ggml-org/llama.cpp

Pull Requests Commits

Merge branch 'cuda-cublas-opts' into gg/phi-2

ggerganov committed 2 years ago

Verified d2f1e0da

ggml : fix NeoX rope to rotate just first n_dims

ggerganov committed 2 years ago

Verified b672c169

Merge branch 'master' into cuda-cublas-opts

ggerganov committed 2 years ago

Verified e75889a9

lora : add support for non-llama models (#3333)

slaren committed 2 years ago

Verified c6c4fc08

phi-2 : scale Q instead of KQ for better precision

ggerganov committed 2 years ago

Verified 0644c3be

convert : revert "added_tokens_decoder" change

ggerganov committed 2 years ago

Verified 0b6ffa58

convert : phi don't add BOS token

ggerganov committed 2 years ago

Verified a878be4c

llama : fix meta KV override bug

ggerganov committed 2 years ago

Verified 5469d82d

py : whitespaces

ggerganov committed 2 years ago

Verified 7500fa2f

phi-2 : use layer norm eps

ggerganov committed 2 years ago

Verified aa5c881a

phi-2 : various fixes

ggerganov committed 2 years ago

Verified a2a3d2c8

llama : sanity checks for access to logits (#4274)

cebtenzzre committed 2 years ago

Verified 8a5be3bd

fix breaking change

eabraham-1 committed 2 years ago

e2076553

phi2 implementation

eabraham-1 committed 2 years ago

12cc80cb

server : add optional API Key Authentication example (#4441)

ShadovvBeast committed 2 years ago

Verified 88ae8952

ggml : group mul_mat_id rows by matrix (cpu only) (#4480)

slaren committed 2 years ago

Verified ee4725a6

ggml : use ggml_row_size where possible (#4472)

slaren committed 2 years ago

Verified 6744dbe9

ggml : remove n_dims from ggml_tensor (#4469)

slaren committed 2 years ago

Verified cafcd4f8

py : add protobuf dependency (#4466)

strutive07 committed 2 years ago

Verified c50e4001

ggml : add ggml_row_size() (fixes llama out of space) (#4461)

LostRuins committed 2 years ago

Verified 20a68a70

ggml : fix OpenCL broadcast requirement for ggml_mul (close #4453)

ggerganov committed 2 years ago

Verified 55e87c37

convert : support loading vocab from fast tokenizer config (#3633)

strutive07 committed 2 years ago

Verified 873637af

readme : update supported model list (#4457)

BarfingLemurs committed 2 years ago

Verified 0353a184

server : fix handling of characters that span multiple tokens when streaming (#4446)

shibe2 committed 2 years ago

Verified 948ff137

sync : ggml (SD ops, tests, kernels) (#4444)

ggerganov committed 2 years ago

Verified 4d98d9a6

build : detect host compiler and cuda compiler separately (#4414)

cebtenzzre committed 2 years ago

Verified 70f806b8

common : add `--version` option to show build info in CLI (#4433)

yusiwen committed 2 years ago

Verified 9fb13f95

readme : update hot topics

ggerganov committed 2 years ago

Verified 113f9942

llama : add Mixtral support (#4406)

slaren committed 2 years ago

Verified 799a1cb1

server : tweak default sampling parameters (#4367)

kalomaze committed 2 years ago

Verified fecac456

Older