ngxson/llama.cpp

Pull Requests Commits

fix small typo for export-lora

ngxson committed 1 year ago

38c7bf27

update help message

ngxson committed 1 year ago

3c965b4f

ngxson committed 1 year ago

cedacf00

examples : remove finetune and train-text-from-scratch

ngxson committed 1 year ago

16c889bc

readme : update UI list [no ci] (#8505)

SommerEngineering committed 1 year ago

Verified 3a7ac530

llama : fix `llama_chat_format_single` for mistral (#8657)

ngxson committed 1 year ago

Verified 96952e71

Re-add erroneously removed -fsycl from GGML_EXTRA_LIBS (#8667)

joeatodd committed 1 year ago

Verified 79167d9e

add llama_lora_adapter_clear (#8653)

ngxson committed 1 year ago

Verified b115105f

examples : Fix `llama-export-lora` example (#8607)

ngxson committed 1 year ago

Verified de280085

server : fix URL.parse in the UI (#8646)

0x4139 committed 1 year ago

Verified b841d074

sycl : Add support for non-release DPC++ & oneMKL (#8644)

joeatodd committed 1 year ago

Verified 64cf50a0

llama : move vocab, grammar and sampling into separate files (#8508)

ggerganov committed 1 year ago

Verified 938943cd

Vulkan IQ4_NL Support (#8613)

0cc4m committed 1 year ago

Verified 751fcfc6

Allow all RDNA2 archs to use sdot4 intrinsic (#8629)

jeroen-mostert committed 1 year ago

Verified 46e47417

contrib : clarify PR squashing + module names (#8630)

ggerganov committed 1 year ago

Verified e7e6487b

[SYCL] fix scratch size of softmax (#8642)

luoyu-intel committed 1 year ago

Verified 063d99ad

llama : fix codeshell support (#8599)

hankeke303 committed 1 year ago

Verified 081fe431

llama : add support for SmolLm pre-tokenizer (#8609)

Stillerman committed 1 year ago

Verified d94c6e0c

*.py: Stylistic adjustments for python (#8233)

jpodivin committed 1 year ago

Verified 566daa5a

llama : allow overrides for tokenizer flags (#8614)

ggerganov committed 1 year ago

Verified 6f11a83e

tests : re-enable tokenizer tests (#8611)

ggerganov committed 1 year ago

Verified e093dd23

llama : add Mistral Nemo inference support (#8604)

iamlemec committed 1 year ago

Verified 50e05353

server : update doc to clarify n_keep when there is bos token (#8619)

kaetemi committed 1 year ago

Verified 62815449

ggml: fix compile error for RISC-V (#8623)

zqb-all committed 1 year ago

Verified 04bab6b7

examples: fix android example cannot be generated continuously (#8621)

devojony committed 1 year ago

Verified b7c11d36

flake.lock: Update (#8610)

ggerganov committed 1 year ago

Verified 45f2c19c

examples : Rewrite pydantic_models_to_grammar_examples.py (#8493)

maruel committed 1 year ago

Verified 22f281aa

gguf-py : fix some metadata name extraction edge cases (#8591)

compilade committed 1 year ago

Verified 328884f4

convert_hf : fix Gemma v1 conversion (#8597)

compilade committed 1 year ago

Verified c69c6303

CUDA: MMQ code deduplication + iquant support (#8495)

JohannesGaessler committed 1 year ago

Verified 69c487f4

Older