Merge branch 'main' into cb-tp-aware-pool - SemanticDiff

Commit

4 days ago

Merge branch 'main' into cb-tp-aware-pool

References

#48294 - [Umbrella] Continuous batching fixes for training-while-generating on one weight copy

#48566 - [CB] Size the KV pool with the per-rank head count under TP

Author

qgallouedec

qgallouedec

Parents

Loading