ggerganov/llama.cpp

Pull Requests Commits

server : pre-calculate EOG logit biases

ggerganov committed 85 days ago

Verified 07908a82

llama : fix parallel processing for plamo2 (#14716)

mitmul committed 85 days ago

Verified e4841d24

server : fix handling of the ignore_eos flag (#14710)

ggerganov committed 85 days ago

Verified 538cc77f

scripts: synthetic prompt mode for server-bench.py (#14695)

JohannesGaessler committed 85 days ago

Verified 5cae7665

convert : only check for tokenizer folder if we need it (#14704)

CISC committed 85 days ago

Verified 4b91d6f7

convert : add pre-computed hashes first to prevent order mishaps (#14701)

CISC committed 85 days ago

Verified cf91f217

llama: add LLAMA_API to deprecated llama_kv_self_seq_div (#14708)

Min-Hua committed 85 days ago

Verified 79e0b68c

gguf-py : dump bpw per layer and model in markdown mode (#14703)

EAddario committed 86 days ago

Verified c81f4192

model : add Kimi-K2 support (#14654)

gabriellarson committed 86 days ago

Verified 4a4f4269

vulkan: fix noncontig check for mat_mul_id splitting (#14683)

jeffbolznv committed 86 days ago

Verified ba1ceb34

vulkan: add RTE variants for glu/add/sub/mul/div (#14653)

jeffbolznv committed 86 days ago

Verified 10a0351a

model : add PLaMo-2 support (#14560)

mitmul committed 86 days ago

Verified 68e37a61

cuda: fix build warnings in set-rows.cu (unused variable) (#14687)

yeahdongcn committed 86 days ago

Verified cbc68be5

sycl: Hotfix for non dnnl codepath (#14677)

ShanoToni committed 87 days ago

Verified bdca3837

ggml : refactor llamafile_sgemm PPC code (#14673)

shalinib-ibm committed 87 days ago

Verified 55c509da

llama-context: add ability to get logits (#14672)

am17an committed 87 days ago

Verified 9c9e4fc6

scripts: benchmark for HTTP server throughput (#14668)

JohannesGaessler committed 87 days ago

Verified 494c5899

SYCL: use 1D kernel for set_rows (#14618)

qnixsynapse committed 87 days ago

Verified 0f4c6ec0

sycl: Batched mulmat rework for oneDNN dispatch (#14617)

ShanoToni committed 87 days ago

Verified 65a3ebb0

llama : add jinja template for rwkv-world (#14665)

MollySophia committed 87 days ago

Verified 0d922676

quantize : fix minor logic flaw in --tensor-type (#14572)

EAddario committed 88 days ago

Verified 982e3472

cuda : add set rows for bf16 (#14664)

CISC committed 88 days ago

Verified 923e3ea2

cuda : add ELU support (#14657)

YavorGIvanov committed 88 days ago

Verified e743cddb

ggml : add build-time message to remind about ggml_set_rows (#14661)

ggerganov committed 88 days ago

Verified 05fec5bd

metal : Add missing unary ops Metal support (#14660)

YavorGIvanov committed 88 days ago

Verified dcf7f2ea

cmake : Add CMake presets for Linux and GCC (#14656)

YavorGIvanov committed 88 days ago

Verified 84b396e0

tests : cover lfm2 cases in test_ssm_conv (#14651)

tdakhran committed 89 days ago

Verified c31e6064

docs : add LFM2 to models section (#14650)

tdakhran committed 89 days ago

Verified 67eade1b

CUDA: add set rows for f32 and f16 (#14551)

am17an committed 89 days ago

Verified 7de5c7ca

ggerganov committed 89 days ago

8eff9554

Older