ggml-org/ggml

Pull Requests Commits

Remove not needed exit check

YavorGIvanov committed 2 years ago

2024e42a

Add check for end token and mark the stream as finished

YavorGIvanov committed 2 years ago

c2058752

ggerganov committed 2 years ago

Verified 63ab3d61

Fix not generating n_predict tokens and fix warn

YavorGIvanov committed 2 years ago

993d226f

gpt-2-batched : simplify kv cache stuff (#574)

ggerganov committed 2 years ago

Verified 898718c0

Free batch at end of example

YavorGIvanov committed 2 years ago

af6a1d94

gpt-2-batched : fix n_kv heuristic

ggerganov committed 2 years ago

Verified d91540a9

Add batched sample output to README and remove hardcoded model path and prompt

YavorGIvanov committed 2 years ago

5ffcbf44

Move gpt-2 batched to a separate target and cpp file

YavorGIvanov committed 2 years ago

42db4049

Add short way of passing n_parallel argument

YavorGIvanov committed 2 years ago

845f39c7

Make it work for n_parallel=1

YavorGIvanov committed 2 years ago

38a17443

Fix crash on trying to use empty embd

YavorGIvanov committed 2 years ago

761db297

Initial attempt to make gpt2 do parallel decoding

YavorGIvanov committed 2 years ago

ce6139c4

ggml : faster ggml_conv_2d using 2-stage op (#483)

leejet committed 2 years ago

Verified 6549d12f

cuda : int counters for device, fix fprintf warning (#560)

iboB committed 2 years ago

Verified faebeb9e

Disable ggml-alloc assert for CPU version of Sam.cpp if the view doesn't have a buffer (#562)

Yavor Ivanov committed 2 years ago

Verified 50d7203b

sync : llama.cpp (Metal + OpenCL + minor alibi) (#558)

ggerganov committed 2 years ago

Verified 2da43c20

fix MSVC build issues (#557)

slaren committed 2 years ago

Verified f225fedd

ggml-alloc : fix crash when used without ggml-backend (#555)

slaren committed 2 years ago

Verified 8b5c5645

readme : mention Metal could be used for gpt-2 (#553)

paschembri committed 2 years ago

Verified 754ea2bd

ggml backends interface v1 (#547)

slaren committed 2 years ago

Verified fc9e9558

ggml : delete duplicate logging macros (#531)

ita9naiwa committed 2 years ago

Verified 4a771d5a

sync : llama.cpp (training, refactoring) (#548)

ggerganov committed 2 years ago

Verified ef336850

cmake : add OPENCL_LIB to solve problem [cannot resolve external symbol clxxxx ] (#527)

wizardforcel committed 2 years ago

Verified 043251ca

ggml : fix batch for ggml_conv_2d (#528)

skirodev committed 2 years ago

Verified 13e2fa69

ggml : add `GGML_OP_CONV_TRANSPOSE_1D` (#524)

PABannier committed 2 years ago

Verified 76d0aca8

ggml : complete implementation of `GGML_OP_CONV_1D` (#523)

PABannier committed 2 years ago

Verified a706d68e

ci : add whisper test (#525)

ggerganov committed 2 years ago

Verified 8ca2c19a

examples : fix compile warnings

ggerganov committed 2 years ago

Verified 1472fed1

sync : whisper.cpp (Metal + ggml sched_yield fix + reduce ggml-alloc size) (#522)

ggerganov committed 2 years ago

Verified 7ddb66d7

Older