ggerganov/llama.cpp

Pull Requests Commits

Force max subgroup size for coopmat shaders

0cc4m committed 322 days ago

fe9235d7

Revert "llama : add Falcon3 support (#10864)" (#10876)

slaren committed 323 days ago

Verified 4da69d1a

Use model->gguf_kv for loading the template instead of using the C API. (#10868)

dranger003 committed 323 days ago

Verified d62b532c

tests: add tests for GGUF (#10830)

JohannesGaessler committed 323 days ago

Verified 081b29bd

ggerganov committed 323 days ago

Verified 5437d4aa

cmake : fix "amd64" processor string (whisper/2638)

ggerganov committed 323 days ago

Verified 78f76676

vulkan : fix soft_max.comp division by zero (whisper/2633)

gn64 committed 323 days ago

Verified 8dd19a48

ggml : remove return from ggml_gallocr_allocate_node (ggml/1048)

danbev committed 323 days ago

Verified 130d0c90

ggml : add check for grad_accs (ggml/1046)

danbev committed 323 days ago

Verified 3919da8e

ggml : update ggml_backend_cpu_device_supports_op (#10867)

ggerganov committed 323 days ago

Verified 0006f5a7

server : fill usage info in embeddings and rerank responses (#10852)

krystiancha committed 323 days ago

Verified 05c3a444

llama : add Falcon3 support (#10864)

mokeddembillel committed 323 days ago

Verified 382bc7f2

readme : update typos (#10863)

ruanych committed 323 days ago

Verified 4f51968a

server : (UI) fix missing async generator on safari (#10857)

ngxson committed 323 days ago

Verified 227d7c5a

vulkan: bugfixes for small subgroup size systems + llvmpipe test (#10809)

netrunnereve committed 323 days ago

Verified 7b1ec53f

rwkv6: add wkv6 support for Vulkan backend (#10829)

zhiyuan1i committed 324 days ago

Verified 160bc039

unicode : improve naming style (#10838)

ggerganov committed 324 days ago

Verified 08ea539d

sampling : refactor + optimize penalties sampler (#10803)

ggerganov committed 324 days ago

Verified 644fd71b

llava : Allow locally downloaded models for QwenVL (#10833)

bartowski1182 committed 325 days ago

Verified 4ddd199f

llama : add Deepseek MoE v1 & GigaChat models (#10827)

Inf1delis committed 325 days ago

Verified a0974156

scripts : change build path to "build-bench" for compare-commits.sh (#10836)

ggerganov committed 325 days ago

Verified 87cf323c

server: (UI) add syntax highlighting and latex math rendering (#10808)

VJHack committed 325 days ago

Verified 5478bbcd

gguf-py : bump to v0.13.0

ggerganov committed 325 days ago

Verified b5ae1ddf

server: Fix `has_next_line` in JSON response (#10818)

MichelleTanPY committed 326 days ago

Verified 89d604f2

nix: allow to override rocm gpu targets (#10794)

kurnevsky committed 326 days ago

Verified e52aba53

llama : add Qwen2VL support + multimodal RoPE (#10361)

HimariO committed 326 days ago

Verified ba1cb19c

Removes spurious \r in output that causes logging in journalctl to treat lines as binary and therefore hidden by default (#10771)

cduk committed 327 days ago

Verified 56eea078

Introducing experimental OpenCL backend with support for Qualcomm Adreno GPUs (#10693)

lhez committed 327 days ago

Verified a76c56fa

Opt class for positional argument handling (#10508)

ericcurtin committed 327 days ago

Verified c27ac678

fix: graceful shutdown for Docker images (#10815)

co42 committed 327 days ago

Verified 11e07fd6

Older