 dm4
|
ea3b0590ee
embedding : free the batch after execution (#7297)
|
hai 1 ano |
 Georgi Gerganov
|
29499bb593
sync : ggml
|
hai 1 ano |
 John Balis
|
48aa8fd1f2
ggml : add `ggml_upscale_ext` (ggml/814)
|
hai 1 ano |
 Johannes Gäßler
|
583fd6b000
server bench: fix bench not waiting for model load (#7284)
|
hai 1 ano |
|
Georgi Gerganov
|
9f773486ab
script : sync ggml-rpc
|
hai 1 ano |
|
Georgi Gerganov
|
e8a7fd4fb0
metal : support FA without mask + add asserts (#7278)
|
hai 1 ano |
|
Georgi Gerganov
|
a5e3fde857
sync : ggml
|
hai 1 ano |
|
Georgi Gerganov
|
f308ea7059
metal : tune soft_max number of threads (whisper/0)
|
hai 1 ano |
|
Georgi Gerganov
|
c3c88f296a
ggml : try fix ppc64 (whisper/0)
|
hai 1 ano |
 Przemysław Pawełczyk
|
182adefcf3
ggml : expose SSE3 and SSSE3 for MSVC when AVX is available (whisper/2128)
|
hai 1 ano |
 Hong Bo PENG
|
0d26d8ccd8
ggml : optimize for ppc64le using VSX intrinsics (ggml/784)
|
hai 1 ano |
 Steve Grubb
|
4f0263633b
server: free sampling contexts on exit (#7264)
|
hai 1 ano |
 Brian
|
1265c670fd
Revert "move ndk code to a new library (#6951)" (#7282)
|
hai 1 ano |
 Radoslav Gerganov
|
5e31828d3e
ggml : add RPC backend (#6829)
|
hai 1 ano |
 slaren
|
541600201e
llama : disable pipeline parallelism with nkvo (#7265)
|
hai 1 ano |
 Elton Kola
|
efc8f767c8
move ndk code to a new library (#6951)
|
hai 1 ano |
 Haggai Nuchi
|
e0f556186b
Add left recursion check: quit early instead of going into an infinite loop (#7083)
|
hai 1 ano |
 Ryuei
|
27f65d6267
docs: Fix typo and update description for --embeddings flag (#7026)
|
hai 1 ano |
 compilade
|
ee52225067
convert-hf : support direct Q8_0 conversion (#7234)
|
hai 1 ano |
|
Georgi Gerganov
|
614d3b914e
llama : less KV padding when FA is off (#7257)
|
hai 1 ano |
 k.h.lai
|
30e70334f7
llava-cli: fix base64 prompt (#7248)
|
hai 1 ano |
|
Johannes Gäßler
|
1c570d8bee
perplexity: add BF16 vs. FP16 results (#7150)
|
hai 1 ano |
 Neo Zhang
|
948f4ec7c5
[SYCL] rm wait() (#7233)
|
hai 1 ano |
 Joan Fontanals
|
9aa672490c
llama : rename jina tokenizers to v2 (#7249)
|
hai 1 ano |
|
Brian
|
b1f8af1886
convert.py: Outfile default name change and additional metadata support (#4858)
|
hai 1 ano |
 Benjamin Findley
|
e586ee4259
change default temperature of OAI compat API from 0 to 1 (#7226)
|
hai 1 ano |
|
Neo Zhang
|
cbf75894d2
[SYCL] Add oneapi runtime dll files to win release package (#7241)
|
hai 1 ano |
|
Neo Zhang
|
0d5cef78ae
[SYCL] update CI with oneapi 2024.1 (#7235)
|
hai 1 ano |
|
Johannes Gäßler
|
dc685be466
CUDA: add FP32 FlashAttention vector kernel (#7188)
|
hai 1 ano |
|
Georgi Gerganov
|
6f1b63606f
cmake : fix version cmp (#7227)
|
hai 1 ano |
