2389)
|
пре 2 година |
 Kawrakow
|
eb542d3932
Add LLAMA_DEFAULT_RMS_EPS so we can change the default (#2384)
|
пре 2 година |
 Xiao-Yong Jin
|
0c06204fb3
main : add `--in-prefix-bos` to prefix BOS to user inputs; keep EOS (#2304)
|
пре 2 година |
 slaren
|
41c674161f
make rms_norm_eps a parameter (#2374)
|
пре 2 година |
 Evan Jones
|
84e09a7d8b
llama : add grammar-based sampling (#1773)
|
пре 2 година |
 Georgi Gerganov
|
e76d630df1
llama : grouped-query attention + LLaMAv2 70B support (#2276)
|
пре 2 година |
|
klosax
|
b5fe67f8c6
Perplexity: Compute scores correlated to HellaSwag (#2312)
|
пре 2 година |
 Guillaume "Vermeille" Sanchez
|
ab0e26bdfb
llama : remove cfg smooth factor as it is only a reparameterization of the guidance scale (#2280)
|
пре 2 година |
|
Georgi Gerganov
|
d01bccde9f
ci : integrate with ggml-org/ci (#2250)
|
пре 2 година |
|
Xiao-Yong Jin
|
6e7cca4047
llama : add custom RoPE (#2054)
|
пре 2 година |
 Bach Le
|
c9c74b4e3f
llama : add classifier-free guidance (#2135)
|
пре 2 година |
 WangHaoranRobin
|
d7d2e6a0f0
server: add option to output probabilities for completion (#1962)
|
пре 2 година |
 Howard Su
|
b8c8dda75f
Use unsigned for random seed (#2006)
|
пре 2 година |
 zrm
|
b853d45601
ggml : add NUMA support (#1556)
|
пре 2 година |
 Didzis Gosko
|
527b6fba1d
llama : make model stateless and context stateful (llama_state) (#1797)
|
пре 2 година |
|
Johannes Gäßler
|
254a7a7a5f
CUDA full GPU acceleration, KV cache in VRAM (#1827)
|
пре 2 година |
 Kerfuffle
|
fa84c4b3e8
Fix issue where interactive mode crashes when input exceeds ctx size (#1789)
|
пре 2 година |
 Willy Tarreau
|
35a84916fb
main: add the possibility to open the prompt cache read-only (#1640)
|
пре 2 година |
|
Johannes Gäßler
|
17366df842
Multi GPU support, CUDA refactor, CUDA scratch buffer (#1703)
|
пре 2 година |
|
Georgi Gerganov
|
ecb217db4f
llama : Metal inference (#1642)
|
пре 2 година |
 Vladimir Zorin
|
337aea1139
examples : add --alias option to gpt_params to set use friendly model name (#1614)
|
пре 2 година |
|
Georgi Gerganov
|
4b7e245adf
minor : fix compile warnings
|
пре 2 година |
 Stephan Walter
|
dc271c52ed
Remove unused n_parts parameter (#1509)
|
пре 2 година |
 András Salamon
|
9560655409
define default model path once, sync path with readme (#1366)
|
пре 2 година |
|
Johannes Gäßler
|
905d87b70a
ggml : GPU-accelerated token generation (#1412)
|
пре 2 година |
|
Evan Jones
|
cf348a60e0
main : add option to save full output to session (#1338)
|
пре 2 година |
|
DannyDaemonic
|
41654efea8
Interface improvements and `--multiline-input` (previously `--author-mode`) (#1040)
|
пре 2 година |
 44670
|
2edbdb0f99
main : add --in-suffix option (#1318)
|
пре 2 година |
