Skip to content

Actions: gpu-cli/zerostart

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
28 workflow runs
28 workflow runs

Filter by Workflow

⌥ + click/return to exclude

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
Bump version to 0.2.0 for release Release #2: Commit 0aa2de8 pushed by lightsofapollo
3m 15s v0.2.0
v0.2.0
3m 15s
Bump version to 0.2.0 for release CI #26: Commit 0aa2de8 pushed by lightsofapollo
39s main
main
39s
Update README with progressive loading, new benchmarks, current archi… CI #25: Commit 9305251 pushed by lightsofapollo
33s main
main
33s
Feat: progressive loading — Python starts while packages install CI #24: Commit cb039b8 pushed by lightsofapollo
42s main
main
42s
Feat: skip nvidia CUDA wheels when system provides compatible libs CI #23: Commit bfb034d pushed by lightsofapollo
37s main
main
37s
Perf: parallel extract workers + download largest wheels first CI #22: Commit d7e6a0d pushed by lightsofapollo
37s main
main
37s
Perf: GET+pipeline architecture, delete Range-request streaming CI #21: Commit a88c2f2 pushed by lightsofapollo
1m 2s main
main
1m 2s
Fix system package isolation and accelerate() ordering in zerostart run CI #20: Commit 742e44f pushed by lightsofapollo
42s main
main
42s
README: clarify cold start speedup is network-dependent, document aut… CI #19: Commit 339f8ae pushed by lightsofapollo
35s main
main
35s
Auto-detect deps from PEP 723, pyproject.toml, and requirements.txt CI #18: Commit 78b4e73 pushed by lightsofapollo
40s main
main
40s
Remove hardcoded path from bench_50.sh CI #17: Commit 1ff0df6 pushed by lightsofapollo
55s main
main
55s
Soften README headline — focus on what zerostart does, not comparisons CI #16: Commit d881de4 pushed by lightsofapollo
55s main
main
55s
Add install script and GitHub release workflow Release #1: Commit 34435c4 pushed by lightsofapollo
4m 59s v0.1.0
v0.1.0
4m 59s
Add install script and GitHub release workflow CI #15: Commit 34435c4 pushed by lightsofapollo
57s main
main
57s
Add disk space guard, shared cache toggle, binary rename CI #14: Commit c43db5e pushed by lightsofapollo
53s main
main
53s
README: add install instructions, honest limitations, when-to-use guide CI #13: Commit f299a9d pushed by lightsofapollo
56s main
main
56s
Document vLLM integration as opt-in in README CI #12: Commit afebcba pushed by lightsofapollo
58s main
main
58s
vLLM loader: use safetensors_load_strategy="eager" on network volumes CI #11: Commit 1f3d10e pushed by lightsofapollo
58s main
main
58s
Fix vLLM --load-format zerostart: rewrite load_format before parent init CI #10: Commit fe408f2 pushed by lightsofapollo
53s main
main
53s
Fix vLLM integration: use register_model_loader API, subclass BaseMod… CI #9: Commit 0450f12 pushed by lightsofapollo
55s main
main
55s
Add diffusers support, ComfyUI and vLLM integrations CI #8: Commit 37affb5 pushed by lightsofapollo
1m 1s main
main
1m 1s
Caching improvements: resolution cache, RECORD-based populate, cache CLI CI #7: Commit 9be4843 pushed by lightsofapollo
58s main
main
58s
Update README with benchmarks and architecture explanation CI #6: Commit cd40eea pushed by lightsofapollo
59s main
main
59s
Make perf knobs configurable via env vars CI #5: Commit 85a7602 pushed by lightsofapollo
57s main
main
57s
Route all wheeled packages through daemon + shared CUDA layer cache CI #4: Commit 115ec5b pushed by lightsofapollo
57s main
main
57s