ngxson/llama.cpp

Pull Requests Commits
add x param to ggml_vec_mad1_f32
ngxson committed 1 year ago
ebbad779
use scalar for __ARM_FEATURE_SVE
ngxson committed 1 year ago
cd1703a3
make code looks more consistent
ngxson committed 1 year ago
4ea74b04
use memcpy for op params
ngxson committed 1 year ago
0d70ca81
rm __ARM_FEATURE_SVE
ngxson committed 1 year ago
50c678f6
vDSP_vsmsa
ngxson committed 1 year ago
563aca0b
fix cann compile error
ngxson committed 1 year ago
265cb435
suggestions from coderabbit
ngxson committed 1 year ago
c8d89317
fix cuda
ngxson committed 1 year ago
b22708fd
will this fix cpu?
ngxson committed 1 year ago
4d019532
opencl
ngxson committed 1 year ago
0e51a0a8
cann (placeholder)
ngxson committed 1 year ago
477a97ad
vulkan
ngxson committed 1 year ago
782b58fa
sycl
ngxson committed 1 year ago
a28df6f0
add CUDA
ngxson committed 1 year ago
92a87384
add more simd
ngxson committed 1 year ago
e427af75
ggml_vec_mad1_f32
ngxson committed 1 year ago
a5ccf168
Merge branch 'master' into xsn/ggml_scale_bias
ngxson committed 1 year ago
7af3fd98
vulkan: optimize flash attention split_k_reduce (#14554)
jeffbolznv committed 1 year ago
Verified 6efcd659
model : fix hunyuan moe chat template (#14584)
stevenkuang-tencent committed 1 year ago
Verified 699f4392
model : add SmolLM3 (#14581)
ngxson committed 1 year ago
Verified 08382869
memory : fix broken batch splits for recurrent cache (#14575)
compilade committed 1 year ago
Verified bb4f7a9e
vulkan : fix rope with partial rotation and non-cont src (#14582)
jeffbolznv committed 1 year ago
Verified b8eeb874
server: Add ability to mount server at prefix (#14544)
oluwandabira committed 1 year ago
Verified 17a1f0d2
model : add hunyuan moe (#14425)
ngxson committed 1 year ago
Verified 8f22dc0a
vulkan: increase timeout for CI (#14574)
jeffbolznv committed 1 year ago
Verified 53903ae6
cuda : fix rope with partial rotation and non-cont src (#14580)
ggerganov committed 1 year ago
Verified 4d0dcd4a
CUDA: add bilinear interpolation for upscale (#14563)
am17an committed 1 year ago
Verified 75c91de6
musa: fix build warnings (unused variable) (#14561)
yeahdongcn committed 1 year ago
Verified 68155c66
llama : fix incorrect minicpm3 v_states shape (#14571)
CISC committed 1 year ago
Verified e1a70590
Terms Of Use Privacy Policy Refund Policy Impressum