ngxson/llama.cpp

Pull Requests Commits
vulkan: fix mul_mat_vec_iq1_s formatting (#18026)
0cc4m committed 17 days ago
Verified 9e6649ec
graph: add f_attn_temp_offset (#18025)
ngxson committed 17 days ago
Verified 0759b09c
common : refactor common_sampler + grammar logic changes (#17937)
ggerganov committed 17 days ago
Verified 254098a2
vulkan: Fix data race/hang in scalar/cm1 flash attention (#17887)
jeffbolznv committed 17 days ago
Verified 3238b140
vulkan: improve mul_mat_vec_iq1_s speed (#17874)
lovedheart committed 17 days ago
Verified 47226716
vulkan: faster q6_k matmul (#17813)
netrunnereve committed 17 days ago
Verified d15d177f
model-conversion : cast logits to float32 (#18009)
ggerganov committed 17 days ago
Verified 77ad8542
models : fix YaRN regression + consolidate logic (#18006)
ggerganov committed 17 days ago
Verified 609a2d02
ggml : arm repack fix build
ggerganov committed 17 days ago
a63cbafb
sync : ggml
ggerganov committed 17 days ago
0e592249
ggml : arm repack fix build (whisper/0)
ggerganov committed 17 days ago
71fdcf06
cmake : set `CMAKE_RUNTIME_OUTPUT_DIRECTORY` for non standalone build (ggml/1394)
HerrCai0907 committed 17 days ago
615655aa
scripts: add script to compare logprobs of llama.cpp against other frameworks (#17947)
ngxson committed 17 days ago
Verified c00ff929
server-models.cpp: add missing <filesystem> (#18000)
barracuda156 committed 17 days ago
Verified 4ed2bae5
llama_context: synchronize before reallocating output buffer (#17974)
jeffbolznv committed 17 days ago
Verified 5266379b
arg: fix common_params_parse not accepting negated arg (#17991)
ngxson committed 18 days ago
Verified 4d5ae24c
cmake: correct scope - link ws2_32 for MinGW/w64devkit builds in cpp-httplib (#17972)
gustrd committed 18 days ago
Verified 66ba5125
vulkan: support get_rows for i32 (#17941)
jeffbolznv committed 18 days ago
Verified 36255a22
vulkan: support GGML_OP_DIAG (#17893)
jeffbolznv committed 18 days ago
Verified 3229a23f
vulkan: Multi-pass softmax for large number of cols (#17892)
jeffbolznv committed 18 days ago
Verified 303f8615
speculative-simple : free batch on exit (#17985)
ggerganov committed 18 days ago
Verified 3c6391e7
common : skip model validation when --completion-bash is requested (#17975)
CISC committed 18 days ago
Verified 8e4d6785
vulkan: Allow non-pow2 n_experts in topk_moe (#17872)
jeffbolznv committed 18 days ago
Verified 07a10c10
add llama-completion to completion-bash executables (#17976)
CISC committed 18 days ago
Verified 2bc94e79
model-conversion : use CONVERTED_MODEL value for converted model [no ci] (#17984)
danbev committed 18 days ago
Verified fd1085ff
common: support negated args (#17919)
ngxson committed 18 days ago
Verified 380b4c98
clip: move model cgraphs into their own files (#17965)
ngxson committed 18 days ago
Verified e39a2ce6
ci : change the cann version and the container pull method (#17953)
xuedinge233 committed 18 days ago
Verified a8c7f33d
docker : include legacy llama-completion binary (#17964)
CISC committed 18 days ago
Verified b7f5f46e
CUDA: fix overflow in MMA kernel without stream-k (#17939)
JohannesGaessler committed 18 days ago
Verified 48221143
Terms Of Use Privacy Policy Refund Policy Impressum