ggerganov/llama.cpp

Pull Requests Commits

llama : cast seq_id in comparison with unsigned n_seq_max

compilade committed 295 days ago

9cddd9ae

llama : more graceful error handling of invalid session files

compilade committed 295 days ago

ffd5117d

llama : fix session file loading

compilade committed 296 days ago

83e6a17d

llama : remove _context suffix for llama_data_context

compilade committed 296 days ago

c8b424fa

llama : various integer type cast and format string fixes

compilade committed 297 days ago

cddc899b

llama : fix uint64_t format type

compilade committed 297 days ago

9e22064a

llama : refactor session file management

compilade committed 297 days ago

8e39037b

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

ngxson committed 299 days ago

Verified de280085

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

0x4139 committed 300 days ago

Verified b841d074

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

joeatodd committed 300 days ago

Verified 64cf50a0

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

ggerganov committed 300 days ago

Verified 938943cd

Vulkan IQ4_NL Support (#8613)

0cc4m committed 300 days ago

Verified 751fcfc6

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

jeroen-mostert committed 300 days ago

Verified 46e47417

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

ggerganov committed 300 days ago

Verified e7e6487b

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

luoyu-intel committed 300 days ago

Verified 063d99ad

llama : fix codeshell support (#8599)

hankeke303 committed 301 days ago

Verified 081fe431

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

Stillerman committed 301 days ago

Verified d94c6e0c

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

jpodivin committed 301 days ago

Verified 566daa5a

llama : allow overrides for tokenizer flags (#8614)

ggerganov committed 301 days ago

Verified 6f11a83e

tests : re-enable tokenizer tests (#8611)

ggerganov committed 301 days ago

Verified e093dd23

llama : add Mistral Nemo inference support (#8604)

iamlemec committed 301 days ago

Verified 50e05353

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

kaetemi committed 301 days ago

Verified 62815449

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

zqb-all committed 301 days ago

Verified 04bab6b7

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

devojony committed 301 days ago

Verified b7c11d36

flake.lock: Update (#8610)

ggerganov committed 302 days ago

Verified 45f2c19c

examples : Rewrite pydantic_models_to_grammar_examples.py (#8493)

maruel committed 302 days ago

Verified 22f281aa

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

compilade committed 302 days ago

Verified 328884f4

convert_hf : fix Gemma v1 conversion (#8597)

compilade committed 302 days ago

Verified c69c6303

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

JohannesGaessler committed 302 days ago

Verified 69c487f4

gguf : handle null name during init (#8587)

ggerganov committed 303 days ago

Verified 07283b1a

Older