12月1月2月3月4月5月6月7月8月9月10月11月
周一周三周五
一年内 260360 次贡献
更少更多
- frostbyte_neo 推送了仓库 github-featured/clacky-ai--openclacky 的 main 分支比较 4 提交 »9123c7df42 chore: bump version to 1.5.1163f53660ec docs: update CHANGELOG for v1.5.111284958311 feat(ext-studio): group published extensions by origin (brand vs marketplace)8ccb9c9ba7 fix: pick up renewed license key on skill decrypt retry
- frostbyte_neo 推送了仓库 github-featured/davila7--claude-code-templates 的 review/project-manager-2026-08-23 分支872db00923 improve: enhance project-manager based on automated review
- frostbyte_neo 推送了仓库 github-featured/davila7--claude-code-templates 的 main 分支比较 2 提交 »5d31328245 chore: Update components and trending datacf4f7d7761 improve: enhance project-manager based on automated review (#831)
- frostbyte_neo 推送了仓库 github-featured/voltagent--awesome-agent-skills 的 main 分支比较 27 提交 »122366cebe Merge pull request #931 from lindblomstefan/mainb776512350 Merge pull request #930 from citlyze-bot/patch-13f76209744 Merge pull request #929 from hedralab/main2272579d91 Update README0175afd44b Merge pull request #940 from liamka/add-apitube-news-api-skill
- frostbyte_neo 推送了仓库 github-featured/lightseekorg--tokenspeed 的 perf/prep-tape-input-tail 分支c044aeabc7 perf(runtime): scrub the padded input tail in one launch
- frostbyte_neo 推送了仓库 github-featured/lightseekorg--tokenspeed 的 hongtaoc/attnres-no-t-limit-103120 分支比较 2 提交 »9d258bf8c7 Merge branch 'main' into hongtaoc/attnres-no-t-limit-10312057324f7355 test(attn_res): cover long Blackwell sequences
- frostbyte_neo 推送了仓库 github-featured/lightseekorg--tokenspeed 的 test/kimi-k3-pdsim-bench 分支比较 2 提交 »5de4bd6980 test(kimi-k3): dedup first turns and guard shrunken samples in pd_sim7a585e7d5d test(kimi-k3): add PD-disaggregation simulation bench (pd_sim)
- frostbyte_neo 推送了仓库 github-featured/lightseekorg--tokenspeed 的 main 分支比较 2 提交 »f992ceb5e1 perf(kda): size the batched commit launches from the grid they actual… (#1217)5f4e37bbb1 feat(runtime): prefill chunk pipeline + event-loop cohesion refactor (#1214)
- frostbyte_neo 推送了仓库 github-featured/lightseekorg--tokenspeed 的 perf/amd-kda-prefill-tuning 分支比较 19 提交 »94604367aa perf(gfx1250): fill the WMMA tile in the KDA state scan91a0b07c08 perf(gfx1250): move KDA to V-major state and retune fused decode2af1bb1971 perf(amd): drop the zero fill for KDA triangular buffers1c0a340299 perf(gfx1250): widen KDA W/U vector kernel and reuse the key load103120bf77 perf(k3): extend latent-tail fusion to M64 with split collectives (#1174)
- frostbyte_neo 推送了仓库 github-featured/jundot--omlx 的 main 分支比较 2 提交 »f244ac0842 fix(dflash): honor repetition context size (#3011)1931c6c875 fix(ane): roll back partial dual dispatch (#2974)
- frostbyte_neo 推送了仓库 github-featured/yamadashy--repomix 的 main 分支比较 2 提交 »f465ad9093 Merge pull request #1816 from yamadashy/renovate/major-root-major-dependencies2c58eac60c fix(deps): update dependency gpt-tokenizer to v4
- frostbyte_neo 推送了仓库 github-featured/yamadashy--repomix 的 gh-pages 分支188aaa4a01 add Repomix Performance (customSmallerIsBetter) benchmark result for f465ad909315a22120636baf03fa5e28701a50cb
- frostbyte_neo 推送了仓库 github-featured/smol-machines--smolvm 的 spec-remote-volumes 分支ae21cd5a7e Carry remote volumes on the embedded machine spec so SDK machines can mount buckets
