One-Step Gradient Delay is Not a Barrier for Large-Scale Asynchronous Pipeline Parallel LLM Pretraining
Paper • 2606.30634 • Published • 18
Egor Petrov
moderntalker
AI & ML interests
None yet
Recent Activity
upvoted a paper about 5 hours ago
One-Step Gradient Delay is Not a Barrier for Large-Scale Asynchronous Pipeline Parallel LLM Pretraining updated a model about 19 hours ago
moderntalker/efficient_pretrain_checkpoints published a model about 2 months ago
moderntalker/efficient_pretrain_checkpointsOrganizations
None yet
upvoted a paper about 5 hours ago