ngxson/llama.cpp

Pull Requests Commits
py : fix persimmon `n_rot` conversion (#5460)
lx200916 committed 2 years ago
Verified dbd8828e
ggml-sycl: Replace 3d ops with macro (#5458)
abhilash1910 committed 2 years ago
Verified 43fe07c1
llava : remove prog parameter from ArgumentParser (#5457)
danbev committed 2 years ago
Verified 4a46d2b7
sync : ggml (#5452)
ggerganov committed 2 years ago
Verified 3b169441
CUDA: mul_mat_vec_q tiling, refactor mul mat logic (#5434)
JohannesGaessler committed 2 years ago
Verified 3bdc4cd0
Add support for BERT embedding models (#5423)
iamlemec committed 2 years ago
Verified 2891c8aa
flake.lock: Update
github-actions[bot] committed 2 years ago
97a33650
vulkan: only use M-sized matmul on Apple GPUs (#5412)
slp committed 2 years ago
Verified c88c74f9
common : use enums for sampler types (#5418)
z80maniac committed 2 years ago
Verified a803333a
server : allow to specify tokens as strings in logit_bias (#5003)
z80maniac committed 2 years ago
Verified 68478014
main : ctrl+C print timing in non-interactive mode (#3873)
ggerganov committed 2 years ago
Verified 85910c5b
common : fix compile warning
ggerganov committed 2 years ago
Verified 139b62a8
ggml : fix compile warnings (unused vars) (#4966)
ggerganov committed 2 years ago
Verified 0f2411f1
ggml : add mmla kernels for quantized GEMM (#4966)
snadampal committed 2 years ago
Verified a07d0fee
lookup: add print for drafting performance (#5450)
JohannesGaessler committed 2 years ago
Verified e4640d8f
server : add llama2 chat template (#5425)
ngxson committed 2 years ago
Verified 907e08c1
metal : use autoreleasepool to avoid memory leaks (#5437)
irbull committed 2 years ago
Verified f026f812
scripts : update sync scripts with new backends
ggerganov committed 2 years ago
Verified cd9aea63
sync : ggml
ggerganov committed 2 years ago
Verified 43b65f5e
ggml : add abort_callback for cpu backend (ggml/725)
Xarbirus committed 2 years ago
Verified 4633d93a
vulkan: Set limit for task concurrency (#5427)
luciferous committed 2 years ago
Verified 4b7b38be
llava : add requirements.txt and update README.md (#5428)
danbev committed 2 years ago
Verified e00d2a62
server : fix prompt caching for repeated prompts (#5420)
ristew committed 2 years ago
Verified 7c777fcd
llama : do not cap thread count when MoE on CPU (#5419)
ptsochantaris committed 2 years ago
Verified e5ca3937
readme : add JavaScript/Wasm repo (#5415)
mtasic85 committed 2 years ago
Verified e4124c24
ggml : fix `error C2078: too many initializers` for MSVC ARM64 (#5404)
Xarbirus committed 2 years ago
Verified b2f87cb6
Fix Vulkan crash on APUs with very little device memory (#5424)
0cc4m committed 2 years ago
Verified 44fbe343
CUDA: more warps for mmvq on NVIDIA (#5394)
JohannesGaessler committed 2 years ago
Verified 8e6a9d2d
llama : do not print "offloading layers" message in CPU-only builds (#5416)
slaren committed 2 years ago
Verified 41f308f5
Fix f16_sycl cpy call from Arc (#5411)
abhilash1910 committed 2 years ago
Verified 6e99f2a0
Terms Of Use Privacy Policy Refund Policy Impressum