ggml-org/llama.cpp

Pull Requests Commits

ggml-alloc : fix assert in debug builds

slaren committed 1 year ago

ee745692

refact : fix convert script + zero out KV cache to avoid nans (#3523)

ggerganov committed 1 year ago

Verified fcca0a70

metal : do not use mul_mm kernels when ne00 < 64 (#3542)

ggerganov committed 1 year ago

Verified dcc09d25

sync : ggml (ggml-backend) (#3548)

ggerganov committed 1 year ago

Verified db3abcc1

ci : add Zig CI/CD and fix build (#2996)

kassane committed 1 year ago

Verified eee42c67

api_like_OAI.py : compat with Microsoft Guidance (#2746)

ryderwishart committed 1 year ago

Verified 8e6716a1

api_like_OAI.py : simplify function (#2796)

arcrank committed 1 year ago

Verified 9c38d181

k-quants : fix comments about block sizing (#3499)

jrudolph committed 1 year ago

Verified a1202a31

ci : enable on obj-c changes + fix metal build (#3540)

ggerganov committed 1 year ago

Verified 94e502df

zig : fix build by introducing train.cpp (#3539)

robertluo committed 1 year ago

Verified 7d8b2493

metal : support MTLGPUFamily < Apple7, formatting, style (#3524)

ggerganov committed 1 year ago

Verified b0ec5218

llama : fix missing break in Persimmon arch case statements (#3535)

KerfuffleV2 committed 1 year ago

Verified 63d3b06a

Fix trying to strip newline from empty prompt and cfg prompt file content (#3534)

KerfuffleV2 committed 1 year ago

Verified a16e89ce

gguf.py : fix CI for publishing GGUF package (#3532)

monatis committed 1 year ago

Verified 4d038332

py : change version of numpy requirement to 1.24.4 (#3515)

lyjia committed 1 year ago

Verified c47066d8

quantize : fail fast on write errors (#3521)

cebtenzzre committed 1 year ago

Verified f1782c68

metal : support default.metallib load & reuse code for swift package (#3522)

jhen0409 committed 1 year ago

Verified c26765a0

llm : support Adept Persimmon 8B (#3410)

phillip-kravtsov committed 1 year ago

Verified 0e797c2f

Fix for #3454 (#3455)

goerch committed 1 year ago

Verified 3a716b4d

readme : update models, cuda + ppl instructions (#3510)

BarfingLemurs committed 1 year ago

Verified 1faaae8c

server : docs fix default values and add n_probs (#3506)

Mihaiii committed 1 year ago

Verified cb13d73a

kv cache slot search improvements (#3493)

KerfuffleV2 committed 1 year ago

Verified 9ca79d5c

prompts : fix editorconfig checks after #3416

ggerganov committed 1 year ago

Verified 0c731ca4

parallel : add option to load external prompt file (#3416)

pudepiedj committed 1 year ago

Verified a8777ad8

server : reuse llama_sample_token common util (#3494)

jhen0409 committed 1 year ago

Verified 97af49fa

llama : correct hparams comparison (#3446)

l3utterfly committed 1 year ago

Verified 16820a5a

ci : fix xcodebuild destinations (#3491)

jhen0409 committed 1 year ago

Verified 04b2f438

convert : update Falcon script for new HF config (#3448)

cebtenzzre committed 1 year ago

Verified 48edda30

build : use std::make_tuple() for compatibility with older GCC versions (#3488)

kenvix committed 1 year ago

Verified 45eba936

common : process escape sequences in reverse prompts (#3461)

staviq committed 1 year ago

Verified acec9eaa

Older