build : use pkg-config for OpenBLAS (#1778)
d5d466c
unverified
Przemysław Pawełczyk
commited on
Spaces:
Running
Running
Commit History
main : add command-style grammar (#1998)
7e6ea10
unverified
make : add grammar parser to common objects
b1f3938
unverified
sync : ggml (#2001)
cbbfa9e
unverified
whisper : improve handling of prompts (#1981)
15949a9
unverified
whisper : improve support for distil-large-v3 (#1982)
749004e
unverified
Sanchit Gandhi
commited on
ruby : fix build (#1980)
20374d7
unverified
docker : libcuda.so.1 in PATH (#1966)
2cd0d06
unverified
Tiago Fassoni
commited on
readme : add Fedora dependencies (#1970)
afc6b1a
unverified
Mohammadreza Hendiani
commited on
whisper : token-level timestamps with DTW (#1485)
ce7ca09
unverified
examples : rename --audio-context to --audio-ctx per help text (#1953)
8e9c985
unverified
whisper : set outputs from conv graph (#1959)
62505d4
unverified
alloc : fix allocation data of pre-allocated leafs
0c378f2
unverified
slaren
commited on
cmake : copy ggml-common.h to bin
f592e46
unverified
gitignore : .vimspector.json
b593d9a
unverified
talk-llama : sync llama.cpp
14e824b
unverified
sync : ggml
1701a5d
unverified
update examples and tests
007ebd7
unverified
slaren
commited on
ggml : add ggml-common.h
8cdfa17
unverified
ggml : designate enum vals for integer types (llama/6050)
0bd0c7a
unverified
metal : build metallib + fix embed path (llama/6015)
27311ef
unverified
Update get version (llama/6025)
9a4e508
unverified
AidanBeltonS
commited on
ggml : reuse quantum structs across backends (llama/5943)
bb0625f
unverified
ggml : fix UB in IQ2_S and IQ3_S (llama/6012)
0c552df
unverified
sycl : update IQ1_S kernels (WIP - not working!) (llama/5995)
16dc72c
unverified
1.5 bit: we can do even better (llama/5999)
36cc71e
unverified
ggml, ci : Windows ARM runner and build fixes (llama/5979)
507b9dd
unverified
Michael Podvitskiy
commited on
Better 1.5 bit quantization (llama/5971)
f3a62cc
unverified
Add q3_s and q1_s (llama/5886)
2957823
unverified
Abhilash Majumder
commited on
metal : move mm_id indices to shared mem (llama/5982)
1350705
unverified
ggml : fix unnecessary f32 -> f16 -> f32 casts (mmla) (llama/5951)
cb8bbaa
unverified
ggml : remove old quantization functions (llama/5942)
11a2545
unverified
ggml : add ggml-common.h to deduplicate shared code (llama/5940)
0a37735
unverified
llama : support Mamba Selective State Space Models (llama/5328)
224fbc2
unverified
compilade
commited on
extra : update sync scripts after ggml-common.h
2e29431
unverified
whisper : document whisper_batch.n_seq_id (#1942)
f08549e
unverified
whisper : improve beam search candidate diversity (#1947)
6e9276c
unverified
bindings/go : add linker flags to make metal work (#1944)
3dee0de
unverified
whisper : make beam candidate sort more stable (#1943)
1316242
unverified
ggml : try fix 32-bit arm compat (#1938)
6ea3354
unverified
talk-llama : use llama_decode instead of llama_eval
301b000
unverified
talk-llama : sync llama.cpp
fe602cb
unverified
talk-llama : sync llama.cpp
802496b
unverified
sync : ggml
bad33a1
unverified
Revert "[SYCL] fix error when set main gpu to non-zero (llama/5901)" (llama/5918)
d7e8525
unverified
Neo Zhang Jianyu
commited on
fix error when set main gpu to non-zero (llama/5901)
829c347
unverified
Neo Zhang Jianyu
commited on
ggml : use SYS_get_cpu if SYS_getcpu is not defined (llama/5906)
909dbdc
unverified
ggml : use `uint8x16_t` return type for `ggml_vqtbl1q_u8` (llama/5894)
7f5bc53
unverified
bobqianic
commited on
add wait() to make code stable (llama/5895)
41c3c12
unverified
Neo Zhang Jianyu
commited on