ggerganov/llama.cpp

Pull Requests Commits
Test-based VRAM scratch size + context adjustment (#2056)
JohannesGaessler committed 2 years ago
Verified befb3a35
cmake : don't force -mcpu=native on aarch64 (#2063)
dsd committed 2 years ago
Verified b2132270
metal : release buffers when freeing metal context (#2062)
apage43 committed 2 years ago
Verified 2f8cd979
convert : add support of baichuan-7b (#2055)
foldl committed 2 years ago
Verified 471aab6e
llama : fix return value of llama_load_session_file_internal (#2022)
ggerganov committed 2 years ago
Verified 463f2f4c
llama : catch llama_load_session_file_internal exceptions (#2022)
randxie committed 2 years ago
Verified cb44dbc7
embd-input : fix returning ptr to temporary
ggerganov committed 2 years ago
Verified 79f634a1
train : fix compile warning
ggerganov committed 2 years ago
Verified 04606a15
ggml : disable GGML_TASK_INIT and GGML_TASK_FINALIZE by default (#1995)
mqy committed 2 years ago
Verified b1ca8f36
Use unsigned for random seed (#2006)
howard0su committed 2 years ago
Verified b8c8dda7
Porting the improved K-Quant CUDA kernels to OpenCL (#1966)
LostRuins committed 2 years ago
Verified 96a712ca
llama : replacing auto &kv with const auto &kv (#2041)
mendax0110 committed 2 years ago
Verified d3494bb8
cuda : remove nchannels_x argument from mul_mat_vec_nc_f16_f32 (#2028)
set-soft committed 2 years ago
Verified 5b351e94
cuda : fix missing const qualifier in casts (#2027)
set-soft committed 2 years ago
Verified 6432aabb
llama : remove shards weight file support (#2000)
howard0su committed 2 years ago
Verified b922bc35
CUDA GPU acceleration for LoRAs + f16 models (#1970)
JohannesGaessler committed 2 years ago
Verified 7f9753fa
llama : support input embeddings directly (#1910)
ningshanwutuobang committed 2 years ago
Verified cfa0750b
fix pthreads setaffinity usage on android (#2020)
Green-Sky committed 2 years ago
Verified 9d23589d
baby-llama : fix build after ggml_rope change (#2016)
howard0su committed 2 years ago
Verified 0be54f75
llama : fix rope usage after ChatGLM change
ggerganov committed 2 years ago
Verified 181e8d97
ggml : add support for ChatGLM RoPE
ggerganov committed 2 years ago
Verified d9779021
readme : add Scala 3 bindings repo (#2010)
donderom committed 2 years ago
Verified d38e4515
ggml : increase max tensor name + clean up compiler warnings in train-text (#1988)
Davidy22 committed 2 years ago
Verified eaa6ca5a
readme : LD_LIBRARY_PATH complement for some Android devices when building with CLBlast inside Termux (#2007)
gustrd committed 2 years ago
Verified aa777abb
ggml : avoid conv 2d kernel round up
ggerganov committed 2 years ago
Verified c824d2e3
ggml : add NUMA support (#1556)
zrm committed 2 years ago
Verified b853d456
k-quants : fix indentation
ggerganov committed 2 years ago
Verified 9225baef
tests : fix quantize perf (#1990)
katsu560 committed 2 years ago
Verified a84ab1da
k-quants : add AVX support to dot functions (#1916)
katsu560 committed 2 years ago
Verified 5743ca80
readme : add link to new k-quants for visibility
ggerganov committed 2 years ago
Verified 412c60e4
Terms Of Use Privacy Policy Refund Policy Impressum