ngxson/llama.cpp

Pull Requests Commits

normalize newlines

CISC committed 8 days ago

Verified ca8d4ca5

check if empty first

CISC committed 8 days ago

Verified 8a88770f

hopefully fix windows

CISC committed 8 days ago

Verified c9a94e77

loosen ensure_vals max count condition, add not impl for map(int)

ngxson committed 8 days ago

e07af2bd

fix array.map()

ngxson committed 8 days ago

25dac2e0

add fuzz func_args, refactor to be safer

ngxson committed 8 days ago

350d87da

CISC committed 8 days ago

Verified 8e1e6ae7

CISC committed 8 days ago

Verified acaf0170

CISC committed 8 days ago

Verified 70d9d9c5

ngxson committed 9 days ago

10a987ae

Merge branch 'master' into xsn/jinja_vm

ngxson committed 9 days ago

78a0112d

fix chat_template_kwargs

ngxson committed 9 days ago

a959ff8b

ngxson committed 9 days ago

abcd7766

ngxson committed 9 days ago

f475f5bc

context : reserve new scheduler when graph topology changes (#18547)

ggerganov committed 9 days ago

Verified 39173bca

CUDA: fix allignment on register spill for FA (#18815)

JohannesGaessler committed 9 days ago

Verified 5c662d21

jinja : static analysis + header cleanup + minor fixes

ggerganov committed 9 days ago

Verified 50568649

ggml-cpu: optimize ggml_vec_dot_bf16 for Power9 (#18837)

shalinib-ibm committed 9 days ago

Verified 8cc0ba95

lora: make sure model keep track of associated adapters (#18490)

ngxson committed 9 days ago

Verified a7e6ddb8

model-loader : support bool array sliding window pattern (#18850)

CISC committed 9 days ago

Verified 2a131801

tests : download models only when running ctest (#18843)

angt committed 9 days ago

Verified ec997b4f

hexagon: support for OP_CPY, host buffers now optional, hvx-utils refactoring and optimizations (#18822)

max-krasnyansky committed 9 days ago

Verified cff777f2

CUDA: Factor out and re-use `block_reduce` function (#18785)

ORippler committed 9 days ago

Verified 36f01324

Restore clip's cb() to its rightful glory - extract common debugging elements in llama (#17914)

pwilkin committed 10 days ago

Verified d98b5481

model : clean up and fix EXAONE-MoE configuration (#18840)

nuxlear committed 10 days ago

Verified 8fb71755

refactor : remove libcurl, use OpenSSL when available (#18828)

angt committed 10 days ago

Verified 516a4ca9

vulkan: Check maxStorageBufferRange in supports_op (#18709)

jeffbolznv committed 10 days ago

Verified 3e4bb296

llama-model: fix unfortunate typo (#18832)

am17an committed 10 days ago

Verified 47f96124

CUDA : fix typo in clang pragma comment [no ci] (#18830)

danbev committed 10 days ago

Verified 01cbdfd7

vulkan: work around Intel fp16 bug in mmq (#18814)

0cc4m committed 10 days ago

Verified 635ef78e

Older