ggerganov/llama.cpp

Pull Requests Commits

metal : fix build and some more comments

ggerganov committed 202 days ago

Verified bd1198a6

ggml: fix zero division in ‘dne’ calculation in CUDA COUNT_EQUAL operator when ‘ne’ is small (#10213)

SongXiaoXi committed 202 days ago

Verified 5b359bb1

ggml : optimize llamafile cpu matrix multiplication for ppc64le (#10156)

amritahs-ibm committed 202 days ago

Verified e8921349

scripts : fix pattern and get n_tokens in one go (#10221)

lhpqaq committed 202 days ago

Verified 8fc393f2

metal : opt-in compile flag for BF16 (#10218)

ggerganov committed 202 days ago

Verified ec450d3b

metal : improve clarity (minor) (#10171)

ggerganov committed 203 days ago

Verified 695ad752

metal : optimize FA kernels (#10171)

ggerganov committed 203 days ago

Verified 841f27ab

swift : exclude ggml-metal-embed.metal (#10211)

jhen0409 committed 203 days ago

Verified d05b3127

server : minor UI fix (#10207)

ngxson committed 203 days ago

Verified 76c6e7f1

server : revamp chat UI with vuejs and daisyui (#10175)

ngxson committed 203 days ago

Verified a71d81cf

scripts : add amx to sync-ggml.sh [no ci]

ggerganov committed 203 days ago

Verified eec4d717

ggerganov committed 203 days ago

Verified 3b088286

scripts : sync update

ggerganov committed 203 days ago

Verified a2c6fd74

ggml : add ggml-cpu.h to the public headers (#10204)

slaren committed 204 days ago

Verified 97404c4a

Remove identical wte/etw logic for jais (#10203)

fmz committed 204 days ago

Verified 60e17ce2

DRY: Fixes clone functionality (#10192)

wwoodsTM committed 204 days ago

Verified 5107e8ce

fix q4_0_8_8 format for corrupted tokens issue (#10198)

snadampal committed 204 days ago

Verified 2319126a

Optimize RWKV6 Operator Naming and Implement Multi-core CPU/ SYCL Acceleration (#10133)

zhiyuan1i committed 204 days ago

Verified 3bcd40b3

metal : add BF16 support (#8439)

ggerganov committed 204 days ago

Verified 5c333e01

server : remove hack for extra parallel slot (#10187)

ggerganov committed 205 days ago

Verified b11f9ba9

metal : fix from ptr buffer name (#10189)

slaren committed 205 days ago

Verified 94d8cb8b

ggml : adjust is_first_call init value (#10193)

ggerganov committed 205 days ago

Verified 1dc04b2d

metal : add quantized FA support (#10149)

ggerganov committed 205 days ago

Verified a1eaf6a9

llama : add <|tool_call|> formatting to Granite template (#10177)

gabe-l-hart committed 206 days ago

Verified b8deef0e

ggml : fix arch check in bf16_to_fp32 (#10164)

slaren committed 206 days ago

Verified a9e8a9a0

Q6_K AVX improvements (#10118)

netrunnereve committed 206 days ago

Verified 34073647

ggml : fix gelu tables initialization (#10172)

slaren committed 206 days ago

Verified d5a409e5

ggml : fix q4xx mat mul, increase ggml_aligned_malloc alignment (#10167)

slaren committed 207 days ago

Verified 401558b7

server : clarify /slots endpoint, add is_processing (#10162)

ngxson committed 207 days ago

Verified 9e0ecfb6

fix build break on arm64 linux (#10166)

snadampal committed 207 days ago

Verified 6a066b99

Older