ggerganov/llama.cpp

Pull Requests Commits

vulkan: correctly report support for OP_CONT (ggml/946)

smeso committed 1 year ago

9cb92608

tests: add gradient tests for all backends (ggml/932)

JohannesGaessler committed 1 year ago

202084d3

ggml: fix ggml_graph_cpy undefined behavior (ggml/943)

JohannesGaessler committed 1 year ago

dbbebcab

cann : fix doxy (ggml/0)

ggerganov committed 1 year ago

ba1cf846

cann : add Ascend NPU support (whisper/2336)

MengqingCao committed 1 year ago

d2d3200b

cuda : mark BF16 CONT as unsupported

ggerganov committed 1 year ago

51d964a4

ggml : fix cont with transposed tensors when one dimension is 1 (ggml/934)

smeso committed 1 year ago

efe6a83e

llama : set attrs of mislabelled EOT/EOM tokens (#9348)

bakkot committed 1 year ago

Verified fbb7fcff

llama.android : fix build (#9350)

ggerganov committed 1 year ago

Verified a5b5d9a1

llama : fix empty ring buffer push (#9358)

ggerganov committed 1 year ago

Verified f12295b8

llama : sanitize invalid tokens (#9357)

ggerganov committed 1 year ago

Verified faf69d42

llamafile : disable sgemm for batch-size 1 (#9330)

netrunnereve committed 1 year ago

Verified e536426d

common : refactor arg parser (#9308)

ngxson committed 1 year ago

Verified 1b9ae518

ggml : always check bounds on get_rows operations (#9354)

slaren committed 1 year ago

Verified e32d0816

llama : refactor sampling v2 (#9294)

ggerganov committed 1 year ago

Verified df270ef7

ggml : fix missing `cpu_set_t` on emscripten (#9336)

ngxson committed 1 year ago

Verified 947538ac

ci : disable rocm image creation (#9340)

slaren committed 1 year ago

Verified 6c89eb0b

server : simplify state machine for slot (#9283)

ngxson committed 1 year ago

Verified 9b2c24c0

llama-bench : log benchmark progress (#9287)

akx committed 1 year ago

Verified 134bc38e

batched-bench : add `--output-format jsonl` option (#9293)

akx committed 1 year ago

Verified 815b1fb2

ggml : fix build break for the vulkan-debug (#9265)

cyzero-kim committed 1 year ago

Verified 409dc4f8

server : fix missing lock (#9334)

ngxson committed 1 year ago

Verified 4a1411b4

Improve Vulkan shader build system (#9239)

mtavenrath committed 1 year ago

Verified 8ebe8dde

ggml-quants : ternary packing for TriLMs and BitNet b1.58 (#8151)

compilade committed 1 year ago

Verified 9bc6db28

Update build.yml (#9184)

awatuna committed 1 year ago

Verified 32b2ec88

CMake fix: host for msvc compiler can only be x86 or x64 (#8624)

Xarbirus committed 1 year ago

Verified 1031771f

cuda : fix defrag with quantized KV (#9319)

slaren committed 1 year ago

Verified 4db04784

llama-bench : fix NUL terminators in CPU name (#9313)

slaren committed 1 year ago

Verified bdf314f3

ggml : AVX2 support for Q4_0_8_8 (#8713)

Srihari-mcw committed 1 year ago

Verified 581c3051

[SYCL] Fix DMMV dequantization (#9279)

OuadiElfarouki committed 1 year ago

Verified 5910ea94

Newer Older