ochafik/llama.cpp

Pull Requests Commits

json: fix zig build

ochafik committed 2 years ago

3af4ac58

json: move json-schema-to-grammar to common lib

ochafik committed 2 years ago

8da85dc2

main: add --json-schema / -j

ochafik committed 2 years ago

1ef7fab7

JSON schema conversion: ⚡️ faster repetitions, min/maxLength for strings, cap number length (#6555)

ochafik committed 2 years ago

Verified ab9a3240

metal : unify mul_mv_id kernels (#6556)

slaren committed 2 years ago

Verified fbbc030b

infill : add download instructions for model (#6626)

danbev committed 2 years ago

Verified 4cc120c7

server : coherent log output for KV cache full (#6637)

phymbert committed 2 years ago

Verified 24ee66ed

llama : add gguf_remove_key + remove split meta during quantize (#6591)

zj040045 committed 2 years ago

Verified 91c73601

chore: Fix markdown warnings (#6625)

reneleonhardt committed 2 years ago

Verified 5c4d767a

imatrix : remove invalid assert (#6632)

ggerganov committed 2 years ago

Verified ef21ce4c

Correct free memory and total memory. (#6630)

MasterYi1024 committed 2 years ago

Verified dee7f8d6

eval-callback: use ggml_op_desc to pretty print unary operator name (#6631)

phymbert committed 2 years ago

Verified 81da18e7

ci : disable Metal for macOS-latest-cmake-x64 (#6628)

ggerganov committed 2 years ago

Verified 9ed2737a

Optimization: eliminate addition of redundant stacks when advancing grammar. (#6616)

HanClinto committed 2 years ago

Verified 04a5ac21

As suggested by @slaren, disabling Metal for test to fix CI build on OSX from #6576 (#6619)

HanClinto committed 2 years ago

Verified f7001ccc

Refactor Error Handling for CUDA (#6575)

nneubacher committed 2 years ago

Verified a474f50e

grammars: 1.5x faster inference w/ complex grammars (vector reserves / reuses) (#6609)

ochafik committed 2 years ago

Verified cbaadc92

ci: download artifacts to release directory (#6612)

Hugi-R committed 2 years ago

Verified 1bbdaf6e

scripts : add --outdir option to hf.sh (#6600)

danbev committed 2 years ago

Verified f4183afe

eval-callback: Example how to use eval callback for debugging (#6576)

phymbert committed 2 years ago

Verified b804b1ef

gguf : add option to not check tensor data (#6582)

danbev committed 2 years ago

Verified 8228b66d

minor layout improvements (#6572)

rsoika committed 2 years ago

Verified b3a96f27

llama : add model types for mixtral (#6589)

slaren committed 2 years ago

Verified 4f407a0a

convert.py : add consolidated.safetensors for mixtral 8x22b (#6587)

slaren committed 2 years ago

Verified 65c64dc3

docs : how to add a model (#6565)

phymbert committed 2 years ago

Verified 67fac4b9

readme : fix ROCm link (#6579)

artem-zinnatullin committed 2 years ago

Verified 29122d32

readme : update UI list (#6560)

ylsdamxssjxxdd committed 2 years ago

Verified b231b37b

readme: fix typo in amdgpu target name (#6573)

Sejsel committed 2 years ago

Verified ba5e134e

BERT tokenizer fixes (#6498)

cebtenzzre committed 2 years ago

Verified 1b67731e

ggerganov committed 2 years ago

Verified c4a3a4ff

Older