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 259919 次贡献
更少更多
- frostbyte_neo 推送了仓库 github-featured/deepspeedai--deepspeed 的 copilot/fix-formatting-checks-again 分支比较 2 提交 »634b7be3af Fix OPSD benchmark formatting checksd7c7142b75 Initial plan
- frostbyte_neo 推送了仓库 github-featured/deepspeedai--deepspeed 的 olruwase/adam_types 分支比较 2 提交 »3ca637140f Merge branch 'master' of github.com:microsoft/DeepSpeed into olruwase/adam_typese51e4d7996 Support optimizer AdamW type
- frostbyte_neo 推送了仓库 github-featured/deepspeedai--deepspeed 的 costineseanu/windows_inference_build 分支87b9ad8513 Merge branch 'master' into costineseanu/windows_inference_build
- frostbyte_neo 推送了仓库 github-featured/deepspeedai--deepspeed 的 olruwase/align_rrg_rs_param_order 分支cb452ccf18 Merge branch 'master' into olruwase/align_rrg_rs_param_order
- frostbyte_neo 推送了仓库 github-featured/deepspeedai--deepspeed 的 copilot/fix-formatting-checks 分支比较 2 提交 »6dea5f64cb Fix formatting for benchmarks/opsd benchmark scripts7fba10ae98 Initial plan
- frostbyte_neo 推送了仓库 github-featured/deepspeedai--deepspeed 的 olruwase/accelerator_abstraction 分支7fbfee64fd Merge branch 'master' into olruwase/accelerator_abstraction
- frostbyte_neo 推送了仓库 github-featured/deepspeedai--deepspeed 的 copilot/fix-dco-failure-again 分支fb731ff9eb Initial plan
- frostbyte_neo 推送了仓库 github-featured/deepspeedai--deepspeed 的 niumanar/gan_optimizer 分支4bee4abd98 gan optimizer
- frostbyte_neo 推送了仓库 github-featured/deepspeedai--deepspeed 的 neox-q-int8 分支77735015dc Merge branch 'master' into neox-q-int8
- frostbyte_neo 推送了仓库 github-featured/deepspeedai--deepspeed 的 copilot/fix-dco-failure 分支比较 2 提交 »39617b927e Fix minor comment/docstring typos in runtime and inference modulescbd679d029 Initial plan
- frostbyte_neo 推送了仓库 github-featured/deepspeedai--deepspeed 的 mz/llama-support 分支比较 5 提交 »aefe074629 revert permissions change43787da969 formatting7a8088a401 Refactored the code to avoid using an except block.fe7503b73e Fixes to support llama fine-tuning in ds-chat056d959d03 Engine side fix for loading llama checkpoint fine-tuned with zero3
