Tag

#lora

Posts tagged with lora

2 posts

Async GRPO training on serverless GPUs swaps NCCL collectives for object storage, a lightweight proxy, and elastic spot rollout workers.

Engineering • Sep 14, 2026 •12 min read

Async GRPO Training on Serverless GPUs Cuts RL Cost

Async GRPO training with LoRA swaps NCCL for a bucket and proxy, turning RL fine-tuning from reserved-cluster spend into elastic spot-GPU economics.

#grpo #lora +4

Deep learning GPU training where gradient accumulation splits an effective batch into micro-batches that change wall-clock time on rented hardware.

Engineering • Aug 19, 2026 •12 min read

The Real Cost of Gradient Accumulation on T4 and L4

Gradient accumulation can make identical batches train at different speeds. Learn why micro-batch shape drives T4 vs L4 wall-clock time and throughput.

#gradient-accumulation #gpu-benchmarks +4