ngxson/llama.cpp

Pull Requests Commits
ggml webgpu: support for backend sampling (#18880)
reeselevine committed 15 days ago
Verified a89002f0
ggml : extend ggml_pool_1d + metal (#16429)
ThoreKoritzius committed 16 days ago
Verified 388ce822
docs : update ops.md for CANN backend (#18654)
hipudding committed 16 days ago
Verified 6ba6a3c7
ggml-blas: hide warnings from included BLAS headers (#18818)
DaAwesomeP committed 16 days ago
Verified 0802d4cf
mtmd : Fix ASR for LFM2.5-Audio-1.5B (#18876)
tdakhran committed 16 days ago
Verified c945aaae
common : implement new jinja template engine (#18462)
ngxson committed 16 days ago
Verified c15395f7
Setting mmap and direct_io to false as default in llama-bench.cpp (#18841)
JTischbein committed 16 days ago
Verified aa1dc377
CANN: Remove unused `ggml_cann_get_device` function (#18625)
rauletorresc committed 16 days ago
Verified 4ea2eaac
CANN: fix an issue where get_env was not fully renamed (#18796)
noemotiovon committed 16 days ago
Verified e20fa27a
CANN: support gated linear attn (#18653)
hipudding committed 16 days ago
Verified baa4ba0a
OpenCL: add SOLVE_TRI op support (#18846)
shaofeiqi committed 17 days ago
Verified 785a7100
cuda : print less debug logs when disabling cuda graphs (#18868)
ggerganov committed 17 days ago
Verified 6e7fc8a1
context : do not reserve scheduler for warmups (#18867)
ggerganov committed 17 days ago
Verified be8e3d95
llama : add adaptive-p sampler (#17927)
ddh0 committed 17 days ago
Verified 13f1e4a9
server: improve slots scheduling for n_cmpl (#18789)
ngxson committed 17 days ago
Verified a04c2b06
context : reserve new scheduler when graph topology changes (#18547)
ggerganov committed 17 days ago
Verified 39173bca
CUDA: fix allignment on register spill for FA (#18815)
JohannesGaessler committed 17 days ago
Verified 5c662d21
ggml-cpu: optimize ggml_vec_dot_bf16 for Power9 (#18837)
shalinib-ibm committed 17 days ago
Verified 8cc0ba95
lora: make sure model keep track of associated adapters (#18490)
ngxson committed 17 days ago
Verified a7e6ddb8
model-loader : support bool array sliding window pattern (#18850)
CISC committed 17 days ago
Verified 2a131801
tests : download models only when running ctest (#18843)
angt committed 17 days ago
Verified ec997b4f
hexagon: support for OP_CPY, host buffers now optional, hvx-utils refactoring and optimizations (#18822)
max-krasnyansky committed 17 days ago
Verified cff777f2
CUDA: Factor out and re-use `block_reduce` function (#18785)
ORippler committed 17 days ago
Verified 36f01324
Restore clip's cb() to its rightful glory - extract common debugging elements in llama (#17914)
pwilkin committed 18 days ago
Verified d98b5481
model : clean up and fix EXAONE-MoE configuration (#18840)
nuxlear committed 18 days ago
Verified 8fb71755
refactor : remove libcurl, use OpenSSL when available (#18828)
angt committed 18 days ago
Verified 516a4ca9
vulkan: Check maxStorageBufferRange in supports_op (#18709)
jeffbolznv committed 18 days ago
Verified 3e4bb296
llama-model: fix unfortunate typo (#18832)
am17an committed 18 days ago
Verified 47f96124
CUDA : fix typo in clang pragma comment [no ci] (#18830)
danbev committed 18 days ago
Verified 01cbdfd7
vulkan: work around Intel fp16 bug in mmq (#18814)
0cc4m committed 18 days ago
Verified 635ef78e
Terms Of Use Privacy Policy Refund Policy Impressum