Skip to content

Actions: JamePeng/llama-cpp-python

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
447 workflow runs
447 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

Optimize longest_token_prefix to use zero-copy NumPy arrays and drop … Tests #553: Commit 5304cba pushed by JamePeng
7m 45s main
main
7m 45s
Free _candidates and large numpy arrays during explicit close() Tests #552: Commit 7b360e2 pushed by JamePeng
9m 48s main
main
9m 48s
Update llama_model_quantize params Tests #551: Commit eedfa52 pushed by JamePeng
8m 14s main
main
8m 14s
Fix custom sampler memory cleanup and improve lifecycle management Tests #550: Commit e7dda03 pushed by JamePeng
8m 5s main
main
8m 5s
Build Wheels (Metal) Build Wheels (Metal) #13: Manually run by JamePeng
19m 13s main
main
19m 13s
fix typo in metal workflow Tests #549: Commit 3d0fd1b pushed by JamePeng
9m 3s main
main
9m 3s
Build Wheels (CU128) for Windows Build Wheels (CU128) for Windows #26: Manually run by JamePeng
4h 30m 22s main
main
4h 30m 22s
Build Wheels (CU124-126) for Windows Build Wheels (CU124-126) for Windows #19: Manually run by JamePeng
8h 4m 47s main
main
8h 4m 47s
Build Wheels (CU130) for Windows Build Wheels (CU130) for Windows #12: Manually run by JamePeng
6h 18m 49s main
main
6h 18m 49s
Build Wheels(CU124) for Linux Build Wheels(CU124) for Linux #66: Manually run by JamePeng
2h 27m 33s main
main
2h 27m 33s
Build Wheels(CU126) for Linux Build Wheels(CU126) for Linux #66: Manually run by JamePeng
1h 13m 25s main
main
1h 13m 25s
Build Wheels(CU128) for Linux Build Wheels(CU128) for Linux #37: Manually run by JamePeng
1h 37m 26s main
main
1h 37m 26s
Build Wheels(CU130) for Linux Build Wheels(CU130) for Linux #11: Manually run by JamePeng
1h 25m 38s main
main
1h 25m 38s
Bump version to 0.3.26 Tests #548: Commit 32f2380 pushed by JamePeng
8m 23s main
main
8m 23s
Update Submodule vendor/llama.cpp c78e682..abb9f3c Tests #547: Commit 6b38182 pushed by JamePeng
8m 33s main
main
8m 33s
feat(loader): extend default library search paths on Linux and macOS Tests #546: Commit 2ebe0eb pushed by JamePeng
8m 48s main
main
8m 48s
optimize(memory): reduce scores buffer size and optimize state saving Tests #537: Commit f9f8669 pushed by JamePeng
8m 7s main
main
8m 7s
Update Submodule vendor/llama.cpp afa6bfe..eeef3cf Tests #536: Commit bb8437a pushed by JamePeng
8m 16s main
main
8m 16s
Copy missing _keep_alive and custom_samplers list to new sampler Tests #535: Commit 645106e pushed by JamePeng
9m 50s main
main
9m 50s
feat: Add explicit memory cleanup for sampling contexts Tests #534: Commit 8f5c9b6 pushed by JamePeng
8m 28s main
main
8m 28s
perf(llama-cpp): optimize LlamaTokenDataArray memory operations Tests #533: Commit c6c85b1 pushed by JamePeng
8m 17s main
main
8m 17s
Update Submodule vendor/llama.cpp cceb1b4..afa6bfe Tests #532: Commit 16175e5 pushed by JamePeng
8m 16s main
main
8m 16s
Implemented Qwen3.5 Tests #531: Pull request #61 opened by alcoftTAO
8m 14s TAO71-AI:main
TAO71-AI:main
8m 14s
build: set GGML_METAL_USE_BF16 option ON for metal build by default Tests #530: Commit 6e52096 pushed by JamePeng
9m 12s main
main
9m 12s
Remove unneed allocation Tests #529: Pull request #60 reopened by roj234
7m 53s roj234:main
roj234:main
7m 53s