ggerganov/llama.cpp

Pull Requests Commits
cuda : fix const ptrs warning causing ROCm build issues (#3913)
ggerganov committed 2 years ago
Verified c7743fe1
cuda : use CUDA memory pool with async memory allocation/deallocation when available (#3903)
young-developer committed 2 years ago
Verified d6069051
gguf : print error for GGUFv1 files (#3908)
ggerganov committed 2 years ago
Verified 4ff1046d
cmake : disable LLAMA_NATIVE by default (#3906)
slaren committed 2 years ago
Verified 21958bb3
gguf : remove special-case code for GGUFv1 (#3901)
ggerganov committed 2 years ago
Verified 2756c4fb
llm : prevent from 1-D tensors being GPU split (#3697)
ggerganov committed 2 years ago
Verified 1efae9b7
build : link against build info instead of compiling against it (#3879)
cebtenzzre committed 2 years ago
Verified b12fa0d1
cuda : check if this fixes Pascal card regression (#3882)
ggerganov committed 2 years ago
Verified 4d719a6d
metal : fix build errors and kernel sig after #2268 (#3898)
ggerganov committed 2 years ago
Verified 183b3fac
cuda : fix RoPE after #2268 (#3897)
cebtenzzre committed 2 years ago
Verified 2fffa0d6
llama : fix llama_context_default_params after #2268 (#3893)
cebtenzzre committed 2 years ago
Verified 0eb332a1
ggml-cuda : compute ptrs for cublasGemmBatchedEx in a kernel (#3891)
slaren committed 2 years ago
Verified d02e98cd
llama : implement YaRN RoPE scaling (#2268)
cebtenzzre committed 2 years ago
Verified 898aeca9
llm : fix llm_build_kqv taking unused tensor (benign, #3837)
ggerganov committed 2 years ago
Verified c43c2da8
llm : fix falcon norm after refactoring (#3837)
ggerganov committed 2 years ago
Verified 523e49b1
metal : multi-simd softmax (#3710)
ggerganov committed 2 years ago
Verified e16b9fa4
common : minor (#3715)
ggerganov committed 2 years ago
Verified ff8f9a88
llm : add llm_build_context (#3881)
ggerganov committed 2 years ago
Verified 50337961
common : allow caller to handle help/argument exceptions (#3715)
bandoti committed 2 years ago
Verified 0e40806c
log : make generating separate log files optional (#3787)
staviq committed 2 years ago
Verified a2758d08
sampling : null grammar field after reset (#3885)
l3utterfly committed 2 years ago
Verified e75dfdd3
ggml : fix UNUSED macro (#3762)
ggerganov committed 2 years ago
Verified 9a3b4f6c
finetune : add -ngl parameter (#3762)
AndrewGodfrey committed 2 years ago
Verified 73bdcb39
scripts : add server-llm.sh (#3868)
ggerganov committed 2 years ago
Verified f0e20932
server : re-enable completion and embedded at the same time (#3876)
a-h committed 2 years ago
Verified ca190bca
llama : refactor graph build code (#3837)
ggerganov committed 2 years ago
Verified 71e3718a
samplers : Min-P sampler implementation [alternative to Top P/Top K] (#3841)
kalomaze committed 2 years ago
Verified 238657db
flake.nix: fix for rocm 5.7 (#3853)
Tungsten842 committed 2 years ago
Verified 07178c98
ggml : move FP16 <-> FP32 code to ggml-impl.h (#3861)
ggerganov committed 2 years ago
Verified 207b5190
Extend llama_kv_cache_seq_rm to allow matching any sequence (#3843)
KerfuffleV2 committed 2 years ago
Verified 6e08281e
Terms Of Use Privacy Policy Refund Policy Impressum