Skip to content

Activity

Showing most recent first

Add example

EricLBuehlerpushed 1 commit to refactor_multiproc_tp • d350620…83b0d56 • 
15 hours ago

Refactor multiproc TP impl

EricLBuehlercreated refactor_multiproc_tp • d350620 • 
15 hours ago

Deploying to gh-pages from @ a691154 🚀

github-actions[bot]pushed 1 commit to gh-pages • d643259…72685ff • 
18 hours ago

Deleted branch

EricLBuehlerdeleted deepseekv3_fixes • 
18 hours ago

DSV3/R1 fixes (#1173)

Pull request merge
EricLBuehlerpushed 1 commit to master • c196ddc…a691154 • 
18 hours ago

Format everything

EricLBuehlerpushed 2 commits to deepseekv3_fixes • 7187fda…eca899c • 
18 hours ago

Works really now

EricLBuehlerpushed 2 commits to deepseekv3_fixes • f7748ea…7187fda • 
yesterday

Deploying to gh-pages from @ c196ddc 🚀

github-actions[bot]pushed 1 commit to gh-pages • 403cae8…d643259 • 
yesterday

Deleted branch

EricLBuehlerdeleted dependabot/cargo/ring-0.17.13 • 
yesterday

Bump ring from 0.17.11 to 0.17.13 (#1179)

Pull request merge
EricLBuehlerpushed 1 commit to master • b73e2e9…c196ddc • 
yesterday

Update build

EricLBuehlerpushed 2 commits to deepseekv3_fixes • 46bd102…f7748ea • 
yesterday

Bump ring from 0.17.11 to 0.17.13

dependabot[bot]created dependabot/cargo/ring-0.17.13 • baeb326 • 
yesterday

Optimize non-mla with cat

EricLBuehlerpushed 1 commit to deepseekv3_fixes • ae67013…46bd102 • 
yesterday

Async ops

EricLBuehlerpushed 1 commit to deepseekv3_fixes • 8d73473…ae67013 • 
yesterday

It actually works

EricLBuehlerpushed 1 commit to deepseekv3_fixes • 794a57d…8d73473 • 
yesterday

Fix launch of blockwise fp8 dequant

EricLBuehlerpushed 1 commit to deepseekv3_fixes • 5b20608…794a57d • 
3 days ago

Just save the progress

EricLBuehlerpushed 1 commit to deepseekv3_fixes • e539a8b…5b20608 • 
4 days ago

Deploying to gh-pages from @ b73e2e9 🚀

github-actions[bot]pushed 1 commit to gh-pages • dd30c81…403cae8 • 
4 days ago

Merge branch 'master' into deepseekv3_fixes

EricLBuehlerpushed 2 commits to deepseekv3_fixes • 5c9960f…e539a8b • 
4 days ago

Deleted branch

EricLBuehlerdeleted refactor_nccl_device_map • 
4 days ago

Refactor NCCL device mappers (#1172)

Pull request merge
EricLBuehlerpushed 1 commit to master • 40ac027…b73e2e9 • 
4 days ago

DSv3 fixes

EricLBuehlercreated deepseekv3_fixes • 5c9960f • 
4 days ago

Refactor nccl device mappers

EricLBuehlercreated refactor_nccl_device_map • 4366208 • 
4 days ago

Deploying to gh-pages from @ 40ac027 🚀

github-actions[bot]pushed 1 commit to gh-pages • 3e26585…dd30c81 • 
5 days ago

Deleted branch

EricLBuehlerdeleted remove_synchronize • 
5 days ago

Remove gpu<>cpu sync for faster long-context (#1170)

Pull request merge
EricLBuehlerpushed 1 commit to master • cf97e8e…40ac027 • 
5 days ago

Some fixes, bump synchronize limit

EricLBuehlerpushed 1 commit to remove_synchronize • 356865b…066b9a9 • 
5 days ago

Remove gpu<>cpu sync for faster long-context

EricLBuehlercreated remove_synchronize • 356865b • 
5 days ago

Deploying to gh-pages from @ cf97e8e 🚀

github-actions[bot]pushed 1 commit to gh-pages • 13a181e…3e26585 • 
6 days ago

Deleted branch

EricLBuehlerdeleted no_extra_cat_rope • 
6 days ago