ggerganov/llama.cpp

Pull Requests Commits
server : do not apply Markdown formatting in code sections (#6850)
mgroeber9110 committed 1 year ago
Verified 3fe847b5
common : revert showing control tokens by default for server (#6860)
K-Mistele committed 1 year ago
Verified 37246b10
Server: fix seed for multiple slots (#6835)
JohannesGaessler committed 1 year ago
Verified 28103f48
ggml : move 32-bit arm compat in ggml-impl.h (#6865)
ggerganov committed 1 year ago
Verified c0d1b3e0
llama : add phi 3 chat template (#6857)
tristandruyen committed 1 year ago
Verified abd33140
convert : add support of codeqwen due to tokenizer (#6707)
JustinLin610 committed 1 year ago
Verified 3fec68be
llama : add phi3 support (#6852)
liuwei-git committed 1 year ago
Verified c8297c6a
server : add help for --flash-attn arg
ggerganov committed 1 year ago
Verified 751591d5
cont
ggerganov committed 1 year ago
Verified d228bf85
llama : fix n_batch requirements
ggerganov committed 1 year ago
Verified 56657e52
llama : prep ALiBi support for BERT models
ggerganov committed 1 year ago
Verified 19e8982f
llama : replace bool need_kq_pos with use_alibi
ggerganov committed 1 year ago
Verified 78d363b0
ggml : add TODO's for F16/F32 mask/pos support in other backends
ggerganov committed 1 year ago
Verified 3864eea4
cuda : try to fix __hgt2_mask
ggerganov committed 1 year ago
Verified c1293697
[SYCL] Windows default build instructions without -DLLAMA_SYCL_F16 flag activated (#6767)
aahouzi committed 1 year ago
Verified 4e96a812
llamafile : improve sgemm.cpp (#6796)
jart committed 1 year ago
Verified 192090ba
cuda : "constexpr dim3" -> "const dim3"
ggerganov committed 1 year ago
Verified c70bfd7b
cuda : uint -> uint32_t
ggerganov committed 1 year ago
Verified 5408d555
ggml : fix calloc argument ordering. (#6820)
airlied committed 1 year ago
Verified e931888d
llama : fix typo in <|im_end|> token text (#6745)
ggerganov committed 1 year ago
Verified 8960fe86
ggml : ggml_soft_max support F16/F32 mask/pos
ggerganov committed 1 year ago
Verified f725ca90
ci: fix job are cancelling each other (#6781)
phymbert committed 1 year ago
Verified c0956b09
flake.lock: Update
github-actions[bot] committed 1 year ago
e9b4a1bf
llama : force disable flash attention for incompatible models
ggerganov committed 1 year ago
Verified c11d05fe
ggml : fix num dimensions in ggml_flash_attn_ext
ggerganov committed 1 year ago
Verified cb76d747
common : print --flash-attn in help
ggerganov committed 1 year ago
Verified a39217d4
`build`: generate hex dump of server assets during build (#6661)
ochafik committed 1 year ago
Verified 5cf5e7d4
llama : add option to render special/control tokens (#6807)
ggerganov committed 1 year ago
Verified 40f74e4d
ggml : fix ggml_backend_cpu_supports_op() for CPY (#0)
ggerganov committed 1 year ago
Verified b9cc76d8
llama : add llama-3 chat template (#6751)
DifferentialityDevelopment committed 1 year ago
Verified 7dbdba56
Terms Of Use Privacy Policy Refund Policy Impressum