ggml-org/llama.cpp

Pull Requests Commits

ggerganov committed 52 days ago

Verified 386892ec

metal : fuse add, mul + add tests (#14596)

ggerganov committed 53 days ago

Verified bf9087f5

graph : fix graph reuse reset of params (#14760)

ggerganov committed 53 days ago

Verified 9fb1042c

parallel : add option for different RNG seeds (#14757)

ggerganov committed 53 days ago

Verified 2adf8d83

cuda : Fix Gemma3n not executed as CUDA_GRAPH on NVGPUs (#14741)

ORippler committed 53 days ago

Verified 021cc28b

graph : avoid huge warm-up graphs for MoE models (#14753)

ggerganov committed 53 days ago

Verified d498af3d

model : fix build after merge conflict (#14754)

ggerganov committed 53 days ago

Verified eacdeb5b

model : add EXAONE 4.0 support (#14630)

lgai-exaone committed 53 days ago

Verified e0cb5c5c

CUDA: set_rows + cpy.cu refactor (#14712)

am17an committed 53 days ago

Verified f9a31eea

graph : refactor context to not pass gf explicitly (#14629)

ggerganov committed 54 days ago

Verified 8f974bc1

graph : Pass the graph placeholder message in debug mode (#14748)

Nexesenex committed 54 days ago

Verified 09651d09

use max work group size for device to replace the magic number (#14732)

NeoZhangJianyu committed 54 days ago

Verified 349ea79f

convert : fix Ernie4.5 MoE without shared experts (#14746)

pwilkin committed 54 days ago

Verified 670e1360

nix : use optionalAttrs for env mkDerivation attrset argument (#14726)

amozeo committed 54 days ago

Verified 760b4484

model: add Ernie 4.5 MoE support (#14658)

pwilkin committed 54 days ago

Verified cb887f1b

kv-cache : fix k-shift for multiple streams (#14742)

ggerganov committed 54 days ago

Verified d6fb3f6b

llama : reuse compute graphs (#14482)

ggerganov committed 54 days ago

Verified 01612b74

llama : fix parallel processing for lfm2 (#14705)

tdakhran committed 54 days ago

Verified 086cf81e

kv-cache : opt mask set input (#14600)

ggerganov committed 54 days ago

Verified d9b69108

batch : fix uninitialized has_cpl flag (#14733)

ggerganov committed 54 days ago

Verified ad57d3ed

ci : disable failing vulkan crossbuilds (#14723)

CISC committed 55 days ago

Verified 1ba45d49

convert : make hf token optional (#14717)

CISC committed 55 days ago

Verified 19e5943d

llama : fix parameter order for hybrid memory initialization (#14725)

dinerburger committed 55 days ago

Verified 496957e1

ggml: Add initial WebGPU backend (#14521)

reeselevine committed 55 days ago

Verified 21c02174

model : support output bias for qwen2 (#14711)

tempstudio committed 55 days ago

Verified b0f0ecc3

llama : add high-throughput mode (#14363)

ggerganov committed 55 days ago

Verified 225e7a14

Support diffusion models: Add Dream 7B (#14644)

am17an committed 55 days ago

Verified ab140198

ggml : add asserts (#14720)

ggerganov committed 55 days ago

Verified 64978340

server : pre-calculate EOG logit biases (#14721)

ggerganov committed 55 days ago

Verified 6ffd4e9c

llama : fix parallel processing for plamo2 (#14716)

mitmul committed 55 days ago

Verified e4841d24

Older